Diana Anderson, D M Conning - Experimental Toxicology-Royal Society of Chemistry (1993) PDF

EXPERIMENTAL TOXICOLOGY
The Basic Issues

Second Edition
EXPERZMENTAL
TOXZCOLOGY
The Basic Issues
Second Edition
Edited by
Diana Anderson
The British Industrial Biological Research Association,
Carshalton, Surrey
D. M. Conning
The British Nutrition Foundat ion, London
A catalogue record for this book is available from the British Library.
ISBN 0-85186-451-1 (Hardback)

0-85 186-461-9 (Softcover)
First published in hardback 1988

First published in softcover 1990
Reprinted 1992
Second edition published 1993
Second Edition 0 The Royal Society of Chemistry 1993
All Rights Reserved

N o part of this book may be reproduced or transmitted in any form or by any means - graphic,
electronic, including photocopying, recording, taping, or information storage and retrieval systems
- without written permission from The Royal Society of Chemistry.
Published by The Royal Society of Chemistry,

Thomas Graham House, The Science Park, Cambridge CB4 4WF
Printed and bound in the UK by Hartnolls Ltd., Bodmin

Contributors
W. N. Aldridge formerly The Robens Institute,

University of Surrey
H. E. Amos Pharmaco U K , Chelmsford
Diana Anderson BIBRA Toxicology International,
Carshalton
W. H. Butler BIBR A Toxicology International,
Carshalton
D. M. Conning British Nutrition Foundation, London
B. Copeland Barry Copeland Architects, London
J. G. Evans Fisons plc, Loughborough
S. D. Gangolli M R C Lab ora tor ies, Carsha1ton
P. Grasso formerly The Robens Institute,
University of Surrey
R. I. Hawkins Coca Cola Northwest Europe, London
A. B. G. Lansdown The Charing Cross and Westminster
Medical School, London
P. N. Lee P. N . Lee Statistics and Computing Ltd.,
Sutton
D. P. Love11 BIBRA Toxicology International,
Carshalton
D. B. McGregor IARC, Lyon, France
G. M. Paddle Department of Occupational Health,
University of Birmingham
J. C. Phillips BIBRA Toxicology International,
Carshalton
P. Rumsby BIBRA Toxicology In ternat ional,
Carshalton
R. A. Riley formerly BIBRA Toxicology International,
Carshalton
F. J. Roe Consultant, 19 Marryat Road,
Wimbledon Common, London
A. B. Wilson Inveresk Research International,
Midlo thian
Contents
Chapter 1 Introduction to Experimental Toxicology

By D. M . Conning 1
Chapter 2 Effects of Physical Form, Route, and

Species
By A . B. Wilson 4
1 Introduction 4
2 Nature of Toxicant 5
3 Routes of Exposure 6
4 Choice of Species 16
5 Conclusions 20
References 20
Chapter 3 Influence of Animal Species, Strain, Age,

Hormonal, and Nutritional Status
By F . J . C. Roe 23
1 Introduction 23
2 Genetic versus Environmental Influences 23
3 Inbred versus Outbred Strains 24
4 Choice of Species 25
5 Numbers of Animals: Safety Evaluation as
Distinct from Mechanism Studies 26
6 Animal Susceptibility 27
7 The Effects of Overfeeding 28
8 Future Developments in Long Term Testing 32
9 Summary 32
10 1993 Update 33
References 34
vii
... Conten&
Vlll
Chapter 4 Experimental Design

B y A . B. Wilson 35
1 Regulatory and Experimental Toxicology 35

2 Object of Study 36
3 Customary Study Designs 39
4 Randomisation and Replication 49
5 Conclusion 53
References 53
Chapter 5 The Biochemical Principles of Toxicology

B y W . N . Aldridge 56
Introduction 56
Types of Chemical Interaction 56
Types of Toxicity 59
Physical and Chemical Properties of Toxic
Chemicals 61
Selective Toxicity and Selectivity 63
Phases of Developing Toxicity 65
Dose-Response Relationships 74
Measurement of Received Dose of Reactive
Chemicals 76
Conclusion 77
References 78
Chapter 6 Animal Husbandry

By A . B. G. Lansdown 82
1 Introduction 82
2 The Research Animal 83
3 Experimental Animal Facilities 87
4 Diet and Nutrition 89
5 Animal Maintenance and Breeding 93
6 Isolator Technology 97
7 Quality Assurance 98
8 Experimental Procedures 100
9 Records 102
10 Discussion 103
References 103
Contents ix
Chapter 7 Inhalation Toxicology

B y R. A . Riley and D. M . Conning 107
1 Introduction 107
2 Physical Classification of Airborne Pollutants 108
3 Methodology of Inhalation Toxicology 108
4 Lung Clearance Mechanisms 114
5 Conclusion 117
References 117
Chapter 8 Histopathology in Safety Evaluation

B y J . G. Evans and W. H . Butler 119
1 Introduction 119
2 The Role of Histopathology 119
3 Examination of Animals and Preparations 120
4 General Histopathology Terms 122
5 Non-neoplastic Histopathology Changes in
Laboratory Animals 124
References 127
Chapter 9 The Metabolism and Disposition of

Xenobiotics
B y S. D . Gangolli and J . C. Phillips 130
Introduction 130
Sites of Absorption 131
Distribution 134
Metabolic Reactions-Bioactivation and
Detoxification Processes 135
Sites of Metabolism 150
Bioactivation and Detoxification Processes 160
Factors Influencing Metabolism and
Pharmacokinetics 164
Concluding Remarks 169
References 170
Chapter 10 Theory and Practice in Metabolic Studies

By J . C . Phillips and S. D . Gangolli 181
Introduction 181
X Contents
1 Theoretical Considerations 181

2 Practical Considerations in Metabolic Studies 190
References 198
Chapter 11 Immunotoxicology-Conceptual Problems

By H . Amos 202
Introduction 202
Inadvertant Immunomodulation 202
Relevance to Man of Immunotoxicological
Models in Animals 206
Hypersensitivity as an Immunotoxic Response 207
Antigen Formation with Haptens 208
Metabolic Conversion and Enzyme
Interactions 209
Alteration of Drugs by Chemical Interaction 210
Pseudo-allergic reactions 210
Conc1usion 21 1
References 21 1
Chapter 12 Perspectives-The Evaluation of Reproductive

Toxicity and Teratogenicity
By A . B. G. Lansdown 214
Introduction 214
Historical 215
Principles of Reproductive Toxicity and
Teratogenesis 217
Epidemiological and Experimental Aspects 220
Mechanisms of Reproductive Toxicity and
Foetal Abnormality 22 1
Legislative Aspects of Reproductive Toxicity
Testing 229
Experiment a1 Considerations 232
Statistical Evaluation 237
General Observations 238
References 239
Chapter 13 Genetic Toxicology

By Diana Anderson 243
1 Introduction 243
2 Mutagenicity Tests 245
Contents xi
3 Usefulness of Some of the Systems for

Screening Purposes 26 1
4 Cell Transformation Assays 265
5 Test Significance and Interpretation 266
6 Strategies for the Protection of Man 267
7 General Conclusions 286
References 287
Chapter 14 Molecular Toxicology

By P. Rumsby 313
1 Introduction 313
2 Techniques of Molecular Biology 313
3 Applications of Molecular Biology to
Toxicology 318
4 Future Developments 328
References 329
Chapter 15 Testing for Carcinogenicity

By P. Grasso 334
1 Introduction 334
2 Methods Employed in Carcinogenic Studies 334
3 Evaluation of Results 349
4 Nature and Type of Tumour Induced 350
5 Factors Affecting Tumour Incidence 351
6 Conclusions 355
References 355
Chapter 16 In Vitro Methods for Teratology Testing

By Diana Anderson 360
1 Introduction 360
2 Sub-mammalian Systems 362
3 Mammalian Systems 364
4 Usefulness of the I n Vitro Methodology 369
References 370
Chapter 17 Assessing Chemical Injury in the

Reproductive System
By S. D . Gangolli and J . C . Phillips 376
1 Introduction 376
xii Contents
2 Anatomy and Physiology of the Male

Reproductive System 377
3 Anatomy and Physiology of the Female
Reproductive System 380
4 Mechanisms Regulating Testicular Function 382
5 Mechanisms Regulating Ovarian Function 385
6 Site Specificity of Chemical Injury by Model
Toxins 386
7 Methods for Studying Reproductive Toxicity
in the Male 390
8 Methods for Studying Reproductive Toxicity
in the Female 394
9 Methods for Assessing Reproductive Function 395
10 Conclusions 397
References 398
Chapter 18 Statistics
By P. N . Lee 405
1 Introduction 405
2 General Principles and Terminology 405
3 Experimental Cesign 412
4 Statistical Analysis 419
References 440
Chapter 19 Risk Assessment of Chemicals

By D . P. Lovell 442
1 Introduction 442
2 Estimating the Risks 444
3 Approaches to Risk Estimation from Animal
Studies 447
4 Interspecies Comparisons 455
5 An Example of Risk Estimation in Practice 457
6 An Example of Risk Estimation Using Mouse
Dominant Skeletal Mutations to Estimate
Human Risk from Radiation 458
7 Risk Evaluation and Management 459
8 Conclusions 462
References 462
...
Contents y-1 I Xlll
Chapter 20 Epidemiology
B y G. M . Paddle 464
1 Introduction 464
2 Sources of Problems 466
3 Guidelines 469
4 Cohort Studies 470
5 Case-control Studies 472
6 Other Types of Study 474
7 Control Groups 476
8 Confounding Factors 478
9 Data Collection 478
10 Data Banks 480
11 Interpretation 480
12 Essential Skills 482
13 Conclusion 482
References 483
Chapter 21 Information and Consultancy Services in

Toxicology
B y D. M . Conning 485
1 Introduction 485
2 Data Acquisition 488
3 Data Interpretation 489
4 Data Dissemination 489
5 Summary and Conclusions 490
Chapter 22 Regulations and Advisory Requirements in

Relation to Food
B y D . M . Conning 49 1
1 Introduction 49 1
2 Evolution of Food Law Exemplified by the
United Kingdom 492
Structure of Food Laws 493
Naturally Occurring Hazards in Food 494
Food Additive Legislation 495
Pesticides 503
Packaging Materials 503
XiV Contents
Chapter 23 The Influence of a Growing

Environmental Awareness on Toxicology
Laboratory Design
By Diana Anderson and B. Copeland 505
1 Introduction 505
2 Animal Laboratories 508
3 General Laboratories 511
References 520
Chapter 24 Good Laboratory Practice

By R . I . Hawkins 523
1 Introduction and Background 523
2 Current Situation 524
3 Specific Requirements 529
4 Conclusions and Summary 538
References 540
Chapter 25 Ethics in Experiments on Animals

By D. B. McGregor 542
1 Introduction 542
2 Human Experiments 544
3 Animal Experiments 546
4 Conclusions 551
References 553
Index 554
CHAPTER 1
Introduction to Experimental
Toxicology
D. M. CONNING
Toxicology is defined, classically, as the study of the adverse effects of

chemicals on living systems. Originally this meant the study of poisons
and that meaning remains the most satisfactory of definitions, essentially
because it embodies the concept of the effect being proportional to the
administered dose. Although the popular concept of poison concerns
poisoning to death, the true study of poisons embraces the induction of
morbid changes that are not necessarily fatal.
In this sense, the study of toxicology came to be regarded as an
extension of the study of pharmacology and it is still so regarded by those
who take a pedantic view of the topic. Of necessity, a pharmacological
assay must explore the optimum dose in relation to the maximum
therapeutic effect, and thus the dosage at which the effect is counter-
productive of the desired result. In many ways the association of
pharmacology and toxicology has been beneficial to the latter as a
burgeoning science because it instilled two basic constraints which
reinforced the scientific nature of the study. First was the recognition that
pharmacology involved the perturbation of physiological function. That
is, it was realised that a variety of detectable changes were compatible
with normal living, and pharmacology sought to enhance those aspects
which would be beneficial in the presence of disease or inhibit others for
the same purpose. Second, the pharmacological study was a study of a
defined function such as heart function, nerve transmission, or renal
reabsorption, and never involved a less specific or more abstruse
objective such as that embodied in the question ‘Do any effects occur?’ In
other words, the pharmacological approach demanded an investigation of
the way a particular physiological function could be chemically modified.
In recent decades, the definition of toxicology has been expanded in a
way which has taken it firmly and, it seems, irrevocably out of the field of
pharmacology. The first and most fundamental change was to add
another purpose to the study. Thus toxicology came to be defined as the
3
2 Introduction to Experimental Toxicology
study of the adverse effects of chemicals on living systems in order to

predict chemical hazard to man. This had the effect of classifying
toxicology as an ancillary to public or community health, and by
extension to preventive medicine, always the poor brother of therapeutic
medicine; and at the same time imposed impossible conditions on its
practice as a science. Not only did the study of toxicology become a study
of the effects of a chemical on any conceivable physiological function,
defined or not, but under any conceivable circumstances because of the
almost infinite variety of human activity and behaviour. Toxicology was
expected thereby to predict the effect of a chemical in systems which
themselves were not capable of being defined.
The problem was compounded by a further expansion of the definition
to include ‘chemicals or other agents’ and the inclusion of ‘man or his
environment.’ Thus toxicology is the study of the adverse effects of
chemicals or other agents on living systems in order to predict hazard to
man and his environment.
A number of very unsatisfactory consequences have resulted from
these developments. The first was the birth of the concept of the
‘no-effect level’ and its embodiment in safety regulations. It is simply not
possible at the present stage of development of toxicological knowledge
to define with any precision the normal values for many biological
activities and thus to define when abnormal values are detected. The best
we can do is to define where the values in treated systems (e.g. the
experimental rat) differ from those in untreated systems maintained in
similar circumstances. We know only rarely whether any observed
differences represent a toxic effect or an adaptive response. Sometimes
we have great difficulty in determining if there are any differences at all, a
problem which has given rise to a massive development in biornetrics.
Our adherence to the ‘no-effect level’ has undermined our faith in
epidemiology. Although the lack of epidemiological data and the relative
insensitivity of epidemiological methods have themselves contributed to
this outcome, it has seemed easier to put our faith in animal results which
can be determined with some precision and therefore appear to be more
easily judged. The result is that attempts to extrapolate the findings in
animals, for example, to predictable effects in man have no basis in
human experience and tend to assume that man’s response will be the
same as that of the animal.
Another consequence has been distortion of the economics of toxicolo-
gical practice in that those who are involved with toxicological experi-
ment spend so much of their time and resources generating data, there is
very little available for scientific interpretation and further experiment.
The construction and testing of hypotheses do not have a prominent role
in toxicology.
All of this has come about for the best possible motives. The
perception of the possible dangers consequent upon our chemical
inventiveness has resulted in the appearance of potent forces to protect
D.M. Conning 3
human communities against such consequences. The diligent pursuit of

detail has extended very considerably the requisite observations before a
‘no effect’ dosage can be determined. It is a matter of profound regret
that much of this invaluable data is not used to further our knowledge of
biological function.
In this book we hope to lay the foundations on which future
toxicologists can build the scientific practice of toxicology. Although we
have acknowledged the demands of modern society and provided the
basis for the career development of the toxicologist charged with the
provision of data on which the acceptability of new chemical or physical
agent can be defined, we hope this has been done in a way which does
not stifle the needs of the enquiring mind. Despite the problems,
toxicology still remains a science which promises real opportunities to
unravel some of the fascinating problems of biology by identifying
chemical and physical tools with which to probe living processes. In the
end, it is this aspect of toxicology that will contribute most to our
understanding of those features which determine the likelihood of human
disease.
CHAPTER 2
Eflects of Physical Form,

Route, and Species
A. B. WILSON
1 Introduction
When investigating the literature in reference texts such as Sax (1979),
which summarise toxicological information on a wide range of com-
pounds, there is the impression of simplicity and unequivocality . More
detailed texts (Clayton and Clayton, 1981) review the toxicology of
classes of compound, indicate gaps in knowledge, and variable,
even contradictory results. Hence there is well founded dispute about the
effects of low doses of heavy metals, whether or not certain chemicals
are carcinogenic, and the mechanisms of action of many compounds. This
is further explored in more general toxicology textbooks (Klaassen,
Amdur, and Doull, 1986; Hayes, 1983; Lu, 1985) and in toxicology
journals such as : Toxicology and Applied Pharmacology (Academic Press,
Duluth), Fundamental and Applied Toxicology (Academic Press, Duluth),
Food and Chemical Toxicology (Pergamon Journals Ltd., Exeter),
Comments on Toxicology (Gordon and Breach Science Publishers,
London), Archives of Toxicology (Springer-Verlag, Germany), Toxicology
(Elsevier, Ireland), Regulatory Toxicology and Pharmacology (Academic
Press, Duluth), and Human and Experimental Toxicology (Macmillan
Press, Basingstoke).
This chapter discusses the major factors that are causes of variability in
results-physical form, vehicles, routes of exposure, and animal species
as encountered in studies designed to screen compounds for general
toxic potential. Few papers are available on the variations that actually
occur in laboratory animal toxicology but information is published by
Gaines and Linder (1986) with regard to acute toxicology of pesticides,
Haseman (1983) with regard to carcinogenicity investigations, Lu (1985)
with regard to a selection of factors, and by Rao (1986), Gartner (1990),
Wollink (1989), and Vogel (1993) with regard to factors affecting
laboratory animals.
4
A . B. Wilson 5
2 Nature of Toxicant
A Physical Form
It can be expected that the physical nature of a material and the route of
exposure will alter toxicity. Experiments in animals have to take account
of this and batches of compound representative of normal production are
usually employed so that purity, particle size, and other factors will
match the real situation.
In general, a reduction in particle size improves solubility, increases
absorption, and is likely, therefore, to increase toxicity. The effects of
particle size are probably greatest on inhalation toxicity and that is
covered separately. Solids will, of course, encounter body fluids in the
gastro-intestinal tract, lungs, eyes, and even the skin and the therefore the
opportunity exists for the material to be dissolved and for absorption to be
improved. Insoluble materials are often regarded as inert and non-toxic.
There are exceptions : asbestos and coal dust induce toxic effects by virtue
of their insolubility and non-absorption which lead to residence in
tissues such as the lung.
The pharmacist ,pharmacologist , and experimental toxicologist are
more likely to be concerned with capsules (perhaps enteric coated to
carry them through the stomach intact), modification or buffering of pH
to alter dissociation, and selection of soluble salts (e.g. soluble aspirin)
to reduce local effects. All these measures and many more can have
potential profound effects upon the toxicity of a compound.
Liquids do not have variations in terms of particle size in quite the
same manner as do solids. However, liquids can become droplets
(aerosols) with varying dimensions, which may be of some relevance if
these are immiscible and of substantial significance if exposure is by
inhalation. Liquids may have significant vapour pressure so they can
sometimes be considered in much the same way as gases.
Gases and vapours present their hazards by being readily inhaled or by
penetrating the skin. They too can be dissolved in liquids or adsorbed
onto solids; indeed some show a remarkable affinity for solid surfaces, a
property exploited in the manufacture of plastic strips containing
pesticides such as dichlorvos.
The physical form of a compound will alter the potential for the
material to enter the body, the extent of likely absorption, and therefore
the eventual toxicity and hazard.
B Vehicle and Concentration

It has already been mentioned that gases,liquids, or solids may be
presented in a vehicle as mixtures, solutions, or suspensions. If the
vehicle is well absorbed, there is a tendency for the solute to be more
rapidly and completely absorbed and vice uersu. Many complex factors
6 Efects of Physical Form, Route, and Species
are involved, including the permeability of the surface to vehicle and
compound, the partition coefficients, the extent of ionisation, and the
effect of the material on local blood flow.
The concentration of material in the vehicle can alter toxicity. Often
the more dilute the solution, the more complete the absorption of the
solute (presuming the vehicle itself is absorbed) though the rate of ab-
sorption may be slower if dilution is substantial, as happens with alcohol. A
rapidly detoxified or excreted compound will show less toxicity if
absorption is slow. If large volumes are administered, it is worth bearing in
mind that the vehicle itself might contribute to the toxicity or variations
between results, for example, the interaction of corn oil in gavage studies
(Nutrition Foundation, 1983). More than one investigator has thought he
was looking at a concentration effect when it was probably vehicle
toxicity.
High concentrations are more likely to cause local reactions such as
irritation, ulceration, altered blood flow, or necrosis, which will them-
selves affect absorption and systemic toxicity, most obviously in dermal
studies but frequently by other routes. The local effects are, of course,
expressions of toxicity in their own right.
The interactions between vehicle and compound are exploited in the
art and science of formulating drugs. By skilful means the therapeutic
action, pharmacokinetics, and toxicity can be balanced to favour the
desirable over the adverse (Prescott and Nimmo, 1979; Prescott and
Nimmo, 1985).
3 Routes of Exposure
A Frequently Encountered Routes
The most frequently encountered exposures to industrial chemicals are by
the dermal and inhalation routes but these predominate in the place of
work where much can be done to limit and control the actual risks. In
fact, most of the experimental toxicology on laboratory animals is carried
out uia the oral route because:
(a) The oral route is the normal model of exposure (e.g. for food
additives and for many pharmaceuticals).
(b) The widest exposure to herbicides and pesticides is likely to be as
residues in crops or meat, which will be ingested.
(c) The oral route is simple to use and expedient. If reasonable evidence
exists to show that differences in toxicity related to route are
quantitative rather than qualitative then the bulk of the investiga-
tions will employ oral administration.
The route by which a chemical encounters the body can alter very
substantially its effects and, in particular, the quantity required to cause
that effect.
The physical presentation of a chemical and the route of exposure
interact to influence greatly whether a chemical will be absorbed, its rate of
A . B. Wilson 7
absorption, tissue distribution, removal, and excretion. The integrity of
the exposed surface, the residence time of the potential toxicant, and the
metabolic activity of the surface are amongst the relevant factors. It may be
reasonable to state that the route is likely to alter primarily the rate of
absorption and the quantity absorbed (hence the dose required to cause
toxicity) rather than the nature of the toxic effect. The exceptions that
occur can generally be attributed to effects at the point of application or to
rapid metabolism in the organs first encountered after absorption, e.g. by
the liver after absorption through the gut.
With regard to route of exposure, coal dust on the skin does not
present much of a hazard, but inhaled (as mentioned above) it can result
in a severe and debilitating disease. However, the skin is not a universal
barrier and many compounds, particularly liquids and gases, will pene-
trate quite readily. It must also be noted that the surface presented to a
compound is particularly vulnerable to local effects as a result of high
concentrations over a potentially small area, such as are likely to occur in
mouse skin carcinogenicity studies (Wilson and Holland, 1982). One
must therefore consider toxicity to the surface as well as through the
surface. The local effects of an irritant or caustic material on the skin or
gut might be quite dramatic but the systemic effects might only be
marginal. Alternatively, compounds such as DDT are likely to have
relatively insignificant local effects upon the skin but can result in very
severe systemic toxicity and death if administered topically.
For completeness, it should also be remembered that toxicity to the
skin or lungs does not necessarily depend upon the chemical being
administered by that route-thallium does not have to be given topically
to cause baldness nor does paraquat have to be inhaled to result in lung
lesions.
The route of administration governs which organ is first encountered by
the compound. That organ will generally meet the highest concentrations
of the material and might therefore be the most seriously affected. It will
potentially metabolise and perhaps detoxify the material, hence reducing
effects on other organs. If propranolol is given intraperitoneally,
it is absorbed through the hepatic-portal circulation and rapidly metab-
olised by the liver so that its toxic (or pharmacological) effects are greatly
diminished. A different route of administration for such compounds (e.g.
intravenous injection) will allow distribution and hence toxicity to other
organs and tissues before metabolism can take place.
An equivalent effect occurs with the insecticide, dichlorvos, which is
very rapidly hydrolysed and detoxified in the stomach. Its oral toxicity is
therefore low but when the vapour is inhaled, hydrolysis and metabolism
are sufficiently delayed for toxicity to be quite significant (though not at
the concentrations likely to be encountered by man).
The tissue or surface first encountered has its influence in other ways
too. The high vasculature of some organs assists the degree and extent of
absorption. Hence absorption via the lungs is extremely rapid and can be
similar in rate and nature to that of intravenous injection. The poorer
8 Effects of Physical Form, Route, and Species
vascularity of the skin is a factor in reducing the extent of absorption and
its speed.
The very rapid achievement of high blood concentrations has substan-
tial effects upon the pharmacokinetics, with the distinct possibility that
peak concentrations for a short time will enable a material to overload
normal metabolic pathways and produce different metabolites or reach
different organs. In recent years there has been a sensible and welcome
increase in the numbers of animal toxicity studies that include measure-
ment of blood (and tissue) levels of compound during their course.
The effects of route of exposure are sources of opportunity for those
concerned with targeting potent drugs to specific organs while minimising
less desirable effects (Prescott and Nimmo, 1979). To the toxicologist
they tend to cause confusion (or at best a challenge and a lesson) when a
material that is toxic by one route is safe by another.
B Oral
Toxicological investigations make extensive use of oral administration.
The material is often administered as a liquid (solution, suspension, or
undiluted) by gavage, i.e. a tube is passed down the oesophagus and the
material is injected into the stomach via the tube. Gavage is a particularly
simple procedure in rats and mice and the method used almost
exclusively in the conduct of oral LD50studies and other short term oral
studies in rodents. Vehicles used frequently for gavage experiments
include: water, vegetable oil (e.g. corn oil), or water and suspending agent
(e.g. carboxymethyl cellulose).
Gavage dosing of rodents over long periods (e.g. daily over two years)
is quite practicable (Nutrition Foundation, 1983). However, incorpora-
tion into the diet is less labour intensive when administration to rodents is
intended to be over weeks or months. The compound is carefully mixed
in the diet to a predetermined concentration and the animal is allowed to
eat in the normal way. For non-rodents, capsules and tablets are obvious
alternatives and are frequently used in studies with dogs and primates,
whether acute or longer term.
The rodent is normally fed ad libitum and eats over a period of several
hours, mainly at night. Therefore intake of a toxicant given in the diet will
be more gradual than by gavage. A rapidly metabolised or short acting
compound might thus have to be given in considerably higher doses than
by gavage to achieve the same effect. For a longer acting, less quickly
metabolised or excreted compound, there might be little difference in dose
required (assuming the compound is equally available to the animal by
both methods). Note that dogs are usually given a fixed amount of food per
day, all of which is likely to be eaten in a short period of time.
The units of dosage by the oral route are normally expressed
as: mg (or g) of compound per kg body weight of animal per day, e.g.
mgkg-lday-l, or parts by weight of compound per million parts by
A . B. Wilson 9
weight of diet, i.e. parts per million (p.p.m.).
When dosing by the diet, the same dietary concentration (p.p.m.) is
sometimes given over a period of months or years but rodents will eat
substantially less per unit body weight as they grow older, effectively
reducing the dose in terms of mgkg-' day? Feeding studies with
herbicides or pesticides are frequently conducted using a constant
concentration in the diet throughout the experiment. Alternatively, the
concentration of the diet can be adjusted regularly according to food
intake and body weight so that a reasonably constant dose (in terms of
mg kg-' day-') is achieved. Convention has it that doses of pharmaceuti-
cal products are adjusted in this way when given in the diet.
A formula for conversion between the two systems is:
Dose
Administered (kg diet eaten day-') x (mg cpd kg diet-')
(mg cpd kg body = Animal's Body Weight (kg)
weight-' day-')
C Dermal
The skin can act very successfully as a barrier but its status in that regard
varies according to its thickness, its vascularity, its degree of hydration
(hydration tends to increase its permeability), whether or not it is intact
(i.e. abraded, ulcerated or otherwise damaged), and, of course, the
compound in question. A fuller account of such factors can be read in
Klaassen et al. (1986).
The skin has a wide selection of glands and other components so
compounds can enter through sebaceous glands, sweat glands, or hair
follicles, as well as via the epidermis itself. The variability of these
components in different parts of the body and from species to species
presents difficulties in predicting accurately the rate of absorption and
quantity of material likely to be absorbed.
Irritation, inflammation, and ulceration are frequent and important
reactions of the skin to insult by foreign chemicals and these can be
investigated using laboratory animals. The rat and the rabbit are the
species most frequently used. An area of the back is shaved of its hair
(generally with clippers) and the compound is applied. The animal has to
be prevented from ingesting the material so a collar may be put on.
Alternatively, the site can be wrapped in an occlusive (impermeable)
dressing. The latter technique normally serves to increase the sensitivity of
the test. In a skin irritation test, the compound is usually applied only once
for a few hours before it is wiped off and the skin is examined. Any
reddening, oedema, ulceration, or other reactions of the skin are carefully
recorded over the next hours and days, to assess both delayed reactions
and the ability of the skin to recover. A typical scoring system is described
by Kay and Calandra (1962).
The most significant variables in this type of test are the concentration
of the compound, whether it is applied as a solid or a liquid, the vehicle
used, the duration of application, and whether or not the skin is occluded.
Within sensible limitations, the dose of compound is less relevant to
irritation than is its concentration.
Unfortunately in vitru studies are not yet at the stage in which they can
satisfactorily substitute for this type of work in animals.
The skin is sometimes used to investigate carcinogenicity-topical or
systemic. Conventionally it is the mouse that is employed, the hair of the
back is shaved off regularly and the material is painted on repeatedly
(e.g. daily or weekly) over a period of months (Toben and Kornhauser,
1982; Slaga and Nesnow, 1985). For some tests a single dose of initiator
(e.g. DMBA) is applied first. The skin and internal organs are examined
for tumours. This type of protocol is most often employed for testing of
compounds of similar chemical structure and making comparisons
bet ween them.
Broadly similar methods are used to investigate toxicity as a result of
absorption through the skin. The rat is the species of choice. While very
frequently it is the toxicity of a single dose that is investigated, sometimes
repeated applications are made over days, weeks, or months. Longer
term dermal experiments pose particular practical problems as they
require the animal to be clipped or shaved regularly (e.g. twice per week)
and occasional skin damage can result from this. If an occlusive dressing
is applied, its frequent application and removal can cause adverse effects.
Other methods of restraint, while appropriate and reasonable for a few
applications, might be difficult to justify over long periods. If no
occlusion or restraint is applied then ingestion through grooming is a
potential complicating factor.
It is sometimes relevant to abrade the skin for toxicity experiments, to
simulate the effect that cuts or grazes might have in man. Damage can
greatly increase permeability and therefore toxicity (Klaassen et al.,
1986).
The variables to be considered in skin toxicity are dose, concentration,
total area of application, vehicle, duration of exposure, number of
exposures, occlusion, and abrasions. Note, however, that a high con-
centration of a compound might prove to be irritant and this would be
likely to improve absorption and hence increase toxicity.
D Inhalation
General Principles of Atmosphere Generation
The subject of inhalation toxicity is exceedingly complex and has resulted
in a sizeable library of chapters and books, addressing themselves to the
relevant theory, practice, and experimental techniques (Willeke, 1980;
Phalen, 1984). This latter publication could be considered to be the
A . B . Wilson 11
standard reference on inhalation toxicology. While it is relatively simple

to prepare a compound for dosing orally or topically and to hope to
approximate the human situation, it is much more difficult by the
inhalation route.
The compound needs to be mixed homogeneously with air so that each
animal is exposed to the same concentration of material. Chamber
designs, cage designs (normally mesh), air flow patterns, and atmosphere
generation systems have all to be considered with this in mind. Even then
it is common to move animals from one position in an inhalation chamber
to another on successive exposures, in order to balance the effects of
position and minimise variations attributable to location.
There are further complications in that inhalation exposure is par-
ticulary relevant to the work place and all too often the hazard there is in
the form of a mixture of materials with different volatility or dispersion
characteristics. Experimental design becomes difficult and atmosphere
generation often requires unusual techniques, for instance, diesel engines
might need to be run to provide exhaust gases.
Ultimately, the atmosphere needs to be measured. On first considera-
tion, one might think that weighing the material used and measuring the
air flow would give the concentration for exposure. However, material
tends to adhere to pipework and the walls of inhalation chambers
considerably reducing the atmospheric concentration during exposure
and sometimes forming a source of toxicant after the exposure has
supposedly finished. For these reasons, analysis of atmospheres is
particularly important. This permits definition of three concentrations:
(a) nominal (i.e. what is desired or nominated),
(b) calculated (i.e. calculated from quantity of compound used, air flow,
etc.), and
(c) actual (i.e. what is present as determined by analysis of the
atmosphere).
Frequently what are described above as ‘nominal’ and ‘calculated’
concentrations are both termed ‘nominal’.
Generation of Gases, Vapours, and Aerosols

Unless a gas has peculiar properties, generation of the desired atmos-
phere can be quite straightforwardly achieved by feeding a supply of the
gas (from a gas cylinder) into the airstream and analysing the attained
concentration.
There are several methods of creating the desired atmosphere from
liquids. Most frequently the air is bubbled through the liquid, droplets
are filtered out, and the resultant concentrated mixture of air and
compound is diluted as required by mixing with more air. Although it is
possible to increase concentration by warming the vessel containing the
liquid, if the saturated vapour pressure is exceeded, the compound will
form droplets and some will condense on the pipework. It will not have
been the first inhalation experiment in the literature to report uninten-

tionally on the toxicity of vapours above their saturated vapour pressure.
Other methods of generation involve blowing air over the liquid which
is adsorbed onto charcoal or filter paper. This avoids droplet formation
and, if the air flow is adjusted correctly, creates an almost saturated
atmosphere of test compound, the concentration of which remains very
stable. This method requires more initial setting up than the former.
The term ‘aerosol’ embraces both droplets of liquid and particles of
solids. The generation of aerosol atmospheres needs to control both
concentration and particle size. It abounds with hardware, such as
atomisers, spinning discs, dust feed mechanisms, cyclones, and fluidised
beds. All too often the hardware has to be adapted to the particular
compound, making accurate comparison of work between laboratories
and test materials even more troublesome than usual.
Within the inhalation chamber different particle sizes follow different
air currents and so homogeneity is difficult to obtain. Deposition upon
surfaces is inevitable, so direct measurement rather than calculation of
concentration is essential. At its simplest, liquids are drawn into a solvent
for analysis, while solids are caught on filters, but again the behaviour of
particles in different air currents means that misleading and selective
sampling of different-sized particles can occur. Other methods are
numerous and include measuring optical density and exploitation of
aerodynamic differences between particles.
Particle size, too, should be measured and care taken that influence of
the sampling method on the particle size recorded is taken into account,
Some techniques depend upon visual microscopic measurement of
samples, others depend upon aerodynamic properties (e.g. cascade
impactors), still others are selective filters (e.g. Andersen), and so on.
Particle size is expressed as ‘mass median diameter’, i.e. the particle size
at which there is the greatest mass (not necessarily number) of particles,
so a few large particles count as much as many more small particles. It is
better, and more normal, to indicate the range of particle sizes and the
proportion by weight of particles in each size range.
Particle Size as a Variable

Particle size matters because different sizes of aerosol deposit in different
parts of the respiratory tract and can, therefore, result in different toxic
responses, sometimes through local effects and sometimes through
differential absorption.
Generally in man, particles above about 20-30pm in diameter are not
inhaled and those between 7 p m and 20pm deposit at various levels in
the upper respiratory tract to be absorbed or swept along by muco-ciliary
motion, coughed up, and swallowed.
Unless particles are below 5-7pm diameter, few or none will
A . B. Wilson 13
penetrate to the alveoli. Below that diameter an increasing proportion

reaches the alveoli but once the size is in the region of 0.1 pm and less, a
very high proportion is exhaled. The dimensions given refer to spherical
particles of unit density. Differently shaped particles will, broadly
speaking, behave similarly to a spherical particle with equivalent aerody-
namic characteristics. Fibres, such as asbestos, can have special charac-
teristics. Whilst the above comments on deposition of particles are true
for man, the respiratory tract and rates of flow of air in the tract are
different in laboratory animals. Thus, absorption and deposition of gases
and specified particles varies between species (Morris et al., 1986).
Inert solids, depositing above the alveoli will be removed by muco-
ciliary action, with little or no significant toxicity. If they reach the
alveoli, muco-ciliary clearance does not occur and it is left to the cells of
the reticulo-endothelial system to remove the particles. This is much less
effective and brings the compound into intimate contact with cells
resulting in many forms of toxicity, attributable to the physical rather
than the chemical nature of the compound. A selection is:
Asbestos-Asbestosis, lung cancer, mesothelioma
Coal Dust-Pulmonary fibrosis
Kaolin-Fibrosis
Talc-Fibrosis , pleural sclerosis
Research on asbestos is apparently showing that fibre length and
diameter have profound effects upon its toxicology (Bernstein, 1982).
Droplets of soluble or chemically active materials may be toxic by
inhalation even if their size is greater than 5-7pm. They can have local
effects and penetrate the nasal, tracheal, or bronchial mucous lining to
cause systemic toxicity. The term ‘respirable’ is commonly used to
distinguish particles that will penetrate to the alveoli, i. e. particles
<5-7pm diameter are classed as respirable to man. This division is
highly relevant in assessing toxic potential of inert solids, but is of less
relevance for other materials, though curiously, some regulatory bodies
have required no inhalation toxicology if aerosol diameter is >7 pm,
whatever the nature of the compound.
Method of Exposure
There are two main methods of exposure, ‘whole body’ and ‘nose only’
(plus occasionally ‘head only’). The terms are reasonably self-explanatory.
For whole body exposure the animals are placed in mesh cages in an
inhalation chamber and a mixture of air and the material to be tested is
passed through the chamber, so not only do the animals breathe the
toxicant but the skin is exposed so dermal absorption, and ingestion by
grooming are also possible. If continuous 24 hours a day exposure is
required, then food and water must be provided and contamination of
these can occur.
‘Nose only’ exposure is based upon restraining the animal so that only
the nose protrudes into the test atmosphere. This reduces contamination
of the skin but is, of course, practicable for only a few hours during each
day. Although rats soon become accustomed to the restraint, the stress of
the first minutes of an exposure could distort results, e.g. the toxicity of
cholinesterase inhibitors is likely to be increased by muscular activity. For
larger animals (e.g. dogs) then either whole body exposure or specially
designed masks are used.
Most inhalation experiments are conducted in a ‘dynamic’ atmosphere,
i.e. one in which there is a continuous flow of atmosphere into (and out
of) the chamber, with little or no recirculation. It has the disadvantage of
requiring large quantities of material and, of course, this has to be
disposed of. Very occasionally a static atmosphere is used, in which case
there is no flow through the chamber. This conserves toxicant (e.g.
radiolabelled compound) but is only practicable for short exposures.
There can be difficulties in maintaining a homogeneous mixture in the
inhalation chamber.
Design of Inhalation Experiments

The variable most peculiar to inhalation toxicology is, therefore, the
duration of each exposure period. Convention is that single exposure of
4-6 hours (either ‘whole body’ or ‘nose only’) is the accepted duration for
investigating the acute toxicity and deriving the LC,, (lethal concentra-
tion to 50% of the animals, the inhalation equivalent of the LD,,). The
Society of Toxicology (1992) reviewed the design of acute inhalation
studies. Regular exposures of 7 to 8 hours a day are taken as being
equivalent to the working day for man and are often selected when there is
a need to simulate the period of an individual’s exposure in the work place.
For the same reason (and also for expediency) these studies are often run
for 5 days out of 7 each week. When the hazard is likely to involve
continuous exposure, as for an environmental contaminant then exposure
for close to 24 hours per day, 7 days each week is necessary.
The units used in inhalation are rather different from oral or dermal
studies :
no. of parts by volume of the compound

parts per million (p.p.m.) =
lo6 parts by volume of air
i.e. 10p.p.m. means 10 ml of the vapour in lo6 ml of air. It is a volume:

volume measurement appropriate to gases and vapours but not to
droplets or particles.
mg m-3 and mg 1-1 are weight:volume measurements suitable for

any materials
A . B. Wilson 15
i.e. 10mgm-3 means 10mg of the compound (vapour, liquid, or

solid) in 1 m3 of air.
The two sets of units can be converted using the following formula:
mgm-3 = Pep .m. x molecular weight

24.5
E Intravenous and Other Routes

The potential for modifying the activity of pharmaceutical products by
altering the route of administration means that many and special methods
need to be developed, most frequently with the object of imitating the
intended method of administration to man. Thus studies are conducted
by intravenous, intraperitoneal, intramuscular, and subcutaneous injec-
tion. Also by the use of implants and suppositories.
Intravenous techniques (bolus injection or continuous infusion) have
come increasingly to the fore with the development of biotechnology
products (Wilson, 1987). Many of these are potent but labile peptides
and proteins which will be rapidly metabolised, therefore requiring
special techniques for administration.
F Dose
The calculation of dosage for oral or parenteral routes of administration
is normally straightforward-merely a measurement of the weight of
compound administered, usually expressed in terms of g of compound per
kg of animal. The extent of absorption is not taken into consideration.
However, dosage calculation presents a very different problem in inhala-
tion studies. It should be remembered that a 100 g rat inhales about 0.07 1
per min which compares with 7.51 per min for a 70 kg man. This means
that it will inhale roughly 6 times as much volume (and therefore toxicant)
per unit body weight as a person. Dosage is difficult to estimate for
inhalation studies (Dahl, Schlesinger, Heck, Medinsky, and Lucier, 1991 ),
so the extent of exposure is stated instead by giving the atmospheric
concentration and the duration of exposure. With some degree of caution,
when these variations are not great, these can be combined in the equation:
concentration x time. It becomes less valid when the time (or concentra-
tion) figures differ significantly. Atherley (1985) discusses this in the
context of human industrial exposure situation.
The results of toxicology studies are frequently used in the calculation
of safety factors. The dose administered (or level of exposure) becomes
relevant but it is not necessarily true that the quantity administered per
unit body weight is a valid basis for extrapolation across species and to
man (Davidson, Parker, and Beliles, 1986), even if factors such as extent
of absorption, route of metabolism, and susceptibility are comparable.
4 Choice of Species
A General Principles
It is obvious that the preferred experimental animal in terms of imitating
man is man. However, even if that were practicable, man would have
many undesirable features as a model: expensive to maintain, needing
large premises, requiring large quantities of test compound, slow growing,
long gestation period, and somewhat variable in nature. The last
mentioned point can be illustrated by racial differences in metabolism
(Bailey, 1983), e.g. lactose metabolism and rate of acetylation.
This indicates that no single strain or species is likely to be ideal and that
an understanding is required for the mechanisms of toxicity so that a
prediction can be made of how different populations will react to
compounds. The routine procedure is for two species, one rodent and one
non-rodent, to be used for most of the pivotal subacute and chronic
studies.
It is assumed that if the two species react in the same way at
approximately the same dose level (and allowing for some skilled
judgement), then man is likely to respond in the same manner. A
reasonable safety factor is usually allowed. When species behave differ-
ently then much more has to be known about metabolism, kinetics,
mechanisms, etc., before there is any confidence in extrapolation to man.
Zbinden (1993) debates this point.
An excellent general reference for information on laboratory animals is
Poole (1987). Relevant journals include Laboratory Animal Science
(Laboratory Animals, UK Ltd .) and Laboratory Animals (American
Association for Laboratory Animal Science, Cordona) .
B Rat
The rat is probably the most commonly used species for regulatory
toxicology studies. Its reactions to chemicals are therefore more likely to
be sensibly interpreted than those of most other species. It is a uniform,
easily bred animal, weighing only a few hundred grams, requiring little
space and an undemanding diet. Its size limits the amount of blood that
can be taken which is a distinct disadvantage and is reflected in the design
of studies, even although modern analytical techniques can be very small
quantities indeed. Toxicokinetic studies on rats generally require animals
to be killed at each time point so that sufficient blood can be obtained for
analysis.
The rat is used extensively in most routine studies-acute and chronic,
oral and parenteral. Its short life span (24-30 months) means that it can be
used for lifespan studies and for investigation of carcinogenicity. How-
A . B. Wilson 17
ever, some strains have high incidences of specific lesions which can
confound interpretation (e.g. mammary tumours in Sprague-Dawley rats
and testicular tumours in the Fischer 344 strain). The nature of the diet
substantially affects longevity and can alter tumour incidence. European
laboratories carrying out lifetime studies (such as carcinogenicity studies)
use specificallyformulated maintenance diets with low levels of protein (e.g.
14%) and a low calorific value.
There is a wide variety of strains. Albino rats of the Sprague-Dawley
(such as the Charles River CD), Fischer 344 (F344),or Wistar strains are
those most frequently encountered but there are many others, sometimes
with specialised applications.
C Mouse
The small size of the mouse (20-30 g) is a major disadvantage and more
than outweighs the virtues of cheapness, ease of housing, and ease of
handling. Although useful quantities of blood can be taken from mice
without causing death, it is probable that the animal’s physiology will be
sufficiently disturbed to be uncertain of the reliability of subsequent
observations. Therefore, the mouse is normally killed if blood samples are
needed.
It has a slightly shorter lifespan than the rat, 21-27 months. The
combination of strengths and weaknesses means that the mouse is used for
acute, LD,, type studies and as a second species in carcinogenicity
investigations. While intermediate length studies are conducted, in most
cases their prime aim is for dose selection for the long term carcinogenicity
work. Its value in carcinogenicity studies has been in doubt since the work
of Thorpe and Walker (1973) indicated its propensity to develop an
increased incidence of liver tumours when dosed with test materials for
which there was little or no other evidence of carcinogenicity. Debate and
discussions continue but a direction was perhaps established at the
November 1991 International Conference of Harmonisation in Brussels
during which presenters indicated that the mouse should not be used
routinely for carcinogenicity studies and that a single species (the rat)
would generally suffice.
The mouse is the species of choice (by custom) for skin carcinogenicity
investigations. Its position may be owed to the use of the mouse in the
1960s as a screen to detect carcinogenic hydrocarbons associated with
certain oils. The suspect material is painted onto the skin regularly for
many months, sometimes even the life time of the mouse, and dermal
tumours are counted. A variation is to administer an ‘initiator’ before
giving the test material and to expect tumours to develop much sooner.
This relates to the ‘initiation’ and ‘promotion’ phases of carcino-
genicity-a concept which was useful in its time but is now probably of
less value. As with rats, a variety of strains are used but the Charles River
CD-1 is probably the most widely used for toxicology. Inbred and
crossbred strains are also employed, e.g. B6C3F 1 for carcinogenicity
studies in the USA National Toxicology Programme (Rao, Birnbaum, Collins,
Tennant, and Skow, 1988). An introduction to the relative merits and
variability of inbred and random bred strains is given by Rice and O’Brien
(1980).
D Guinea Pig
The most frequent use of this species in toxicology is for the prediction of
Delayed Hypersensitivity reactions, Type IV allergic reactions , by such
test methods as described by Magnusson and Kligman (1969) and
Buechler (1965). The mouse is also beginning to be used for this work
(Botham, Basketter, Maurer, Mueller, Potokar, and Bontinck, 1991) and
is gaining acceptability with regulators. It remains to be seen whether it
will replace the guinea pig entirely.
The guinea pig’s lifespan is too long for it to be employed in
carcinogenicity work and its gestation period also long (about 9 weeks)
which means it is hardly ever used in reproduction investigations.
E Rabbit
Albino rabbits (usually New Zealand White) are used to investigate the
irritant potential of compounds to the skin and to the eyes. The large,
exposed, unpigmented eyeball and low blink rate mean that changes can
be readily detected and an adequate safety factor incorporated in the
experimental study .
New Zealand White or other strains of rabbits are frequently used in
teratogenicity investigations. Rabbits show clearly the teratogenic effect of
thalidomide, modelling its effect in man. It is, however, very difficult to
display thalidomide teratogenicity in the rat. It might be that the fewer
layers in the rabbit placenta have led experimenters to believe it will allow
toxicants to pass more readily to the foetus. One way or another, it is firmly
established as the second, (non-rodent) species for teratogenicity work.
F Dog
Virtually all dogs used in toxicology studies are purpose bred Beagles. It is
alleged this breed was chosen because sizeable groups were available, they
accepted confinement very readily, they rarely fought, were easy to handle,
they were short haired, and about the right size. They can be readily
trained to wear face masks for inhalation studies or to stand in slings
without apparent distress. Even if starting afresh, it would be difficult to
think of a more appropriate breed and certainly the extensive background
knowledge now available means it is unlikely to be replaced.
A . B. Wilson 19
The Beagle fills the role of a second species as a non-rodent in toxicity
studies lasting from 1 month to 2 years. It is very rarely employed in acute
work (except for selection of doses for longer term studies) and since its
lifespan is 11-1 5 years it is impracticable to use it in crcinogenicity work.
The latter has been done though more often than not the studies have
turned out to be unsatisfactory.
Repeated blood samples, several ml in volume, can easily be taken while
causing minimal physiological disturbance or stress. This very valuable
asset means the dog is used frequently in metabolism and toxicokinetic
work. The dog does vomit particularly readily and therefore presents
practical difficulties of certain compounds particularly by the oral route.
G Primates
Whilst virtually all other species used in toxicology are bred specially for
the purpose, it is regrettable that the primate was so frequently caught in
the wild and imported for the purpose. This situation is changing rapidly.
Breeding colonies of Macaques have been established in China, the
Philippines, Mauritius, and USA. Supplies of purpose bred animals are
now becoming available, so that studies can be conducted using clean
animals of known parentage and history.
Marmosets have been used but are rather smaller than is ideal, being
only a few hundred grams in weight. This limits the volume of blood
available and considerably restricts their use in kinetic investigations.
Supplies available to toxicology were very restricted during the 1980s so
only limited background data are available.
The Cynomolgus monkey is the species that has been used frequently.
The role of the primate in toxicology is similar to that of the dog (i.e. as a
second species for subacute and chronic work) though, in addition, they
are occasionally used in teratology and some other reproductive studies. It
is curious that primates are almost never chosen for toxicology studies on
herbicides, pesticides, or industrial chemicals, but are chosen on a very
substantial number of occasions as the non-rodent species in subacute or
chronic studies on pharmaceutical products.
Primates can be difficult to handle, used to be of unknown parentages or
history, and were often diseased, all of which tended to militate against
their use. Furthermore, the diseases include very significant zoonoses,
tuberculosis, salmonellosis, virus B, and rabies, which complicates
matters even more. The primate is a higher species and there is concern
that its use in toxicology is carefully justified on each occasion. It is
assumed, and possibly correctly, that primates are more similar to man in
their physiology and metabolism than are other species. However, if
studies on the rat and dog display substantially different toxicology in
each, it would be a brave man to assume that the results in primates
would provide the definitive answer-at least not without knowing much
more about the mechanisms of action of the compound.
H Other Species
While species other than the above are occasionally used, it is usually for
a specific reason related to the actions or metabolism of a particular
compound and because some laboratories are a little more adventurous
in their choice. One can find the ferret used frequently in the occasional
laboratory as an alternative to dogs and primates (it is a non-rodent),
Syrian hamsters used for carcinogenicity work instead of mice, Chinese
hamsters employed in cytogenetic studies (they have only 11 pairs of
chromosomes), and hens used to investigate delayed neurotoxicity
associated with organophosphorus compounds. The mini-pig and micro-
pig have been investigated over many years and are being used more
frequently (but not commonly) as non-rodent species.
Background data are very valuable in aiding the interpretation of
ambiguous and unusual results so there is a built-in inertia which
discourages the use of more unusual species.
Last, but by no means least, it is vital to be aware of the significant
advances being made in the use of bacteria, yeasts, cell cultures, and
other methods by which the use of animals can be appreciably reduced.
They are widely exploited in the prediction of genotoxic effects and much
research is on-going to develop suitable techniques for skin and eye
irritation and for teratology (Balls, Riddell, and Worden, 1983). Each test
tends to work satisfactorily for a particular group of materials and may be
sensitive to a specific mode of action (Green, Chambers, Gupta, Hill,
Hurley, Lambert, Lee, Lee, Liu, Lawther, Roberts, Seabaugh, Springer,
and Wilcox, 1993). They currently show most potential for classifying
closely related chemicals and for indicating potential activities.
5 Conclusions
Thus, it can be appreciated that seemingly very different toxicological
results can be obtained for what is apparently the same material,
depending upon its nature when presented to the animal, the route by
which it is presented, and the species used.
Haseman (1983) summarises a selection of the results from long term
studies conducted for the USA National Cancer Institute and produces
data on the variations between laboratories and between animal sup-
pliers. This illustrates how much there is to be done to reduce variability
in animal toxicology studies. Biologists tend to appreciate and accept
this, perhaps too apathetically. The chemist and the statistician find it
puzzling and frustrating, as illustrated by Horowitz (1984). Toxicological
studies will be extrapolated to man in whom there will be yet another
source of variability (Bailey, 1983).
References
Atherley, G. (1985). A critical review of time-weighted average as an index of
exposure and dose, and of its key elements. Am. Ind. Hyg. Assoc. J . , 46,
481-487.
A . B. Wilson 21
Bailey, K. (1983). Physiological factors affecting drug toxicity. Regulat. Toxicol.

Pharmacol, 3, 389-398.
Balls, M. J., Riddell, R. J . , and Worden, A. N. (1983). (Eds). ‘Animals and
Alternatives in Toxicity Testing’. Academic Press, London.
Bernstein, D. M., Drew, R. T., and Kuschner, M. (1982). The pulmonary
response to sized glass fibres in rats. Toxicologist, 2 , 59.
Botham, P. A., Basketter, D. A., Maurer, T., Mueller, D., Potokar, M., and
Bontinck, W. J. (1991). Skin sensitisation-a critical review of predictive test
methods in animals and man. Food Chem. Toxicol., 29, 275-286.
Buehler, E. V. (1965). Delayed contact sensitivity in the guinea pig. Arch.
Dermatol, 91, 171-175.
Clayton, G. D. and Clayton, F. E. (1990). ‘Patty’s Industrial Hygiene and
Toxicology’, 4th Edn., Wiley (Interscience), New York.
Dahl, A. R., Schlesinger, R. B., Heck, H. d’A., Medinsky, M. A., and Lucier, G . W.
(1991). Comparative dosimetry of inhaled materials: differences among
animal species and extrapolation to man. Fundam. Appl. Toxicol., 16, 1-13.
Davidson, I. W. F., Parker, J. C., and Beliles, R. P. (1986). Biological basis for
extrapolation across mammalian species. Regulat. Toxicol. Phurmacol., 6 ,
211-237.
Gaines, T. B. and Linder, R. E. (1986). Acute toxicity of pesticides in adult and
weanling rats. Fundam. Appl. Toxicol., 7 , 299-308.
Gartner, K. (1990). A third component causing random variability beside
environment and genotype. Lab. h i m . , 24, 71-77.
Goodman, J. I., Ward, J. M., Popp, J. A., Klaunig, J. E., and Fox, T. R. (1991).
Symposium overview. Mouse liver carcinogenesis: mechanisms and
relevance. Fundam. Appl. Toxicol., 17, 651-665.
Green, S., Chambers, W. A., Gupta, K. C., Hill, R. N., Hurley, P. M.,
Lambert, L. A., Lee, C. C., Lee, J. K., Liu, P. T., Lowther, D. K.,
Roberts, C. D., Seabaugh, I. M., Springer, J. A., and Wilcox, N. L. (1993).
Criteria for in vitro alternatives for the eye irritation test. Food Chem.
Toxicol., 31, 81-85.
Haseman, J. K. (1983). Pattern of tumour incidence in two-year cancer bioassay
feeding studies in Fischer 344 rats. Fundam. Appl. Toxicol., 3 , 1-9.
Hayes, A. W. (1983). (Ed.). ‘Principles and Methods of Toxicology.’ Raven
Press, New York.
Horowitz, W. (1984). Decision making in analytical chemistry. Fundam. Appl.
Toxicol., 4, S309-S317.
Kay, J. H. and Calandra, J. C. (1962). Interpretation of eye irritation tests. J . Soc.
Cosmet. Chem., 13, 281-284.
Klaassen, C. D., Amour, M. D., and Doull, J. (1986). ‘Cassarett and Doull’s
Toxicology’. 3rd Edn. Macmillan, New York.
Kimber, L. and Botham, P. A. (1993). Results with OECD recommended positive
control sensitizers in the maximisation, Buehler and local lymph node assays.
Food Chem. Toxicol., 31, 63-67.
Lu, F. C. (1991). ‘Basic Toxicology’. Hemisphere Publishing Corporation, New
York.
Magnusson, B. and Kligman, A. M. (1969). The identification of contact
allergens by animal assays: the guinea pig maximisation test. J . Invest.
Dermatol., 52, 268-276.
Nutrition Foundation Report of the Ad HOC Working Group on Oil/Gavage in
Toxicology. The Nutrition Foundation, Washington (1983).
Morris, J. B . , Clay, R. J., and Cavanagh, D. G. (1986). Species differences in
upper respiratory tract deposition of acetone and ethanol vapours. Fundam.
Appl. Toxicol., 7 , 671-680.
Phalen, R. F. (1984). ‘Inhalation Studies, Foundations and Techniques.’ CRC
Press.
Poole, T. (1987). ‘The UFAW Handbook on the Care and Management of
Laboratory Animals’, 6th Edn. Longman Scientific and Technical, Harlow.
Prescott, L. T. and Nimmo, W. S. (1979). ‘Drug Absorption Proceedings of the
Edinburgh International Conference’. ADIS Press, New South Wales.
Prescott, L. T. and Nimmo, W. S. (1985). ‘Rate Control in Drug Therapy’.
Churchill Livingstone, Edinburgh.
Rao, G. N., Birnbaum, L. S., Collins, J. J., Tennant, R. W., and Skow, L. C.
(1988). Mouse strains for chemical carcinogenicity studies: overview of a
workshop. Fundam. Appl. Toxicol. , 10, 385-394.
Rice, M. C. and O’Brien, S. J. (1980). Genetic variance of laboratory outlined
Swiss mice. Nature (London), 283, 157-161.
Sax, I. N. (1979). ‘Dangerous Properties of Industrial Materials’, Van Nostrand
Reinhold Co., New York.
Slaga, T. J. and Nesnow, S. (1985). In: ‘Handbook of Carcinogen Testing’.
Noyes Publications, New Jersey.
Society of Toxicology, Technical Committee of the Inhalation Specialty Section.
(1992). Recommendations for the conduct of Acute Inhalation Limit Tests.
Fundam. Appl. Toxicol., 18, 321-327.
Tobin, P. S., Kornhausser, A., and Scheuplein, R. J. (1982). An evaluation of
skin painting studies as determinants of tumorigenesis potential following
skin contact with carcinogens. Regulat. Toxicol. Pharmacol. , 2, 33-37.
Thorpe, E. and Walker, A. I. T. (1973). The toxicology of HEOD, 11. The
comparative long term oral toxicity studies in mice with dieldrin, DDT,
phenobarbitone, beta-BHK, and alpha-BMC. Food Cosmet. Toxicol., 11,
433-442.
Vogel, W. H. (1993). The effect of stress on toxicological investigations. Hum. E x p .
Toxicol.. 12. 265-271.
Willeke, K. (1980). ‘Generation of Aerosols and Facilities for Exposure
Experiments.’ Ann Arbor Science, Ann Arbor.
Wilson, A. B. (1987). The toxicology of the end products from biotechnology
processes. Arch. Toxicol., Suppl. 11, 194-199.
Wilson, J. S. and Holland, L. M. (1982). The effect of application frequency on
epidermal carcinogenesis assays. Toxicology, 24, 45-53.
Wollink, F. (1989). Physiology and regulation of biological rhythms in laboratory
animals: an overview. Lab. h i m , 23, 107-125.
Zbinden, G. (1993). The concept of multispecies testing in Industrial Toxicology.
Regulat. Toxicol. Pharmacol., 17, 85-94.
CHAPTER 3
Znfluence of Animal Species,

Strain, Age, Hormonal, and
Nutritional Status
F. J. C. ROE
1 Introduction
The use of laboratory animals in toxicology and safety evaluation is based
on the assumption that they are models for man. For many potent toxins
this assumption is well founded in respect of qualitative findings. For
weak toxins, for the purposes of safety evaluation and for quantitative
prediction of toxic risk, the value of tests in laboratory animals is much
more questionable. This is because the animals used are intrinsically poor
models and because tests are undertaken under variable conditions which
influence the results. The present chapter addresses these problems.
There is, throughout evolution, a continuity in the mechanisms that
underlie life processes, and the extent of similarity in body structure
between different mammals is striking. Nevertheless, there exist huge
differences between species, particularly in the spectrum of foods which
they eat and their lifespan. For example, a three year old rat or mouse is
equivalent to a human centenarian and yet, within their relatively short
lifespans, those species manage to manifest many of the diseases to which
mankind is prone, including some degenerative conditions and a wide
array of cancers. Higher metabolic rate and faster turnover of cells have
been evoked to explain differences in longevity between species. How-
ever, these are unlikely to be complete explanations since cells derived
from laboratory rodents, for example, behave similarly in in vitro cell
cultures, and have similar growth requirements, to human cells.
2 Genetic Versus Environmental Influences

There has been much investigation of the ageing process but no
satisfactory analysis or explanation of the process yet exists. In man, it is
23
24 1n.uence of Animal Species
a matter of common observation that age-associated changes occur
earlier in some individuals than others. Moreover, longevity and early
and late ageing seem to be familial. Nevertheless, it is not clear whether
genetic or environmental factors are primarily responsible for these
differences. Ageing is not a single entity. It is, rather, a progressive
accumulation of structural faults and defects in bodily functions. Al-
though either genetic or environmental factors may determine the time of
onset of these faults and defects, it is more probable that genetic and
environmental factors interact as determinants of the time of onset of
age-associated diseases. Undoubtedly, genetic constitution often deter-
mines susceptibility to particular kinds of ageing defect. But environmen-
tal influences serve to reveal this susceptibility. Thus, al-antitrypsin
deficiency renders humans susceptible to emphysema, but if deficient
individuals totally avoided exposure to lung-damaging agents they may
never develop the disease.
In the past, premature ageing and a wide variety of animal diseases
have been attributed to defective genetic constitution but it is now clear
that these may have been false attributions. By changing the quality of
the laboratory environment, for example, many of these diseases can be
made to disappear. The first major step in this direction came with the
development of specified pathogen-free conditions of husbandry.
Diseases such as ectromelia in mice and debilitating chronic respiratory
disease, complicated by bronchiectasis, in rats markedly reduced in
incidence or disappeared. (Bronchiectasis was once regarded as a
strain-characteristic and as a ‘normal’ age-associated change by some
experimentalists). The second major development has been a growing
recognition that many spontaneous and age-associated diseases of labora-
tory animals may be due to gross over-feeding. This is, for instance, true
of the common chronic nephropathy of rats and of many types of tumour
of both rats and mice. Unfortunately, this growing recognition has not
yet led to any major changes in diet formulation or feeding schedules in
experimental laboratories. But it is certain that such changes must be
made. The distinction between genetic and environmental causes of
diseases in laboratory animals is difficult and unreliable. The influence of
environment on the risk of developing various diseases continues to be
underestimated.
3 Versus Outbred Strains

Sequential inbreeding serves to reveal genetic faults by progressively
increasing the risk that offspring inherit defective genes from each parent.
Similarly, first generation hybrids of two pure inbred strains are often
more vigorous and longer-lived than their parents because hybridisation
reduces the chances of inheriting pairs of defective genes.
The experimentalist seeking a model has to choose between two needs.
On the one hand, he will desire to improve the experimental design by
F. J . C. Roe 25
using animals which are genetically identical, since the probability that
effects are due to test agents is improved if ‘like’ is being compared with
‘like’, except for exposure to the test agent itself. On the other hand, the
experimentalist knows that man is not inbred and that if he chooses a
particular inbred strain it may, because of its peculiar genetic make up,
give quite misleading information concerning the effects of a particular
test agent.
For the purposes of evaluating new chemicals for safety for man, it has
been argued that it is better to use outbred strains in the hope that at
least a few of the animals included in the study will respond in a way that
is predictive for man. This is a fallacious argument since there is no way
of predicting which observation is pertinent to man. Since different
experiments have different objectives, there is no single answer to the
problem.
4 Choice of Species
Absolute qualitative differences in the way species respond to chemicals
are uncommon but quantitatively very big differences are well docu-
mented. For the most part such differences are due to differences in
absorption and metabolism, dependent on the presence of particular
enzymes and on other conditions which pertain in the gut, the liver, the
kidneys, and other organs. Where there is adequate information from
studies in several species, which is rarely the case, it is common to find
that different species share a spectrum of metabolic pathways but that the
predominant pathway in one species is different from that in another.
The reasons for such differences are not necessarily genetic. Lead acetate
can be fed to rats throughout their lives at dose rates well above ones that
would be lethal for man (Van Esch, Van Genderen, and Vink, 1962).
This is because the proportion of lead absorbed from the gut is very low
in rats fed a standard chow. Changing to a high milk diet multiplies the
extent of lead absorption 40-fold and renders lead almost as toxic for the
rat as for man (Kostial and Kello, 1979).
Ideally, for safety evaluation purposes, one should choose the species
that most closely mimics man in the way that it handles a test material
metabolically. In practice, this ideal may be quite unattainable, Firstly,
metabolism studies can be difficult, time-consuming, and costly, even if
interest is confined to a short list of regularly used laboratory species.
Secondly, the studies undertaken may show that man is seemingly unique
and that there is nu good laboratory animal model. Thirdly, it may simply
not be known whether other strains behave similarly or how much the
available observations are dependent on factors peculiar to the set of
conditions of the experiment. Regrettably, much of the literature on
comparative metabolism is inadequate and, in addition, the information
available for man is usually scanty and not comparable with that for the
26 Influence of Animal Species
laboratory species studied. Thus, for ethical reasons the only human
studies performed are often for single, very small doses, whereas there
are data for animals which have been exposed to larger and repeated
dosage.
Furthermore, in tests for chronic toxicity and carcinogenicity, the
choice of species is limited if there is to be a reasonable chance of
detection of more than the strongest effects. To achieve sensitivity,
relatively large numbers of animals (e.g., groups of 50 males and 50
females) are required, particularly if effects in animals exposed to at least
2, and preferably 3, different dose levels are to be compared. This results
in huge experiments that cannot even be contemplated except for small
animals such as rats, mice, and hamsters. Since weak toxic and
carcinogenic effects may not become manifest until animals are well
through their available lifespan, it is usually necessary to choose
short-lived species. In practice this tends, again, to limit the choice for
chronic toxicity testing to rats, mice, and hamsters.
Finally, there is yet one further important consideration. Meaningful
interpretation of tests on laboratory animals depends on the existence of
adequate information on the array of ‘spontaneous’ lesions to which the
test species is prone. The pathologist trying to evaluate what he sees at
necropsy and later during the histological examination of tissues, may not
know whether particular appearances are a result of exposure to the test
material or manifestations of a spontaneous disease too uncommon to
have appeared in the small number of controls.
Thus, the choice of species for the ideal experiment can be very
difficult. Undoubtedly, man is the best model for man, but even here,
intraspecies variation in response plus interference from environmental
factors prevent really confident extrapolation from one man to another.
These sources of uncertainty are magnified by species difference but
somewhat reduced by the use of large numbers of animals which is
possible in the case of small rodents. Comparative metabolism studies are
of little practical value except to identify an animal species as
inappropriate as a model for man.
5 Numbers of Animals: Safety Evaluation as

Distinct from Mechanism Studies
Most toxicologists would regard it as more important and wor hwhile
(and certainly more interesting) to study mechanisms of toxicity han to
screen new chemicals for possible toxicity. Regulatory authorities on the
other hand (under pressure from an inadequately informed public)
require extensive safety testing which utilises most of the available
resources. To make matters even worse, it can be very difficult to
persuade a regulatory authority that a substance which is in some way
toxic for laboratory animals is safe for man, even if some studies of
mechanisms have been completed.
F. J. C. Roe 27
The experimental requirements for mechanism studies and safety

evaluation differ substantially. The former may require no more than a
few animals in which specific measurements are made using sensitive
methods. Arguably, studies of mechanisms are by far the most reliable
way to characterise underlying changes that result in toxicity and to
establish the basis of reliable prediction of human effects. The field is one
that is ripe for fruitful expansion. As an example, there is in progress
rapid expansion in our knowledge of regulatory peptides (Bloom, Polak,
and Lindenlaub, 1982) and the mechanisms involved in homeostasis. This
new knowledge is being turned to good effect by the pharmaceutical
industry but has yet to be assimilated by toxicologists. Most of the
observations routinely made by toxicologists (e.g., haematological, win-
alysis, the examination of haematoxylin and eosin-stained sections of
tissues, etc.) are no more than fairly crude and insensitive screens for
possible effects of xenobiotic agents.
In contrast, to establish the apparent safety of a substance for human
use it is usually necessary to study large numbers of animals because of
the insensitivity of the methods used, and yet the results are regarded as
more pertinent than knowledge of toxic mechanisms. In the case of
teratogenic and carcinogenic effects, for example, a positive response at
any dose level may lead to the test substance being proscribed for human
use. Thus the test is essentially designed to demonstrate the absence of
activity, i.e., to prove a negative eflect-which is an impossibility. In
practice, the negative results required are listed, along with details of the
methods to be used, as ‘Guidelines’ by various authorities (e.g., OECD
Guidelines, 1981).
One consequence of the development of new and more sensitive tests,
many of them applicable to living animals, will be to highlight the
problem of how to distinguish between an ‘effect’ and a ‘toxic effect’.
Animal Susceptibility
Rats and mice in the wild carry microbial and parasitic diseases and, like
humans, are prey to epidemics of infectious diseases. Before the
development of barrier conditions within which disease-free laboratory
animals could be given pathogen-free food, endemic and epidemic
disease rendered long-term experimentation frustratingly difficult. The
cleaning up of animal laboratories has not, however, completely solved
the problem of there being an unacceptable level of background disease,
although the spectrum of diseases that most commonly occur has
changed.
The inability to eradicate natural disease processes raises the question
as to how far such conditions render the experimental animal more (or
less) susceptible to the adverse effects of chemical treatments and how
this affects the process of extrapolation to man.
It is possible that a high incidence of a particular disease in untreated
animals is indicative of susceptibility to the induction of that disease by
exogenous agents. This is of particular relevance to specified types of
tumour. For example, it is seemingly easier to induce adenomatous
tumours of the lung in strains of mice that exhibit a high incidence of
such tumours ‘spontaneously’ than in low ‘spontaneous’ incidence strains.
At one time breeding programmes, aimed at developing high tumour
incidence strains of mice (e.g., the B6C3F1 hybrid) in the hope that such
strains would prove to be sensitive tools for detecting carcinogenicity,
were initiated. There is no good evidence that such efforts improve safety
prediction for man.
Ideally an animal model should exhibit the spectrum of morbid and
fatal diseases of man (e.g., high incidences of cardiovascular disease and
cancers of epithelial cell origin, etc.). At present no such animal model is
available. Many strains of the laboratory rat show very high incidences of
endocrine tumours (e.g., 100% incidence of mammary tumours in
females, 100% incidence of Leydig cell tumours in males, up to 80%
incidence of pituitary tumours in both sexes, high incidences of adrenal,
thyroid, and parathyroid tumours, etc.) (Roe, 1981); and many strains of
laboratory mouse, including the B6C3F1 hybrid referred to above, exhibit
high incidences of malignant lymphoma, liver, and lung tumours. In
addition, rats aged two years or more often develop severe progressive
nephropathy which affects both glomeruli and tubules to varying degrees,
and which is associated with increased low-molecular weight proteinuria.
Nor does the hamster offer a way of avoiding these deficiencies because
under laboratory conditions, these animals tend to develop very high
incidences of amyloid degeneration of the kidney and of septic atrial
thrombosis.
Overall, the position is unsatisfactory in that animals in control groups
commonly develop high incidences of diseases which are uncommon or
even rare in man and very few animals die spontaneously of any of the
diseases which most commonly affect humans. There are two serious
corollaries of this. First, if the aim is to reduce man’s burden of disease,
then the tests should aim at detecting environmental factors or agents
which induce or exacerbate the common diseases of man (e.g. arthritis,
heart disease, stroke). Secondly, if a high spontaneous incidence of a
disease is indicative of easy inducibility, the animal models we presently
use are of little value for this purpose and others should be sought.
Moreover, the high incidence of irrelevant spontaneous diseases render
animal models unsuitable for detecting subtle manifestations of toxicity
of certain kinds. It would not, for instance, be possible to detect a weak
chronic toxic effect on the kidney under experimental conditions in which
all the untreated control rats develop serious spontaneous renal disease.
The Effects of Overfeeding

When the spectra of neoplastic and non-neoplastic diseases that have
been encountered in untreated animals in conventional chronic
F. J . C. Roe 29
Table 1 A few of the many kinds of background

pathology encountered in a group of 60 untreated
male Sprague-Da wley rats that constituted the
controls in a carcinogenicity study of 2 years
duration
(Unpublished data from personal files)
%
Moderate to severe
chronic progressive nephropathy 67
Parathryoid hyperplasia 67
Calcification of aorta 34
Adrenal medullary
- hyperplasia/neoplasia 32
- neoplasia 20
Chronic fibrosing myocarditis 83
toxicity/carcinogenicity tests (Tables 1-3) are compared with the effects

of controlling dietary intakes that have been observed in different studies
(Tables 4-6), it is clear that much of the background disease which is
presently such a prominent feature of long-term rodent studies is
associated with overfeeding and is probably avoidable. At present, the
Table 2 Incidence of certain non -neoplastic and neoplastic

diseases in groups of 86 male and female untreated Sprague-
Dawley rats
(Reproduced by permission from Kociba et al., 1979,
Food Cosmet. Toxicol., 17, 205-221)
% of rats with: Males Females

Mineral deposition kidney - 29
Moderate or severe renal disease 65 7
Mineral deposition in gastric
mucosa and muscularis
secondary to kidney disease 29
Multiple foci of hepatocellular
alteration 17
Focal hepatocellular cytoplasmic
vacuolation 40
Periarteritis 23
Neoplasms of:
Pancreas - exocrine 33 -
Pancreas - endocrine 16 9
Pituitary - pars distalis 31 62
Adrenal - medulla 51 8
Thyroid - C-cell 8 8
Mammary gland - fibroadenoma 1 76
- adenocarcinoma 3 8
Any site 88 97
Table 3 Percentages of rats with certain endocrine tumours in the control

groups in 3 separate 2-year carcinogenicity studies on a prolactin -releasing
drug. (The arrows in brackets indicate the effect of the drug on tumour
incidence, if any *)
(From confidential data in personal files)
Study No. 1 2 3
Sprague-
Strains of rat Wistar Wistar Dawley
Sex 8 9 6 0 6 0
Pituitary 17 53 41 46
Benign mammary 2 11 S(.T> 77
Malignant mammary 0 0 2 0
Phaeochromocytoma 0 0 2 0
Adrenal cortex 0 0 0 3
Thymoma (endocrine type) 0 0 0 0
Thyroid - follicular 0 0 0 0
- C-cell 9 5 1 0
Pancreas - islet cell 4(?) 4(?) 3(T) 4 5
Parathyroid hyperplasia 25 0 0 0 0
* If only study No. 3 had been carried out, the drug would have run into fewer regulatory
problems than it did!
simplest way of avoiding overfeeding is to limit the time during which

food is freely available. Whether the same effect can be achieved by
reducing the nutritive value of the diet provided throughout the 24 hours
of the day, is not yet certain. Of special interest is the observation (Ross,
Lustbader, and Bras, 1982) that the amount and kind of food an animal
eats shortly after weaning has a seemingly indelible effect on its
subsequent body growth and the incidence of tumours. Clearly the effects
Table 4 Effect of restricting the access of Wistar

rats to a standard laboratory chow ( P R D formula)
from 24 hours per day to 64 hours per day on
incidence of chronic progressive nephropathy at 2
years
(Reproduced from Salmon et al., 1990)
Sex: Male Female

Hours of access to
food per day: 24 6; 24 6$
% of rats with
moderate or
severe nephropathy : 65 5 60 0
F. J. C. Roe 31
Table 5 Effect of 20% diet restriction by weight on tumour incidence

before 2 years in groups of 50 male and female ICI Wistar rats
(Reproduced by permission from Tucker, 1979, Int. J . Cancer, 23,
803-807)
Male Female
20% 20
A d lib. Restricted A d lib. Restricted
% survival to 2 years 72 88 68 88
% of rats which developed
tumours at any site 66 24*** 82 57**
% of rats which developed
tumours at more than one site 22 2** 26 10"
Total number of all sites of
tumours in group of 50 rats 47 14** * 59 37**
% of rats that developed
mammary tumours 0 0 34 6***
% of rats that developed
malignant mammary tumours 0 0 12 2
% of rats that developed pituitary
tumours 32 O*** 66 38**
* p <0.05
** p < 0.01
* * * p< 0.001
of diet composition, food-scheduling , and eating patterns on spontaneous

disease in laboratory animals requires more intensive investigation.
The data presented illustrate the extent to which environmental factors
influence the incidence of diseases which, in the past, many experimen-
talists considered to be primarily determined by genetic constitution.
Table 6 Effect of diet restriction on longevity and tumour incidence in

mice
(Reproduced by permission from Conybeare, 1980, Food Cosmet.
Toxicol., 18, 65-75)
Restricted Restricted
(75% of (75% of
A d lib. ad lib.) A d lib. ad lib.)
Number of mice 160 160 160 160
% survival to 83 weeks 58 66 62 77"
% Lung tumours 19 12" 15 5* *
% Liver tumours 29 8**+ 4 0.6"
% Any tumour at any site 44 23*** 31 11*
% Any malignant tumour at
any site 11 4 14 4**
* p <0.05
* * p <0.01
*** p <0.003
However, it is clear that the problem of high incidences of background

disease cannot be fully resolved by attention to diet. Although diet
restriction has been shown to have major beneficial effects on the
incidence of kidney disease, endocrine changes, and neoplasia of various
kinds, etc., there remains in the restricted-fed animals an unacceptably
high incidence of such lesions (particularly hyperplasia and neoplasia of
the anterior pituitary gland in rats, and malignant lymphoma in mice).
This suggests that the contributions of other aspects of the laboratory
environment to the causation of disease in untreated animals should be
explored. The suggestion that enforced celibacy may be detrimental
(Roe, 1981) has not been supported by a recent small-scale study (Salmon
et al., 1990). It is probable that lack of exercise or general boredom will be
found to be important. A study of the effects of providing an exercise
wheel on the development of ageing and other diseases in mice was
reported by Conybeare (1988). Behavioural scientists are probably best
placed to devise protocols for testing the effect of general boredom!
8 Future Development in Long-term Testing

At present, the laboratories that carry out tests and the regulatory
authorities who judge the results are understandably reluctant to use, or
recommend the use of, diet restriction in the design of chronic
toxicity/carcinogenicity tests. A fundamental change of this kind might
devalue the banks of accumulated data derived from earlier
conventionally-designed tests, and there is the fear that by reducing
background disease, the model will be made less sensitive to the
induction of those diseases relieved by dietary control. It is of critical
importance that these objections are examined by properly conducted
scientific experiment. It may then be possible to devise conditions which
maintain laboratory animals in good health, free of infection, obesity,
endocrine disorder, and many forms of neoplasia, and which are all
properly monitored and controlled. It will then only be a short step for
regulatory authorities to require that all tests for toxicity and car-
cinogenicity shall be undertaken in hormonally normal animals rather
than in animals that display a mixed medley of laboratory artefacts.
9 Summary
This chapter has addressed a number of inter-related aspects of basic
toxicology.
(a) There is no clear understanding of the ‘ageing’ process because of
widespread confusion between diseases that are simply more
common in old age (Le., intrinsically due to an ageing process) and
diseases which are caused by avoidable environmental exposures.
F. J . C. Roe 33
(b) Genetic constitution influences longevity where inbreeding results
in offspring receiving defective genes from both parents. Hybrid
vigour illustrates an escape from this handicap.
(c) Many so-called ‘strain characteristics’ are wholly or partly environ-
mentally determined.
(d) The advantages and disadvantages of using inbred as distinct from
outbred strains must be considered in relation to the type and
purpose of the experiment.
(e) The use of a species and/or strain of animals that handles the
chemical to be tested in the same way as man is usually
impracticable, either because the comparative metabolic data do
not exist, or because no species which it is feasible to use is
sufficiently akin to man.
(f) The principles underlying the investigation of mechanisms of
toxicity are quite different from those underlying the design of
safety evaluation tests. More studies of mechanisms are necessary
to improve interspecies extrapolation.
(8) The spectrum of diseases which afflict man is quite different from
the spectra which afflict rodent species. The relevance of rodents
as models for the more common diseases of man merits further
study.
(h) The high incidence of obesity and of many kinds of background
disease in the laboratory rodents used in chronic and car-
cinogenicity testing today and, in particular, the high incidence of
various endocrine disturbance and neoplasia of endocrine glands
are unacceptable. Over-feeding appears to be a major factor in the
causation of chronic progressive nephropathy , periarteritis, car-
diomyopathy, and various endocrine tumours in rats and of
malignant lymphoma, liver, and lung tumours in mice.
(i) Further research into ways of maintaining laboratory animals in
normal endocrine status throughout their natural lives is urgently
needed.
10 1993 Update
Since the above chapter was written for the first edition of this book two
major developments have occurred in relation to the role of non-genotoxic
mechanisms in carcinogenesis.
In relation to the first of these developments, the reader is referred to
two books: Weindruch and Walford (1988) and Fishbein (1991). In the
latter, there is a report of the results of an experiment involving 1200 rats
exposed to various dietary regimes (Roe, 1991) and some of the findings in
this experiment have been summarised in Roe (1993). Particularly
noteworthy are the highly significant correlations between body weight at
the age of 29 weeks and risk of premature death and/or malignant tumour
development before the age of 133 weeks.
The second development concerns the growing recognition of the role of

endogenous mutagens and the importance of increased cell turnover rates
as determinants of premature ageing and increased cancer risk. This
development is also discussed briefly in Roe (1993).
References
Bloom, S. R., Polak, J. M., and Lindenlaub, E. (1982). ‘Systemic Role of
Regulatory Peptides.’ pp. 1-547. Schattauer Verlag, Stuttgart/New York.
Conybeare, G. (1980). Effect of quality and quantity of diet on survival and
tumour incidence in outbred Swiss mice. Food Cosmet. Toxicol., 18, 65-75.
Conybeare, G. (1988). Modulating factors: challenges to experimental design. I n :
‘Carcinogenicity : the design, analysis and interpretation of long-term animal
studies’. Eds H. C. Grice and J. L. Ciminera. ILSI Monograph. Springer-
Verlag, New York, pp. 149-172.
Fischbein, L. (1991). ‘Biological effects of dietary restriction’. Springer-Verlag,
Berlin, 349pp.
Kociba, R. J., Keyes, D. G., Lisowe, R. W., Kalnins, R. P., Dittenber, D. D.,
Wade, C. E., Gorzinski, S. J., Mahle, N. H., and Schwetz, B. A. (1979).
Results of a two-year chronic toxicity and oncogenic study of rats ingesting
diets containing 2,4,5-trichlorophenoxyaceticacid (2,4,5-T). Food Cosmet.
Toxicol., 17, 205-221.
Kostial, K. and Kello, D. (1979). Bioavailability of lead in rats fed ‘human’ diets.
Bull. Enuiron. Contam. Toxicol., 21, 312-315.
OECD Guidelines for Testing of Chemicals (1981).
Roe, F. J. C. (1981). Are nutritionists worried about the epidemic of tumours in
laboratory animals? Proc. Nutr. SOC., 40, 57-65.
Roe, F. J. C. (1991). 1200-Rat Biosure Study: Design and overview of results.
Chapter 26 in: ‘Biological effects of dietary restriction’. Ed. L. Fishbein.
Springer-Verlag, Berlin, pp. 287-304.
Roe F. J. C. (1993). Recent advances in toxicology relevant to carcinogenesis:
seven cameos. Food Chem. Toxicol., 31, in press.
Ross, M. H., Lustbader, E. D., and Bras, G. (1982). Dietary practices in early
life and spontaneous tumours of the rat. Nutr. Cancer, 3, 150-167.
Salmon, G. K., Leslie, G., Roe, F. J. C., and Lee, P. N. (1990). Influence of food
intake and sexual segregation on longevity, organ weights and neoplastic
diseases in rats. Food Chern. Toxicol., 28, 39-48.
Tucker, M. J. (1979). The effect of long-term food restriction on tumours in
rodents. Int. J. Cancer, 23, 803-807.
Van Esch, G. J., Van Genderen, H., and Vink, H. H. (1962). The induction of
renal tumours by feeding of basic lead acetate to rats, Br. J . Cancer, 16,
289-297.
Weindruch, R. and Walford, R. L. (1988). Retardation of ageing and disease by
dietary restriction. C . C. Thomas, Springfield, Illinois, 436pp.
CHAPTER 4
Experimental Design
A. B. WILSON
1 Regulatory and Experimental Toxicology

Regulatory toxicology includes a wide range of procedures (Tables 1 and
2), conducted with the main objective of screening compounds and
meeting the routine demands of regulatory bodies in a variety of countries.
Chapter 22 will highlight the nature of these requirements for food but
they result in a spectrum of studies of fairly standard design. In practice,
most government authorities provide considerable scope for variations to
provide for the inevitable differences between compounds, their uses and
effects. This chapter will outline some of the standard designs used and
discuss a selection of design problems associated with these
investigations.
Investigational toxicology arises from a desire to study toxic effects
more deeply and specifically than routine studies permit, Whilst regula-
tory toxicology is mainly a screen and a first stage, investigative work often
starts from there. The screen identifies an effect which is of such a nature or
exposure level that further investigations are warranted. Sometimes these
will be adequately covered by other parts of the screen, e.g. use of a second
species, studies on the metabolism of the compound in man and animals,
and pharmacokinetics. On other occasions, the designs of screening
studies will be adjusted to provide additional information on particular
effects, e.g. electron microscopy of the liver after hepatoxicity or
behavioural tests after observing CNS signs.
The potential range of disciplines involved in investigative toxicology is
extremely wide, and can involve tissue culture and tissue homogenates,
electron microscopy and histochemistry, autoradiography and pharmaco-
kinetics, genetics and molecular biology, epidemiology and evaluation of
information from accidental exposures, agonists and antagonists, and
many other approaches. One of the skills is to select which of these are
really likely to provide useful information, but further investigations
generally depend upon the development of a pertinent model of the lesion
or effect that is to be explored.
35
36 Experimental Design
In conducting these studies one can only recommend sound scientific
practice as required by Good Laboratory Practice (GLP) Regulations:
evidence of thought, reflection, justification, and planning of a study;
adequate resources (people, equipment, premises) both in quantity and
in quality; and an ability to reconstruct the experiment and the process
leading from the original data to the conclusions.
Chapter 24 will explore how these are achieved in the context of
toxicology but the fundamentals, as stated above, are relevant to most
branches of science and certainly to both regulatory and investigational
toxicology. The intelligent application of GLP has transformed toxicol-
ogy into a branch of science with an exceptionally high standard of
experimental conduct.
2 Object of Study
A Reflection
Clark (1977) in his chapter on inhalation toxicology describes reflection
as the most necessary but most neglected first step in designing a study.
This corresponds with the emphasis GLP places upon planning and
justification for actions. It has been said that organisation is the enemy of
research and, while it is certainly true that organisation blunts spon-
taneity, there are a great many spontaneous actions which are later
regretted. Certainly, in regulatory toxicology, an orderly and disciplined
approach is of more value than bursts of occasional brilliance.
The first step, before becoming more involved with experimental
design, is to assess the knowledge already available on the compound in
question, i.e. the problem being addressed and the techniques likely to be
pertinent. The wide range of computer-based literature searching facilities
available (Wexler, 1987) provides a good basis from which to start so
that, on a novel compound, structure-activity relationships, properties of
similar compounds, and strengths and weaknesses of the test methods can
receive a preliminary assessment.
B Definition of Object
The next step, or perhaps in parallel with the first, is to define the object
of the investigation. This is not as easy as it first sounds and the omission,
ambiguity, or superficiality of many of the ‘Object’ paragraphs of
regulatory toxicology protocols and reports too often reflect inadequate
considerations of the matter and can result in a poor or wasted
experiment. A few examples might be relevant. It is common practice to
conduct animal studies involving dosing for 90 days, embracing a range of
clinical, haematological, clinical chemical, and pathological examin-
ations, the object being described as ‘to investigate the effects of dosing
compound X for 90 days). This may be only half the story. The object may
A . B. Wilson 37
also be to satisfy a perceived regulatory requirement, which would explain

a few of the observations required which may otherwise be viewed with
curiosity.
Sometimes special investigations (e.g. for effects upon the kidney) are
coupled with such studies. This is when complications arise, since there
would then be 3 probable objects:
(a) to investigate the toxicity of compound X over 90 days,
(b) to satisfy regulatory bodies, and
(c) to gather further data on the renal toxicity of compound X.
The last mentioned object is very loosely phrased but is often what
occurs and might indeed be an adequate description. On many occasions,
certain specific techniques are considered which might compromise other
declared objects. The combining of several objects can result in a
potentially unhappy compromise through which none is particularly well
satisfied.
The type of conflict that occurs in the above example can arise from
interference with an animal for certain observations, and hence cause a
change. Anaesthesia, restraint, withdrawal of food, and frequent blood
sampling (Cardy and Warner, 1979), can confound routine observations
fundamental to the other objects. Tests for behavioural changes (Alder
and Zbinden, 1983) are particularly difficult to incorporate for these
reasons. The more objects there are, the more frequently such compro-
mises will occur.
Many toxicology experiments are organisationally somewhat compli-
cated, involving the co-operation and integration of a wide variety of
disciplines over months and even years, resulting in tens and hundreds of
thousands of items of data recording. Any attempt to further complicate
the issue is to court disaster since the organisational system might not
cope.
It is essential to keep study designs sufficiently simple and robust for
successful accomplishment of the original objects to be assured-
scientists are not always the best managers or organisers and vice versa.
There are, thus, two reasons for keeping study objects simple:
(1) to avoid unsatisfactory compromises in scientific approach, and
(2) to allow the organisation to cope.
It is, therefore, far better to divide a complicated study into different
portions, each simple, thorough, and relevant, than to try to create a
complicated experiment. As an experienced researcher remarked,
“There is always time to repeat an experiment but never time to do it
once properly. ”
For example, in the field of reproductive toxicology there are two
conventional approaches. One, favoured by regulatory bodies involved
with agricultural chemicals, is to breed two or three successive gener-
ations of rodents, looking for effects on fertility, foetal development,
post-natal development, and sometimes incorporating 90 day toxicology.
Reviews of such studies show, not surprisingly, that very few have been
pivotal to the assessment of a compound (Food and Drug Administra-

tion, 1970). The alternative approach, generally used for testing of
pharmaceutical products, is to treat each portion separately-a fertility, a
teratology, and a peri-post-natal study. This latter approach uses no more
animals, takes less time (studies can be overlapped), results from one
study can be used to modify the design of the succeeding study, and if
there are practical problems with one component it does not mean
repetition of the full sequence of studies to rectify matters.
C Confirmation of Design
Having attained a provisional object and study design, it is as well to
check what action will be taken if various combinations of results occur.
Thus, special investigations of renal function within a 90 day screening
study may result in the following: tests negative, perhaps not sufficiently
sensitive (design compromise) so a specifically designed repeat study is
required, or tests positive, therefore specifically designed repeat studies
are justified, i.e. the same endpoint may be reached whatever the result,
making one question the need for the original experiment.
Three questions can be asked when the results of a study appear.
(a) Do I believe this effect would recur if I repeated the experiment in
the same way, i.e. is it reproducible?
(b) Do I believe the effect was attributable to the compound?
(c) Do I believe the effect is biologically significant?
It is sensible to look at the experimental design in anticipation of these
questions and to try out various combinations of results to check what
conclusions or further actions will follow. In practice most major studies
follow well established patterns. Study design difficulties occur more
frequently with shorter studies, preliminary studies, and special investiga-
tions. Though smaller in terms of numbers of animals used and cheaper in
terms of man hours, they are more demanding in terms of forethought.
D Historical Data
Data from previous studies are frequently and rightly used to aid in the
interpretation of study results. There needs to be a degree of caution
since work carried out, even under apparently similar circumstances, can
produce different results. However, they can indicate whether (for
example) performance of the control group on a study was in any way
exceptional-perhaps having a lower incidence of a particular tumour
and giving the spurious impression of an increased incidence caused by
the compound in question.
Background data form the basis of ‘normal ranges’ and these are
usually taken to mean the range of values generally encountered in
equivalent control animals maintained under similar conditions. If a
statistically significant change occurs in a study but the control and test
A . B. Wilson 39
values both lie within ‘normal limits’, it can be suggested that the change
is, of itself, of doubtful toxicological significance, so long as it does not
point to another more profound effect. For example, a slightly lower
haemoglobin content might lie within normal limits for that laboratory
and so it could be suggested that those animals could live and function
entirely satisfactorily with that slight change. However, the change might
be the result of hepatotoxicity, bone marrow disfunction, or another
serious effect of the chemical under investigation. The mere fact that the
change lies within normal limits does not mean it can be dismissed as
irrelevant.
3 Customary Study Designs

A Principles
Most routine studies, orientated towards general, non-specific investiga-
tions and satisfying regulatory bodies, have a similar pattern: they
attempt to imitate the routes of exposure likely to be experienced by
man; they tend to exaggerate the likely exposure, both in terms of
duration and dosage; and they have a control group and three to five
dose groups with increasing dose levels. The question of route of
exposure is covered more fully in Chapter 2.
Exaggeration of exposure works on the simple principle that a high
dose is more likely to show up the potential effects, although effects
never likely to occur in man could also be encountered. While it is
probably a valid approach, the process of exaggeration is similar to using
a magnifying glass-it often distorts the image and it focuses upon detail.
It takes an expert judgement to decide whether the detail is relevant and
to allow for the distortion.
The highest dose used in longer term and carcinogenic studies is
generally the ‘Maximum Tolerated Dose’ (MTD). The definition varies
inevitably with the circumstances of each study (Haseman, 1986; Feron
and Kroes, 1986). It reflects an attempt to administer as high a dose as
possible without the toxicity of the compound distorting the animal
model excessively. However, the effects contributing to a judgement of
M T D are often clinical (perhaps bodyweight or laboratory tests or
pathology). Therefore, some toxic effects in longer term studies may arise
from overload of the normal routes of metabolism of a compound with
resultant unrealistically high concentrations of metabolites (or the parent
compound). The International Conference on Harmonisation (199 1)
recommended greater attention to toxicokinetics in the selection of dose
levels for carcinogenicity studies. Coupled with a series of pertinent
publications (Carr and Kolbye, 1991; Goodman and Wilson, 1991 ;
Monro, 1992) it seems possible that there will be a shift in thinking so far as
carcinogenicity studies are concerned.
Again, it is important to address the original object of the study,
whether it was to show the toxic potential of a compound or to help in
assessing safety. The differences between these can be critical to the
experimental design, particularly with compounds that are relatively
non-toxic at high dose levels. Under such circumstances, it is reasonable
to expect a safety factor of perhaps 100 to 1000 fold to be adequate, and
dosing beyond that to be irrelevant to the assessment of safety, though
not irrelevant to the investigation of potential toxicity.
This is illustrated in the different approaches of regulatory bodies. Some
tend to request that sufficient compound be administered in a rodent
carcinogenicity study to provoke a clear effect, even if that dose is several
hundred times the human exposure. The regulatory body is, therefore,
investigating the carcinogenic potential of the compound, but may
presumably choose to interpret carefully a toxic result at a high dose level.
Other bodies, in addressing themselves to the problem of the safety of the
compound, do not believe that heroic doses can normally be justified in
such an experimental design.
B Control Groups
In many conventionally designed toxicology studies, the compound is
given to animals in various treatment groups and results are compared
with the results of a control group. A control group is not always required
where its response can be predicted with adequate certainty. For instance,
in oral rat LD,, studies, where clear clinical signs and deaths are expected,
most laboratories will have sufficient experience to know that rats and
most other species rarely die or show clinical signs under those conditions
unless the compound is responsible. If, however, an unusual vehicle is
being used or more subtle observations made, comparison with back-
ground data may be inadequate and control animals would then be
needed.
It is apparent that a control group can take more than one form.
Normally the control group is treated in precisely the same manner as
treatment groups, except for administration of the test compound, i.e.
animals are handled, weighed, observed, fed, and sham dosed (generally
with vehicle) in an identical manner to dosed animals. Thus it is possible
to identify effects attributable to the compound. On occasions, the
manipulation of control animals is likely to have a substantial effect upon
them, as can occur in restraint for nose-only exposure of rats (Chapter 2),
the feeding of an artificially high protein diet to imitate the protein
content of an enzyme (contributing protein to the diet) under investiga-
tion, or the dosing by gavage with substantial quantities of vegetable oil
as the vehicle. In such instances a second control group, maintained
under more normal undosed or unrestrained circumstances, can be
considered. Its value is limited since its performance is compared (usually
subjectively) to background control performance, with the object of
assessing whether the routine factors of the study were being conducted
satisfactorily and showing the extent to which the experimental technique
altered the animal model. A placebo group is another type of control but
appropriate only to human studies.
A . B. Wilson 41
The control groups described above are sometimes termed ‘negative

controls’ since they are not treated with compound. If the group is dosed
with the vehicle used for suspending/dissolving the test compound, then
it is frequently termed the ‘vehicle control’. A ‘positive control’ refers to
the administration of a compound to a further group of animals. The
compound is selected as having the effect being investigated, for
example, mutagens such as 2-amino anthracene or sodium azide are used
in bacterial mutagenicity studies. Alternatively, the control compound
might be chosen because it has either similar biological activity to the test
compound or a similar structure, in both instances its toxicology would
normally be known. In such circumstances it might be more correctly
termed a ‘reference compound’. Positive controls give an indication as to
whether or not the test procedure has been succesfully carried out and are
more appropriate to studies for specific effects (e.g. mutagenicity) than for
general screening investigations. There are other methods of achieving the
same end so positive controls are not used all that frequently, except in
short term tests for identifying mutagenic or genotoxic compounds
(generally ‘in vitro’ tests). The use of higher dose levels in screening studies
may be regarded as equivalent to a positive control in many respects.
The value of a positive control group is sometimes misjudged. For
instance, the inclusion of a group of animals dosed with CC14 (a
hepatotoxin) will give some added confidence that liver toxicity would
have been detected if present. It does not, however, prove the suscep-
tibility of that strain of animal or suitability of that test protocol for all
potential hepatotoxins.
The control group, particularly the ‘negative control’, is crucial since
each dose group is compared to it statistically as well as biologically. If the
control group ‘misbehaves’ in some way it could give rise to misleading
conclusions, A subjective assessment therefore indicates that the control
group should be larger than any one of the treatment groups (see Chapter
18).
Despite the statistical arguments and basic logic, it is the exception
rather than the rule for negative control groups in toxicology studies to be
larger in number than treatment groups. Rodent carcinogenicity studies
do, sometimes, include a large (often double size) control group, and dog
or primate experiments where group sizes are generally very small (3 to 6
per sex per group) sometimes have larger control groups.
C Selection of Dose Levels

As previously described, the vast majority of toxicology studies comprise
a control group and three or four dose groups, e.g.:
Group Dose (mg kg-’)

Control 0
Low Dose 1
Intermediate Dose 2
High Dose 4
The first stages of dose selection are covered under sections 2A and 2B of
this chapter. For instance, there should be a review of the available
information and, in many instances, preliminary studies will have been
carried out with the express object of discovering at what dose levels
effects are likely to occur or not to occur. These might well have
commenced with a form of acute investigation and progressed through
various study durations as described later.
High Dose
Potency, palatability, solubility, and practicalities of administering a
larger dose generally restrict the maximum dose that is administered. It is
normally essential that this dose is sufficiently high to cause a clear and
indisputable toxic effect. The effect might be reduced body weight gain,
unusual clinical signs, or more subtle effects observed by the pathologists,
haematologists, or clinical chemists. Sometimes it is governed by meta-
bolic capacity or kinetics.
Views differ as to the desired severity of the effect and, of course, the
object of the experiment might be a determinant in that regard, e.g. an
LDS0 study requires death to be the effect. A key question is, ‘If no
effects are seen at the highest dose level will the experiment need to be
repeated?’.
There are pressures to limit as the highest dose given, to the lowest
dose level that will cause a significant and unambiguous effect. Keeping
the dose low reduces the span of doses to be covered in the study.
Low Dose
It is normal to select a dose at which no effect is likely to be discerned
and to wish to place this as high as possible, consistent with no effect
being observed. Another frequent determinant is the minimum accept-
able safety factor. If the human dose of a drug is to be 1mg kg-’ it might
be that a minimum safety factor of 5 is acceptable for certain parameters
in which case there is little point in selecting a dose under 5 mg kg-’. If
this sort of judgement is made, the scientist in the example has to be sure
that if a marginal effect occurs at 5mgkg-’ he will not be obliged to
repeat the experiment at a still lower dose level. The ratio between the
toxic dose and the therapeutic dose is the ‘therapeutic index7-the
pharmaceutical equivalent of the ‘safety factor’.
Intermediate Dose
In practice, it is usual to select both high and low dose before the
intermediate dose. Doses may be spread arithmetically, so that the
intermediate dose is the mean of the low and high dose (relatively
uncommon). More frequently an attempt is made to spread the doses
A . B. Wilson 43
geometrically, so that the intermediate dose is the same multiple of the
low dose as the high dose is of the intermediate dose. The statistician
may plead for a logarithmic spread of dose levels; however, the scientist
may not be interested in constructing a dose-response curve. He is often
more interested in setting the highest dose he can at which no effect is
discerned, in which case he may place the intermediate dose group rather
lower than otherwise. The intermediate dose group is often an attempt to
salvage something of value from the experiment when the high dose is
too toxic or the low dose too low. A dose-response curve can be of
considerable assistance in judging the ‘safety factor’. A steep curve
implies the effects will occur over a narrow range of dose levels so the ‘no
observed effect level’ can be defined more precisely than a shallow dose-
response curve would permit.
Experiments may be designed to display the carcinogenic or teratogenic
potential of the compound. The top dose is normally selected to have a
slight toxic effect on the animals (the MTD, vide supra), but a significant
toxic effect could lead to teratogenic effects secondary to significant
maternal toxicity or carcinogenic effects related to altered growth rate or
maturity. Under such circumstances, the intermediate dose may be
placed quite high, as a reserve high dose in case the high dose chosen
proves to be so toxic that results are difficult to interpret. The relevance
of the MTD and of safety factors in reproductive toxicology is well
dicussed by Johnson (1988).
In summary, for most animal toxicology studies, the experimenter has
to be very sure that the high dose will have some distinct and obser-
vable effect and the low dose will have none. It is probably true to say
that many more errors are made by having doses too close together than
by placing them too far apart.
It is curious that some regulations and guidelines request that lower
dose levels be set as fixed fractions of the top dose level. Such dictates
certainly curtail discussion of dose levels but they make no acknowl-
edgement of steep or shallow dose-response curves, or of the arguments
produced above, and place conformity before good sense (Mackay and
Elliott, 1992).
D Selection of Species and Number of Animals

Chapter 2 provides some basic information on the merits, limitations, and
conventions in the use of specific species of laboratory animals in toxi-
cology screening studies. Familiarity and availability have a great deal to do
with the initial selection of the species for testing. Following that, the
selection should preferably be for the one most likely to be relevant to man,
not necessarily the most susceptible species. There are substantial
differences between species in their anatomy, physiology, and metabolism,
any or all of which might render the species either anomalous or the species
of choice. Homburger (1987) presents a series of interesting examples as
well as some of the strengths and limitations of in vitro studies. However,

if the species turns out to be a somewhat exotic animal, which is difficult
to use and is poorly understood, then there are sound arguments for
remaining with the more normal species and interpreting results more
carefully. Frequently, the species selected for testing of pharmaceuticals
is the species which shows similarity to man in its pharmacological
reaction to the drug. Primates are often selected for studies on products of
biotechnology (e.g. hormones and monoclonal antibodies) because of
species specificity as well as reduced antigenicity. Similarities regarding
pharmacological activity do not mean that species will necessarily react
appropriately regarding other actions of the drug.
The choice of numbers of animals in a study provides some interesting
contradictions. On first consideration one might select the number based
upon the degree of change to be detected and its background variability
(Plaa, 1982). Tables 1 and 2 indicate the numbers used in routine studies
and they do not entirely tally with this logic. Dogs and primates have
much more variability than rats or mice, yet substantially fewer are used
in each dose group, 3 to 6 per sex, whereas for an equivalent study 10 to
25 rats are used. Expense, expediency and convention are the main
reasons.
If very large numbers of animals are employed, precision should
increase, but the mean values might be no different (Muller and Kley,
1982) and more changes will be identified which are statistically sig-
nificant. However, each change of statistical significance has to be
interpreted biologically. If the finding is slight (e.g. 2% alteration in body
weight gain), this will probably be quite uninterpretable on its own. There
is, therefore, little point in identifying such a change (except that it might
be part of a larger picture or an indicator of more significant effects in
other organs). It is, therefore, unusual in screening studies to take blood
samples from more than about 10 or 12 rats per sex per group for routine
haematology or clinical chemistry. Histopathology is usually adequate
from 10 to 15 per sex per group, unless a low but relevant incidence is
suspected, e.g. tumours. Organ weight data are not easy to interpret in
any case, so it is difficult to justify weighing organs from above 10 per sex
per group. If a compound is deemed to have caused the change, but the
change is interpreted as being of no biological or toxicological relevance,
then the merit of identifying it has much diminished. Chanter et al. (1987)
used data from a wide range of routine toxicology studies to indicate the
precision associated with many routine parameters and, in particular, the
likelihood of false negative results.
Debate over the precision necessary in LD50 studies has highlighted
these points. Muller and Kley (1982) and many others confirmed that the
use of fewer animals does not generally alter the mean value obtained for
the LD50,so it is difficult to justify the use of more than 5 animals per sex
per group under most circumstances. Even then, it is reasonable to obtain
the value for one sex before investigating the other-females are, in
?
Table 1 Rat and Mice: Typical Study Designs for Toxicity Studies 9
Number of Animals/ Haematology/ Months from 3
g
groups inc. sex/ Body Food clinical Gross autopsy to 3
Duration Purpose control group weight consumption chemistry autopsy Histopathology report
14 days To select 4-6 5-10 Weekly Weekly At end All Major organs 1-3 mth
or doses for animals and target
28 days longer organs*
studies
28 days To provide 4-5 10 Weekly Weekly At end All Wide selection 1-3 mth
definitive animals of organs"
information
90 days To provide 4 15 Weekly Weekly Week 6 and All Wide selection 3-6 mth
definitive end animals of organs?
information
26 weeks To provide 4 15-20 Weekly for 13 weeks Weeks 6, 12, All Wide selection 4-9 mth
or definitive then every 2-4 weeks. 26 and end animals of organst
52 weeks information
18-24 mth Carcino- 4 50 or Weekly for 13 weeks - All Wide selection 7-15 mth
(mouse) genicity more then every 2-4 weeks. animals of organs?
24-30 mth
(rat)
* Histopathology on control, top dose and selected animals/organs from other groups
t Histopathology on all animals
The above designs are intended only as examples.
Table 2 Dog and Primates: Typical Study Designs for Toxicity Studies
Number of Animals/ Haematology/ Months from

groups inc. sex/ Body Food clinical Gross autopsy to
Duration Purpose control group weight consumption chemistry autopsy Histopathology report
28 days To provide 4 3-4 Weekly Weekly Pretrial All Wide selection 1-3 mth
definitive and at end animals of organs on
information all animals
90 days To provide 4 3-6 Weekly Weekly Pretrial, All Wide selection 1-3 mth
definitive weeks, 6 animals of organs on
information and 12 all animals
26 week Toprovide 4 4-6 Weekly Weekly Pretrial, All Wide selection 3-6 mth
or definitive weeks , animals of organs on
52 week information 6, 12, 26, all animals
and 52
A . B. Wilson 47
general, a little more susceptible than males. The lack of precision makes
it more difficult to compute a dose-response slope with confidence.
However, a great many LD50 tests are carried out either for labelling
purposes (just for classification of a chemical) or for selection of dose
levels for longer studies (increasingly flexible approaches are now being
used and LD50 studies as dose ranging investigations are relatively rare).
In the former instance, dose response is not required and, in the latter
instance, it would often be more useful to extend the range of doses
selected at the next phase of testing than to increase the precision of the
acute study by doubling animal numbers. It is, as ever, important to be
quite clear about the object of a study so that the design can produce the
most pertinent results from the minimum numbers of animals: Zbinden
and Flury-Roversi (1981) discuss this and other topics related to the
conduct and interpretation of LD,, studies most thoroughly. Alternatives
to the LD,, are gaining credibility (van den Heuvel, Clark, Fielder,
Koundakjian, Oliver, Pelling, Tomlinson, and Walker, 1990; Yam, Reer,
and Bruce, 1991).
E Study Duration
Ideally, a study should be conducted for as long as is required to identify
the toxicological events of significance to man. Short term studies of high
dosage give valuable indications of the effects of acute poisoning, which
may be of value in the treatment of poisoning. Often the exposure of man is
over a much longer period and to smaller dosages. The response may then
be very different and there is no fixed relationship between the dosage
causing an acute effect and that causing a chronic effect (Plaa, 1982). It is
therefore necessary to conduct both single and multiple dose studies in
animals. The rat and mouse may be dosed for their full lifespan (1.5 to
2.5 years) on the assumption that this will equate to man. The evidence is
tenuous but the conclusion convenient.
Two factors are generally relevant to the selection of study duration:
(a) convention has dictated a series of arbitrary study durations from
acute to chronic, and (b) the shorter studies are required for selection of
dose levels before longer studies can be conducted.
Following a single dose (acute) study, sub-acute is the next duration
frequently used, i.e. 14 or 28 days of daily dosing (or such intervals of
dosing as are indicated by the likely use). The 14 day period is rarely
mentioned in regulations and is generally used only for screening studies
or as a preliminary to longer studies. However, it may be the appropriate
duration of study for a material to which man is likely to be exposed very
infrequently and for a very short duration, e.g. a vaccine or diagnostic aid.
Twenty-eight day studies (4 weeks, 1 month) are mentioned in
regulations and are often the longest study carried out for many industrial
chemicals with limited production volume and limited numbers of people
exposed. While there is nothing to prevent a study being designed with a

length between 1 month and 3 months, convention dictates that 90 days, (3
months or 13 weeks) is the next step, often called sub-chronic.
Six months (26 weeks) or 12 months studies, the duration again dictated
by convention and regulations, are regarded as chronic toxicity studies. It
is unusual for a toxicity study (as distinct from a carcinogenicity study) to
be extended beyond 12 months-indeed there are very few compounds
that show effects after 6 months or 12 months that do not show evidence
of the effect at 3 months (Heywood, 1983; Walker, Schuetz, Schuppan,
and Gelzer, 1984). Ocular toxicity, and cataracts in particular, represents
one of the few changes requiring study durations to extend beyond 3
months. European countries rarely request toxicity studies longer than 6
months, though Japan and the USA prefer 12 month studies. This is a
subject that was debated at the International Conference on Harmonis-
ation (1991). Whilst there was much sympathy for reducing the duration
of chronic studies to 6 months, there is, in practice a reluctance from
regulators to put this into practice.
The above implies a distinction between carcinogenicity testing and
toxicity testing. Without wishing to discuss the justification of that
separation, it is true that the jargon of toxicology does draw such a
distinction. Carcinogenicity studies are designed with the main object of
assessing the carcinogenic potential of a compound and normally last
most of an animal’s lifespan-in mice that means 18 months to 24
months and in rats 24 months to 30 months. At 18-30 months, rodents
are developing a great many lesions associated with old age and therefore
it can be very difficult to distinguish between toxic changes and
background lesions. This is one reason why toxicity studies of this length
are of limited value. There is a ‘Catch 22’ for carcinogenicity studies: it is
assumed that the animals must reach a certain maturity (i.e. be approach-
ing death) before a judgement on lack of carcinogenicity can be made. This
means that minimal durations of 18 months (mouse) and 24 months (rat)
are normally specified. However, various regulatory bodies prefer a high
survival at that time and some pathologists (e.g. Roe, 1983) complain of
the background lesions present. The animal husbandry experts, therefore,
improve conditions, lower dietary protein (Tannenbaum, 1940; Ford and
Ward, 1983; Ross, 1959), select longer lived strains of animals, and adjust
husbandry methods to achieve greater longevity. The regulators then say
the survival at 18 months or 24 months is too high so that studies must be
lengthened to ensure the animals are near their maximal lifespan, which
results in a high incidence of background lesions, and so on. The solution
to the dilemma lies in better characterisation of the end points to be
identified and for pathologists and toxicologists to define what they want
the ageing animal to look like or die of. This will involve an acceptance
that not all biological events are of toxicological significance and studies
should not be designed on the assumption that they are.
A . B. Wilson 49
F Study Designs
Preceding paragraphs have alluded to various conventional study designs
employed for screening purposes in animal toxicology. Tables 1 and 2 are
brief outlines of examples of this approach. The examples are not meant
to be exhaustive. Table 3 gives typical lists of haematology, clinical
chemistry, urinalysis, and pathology observations and again there are
many variations possible. Lumley and Walker (1983, Heywood (1981)
and Grandjean, Sandoe, and Kimbrough (199 1) provide information on
the organs and observations that most frequently show evidence of a toxic
response.
4 Randomisation and Replication

It is generally advised that the statistician be involved at the very outset
of a study’s design but this is often ignored because:
(a) the design may be standard or be dictated by other factors;
(b) the scientist has not defined his objects clearly enough to consult a
statistician;
(c) the statistician talks a ‘foreign language’; and
(d) the statistician is likely to explain that the study requires twice as
many animals to meet its objectives.
There are, however, a few basic principles relating to the avoidance of
bias that should be followed, replication and randomisation being
amongst them. If animals are not formally randomised (as opposed to
haphazardly allocated) to groups, then bias can occur-perhaps due to
placing all the heavy animals in one group, all the fast moving animals in
another, etc. If small numbers of animals are being used it is very possible
that, despite formal use of random number tables, etc., the various
groups will be obviously different from each other in one or more key
parameters.
Two ways round this are to re-allocate animals and/or substitute a few
spares for animals already allocated, and thus balance the groups. These
solutions are not entirely satisfactory since animals will not have been
allocated randomly and the basic assumption behind the statistical tests
applied is that random allocation has occurred. Another solution is to
measure the critical parameter(s), say body weight, and rank the animals
in body weight order or place them in a series of batches each containing a
specific range of body weights. The animals may then be randomised by
mixing one member of one batch with another member from each other
batch such that each treatment group will have similar numbers of animals
from each weight range. The means of each treatment group will thus be
similar. This is the first stage of ‘replication’ whereby blocks of animals, or
replicates, are identified.
The practicalities should not be forgotten and, if 600 rodents are to be
allocated to several different treatment groups, they will probably have to
Table 3 Typical Lists of Requirements for Laboratory

Investigations and Pathology in Toxicology Studies
HAEMATOLOGY PATHOLOGY
Erythrocyte count Adrenal*
Haemoglobin content Any abnormal lesions
Packed cell volume Aorta
White blood cell count Bone with Marrow (Sternum)
Differential white cell count Brain*
Platelet count Caecum
Colon
Selected measurements
of clotting function Eye
Heart*
Intestine (Small and Large)
CLINICAL CHEMISTRY
Kidney*
Blood urea nitogen Liver*
Total protein Lung and Bronchus*
Albumin Lymph Nodes (Selected)
Albumin/Globulin ratio Muscle
Aspartate aminotransferase (AST) Oesophagus
Alanine aminotransferase (ALT) Pancreas
Alkaline phosphatase (AP) Pituitary*
Sodium Prostate/Uterus
Chloride Salivary Gland
Potassium Sciatic Nerve
Blood glucose Skin with Mammary Gland
Spinal Cord
URINALYSIS Spleen*
Stomach
Volume Testis/Ovary *
Appearance Thymus
PH Thyroid with Parathyroid
Sediments Trachea
S.G.
Blood pigments
Protein
Bilirubin
Ketones
Glucose
* Organs frequently weighed.
These lists can be extended very considerably if justified by t.,e nature of 1 ie
compound and its likely effects.
be numbered provisionally, randomised, moved to different cages, given

new numbers, allocated to groups, and labelled accordingly. One has to
be sure that a refined and elaborate randomisation system will not cause
confusion in its operation. If it does, a similar but less ‘pure’ method may
be better. Some practical methods of randomisation are described in one
of a series of volumes on Standard Operating Procedures in Toxicology
(Paget and Thomson, 1979).
Having randomised animals to groups it is important to consider
location of cages or pens. Different locations will have different tempera-
A . B. Wilson 51
tures, different light intensities, different times of dosing, different times

of observation, different noise levels, and different airflows. Even in a
well designed animal room, the top and bottom of a rack of rat cages can
differ in temperature by 5°C so the method sometimes encountered of
putting each dose at a different level in a rack cannot normally be
justified. Usually the randomisation process used to allocate animals to
groups also serves to place them in specific locations; there is no
significant advantage in employing two separate procedures.
For larger rodent studies, using several hundred animals, there is often
considerable debate as to whether dose groups should be scattered
randomly within each rack of cages or whether one rack should contain
only one dose group. The former is better, so long as the random
positioning does not cause confusion to the operator and result in
administration of the wrong dose to a particular animal or contamination
by dust in feeding studies. The latter randomisation method probably
presents less chance of a dosing error, although it is arguable that if it
occurs, all the animals in at least one group will be misdosed. A
compromise is to employ one rack per dose group and to move the racks
to different positions in the room at regular intervals. Haseman,
Winbush, and O’Donnell (1986) reviewed data from a large number of
rodent carcinogenicity studies with dual control groups in order to assess
the incidence of false positive results. One of the two main recommenda-
tions was to distribute dose groups systematically or randomly.
Similarly, there is debate as to whether or not animals should be dosed
in a random order. This is, of course, desirable, but if large numbers are
involved, the chances of dosing an animal with the wrong material are
much increased. Often a compromise is reached whereby each group is
dosed in its entirety but the order of dosing is changed on different days
or weeks, i.e. the potential operator errors inherent in the better system
are deemed greater than the bias introduced by dosing all of one group at
one time.
Rodents are often housed about 4 or 5 animals to a cage. If a study
requires only 10 animals per group the numbers per cage should normally
be reduced otherwise ‘cage effects’ rather than compound effects might
be measured.
There are several other reasons for considering the use of replicates in
a study. Replication ensures that fortuitous allocation of animals in an
unbalanced manner should not occur, e.g. one could, by chance, have all
of one dose group placed in adjacent cages when in fact one had hoped to
distribute them fairly evenly. In large studies it is not always possible to
carry out all of a day’s procedures within a single 24 hour period, e.g.
taking and examining blood samples from 150 animals or carrying out a
detailed autopsy on 500 rats. If replicates are used then a suitable number
of blocks can be observed, bled, or autopsied each day, whilst ensuring
that a balanced number of each group is present and in a random order
within each block. The second of the recommendations by Haseman et al.
(1986) addressed pathology and included the suggestion that his-

topathological evaluation should contain an element of evaluation in
which the pathologist is unaware of the group from which the slide derived.
This emphasises the importance of evaluation being a random order and
‘blind’. That is, ‘blind’ in relation to the group, not necessarily in relation
to the lesion in question. Aspects of experimental design such as
replication, factorial arrangement of treatments, and strain variability are
reviewed by Festing (1992).
Some experiments have particular problems. Inhalation studies often
require each dose level to be in a separate inhalation chamber. If there
are environmental differences between the chambers, these could confuse
the results. For instance, efforts are often made to generate atmospheres
of test material with very clean, oil free, dry air and diluting this
concentrate with room air. In high dose groups, the dilution is likely to
be much less and the relative humidity in that group’s chamber will be
lower. In extreme circumstances this can cause ringtail in rats. The
scientist thus has to strive to take account of the potential differences and
attempt to minimise them. In this respect, it is important to house or
restrain control animals in the same manner as the test groups if
observations much beyond gross clinical signs are to be measured.
Normally, the range of doses from low to high is within a span of about
10 fold and it is most exceptional for the range to be greater than 100 fold.
If this range is very wide (e.g. 1000 fold) there is a real possibility of
contamination (from vapour, dust, or compound in excreta) of control
groups from high dose groups to a biologically significant extent. If the
dose range is within the more usual limits then the contamination is likely
to be well below levels of any consequence, unless the compound is in the
diet and is exceptionally volatile or dusty. Under such circumstances each
dose group can be held together and separated from the other groups, the
errors likely to be induced by contamination being greater than the bias
likely to arise from lack of positional randomisation.
Multigeneration reproduction studies and similar studies illustrate
other difficulties. If there is a slight chance effect that alters one factor at an
early stage, the knock-on effects can continue throughout the study. Thus
a reduction in litter size could increase weight gain in the young, with
possible effects on age of reaching sexual maturity and effects on
haematology, clinical chemistry, and organ weights. An example of
differences in F, body weights causing confusion in a carcinogenicity
study is given in Olsen, Wurtzen, and Meyer (1983). The statistician will
feel uncomfortable because F, and succeeding generations will not have
been allocated randomly to their dose groups but will have been allocated
according to their parents’ dose group.
Teratology study results are handled in two ways: (a) using the litter as
the unit, i.e. comparing the number of litters containing abnormal young,
or (b) using the foetus as the unit, i.e. comparing the number of
abnormal foetuses. Both have validity and, in general, both approaches
are used in the evaluation of results in any one study.
A . B. Wilson 53
5 Conclusion
It should be remembered that the toxicological investigation of a
compound rarely rests upon one study on its own. Although each
experiment has to stand alone, the assessment of toxicity will mean an
overview of the full range of studies: acute and chronic, rodent and
non-rodent, in vitro work and human experience, metabolism and
literature searching, reproductive effects and pharmacokinetics.
Regulatory bodies have a dominant influence upon the design of
toxicological investigations for new and existing compounds. It is pleasing
to note that there is now a very high degree of accord between different
countries in the studies and study designs they require. This has almost
eliminated repetition of animal work for idiosyncratic rather than
scientific reasons. However, achievement of improvement at any speed is
likely to compromise this hard earned uniformity and interesting conflicts
will thus arise. It is possible that the novel challenges presented by the
toxicological evaluation of peptides and proteins resulting from biotech-
nology will be a catalyst for a constructive evaluation (Petricciani, 1983;
Lasagna, 1986, and Wilson, 1987).
In the context of toxicology, the experimenter would do well to consult
the papers of Brusick (1987), Oser (1987), and Zbinden (1991) before
making judgement as to whether toxicological screening studies as
currently conducted can match the objectives they set out to achieve and
whether substantial improvement is necessary.
References
Alder, S. and Zbinden, G. (1983). Neurobehavioural tests in single and repeat
dose toxicity studies in small rodients. Arch. Toxicol., 54, 1-29.
Brusick, D. (1987). Current approaches to toxicity testing-status and prospects,
Comments Toxicol. 1, 257-264.
)
Cardy, R. H. and Warner, J. W. (1979). Effect of sequential bleeding on body

weight gain in rats. Lab. Anim. Sci., 29, 179-181.
Chanter, D. O., Tuck, M. A . , and Coombs, D. W. (1987). The chances of false
negative results in conventional toxicology studies with rats. Toxicology, 43,
65-74.
Clark, D. G. (1977). Long-term inhalation toxicology studies. I n : ‘Current
Approaches in Toxicology’, Ed. B. Ballantyne. John Wright and Sons,
Bristol, pp. 105-114.
Feron, V. J. and Kroes, R. (1986). The long-term study in rodents for identifying
carcinogens: some controversies and suggestions for improvements. J . Appl.
Toxicol., 6 , 307-31 1.
Festing, M. F. W. (1992). The scope for improving the design of laboratory
animal experiments. Lab. h i m . , 26, 256-267.
Food and Drug Administration Advisory Committee on Protocols for Safety
Evaluation: Panel on Reproductive Report on Reproductive Studies in the
Safety and Evaluation of Food Additives and Pesticide Residues (1970).
Toxicol. Appl. Pharmacol., 16, 264-296.
Ford, D. J. and Ward, R. J. (1983). The effect on rats of practical diets

containing different protein and energy levels. Lab. Anim., 17, 330-385.
Grandjean, P., Sandoe, S. H., and Kimbrough, R. D. (1991). Non-specificity of
clinical signs and symptoms caused by environmental chemicals. Hum. Exp.
Toxicol., 10, 167-173.
Haseman, J. K. (1986). Issues in carcinogenicity testing: dose selection. Fundam.
Appl. Toxicol., 5, 66-78.
Haseman, J. K., Winbush, J. S., and O’Donnell, M. W. (1986). Use of dual
control groups to estimate false positive rates in laboratory animals car-
cinogenicity studies. Fundam. Appl. Toxicol., 7 , 573-584.
Heywood, R. (1981). Target organ toxicity. Toxicol. Lett., 8, 349-358.
Heywood, R. (1983). In: ‘Animal and Alternatives in Toxicity Testing.’ Eds M.
Balls, R. J. Riddell, and A. N. Worden. Academic Press, London, pp.
79-93.
Homburger, F. (1987). The necessity of animal studies in routine toxicology.
Comments Toxicol., I, 245-255.
Johnson, E. M. (1988). Cross-species extrapolations and the biological basis for
safety factor determinations in developmental toxicology. Regulat. Toxicol.
Pharmacol., 8, 22-36.
Lasagna, L. (1986). Clinical testing of products prepared by biotechnology.
Regulat. Toxicol. Pharmacol., 6, 385-390.
Lumley, C . E. and Walker, S. R. (1985). The value of chronic animal toxicology
studies of pharmaceutical compounds: a retrospective analysis. Fundam.
Appl. Toxicol., 5, 1007-1024.
Mackay, J. M. and Elliott, B. M. (1992). Dose ranging and dose setting for in vivo
genetic toxicology studies. Mutat. Res., 271, 97-99.
Muller, H. and Kley, H. P. (1982). Retrospective study on the reliability of an
‘Approximate LD507determined with a small number of animals. Arch.
Toxicol., 51, 189-196.
Olsen, P., Wurtzen, G . , and Meyer D. (1983). BHT, Long-term toxicity.
Toxicol. Forum, Geneva, Oct 18-22, 1983.
Oser, B. L. (1987). Toxicology then and now. Regulat. Toxicol. Pharmacol., 7,
427-443.
Paget, G. E. and Thomson, R. (1979). ‘Standard Operating Procedures in
Toxicology’, Vol I. , MTP Press, Lancaster.
Petricciani, J. C. (1983). An overview of safety and regulatory aspects of the new
biotechnology . Regulat. Toxicol. Pharmacol., 3, 428-433.
Plaa, G. L. (1982). Principles of toxicology. In: ‘Survey of Contemporary
Toxicology’, Vol 2, Ed. A. T. Tu. John Wiley, New York, pp. 203-225.
Roe, F. J. C. (1983). Carcinogenicity testing. In: ‘Animals and Alternatives in
Toxicity Testing’. Eds M. Balls, R. J. Riddell, and A. N. Worden. Academic
Press, London, pp. 127-130.
Ross, M. H. (1959). Protein, calories, and life expectancy. Fed. Proc., 18,
1190-1207.
Tannenbaum, A. (1940). The initiation of the growth of tumours. 1. Effects of
underfeeding. Am. J . Cancer, 38, 335-350.
Van den Heuvel, M. J., Clark, D. G., Fielder, R. J., Koundakjian, P. P.,
Oliver, G. J. A., Pelling, D., Tomlinson, N. J., and Walker, A. P. (1990). The
international validation of a Fixed-dose Procedure as an alternative to the
classical LD,, test. Food Chem. Toxicol., 28, 469-482.
A . B. Wilson 55
Walker, S. G., Schuetz, E., Schuppan, D . , and Gelzer, J. (1984). A comparative
retrospective analysis of data from short- and long-term animal toxicity
studies on 40 pharmaceutical compounds. Arch. Toxicol., Suppl 7 , 485-487.
Wexler, P. (1987). ‘Information Resources in Toxicology’, 2nd edn. Elsevier,
New York.
Wilson, A. B. (1987). The toxicology of the end products from biotechnology
Processes. Arch. Toxicol., Suppl 11, 194-197.
Yam, J., Reer, P. J., and Bruce, R. D. (1991). Comparison of the up-and-down
method and the Fixed-dose Procedure for acute oral toxicity testing. Food
Chem. Toxicol., 29, 259-263.
Zbinden, G . and Flury-Roversi, M. (1981). Significance of the LD,, Test for the
toxicological evaluation of chemical substances. Arch. Toxicol., 47, 77-99.
Zbinden, G. (1991). Predictive value of animal studies in toxicology. Regulat.
Toxicol. Pharmacol., 14, 167-1 77.
CHAPTER 5
The Biochemical Principles

of Toxicology
W. N. ALDRIDGE
1 Introduction
Toxicology is the science of the action of substances which when
introduced into an organism cause change in its structure and/or
function. In this chapter, the substances discussed are restricted to low
molecular weight chemicals (proteins, etc., are excluded) and to mam-
mals or their constituent parts.
Toxicology for its scientific development requires the integration of
information acquired by the use of techniques used in many different
scientific disciplines (see Figure 3). To speak of the ‘biochemical
principles of toxicology’ is thus in some way a contradiction. Toxicologi-
cal problems can be tackled by applying biochemical techniques or
biochemical hypotheses to many biological systems. Advances in toxicol-
ogy result as the information is shown to be relevant to toxicity in whole
animals.
This chapter, therefore, is concerned with those hypotheses (prin-
ciples) which have evolved from research aiming to solve questions
concerning the molecular basis for toxicity. It should be self-evident that
the action of a chemical on a biological system ought to be explainable in
chemical terms. That it is not always so is a reflection that many areas of
normal biology are not yet understood in molecular terms (Aldridge,
1980, 1981).
2 Types of Chemical Interaction

The receptor theory in pharmacology is now supported by the isolation of
receptor proteins. In toxicology there are primary reactions with macro-
molecules which initiate a toxic reaction; these are sometimes called
targets and may be proteins, nucleic acids, components of membranes,
enzymes, etc. The reactions between a toxic chemical and macro-
56
W. N . Aldridge 57
molecules are chemical reactions and can be examined by the normal

methods of chemistry.
Reactions can be classified into two categories: reversible or covalent
interactions:
M + AB k+l
k-1
M AB
M+ A B LB + MA----*
M = macromolecule AB = toxic chemical
These two types of reaction govern dose-response relationships at the

target. The kinetic behaviour expected depends theoretically on a
knowledge of the concentration of both reactants. As will become clear,
these concentration terms are influenced by many factors.
A Reversible Reactions
The relationship between binding and concentration of reactant in
reaction (1) is shown by:
The affinity constant (Kaff)defines the concentration for 50% reaction

to take place. The relationship between percentage binding and con-
'
centration when Kaffis lo5 M - is shown in Figure 1A under the common
condition when the initial concentration of the chemical ([AB],) is much
greater than that of the macromolecule ([MI,). From this relationship and
for macromolecules where there may be more than one binding site
several graphical methods have been derived for handling experimental
data. The Scatchard (1949) method is perhaps the most common and is
based on the following equation:
B
-= nK,, - BKaff (4)
F
where
B = bound chemical expressed in various concentration terms such as
mol rnol-l, mol mg-l protein, etc.
F = free (unbound) chemical expressed as a concentration.
n = the maximum concentration of binding sites of affinity, K , and
expressed as n mol-', nmol mg-' protein, etc.
58 The Biochemical Principles of Toxicology
100
15
-
I-
s
.-5 50
c
U
ru
lx
W
25
0
1 2
B o u n d Liqand ( B ; mol/rnol-l)
2.0 0.25 r
1.5
1.0
5 10 15
Time / m i n
LA BI 0
Figure 1 Reversible and covalent interactions. @ shows the extent of a reversible
reaction when [AB], >> [MIo (in text, Equation 3) and Kaffis lo5 M-'.
E
(J is a Scatchard plot for a reversible reaction of AB with a protein containing
two binding sites with a KaEof lo5M-' (in text, Equation 3).
0and @ represent the rate of a covalent reaction when [AB], >> [MI,, and the
second order rate constant k is 4.6 X lo31mol-'min-' (in text, Equation 5). 0is
the rate of reaction for increasing [AB] and @ is the relationship between the first
order rate constants ( k derived from 0) and [AB].
A simple case where there is one class of binding site (2 sites with the
same affinity for the ligand) on a macromolecule is shown in Figure 1B.
For many situations the degree of binding of a toxic chemical cannot be
measured directly. Changes in biochemical and/or physiological para-
meters are often the only quantitative measures of binding to a target. In
the worst case the concentration of chemical, [A] related to the binding
to the target has to be assumed to be:
[A] = f [dose administered]

W.N . Aldridge 59
The exact mathematical relationship is always complicated since in whole

animal studies, steady state conditions are rarely attained; only if input
equals the output of the chemical is this so.
The most important feature of reversible interactions is that the
substance AB is not changed chemically by the interaction; if the
concentration of AB is reduced by excretion and/or metabolism the
reaction is reversed and M AB dissociates by the same route ( k - J as it
was formed (k+l).
For a more extensive discussion of reversible interactions see Gold-
stein, Aronow, and Kalman (1974).
B Covalent Interactions
The covalent interaction shown in Equation (2) is progressive and will be
defined by the second order kinetics of a biomolecular reaction.
It is often the case, as for reversible interactions, that [ABo]>>[M,]
and under these conditions the second order equation simplifies to
k, =-
1 No1
[AB,]t In [MI
in which k , is the bimolecular rate constant (1 mol- I ) , [AB,] and [Mo] are
the initial concentrations of AB and M, and t is the time. The
relationships between extent and rate of reaction to time and concentra-
tion of AB is shown in Figure 1C and D.
It is obvious that under this condition ([AB,] >> [M,]), and if the time is
prolonged, then almost all of M will become MA (Equation 2). This
condition may apply in in uitro systems but in vivo there are often
processes which remove AB.
An important distinction between reaction (2) and the reversible
reaction (1) is that when one molecule of M reacts with one molecule of
AB to form MA, a molecule of AB is destroyed. Formation of M from
MA is almost always slow and is not by a reversal of reaction (2) but by
an entirely different reaction.
For a more extensive discussion of covalent interactions see Aldridge
and Reiner (1971).
3 Types of Toxicity
Definitions of acute and chronic toxicity should be stated in a form which
reflects what we know of molecular mechanisms and distinguished from
operational criteria, such as the mode of administration.
A Acute Toxicity
Acute toxicity is that toxicity which arises soon after administration and,
unless death occurs, recovery is complete. Examples are the toxicity of
cyanides or organophosphorus compounds. The interaction of cyanide

with cytochrome oxidase is dissociable and recovery is rapid as cyanide is
detoxified or excreted. Organophosphorus compounds react covalently
with acetylcholinesterase. Recovery may be complete but is slow, since in
many cases it depends on resynthesis of new enzyme.
B Chronic Toxicity
Chronic toxicity may be divided into two main types
(a) that which requires prolonged or repeated administration of the
compound before the toxicity becomes apparent; and
(b) that which is long lasting but can be brought about by one or very
few doses of the chemical.
Prolonged Administration of the Chemical

Chronic toxicity, i.e. that toxicity which may take a long time to appear,
can be subdivided into several types. Repeated or continuous dosing of a
substance may be required to increase its concentration or a metabolite
of it at its site of action. Cadmium is a good example, where over several
years its concentration may increase in the kidney cortex up to 300 pg 8-l
and thus cause functional deficiency (see page 72). Cadmium will
produce nephropathy after an acute dose but since it is highly cumulative
in mammals, it can produce toxicity to the kidney after exposure to low
concentrations for a long time.
In many instances of the above type of chronic toxicity, the tissue
repairs and function may become, if not normal, then adequate when the
exposure to the chemical ceases. However, if the tissue contains cells
which do not undergo mitotic division then a chemical insult which causes
death of these cells will lead to a permanent deficit. The consequences
may not be severe. Death of striated muscle cells, for example, may not
be very important, since, although they will not be replaced, those
remaining will hypertrophy so as to maintain normal function. If the cells
dying are neurons then a permanent deficit remains for the rest of the
animal’s life (see page 64).
Repeated administration of a chemical may cause necrosis of cells, e.g.
carbon tetrachloride causes necrosis of hepatocytes. The remaining
hepatocytes divide so as to replace those damaged but, in addition, the
systems concerned with the structural elements in the liver are activated
and lead to the laying down of an abnormal proportion of collagen (liver
cirrhosis; Cameron and Karunaratne, 1936). The different molecular
forms of collagen (types I, 111, and p ) are all increased in concentration
(Rojkind and Martinez-Palomo, 1976; Biempica et al., 1980; Seyer,
1980). In this way the functional capacity of the liver is permanently
reduced. For example , dimethylnitrosamine which, besides being a
potent carcinogen (Magee and Barnes, 1956) also causes centrilobula;
W. N . Aldridge 61
necrosis in the liver, was first suspected of being a toxic compound
because of the diagnosis of liver cirrhosis in men working in a pilot plant
for its manufacture (Barnes and Magee, 1954). Another example is the
pulmonary fibrosis diagnosed in some coal miners after many years
exposure to coal dust.
One or Few Doses of the Chemical

Chronic toxicity may be a long lasting biological effect produced by one
or very few doses of unstable compounds. The chemical, although
present for a very short time, initiates a biological ‘cascade’ leading to
biological change sometime later. Dimethylnitrosamine given as a single
oral dose to rats on a particular dietary regime produces tumours in the
kidneys of all the rats many months later (Swan and McLean, 1968).
Delayed neuropathy produced by organophosphorus compounds is
another example (see page 70).
4 Physical and Chemical Properties of Toxic

Chemicals
The essence of organised biological systems is the integrated structure in
which various functions are exercised by specific organs; within organs
chemical reactions occur in specific compartments. Their functions and
biochemistry involved could not be smoothly organised without a
multitude of compartments surrounded by membranes. While these
membranes are essential for the orderly function of the organism, they
also provide barriers to extraneous chemicals. The properties conducive
for the penetration of membranes containing protein and lipid layers
have been well worked out. Rapid penetration of chemicals is best
achieved when the substance has both lipophilic and hydrophilic pro-
perties. Two substances serve to illustrate this point. Dimethylnitrosa-
mine, a potent carcinogen, is soluble in lipid and water and on oral
administration is rapidly absorbed from the gastrointestinal tract and
mixes with body water (Magee, 1956). O,S,S-Trimethyl phosphorodi-
thioate, which causes death due to lung damage, has the same physical
properties, is very rapidly absorbed after oral dosing, and mixes so
rapidly with the body water that the first passage through the liver where
it is metabolised does not appreciably influence its final concentration in
the blood and tissue. There is no doubt that both of these substances will
be taken up rapidly by other routes, e.g. skin, lungs, mouth, etc. (Aldridge,
Verschoyle, and Peal, 1984).
For substances either more lipophilic or more hydrophilic, constraints
on the rate of absorption are found. For example, it is difficult, if not
impossible, to administer by the oral route a toxic dose of the lipophilic
DDT as a water suspension of finely divided compound; it is, however,
absorbed rapidly if given as a solution in vegetable oil. DDT is soluble in
oil but has a very low solubility in water.
Negatively charged substances are not absorbed well but may be

absorbed from the acid conditions in the stomach where sufficient of the
compound is in the uncharged form, e.g. the uncoupler , hexachlorophane
(pK of hydroxyl groups 5.4 and 10.85; Mahler, 1954). Positively charged
species are not well absorbed. Nevertheless, none of these rules are
absolute and the fact that compounds such as paraquat are absorbed from
the gut may well be due to specialised transport systems (Figure 2).
After substances have been absorbed, there are many membranous
impediments to their penetration into different tissues. For example,
substances which are positively charged do not easily penetrate the
central nervous system (CNS), e.g. paraquat does not penetrate cells in
the central nervous system and other tissues. It is toxic to the lung
because Type 1 and Type 2 pneumocytes possess a transport system for
putrescine and this transport system can be utilised by paraquat (Smith
and Wyatt, 1981). From work on tissue slices it appears that paraquat is
actively taken up by cells of the brain. However, in the intact animal
these cells are protected because paraquat does not penetrate the blood
brain barrier.
Paraquat: N , N ' - d i m e t h y l - 4 , 4 '

OSS - Trimethyl - bipyridilium chloride
phosphorodithioate
CH3,
N -NO
CH3/
CI CI
Dimethyl nitrosamine
Hexachlorophane : 2, 2 I -methylenebis-
( 3 , 4, 6 - trichlorophenol)
N
I
CH 3
Trietnyltin hydroxide and chloride
Pralidoxime : 2 -formy1 - I -methyl
pyridinium salt
Figure 2 Structural formulae of chemicals showing difering penetration of

animals, organs, and cells
W. N . Aldridge 63
The positively charged organophosphorus compounds produced by

storage of metasystox in water (Heath and Vandekar, 1957) do not
penetrate the CNS as shown by lack of inhibition of acetylcholinesterase.
Death occurs due to inhibition of this enzyme in the peripheral nervous
system. The therapeutic agent, pralidoxime, which is used for the
treatment of poisoning by organophosphorus compounds and which
reactivates the phosphorylated acetylcholinesterase, is ineffective for the
enzyme in the brain because it does not penetrate the central nervous
system.
Organotins, such as triethyltin sulphate, under mildly acid conditions,
are ionised into the positively charged triethyltin ion. Nevertheless it is
rapidly absorbed on oral administration because the hydroxide and
chloride (Kaff for hydroxyl ion: 2 x lo7 M-' and for chloride: approxim-
ately 10'm-'; Aldridge, Street, and Skilleter, 1977) are lipid and water
soluble.
The way properties of related compounds influence their distribution is
well shown by mercury and the organomercurials. Mercuric ions (Hg2+)
do not penetrate the nervous system but concentrate in and damage the
kidney. Methyl- and phenylmercuric salts, however, are lipid soluble and
do penetrate the central nervous system. Although methylmercuric salts
are much more stable than phenylmercuric salts, and while methylmer-
cury salts are toxic to the nervous system, phenylmercuric salts, by virtue
of the mercuric ions released by its metabolism, are toxic to the kidney.
Metallic mercury is lipophilic. Its penetration of the nervous system is
dependence on the supply of blood to the brain and its rate of oxidation in
the blood to mercuric ion (Magos, 1981; 1982).
Thus, there are physico-chemical properties of chemicals which facili-
tate their penetration into an organism and its constituent cells. Selective
toxicity, however, may also be due to special uptake mechanisms which
are there to transport an essential substance into cells but which can also
accommodate the foreign chemical.
For a more extensive discussion of the influence of physico-chemical
properties on toxicity see Albert (1985).
5 Selective Toxicity and Selectivity

The term selective toxicity is often used to indicate a differing toxicity to
different organisms or species (Albert, 1985). Thus, penicillin is selec-
tively toxic to certain bacteria and has a low toxicity to mammals.
Malathion is toxic to many insects and yet has very low toxicity to
mammals.
Within an animal, the damage a chemical induces is often pre-
dominantly to one organ. Beryllium compounds, for example, produce
necrosis of the liver, less damage in the spleen and kidney, and no
changes in other organs. This is because, although beryllium is toxic to
other organs, it does not enter the cells. Similarly, paraquat damages the
lung because certain cell types concentrate it-paraquat is toxic to all
mammalian cells but is concentrated in very few. Trimethyltin is
selectively toxic to the brain of mammals, primarily in the hippocampus,
with less damage in the amygdaloid nucleus and pyriform cortex (Brown,
Aldridge, Street, and Verschoyle, 1979). In this case, the selectivity is
Figure 3A Neuronal necrosis in the rat hippocampus produced by trimethyltin.

Control rat hippocampus Haematoxylin and eosin X 42
Figure 3 Neuronal necrosis in the rat hippocampus produced by trimethyltin. Ten

weeks after a single dose of trimethyltin chloride (10 mg kg-'). Shrunken hip-
pocampus with marked neuronal loss in the pyramidal cell band (arrows), less
damage in the fascia dentata and portion of the Sommer sector ( S ) . Haematoxylin
and eosin x 46
W. N . Aldridge 65
even greater for it is only neurons in particular locations in the
hippocampus which are affected (Figure 3). The evidence currently
available suggests that this is not due to selective accumulation of
trimethyltin in these areas of the brain.
Thus, selective toxicity is mediated by changes brought about in only
one organ and often by certain cells in that organ. As explanations are
sought for such selective action, an affinity of the chemical for certain
subcellular organelles, enzyme system, or membranes may be found;
once the target has been identified this selectivity at the molecular level
may allow the structure-activity relationships for the target to be
disentangled from that for the whole animal. For a discussion of some of
the principles and definitions of selective toxicity, see Aldridge (1987).
6 Phases of Developing Toxicity

It is convenient to consider toxicity as four interconnecting phases
(Figure 4). They are, in order: (1) all systems which affect the delivery of
the chemical to its site(s) of action; (2) its primary reaction with
macromolecular targets; (3) the cascade of biological (biochemical,
physiological, morphological) responses resulting from reaction with the
primary targets; and (4) the final clinical signs, symptoms, or syndrome in
animals or in man. This scheme is highly simplified and many examples
can be found where more elaboration of the diagram might be desirable.
DELIVERY PRIMARY BIOCHEMICAL- CONSEQUENCE

REACTION PHYSIOLOGICAL T O ORGANISM
CHANGES
+ No toxicity
.-------------- * No toxicity
-TOXICITY
(Clinical signs,
symptoms or
syndrome)
Figure 4 Scheme illustrating phases of developing toxicity following entry of

chemicals into an organism. Heavy lines and letters lead to toxicity whereas light
lines are events which are unproductive. T,, T2,T3 and MT1, MTZ,MT, are targets
before and after modiJication by the chemical, respectively, and
A, B, C, X, Y, Z, a, b, c represent an undefined and sometimes unconnected num-
ber of changes which may or may not be sequential. Acute and chronic toxicity
may be diferentiated by the time in one or several of the phases. Recovery f r o m
changes in function or morphology can occur by regeneration of T, and by
disposal and repair of modiJied tissue. Prophylaxis or therapy is theoretically
possible in several of the phases.
(Reproduced by permission from Aldridge, 1981, Trends Pharmacol. Sci., 2,
228-231).
It does, however, provide a framework for thought and perhaps to

suggest testable hypotheses for experimental studies.
A Delivery
The properties of chemicals which allow their ready absorption through
the gastric mucosa, lungs, intestine, skin, mouth, etc. have been
previously discussed (p. 61). Once a chemical enters the circulation its
physical properties determine its penetration of cells. A major deter-
minant of the toxicity of the compound is the rate at which it is
metabolised.
A discussion of the action of Malathion in the rat will illustrate the
principles involved. Malathion is a very effective organophosphorus
insecticide and when pure has an oral LDS0to rats of 10 g kg-' (Figure 5).
The rats die with symptoms caused by inhibition of acetylcholinesterase
(salivation, muscular fasciculation, red tears, respiratory distress).
Malathion does not inhibit acetylcholinesterase in vitro. The thionophos-
phorus grouping (P=S) is oxidised to the oxon (P=O) predominantly by
the liver. The product, Malaoxon, is a direct inhibitor of acetylcholines-
terase and has an oral toxicity of approximately 160 mg kg-l. The major
reason for the low toxicity of Malathion is that in the liver there is a
carboxylesterase which removes predominantly the &-ethyl group to
produce a non-toxic ionised compound (Figure 5 ) . This system is highly
effective since when this carboxylesterase is inhibited by certain other
organophosphorus compounds, thereby preventing the detoxification of
malathion, its toxic dose will be as low as 20 mg kg-l. The rat becomes as
sensitive to malathion as insects which do not possess this detoxification
pathway. These findings indicate that from an oral dose only about 1
molecule of malathion in 500 is converted to malaoxon to cause toxicity
(for discussion and references, see Aldridge, Miles, Mount, and Vers-
Carboxylesterase . inhibition
malaoxon : direct by some organophosphorus
nhibitor of acetyl- cH3O!OCH3 compounds prevents
I cholinesterase < detoxification and increases
I Toxicity 160 mg kg-'I 1 toxicity of malathion ( 2 3 mgk g j
Malathion : does CH3 0 Non toxic : either

not i n h i b i t acetyl - P = S not oxidised or
choli nesterase P = O compound is a poor
Toxicity 109 kq -1 inhibitor
Figure 5 Pathways of metabolism of malathion which influence its toxicity and

biological activity
W.N . Aldridge 67
choyle, 1979). This example illustrates the two general consequences-

that the toxicity of a chemical may be reduced or may be increased by
metabolism.
DetoxiJication by Metabolism
This area of biochemistry has been well studied for many years. There
are two classes of metabolism, Phase 1 and 2. Phase 1 are mainly
oxidative, reductive, or hydrolytic processes and produce substrates for
Phase 2 conjugations.
Oxidative Phase 1 reactions are usually carried out by the cytochrome
P-450 mono-oxygenase system of enzymes. Recent work indicates that
this is a family of enzymes (proteins) which can be shown by analytical
and immunological techniques to be different proteins. They are unusual
in that, unlike many other enzymes, they accept a wide range of
substrates of differing chemical structures. The enzymes in many tissues
are located in the microsomal fraction (endoplasmic reticulum) of
lipoprotein membranes in which the cytochrome P-450 is incorporated. In
general, lipophilic chemicals are accepted substrates and their oxidation
yields compounds which are more water soluble. Microsomal oxidations
include aromatic, aliphatic, alicyclic, and heterocyclic hydroxylations,
and the removal of alkyl groups attached to nitrogen or sulphur. The final
products of many of these reactions are suitable for conjugation (Phase 2)
with glucuronic acid, sulphate, glutathione, and glycine. These products
are water-soluble and are either excreted in urine or are further
metabolised to equally water-soluble compounds prior to excretion.
Other Phase 1 reactions include amine oxidase and hydrolytic reactions
catalysed by esterases, amidases, etc.
Although the foregoing is a short appraisal of a large field, it serves to
illustrate the principles involved. There are in most animals many systems
which serve to protect them from the deleterious actions of chemicals by
converting them rapidly to less toxic substances, followed by excretion.
Although these systems exist in all mammals, their efficacy against a
particular compound varies between different species. If large differences
in toxicity are found for a chemical in several species of mammal,
provided absorption occurs, the first hypothesis should be that the
compound is detoxified at different rates. Thus, for example, the oral
LDSOs (mg kg-' body weight) for chlorfenvinphos [O, O-diethyl-0-2-
chloro-l-(2,4-dichlorophenyl)vinyl phosphate] are for the rat 10, mouse
100, rabbit 500, and dog 12,000. These differences are mainly accounted
for by differences in the rate of removal of one ethyl group by
microsomal oxidation (Donninger, 1971).
IntoxiJicationby Products of Metabolism

Many examples are now known of increases in toxicity of chemicals due
to metabolism. As the chemical mechanism of Phase 1 has been
CH3, + *
,N.NO+CH3 Magee a n d F a r b e r (1962)
c H3 L y j i n s k y , L O O ,a n d R o s s
D i m e t h y l n it rosamine (1968)
-
p
:=s - fP=O A f l a t o x i n 8, A f l a t o x in-Z13-ox ide
-
Organophosphorus 0 xon Swensen, Miller,and
thion M i l l e r (1974)
Neal (1980)
M a r t i n a n d G a r n e r (1977)
cs, d cos co2
Carbon Carbon
disulphide oxysulphide
E d g a r , Smith a n d
I
C u l v e n o r (1970)
0 Mattocks (1972)
II H e l i o t r i n e R=CH,OCOC ( t s o - C 3 H 7 ) ( O H ) CH(OCH,)CH,
C6H50-P-OC6H
I
0
C H3 CH2 CH2CH 2 C H 2 CH,
Eto,Casida
Hexane CH3C OC H2CH2C OC H,
a n d E t o (1962)
CH3COCHZCH2CH 2CH3
2,5-h e x a n e d i o n e
Mono-2-tolyl-diphenyl Phenyl saligenin Met h y l - n - b u t y l k e t o n e Divincenzo , Krasavage,and
-
phosphate phosphate O'Donoghue (1980)
(C, H5I4P b (C, H5I3P b' Creme r (1 9 5 9 H 5 C 2 0 0 C 3COOC2H5

P r o t o p o r p h r i n - N-CH3
Tet raet h y llead Triethyllead
CH3 H CH3
3,5 - d i e t h o x y c a r b o n y I-1,4 - D e M a t t e i s , G i b b s . Farmer,

dihydro-2,4,6-trimethylpyridine L a m b , a n d H o l l a n d s (1982)
(DDC)
Figure 6 Chemicals whose metabolism leads to increased toxicity
* These structures have not been isolated
W. N . Aldridge 69
elucidated it is clear that, although the major throughput of products is

often towards detoxification and excretion, intermediates are chemically
reactive and toxic. The enzymes normally present in animals accept
chemicals as substrates and metabolise them if they fit into the right
conformation in the catalytic centre. It must be recognised also that the
concentration or amount of reactive chemical necessary to cause toxicity
if produced in vivo may be very small. Thus it may be calculated that out
of 2 g (6mmol) Malathion given to a 200g rat (LD,,; lOgkg-’) only
approximately 100 pmol of the active metabolite malaoxon will phos-
phorylate the acetylcholinesterase in the brain, i.e. approximately 2 x
of the dose administered.
In examples of a metabolism which increases the toxicity of compounds
(Figure 6), the site of toxicity depends on the stability of the metabolite
and whether it can escape from the organ where it was produced.
Dimethylnitrosamine is metabolised in the liver and causes liver necrosis.
Vinyl chloride is metabolised in the liver and causes angiosarcoma in the
liver. Carbon disulphide is metabolised in the liver and causes hydropic
degeneration and destruction of the cytochrome P-450system in the liver.
Tetraethyllead is metabolised in the liver and causes lesions in the brain.
Organophosphorus insecticides containing thiono phosphorus groups
(P=S) are oxidised to the oxon form (P=O) and are toxic by inhibiting
acetylcholinesterase in the nervous system. Diphenyl-2-tolylphosphate is
metabolised in the liver and causes delayed neuropathy due to ‘dying
back’ of long axons in the central and peripheral nervous system.
Although glutathione is a powerful protectant of cells from attack by
electrophilic chemicals, reactive oxygen species, etc., there are now
examples of glutathione adducts which are themselves toxic. If potential
reactive groupings are retained in the adduct then further metabolism may
yield products which are selectively toxic. Thus, hexachloro- 1,3-butadiene
reacts with glutathione in the liver (catalysed by glutathione S-trans-
ferase), is excreted via the bile, metabolised in the gut, reabsorbed and
taken up by the kidney, and metabolised by P-lyase to give a reactive
compound which causes toxicity in the tubular elements of the kidney
(Lock and Ishmael, 1979). For reviews in this rapidly expanding field see
Dekant et al. (1992) and van Bladeren (1988).
For a recent review of biologically active metabolites, see Gillette
(1982); Anders (I 985).
B Primary Reaction with Target

The identification of the primary target is an important task achieved in
only a few instances. Two examples of reversible interactions and one
example of covalent interaction illustrate the principles involved.
Reversible Interaction
Cyanide gains entry as the uncharged volatile hydrocyanic acid via the
lung or as the acid formed from salts such as sodium cyanide in the
stomach. Absorption and distribution is rapid. The animal dies in
convulsions with the oxygen tension in venous blood almost as high as in
arterial blood. These clinical signs are caused by the inhibition of
cytochrome oxidase in mitochondria, the terminal step in the electron
transport chain for the utilisation of oxygen for the oxidation of
substrates. The affinity of cyanide for cytochrome oxidase is 3 x lo5 M-'
(Wainio and Greenlees, 1960)- The circulating concentration of cyanide
required to cause death is approximately 5 pg ml-' (1.8 x m) (Ald-
ridge and Lovatt Evans, 1946). Assuming a uniform distribution of
cyanide in brain and plasma, this concentration would produce 98%
inhibition of cytochrome oxidase. The interaction of cyanide with
cytochrome oxidase is reversible and, if death does not occur, recovery is
rapid due to the excretion of HCN by the lung and, more importantly, by
its conversion in the liver to thiocyanate by the enzyme rhodanase.
Carbon monoxide is absorbed from the lung rapidly and binds to
haemoglobin. The affinity of carbon monoxide for haemoglobin is 250
times that of oxygen so that exposure to 0.08% CO in air (20% 0,) will
yield a 50:50 ratio C O : 0 2 bound to the protein. The oxygen in
haemoglobin is replaced by carbon monoxide and, when 50% saturation
is exceeded, death may occur. Comparison of men whose haemoglobin is
+
half-saturated with carbon monoxide (50% CO 50% 0,) with those
with 50% oxyhaemoglobin (either through haemorrhage or reduced
oxygen tension) is instructive. The latter are little affected but the former
are near collapse. The explanation of this lies with the properties of
haemoglobin. Four molecules of oxygen combine with the four sub-units
in haemoglobin. As each molecule of oxygen combines with each
sub-unit, it induces an allosteric change in the remaining sub-unit so that
the affinity of successive sites for oxygen are increased. Thus, under
normal conditions, rapid saturation of haemoglobin with oxygen during
passage through the lung is ensured. For the release of oxygen during
circulation through the tissues, the reverse process takes place. Release
of oxygen bound to one sub-unit with low affinity leads to the reverse
allosteric change so that the affinity of one of the remaining oxygens is
reduced, and so on. Carbon monoxide binds tightly to haemoglobin and
also mimics oxygen, increasing the affinity of the remaining sites for
oxygen. Thus, with two sites occupied by carbon monoxide (50%
saturation), the oxygen on to the remaining two sites is bound so tightly
that it cannot be released to the tissues (Roughton and Root, 1944). The
high affinity of carbon monoxide also explains why recovery by its
exhalation is so slow.
In these two examples, HCN and CO, the capacity of the tissues to
utilise oxygen is impaired, but brought about by different mechanisms.
Symptoms mainly originate from the central nervous system, even though
in only one case (HCN) is a primary biochemical interaction present in
the tissue affected. The onset of, and recovery from, symptoms is
dominated by the pharmacokinetics and/or metabolism of the chemical
itself.
W. N . Aldridge 71
0
NTE-P
5,ll
\
X R2
0 0
x
Figure 7 Initiation of delayed neuropathy by organophosphorus compounds.
A. Delayed neuropathy not produced by phosphinates. Phosphinylated neuro-
toxic esterase cannot lose a group (R,)because the phosphorus-carbon bond is
stable.
€3. Delayed neuropathy produced by phosphates and phosphonates. Phosphory -
lated and phosphonylated neurotoxic esterase can lose a group (R,)because the
phosphorous-oxygen or the carbon-oxygen bounds can be broken (the NTEIugeing
hypothesis)
Covalent Interaction
Certain organophosphorus compounds, while causing few or early
symptoms due to inhibition of cholinesterase, cause a condition of
delayed neuropathy in which symptoms of unsteadiness, progressing to
paralysis, do not appear until 10-14 days after exposure. Man is a
sensitive species for this condition and the chicken is the best experimen-
tal animal.
The primary target has now been identified as an esterase present in
the nervous system and tightly-bound to membranes (Johnson, 1975a and
b; 1982). It has been called ‘neurotoxic esterase’ or ‘neuropathy target
esterase’ (NTE).
The essential reactions (Figure 7) for those compounds causing delayed
neuropathy is phosphorylation or phosphonylation of the esterase
followed by the release of one of the groups attached to the phosphorus
(this process is called ageing because it is often a slow process). This
group becomes attached to the protein, in contrast to the ageing of other
phosphorylated or phosphonylated esterases when the group is released
into the water. The significance of this unique behaviour of ‘neurotoxic
esterase’ is not understood.
It is known that the reaction of di-iso-propylphosphorofluoridate
(DFP), with ‘neurotoxic esterase’ followed by ageing takes place very
rapidly (probably within 1 hour) and the excess DFP is detoxified. Signs
of delayed neuropathy do not appear for 10-14 days. Therefore this
chronic condition (often permanent paralysis of the limbs) is initiated by
a rapid chemical reaction. The nature of the biological cascade initiated

by this reaction, which leads to the clinical symptoms, is not known.
Presumably by synthesis of new protein, neurotoxic esterase activity
returns over the 10-14 days before symptoms appear. The ‘dying back’ of
long axons in the peripheral nerves can be repaired, but in the spinal cord
they cannot.
This example illustrates the principle that chronic diseases can be
initiated by a chemical reaction which is complete long before evidence of
disease is apparent. Presumably biological systems are switched on or off,
and the changes, once initiated, proceed and cannot be prevented.
Recent research has provided two observations which are difficult to
interpret at present and may require some modification of the ‘NTE/
ageing hypothesis’ (Lotti et al., 1991; Johnson and Safi, 1992).
C Biochemical, Physiological, or Morphological

Consequences of the Primary Reaction with Target(s)
In this phase of toxicity we are concerned with the consequences of
interaction of the toxic chemical with the target (see Figure 4). Many
changes may be induced; some of these lead to the symptoms, signs, or
syndrome of poisoning. This phase is, therefore, the link defined in
biochemical, physiological, or morphological terms between the primary
chemical interaction with the target and the changes in the function of the
whole organism. As illustrated in Figure 4, some reactions of chemicals
with macromolecules may take place which have either no biological
consequences or induce changes which do not lead to toxicity.
Organophosphorus compounds and carbamates inhibit acetylcholines-
terase by phosphorylating the active centre. This inhibition of the enzyme
in the nervous system is causally related to increases in the concentration
of acetylcholine and this increase is in turn related to the symptoms and
eventual death (often by respiratory failure). Acetylcholinesterase, with
identical properties to that in the nervous system, is present on the
membranes of erythrocytes. For most purposes inhibition of this enzyme
in the blood is an excellent monitor for exposure and the degree of
inhibition has the same relation to symptoms as the inhibition of the
enzyme in the brain. Thus, although inhibition of acetylcholinesterase in
human erythrocytes causes no toxicity, its inhibition by most anti-
cholinesterases is a mirror of changes in the activity of the enzyme in the
nervous system (Vandekar, 1980).
On chronic administration to man, cadmium accumulates in the liver
and kidney. Continuous exposure to small concentrations will result in
larger concentrations in these tissues because the half-life for cadmium is
16-30 years (Figure 8)- The rather simple arithmetic calculations (Figure
€9,using values for intake of cadmium and its half-life in vivo, show that
concentrations in the kidney and the daily excretion can be predicted.
This is a highly simplified calculation and concerns only one compart-
W. N . Aldridge 73
1
DOSE kl k2
30-60 pg 0.01 - 0 . 0 5 ' t i 16-30 years
1 day day-1 '2= 4 . 5 - 8 . 5 ~ 1 0 - 5d a y - ?
X is concentration i n kidney
Kidney i s 2 8 0 g ( 7 0 k g m a n ) - '
K i d n e y c o n t a i n s l/3 of t h e b o d y b u r d e n
C a l c u l a t e d e x p e c t e d mean c o n c e n t r a t i o n of
c a d m i u m i n t h e kidney.
45 x 0 . 0 3 = 2 5 pgg-1
6.5 x 10-5 x 3 x 280
P u b l i s h e d v a l u e s 20-30 clg 9-1
-
Expected d a i l y e x c r e t i o n of c a d m i u m
25 x 280 x 3 x 6 x = 1.3 pg day-1
Figure 8 Relationships between intake and output of cadmium and concentrations
in the kidney
(For a more extensive discussion consult Webb, 1979; Shank and Vetter, 1981)
ment, the kidney. Other, much more complicated systems have been
derived (Shank and Vetter, 1981).
Cadmium is attached to metallothionein in both the liver and kidney,
and cadmium complexed in this manner does not cause tubular necrosis
unless the concentration in the kidney cortex exceeds 300-400 pg g-' (i.e.
overall average concentration 200 pg g-'). Identification of the binding
protein has allowed the development of immunoassay techniques for the
determination of cadmium thionein in the urine (Tohyama, Shaikh,
Nogawa, Kobayashi, and Houda, 1982; Vander Mallie and Garvey,
1980; Brady and Kofka, 1979; Chang, Vander Mallie, and Garvey, 1979).
It is now established that when cadmium causes necrosis in the kidney,
proteins, including cadmium thionein, are excreted in the urine.
The determination of soluble cytoplasmic enzymes released into the
circulation from dying cells, e.g. lactic dehydrogenase, sorbitol de-
hydrogenase, or the transaminases can be used for the detection of cell
necrosis.
The importance of measurements of indicators causally related or
proportional to the disease process cannot be over-emphasised. Dose-
response relationships can be compared with the reaction with the
primary target (if this is known) or with the dose-response of the whole
animal. Comparison of the relationships in different species with, in
addition, the pharmacokinetic behaviour of the chemical, can provide
valuable information and may lead to the possibility of measurement of
such processes in human tissue and thereby allow an estimate of risk of
exposure to man.
3 .O
1.0
h
c
I
0.5 cn
- n
I
Y
or
0 2.5 E
Y U
0 Q,
v)
€
U
0 0
0
QI d
v) b
0 *
m 0
*
:I-0.5 0
0
d
-.
2.0
-1 .o
1.6 2 .o 2.4 2.8 3.2

l o g i n d u c t i o n time (days)
Figure 9 Chronic administration in the drinking water of diethylnitrosamine to
rats. During the course of the experiment in the seven groups receiving the highest
doses, all rats developed tumours. In the two groups receiving 0.15 and
0.075mgkg-' day-' 27/30 and 517 developed tumours by the end of the
experiment
7 Dose-Response Relationships
When making decisions about the safe use of a compound, it is vital to
have information on whether there is a dose threshold, below which no
toxicity is found, even with prolonged and continuous exposure.
Chemical carcinogenesis presents this problem in its most acute form,
if the concept that a single lesion (one hit) on DNA is all that is necessary
to initiate the growth of a tumour is accepted. With the increasing
information of the many steps which may be involved in the carcinogenic
process, the hypothesis is perhaps less certain; nevertheless, results are
available which indicate that a safe dose of some chemicals, even if it
exists, must be very small (Figure 9; Schmahl, 1979; Hoel, Kaplan, and
Anderson, 1983).
Knowledge of the primary target sometimes allows a fundamental
rather than an empirical approach to the solution of the problem as to
whether a threshold dose exists. Measurement of the degree of reaction
W. N . Aldridge 75
-s 100 12.5mg kg-l daily

'. 1
'\
"\
'
'\
\
0
; 40
I
3
Q,
t
.-t 20
a
I
m
I I I 1 I
0 2 4 6 8 10
Time (weeks)
Figure 10 Mono-2-tolyl diphenylphosphate and delayed neuropathy in the hen.
The solid lines indicate the activity of brain neurotoxic esterase as a percentage of
control and the dashed lines the relation between dose and clinical condition.
Similar results are reported for spinal cord.
(Reproduced by permission from Lotti and Johnson, 1980, Arch. Toxicol., 45,
263-271 .)
of the chemical with the target indicates, under all circumstances, the
amount of chemical delivered to its site of action. For example, it was
known that a single dose of DFP caused a 70-80% inhibition, by
phosphorylation, of the 'neurotoxic esterase' in the nervous system of
chickens and that this was followed by the development of permanent
neuropathy. It was not known if repeated exposure, to a smaller dose
leading to a maintained but lower inhibition, would gradually build up
damage until clinical symptoms would appear. The experiment shown in
Figure 10 demonstrates that, in the chicken, inhibition may be maintained
at 50% for 10 weeks with no signs ofneuropathy (Lotti and Johnson, 1980).
Increase in the dosage administered led to symptoms, showing that the
animals do not become tolerant. This experiment answers the question in a
fundamental way because delivery of the toxic chemical to the target
has been established-inhibition of the neurotoxic esterase shows that the
target has been reached. It seems, therefore, that 50% inhibition of
neurotoxic esterase by the mechanism shown in Figure 7 can be tolerated
indefinitely. To transfer this decision to man requires assumptions about
the relationship between the degree of inhibition and toxicity, but it
seems very likely that a threshold does exist in man as it does in the hen.
Opportunities may be found to acquire information in man after
accidental exposure since the same or a similar esterase to 'neurotoxic
esterase' is found in lymphocytes.
For an extensive discussion of thresholds and the principles involved in

their determination, see Aldridge (1985). One way of bridging the
animal-man gap has recently been demonstrated (Lotti and Johnson,
1978). Both acetylcholinesterase and neurotoxic esterase are inhibited in
vitro by many organophosphorus compounds and they must be close to
each other in the cell since they sediment in the same subcellular fraction. It
is impossible to predict even the acute toxicity from the rate constants for
the reaction with acetylcholinesterase because the rates of disposal by
metabolism may exert a dominating influence (cf p. 67). If, however, the
compound is delivered to the two enzymes, acetylcholinesterase and
neurotoxic esterase, at the same concentration, then the following
relationship should hold.
when: LD50 is the dose for 50% lethality (acute toxicity).

ED,, is the effective dose to produce delayed neuropathy in 50%
of the animals.
D,,(AChE) and D,,(NTE) is the dose required in vivo to cause
50% inhibition of acetylcholinesterase and neurotoxic esterase,
respectively.
k,(NTE) and k,(AChE) is the bimolecular rate constant for
inhibition in uitro of neurotoxic esterase and acetylcholin-
esterase, respectively.
I,,(AChE) and I,,(NTE) is the concentration necessary for 50%
inhibition in uitro (both under the same defined conditions) of
acetylcholinesterase and neurotoxic esterase, respectively.
It has been shown that this relationship does hold in chickens (Lotti and
Johnson, 1978). Thus it is possible to determine the ratio of the in vitro
inhibitory potency for direct inhibitors of these enzymes using postmor-
tem human tissue. From this ratio we can predict the potential for
causing delayed neuropathy compared with acute toxicity. As a rough
guide, when a ratio greater than 0.05 is obtained, the compound should
be suspect.
8 Measurement of Received Dose of Reactive

Chemicals
The dose of a chemical to an animal or man is usually expressed relative
to body weight. For experimental studies this is an accurate statement
although there is often much doubt about the amount of a toxic chemical
that is absorbed by different routes of administration. Measurement of
exposure of man in a working environment is extremely difficult; analyses
of the concentration in the air are difficult to relate to intake by
individual workers.
W. N . Aldridge 77
A new approach has been made for electrophilic chemicals which may
be genotoxic. Measurement of the extent of the reaction of such
electrophiles with haemoglobin in viuo allows the internal exposure to be
derived (Ehrenberg and Husain, 1981). Haemoglobin is a suitable protein
because it is available in high concentration and has a relatively long life
in viuo. Thus measurement of, for example, alkylated residues in
haemoglobin can provide an integral of the dose received over a
relatively long previous exposure. For epidemiological studies, for risk
assessment, and for the identification of dangerous working practices, this
is valuable since the measurement is made on individual workers and is
not some vague assessment of average exposure.
Analytical techniques are developing so fast that detection of a low
degree of reaction with, for example, cysteine, histidine, valine, or lysine
are becoming possible (Farmer, Gorf, and Bailey, 1982; Farmer, Bailey,
Lamb, and Connors, 1980). From the known reaction rates with these
residues, determined in uitro (biomolecular rate constants), the inte-
grated dose (concentration x time) can be readily calculated.
Since reaction with DNA is directly related to the dose of some
carcinogens (Weiland and Neumann, 1978; Gangler and Neumann, 1979;
Gangler, Neumann, Saubner, and Scribner, 1979) it seems likely that is
will be possible for some compounds to calibrate the reaction of
haemoglobin with that of DNA.
There are, of course, many problems to solve, and particularly for
chemicals which are bioactivated to electrophiles, but this work illustrates
the use of modern developments in analysis (separative method, mass
spectrometry, etc.) to move forward to more reliable estimates of
exposure (biological monitoring). It is an accepted principle that the
toxicity of a chemical will be directly related to the concentration that
reaches the target. The approaches described above are leading towards
an estimate of that concentration. It is also important to appreciate that
this advance is due to the detailed study of reactions with a macro-
molecule which has no direct relevance to the toxicity of the chemical.
9 Conclusions
The ideal of the toxicologist is to be able to predict the toxicity of a
chemical by looking at its chemical structure. Knowledge of the biochem-
ical principles of the many factors influencing toxicity and the mechan-
isms whereby the deleterious biological consequences are initiated allow
a much more certain prediction of the type(s) of toxicity expected from a
particular molecule. Quantitative prediction of toxicity (expected dose-
response) is still difficult. But those working in biological research should
not despair; prediction of the expected rate of a chemical reaction from
the structure of the reactants is just as difficult.
In this chapter examples of biochemical principles, some accepted and
some less well established, have been discussed. Perhaps the most
important biochemical principle is that we must strive to ask molecular

questions about toxicity (see DeMatteis and Lock, 1987). The answers,
as they move from suggestions to hypotheses and concepts, will provide
the framework for the assessment of hazard and the safe use of
chemicals.
References
Albert, A. (1985). ‘Selective Toxicity’. The physicochemical basis of therapy.
Chapman and Hall, London, pp. 1-750.
Aldridge, W. N. (1980). The need to understand mechanisms. In: ‘The Scientific
Basis of Toxicity Assessment’, Ed. H. Witschi. Elsevier/North-Holland
Biomedical Press, Amsterdam. pp. 305-319.
Aldridge, W. N. (1981). Mechanisms of toxicity. New concepts are required in
toxicology. Trends Pharmacol. Sci., 2, 228-231.
Aldridge, W. N. (1985). The biological basis of thresholds. Annu. Rev.
Pharmacol. Toxicol., 26, 39-58.
Aldridge, W. N. (1987). Selective toxicity and specific interactions. In: ‘Seminars
of Toxicity Mechanisms’ Vol. 1. Eds K. Hashimoto and M. Minami. Center
of Academic Publications Japan, Tokyo, pp. 9-27.
Aldridge, W. N. and Lovatt Evans, C. (1946). The physiological effects and fate
of cyanogen chloride. Q. J. Exp. Physiol., 33, 241-266.
Aldridge, W. N., Miles, J. W., Mount, D. L., and Verschoyle, R. D. (1979). The
toxicological properties of impurities in malathion. Arch. Toxicol., 42,
95- 106.
Aldridge, W. N. and Reiner, E. (1971). ‘Enzyme Inhibitors as Substrates’.
Interaction of esterases with esters of organophosphorus and carbonic acids.
North Holland Publishing Co., Amsterdam, pp. 1-328.
Aldridge, W. N., Street, B. W., and Skilleter, D. N. (1977). Oxidative
phosphorylation. Halide dependent and halide independent effects of tri-
organotin and triorganolead compounds on mitochondria1 function.
Biochem. J., 168, 353-364.
Aldridge, W. N., Verschoyle, R. D., and Peal, J. A. (1984). O,S,S-Trimethyl-
phosphorodit hioate and 0,S , S-Triethylphosphorothioate: pharmacokinetics
in rats and effect of pretreatment with compounds affecting the drug
processing systems. Pestic. Biochem. Physiol., 21, 265-274.
Anders, M. W. ( E d . ) (1985). ‘Bioactivation of foreign compounds’. Academic
Press, Orlando, pp, 1-555.
Barnes, J. M. and Magee, P. N. (1954). Some toxic properties of dimethylnitros-
amine. Br. J. Ind. Med., 11, 167-174.
Biempica, L., Morecki, R., Wu, C. H., Giambrone, M. A., and Rojkind, A.
(1980). Immunochemical localisation of type B collagen. Am. J. Puthol., 98,
591-602.
Brady, F. 0. and Kafka, R. L. (1979). Radioimmunoassay of rat liver
metallothionein. Anal. Biochem., 98, 89-94.
Brown, A , W., Aldridge, W. N., Street, B. W., and Verschoyle, R. D. (1979).
The behavioural and neuropatholagical sequelae of intoxication by trimethyl
tin compounds in the rat. Am. J . Pathol., 97, 59-82.
Cameron, G. R. and Karunaratne, W. A. E. (1936). Carbon tetrachloride
cirrhosis in relation to liver regeneration. J . Pathol. Bacteriol., 42, 1-21.
W. N . Aldridge 79
Chang, C. C., Vander Mallie, R. J., and Garvey, J. S. (1980). A radio

immunoassay for human metallothionin. Toxic01 Appl. Pharmacol., 55,
94-1 02.
Cremer, J. E. (1959). Biochemical studies on the toxicity of tetraethyllead and
other organolead compounds. Br. J . Ind. Med., 16, 191-1 99.
Dekant, W., Vamvakas, S., and Anders, M. W. (1992). The kidney as a target for
xenobiotics bioactivated by glutathione conjugation. In: ‘Tissue specific
toxicity: biochemical mechanisms’, Eds W. Dekant and H.-G. Neumann.
Academic Press, London, pp. 163-1 94.
DeMatteis, F., Gibbs, A. H., Farmer, P. B., Lamb, J. H., and Hollands, C.
(1982). Liver Haem as a target for drug toxicity. In: ‘Advances in
Pharmacology and Therapeutics II’, Vol. 5. Eds H. Yoshida, Y. Hagihara,
and S. Ebashi. Pergamon Press, Oxford, pp. 131-138.
DeMatteis, F. and Lock, E. A. (Eds) (1987). ‘Selectivity and Molecular
Mechanisms of Toxicology’. Macmillan Press, Ltd., Basingstoke, UK,
pp. 1-302.
Divincenzo, G. D., Krasavage, W. J., and O’Donoghue, J. L. (1980). Role of
metabolism in hexacarbon neuropathy. In: ‘The Scientific Basis of Toxicity
Assessment’, Ed. M. Witschi. Elsevier North-Holland Biomedical Press,
Amsterdam, pp, 183-200.
Donninger, C. (1971). Species specificity of phosphate triester anticholin-
esterases. Bull. W. H. O . , 44, 265-268.
Edgar, J. A., Smith, L. N., and Culvenor, C. C. J. (1970). Metabolic conversion
of heliotridine-based pyrrolizidine alkaloids to dehydroheliotridine. Mol.
Pharmacol., 6, 402-406.
Ehrenberg, L. and Husain, S. (1981). Genetic toxicity of some important
epoxides. Mutat. Res., 86, 1-113.
Eto, M., Casida, J. E. and Eto, T. (1962). Hydroxylation and cyclisation reaction
involved in the metabolism of tri-o-cresyl phosphate. Biochem. Pharmacol.,
11, 337-352.
Farmer, P. B., Gorf, S. M., and Bailey, E. (1982). Determination of hydroxy-
propyl histidine in haemoglobin as a measure of exposure to propylene
oxide using high resolution gas chromatography-mass spectrometry.
Biomed. Mass Spec., 9, 69-71.
Farmer, P. B., Bailey, E., Lamb, J. H., and Connors, T. A. (1980). Approach to
the quantitation of alkylated amino acids in haemoglobin by gas
chromatography-mass spectrometry. Biomed. Mass Spec., 7, 41-46.
Gangler, B. J. and Neumann, H. G. (1979). The binding of metabolites formed
from aminostilbene derivatives to nucleic acids in the liver of rats. Chem
Biol. Interact., 24, 355-372.
Gangler, B. J., Neumann, H. G., Saubuer, N. K., and Scribner, J. D. (1979).
Identification of some products from the reaction of trans-4-aminostilbene
metabolites and nucleic acids in uivo. Chem. Biol. Interact., 27, 335-342.
Gillette, J. R. (1982). The problem of chemically reactive metabolites. Drug
Metab. Rev., 13, 941-961.
Goldstein, A,, Aronow, L., and Kalman, S. M. (1974). ‘Principles of Drug
Action. The Basis of Pharmacology.’ John Wiley and Son, New York, pp.
1-854.
Heath, D. F. and Vandekar, M. (1957). Some spontaneous reactions of
0,O-dimethyl-S-ethylthioethylphosphorothiolateand related compounds in
water and on storage and their effects on the toxicological properties of the
compounds. Biochem. J . , 67, 187-201.
Hoel, D. G . , Kaplan, N. L., and Anderson, M. W. (1983). Implication of non-
linear kinetics on risk estimation in carcinogenesis. Science, 219, 1032-1037.
Johnson, M. K. (1975a). The delayed neuropathy caused by some organophos-
phorus esters: mechanism and challenge. Crit. Rev. Toxicol., 3, 289-3 16.
Johnson, M. K. (1975b). Organophosphorus esters causing delayed neurotoxic
effects. Mechanism of action and structure/activity studies. Arch. Toxicol.,
34, 259-288.
Johnson, M. K. (1982). The target for initiation of delayed neurotoxicity by
organophosphorus esters: biochemical studies and toxicological applications,
Rev. Biochem. Toxicol., 4, 141-212.
Johnson, M. K. and Safi, J. M. (1992). Organophosphoramidation of neuropathy
target esterase (NTE) is sufficient to initiate delayed neuropathy (DN)
without the necessity of an ‘ageing reaction’. Toxicologist, 12, Abstract 63.
Lijinsky, W., Loo, J., and Ross, A. E. (1968). Mechanism of alkylation of
nucleic acids by nitrosodimethylamine. Nature (London), 218, 1174-1175.
Lock, E. A. and Ishmael, J. (1979). The acute toxic effects of hexachloro-
1,3-butadiene on the rat kidney. Arch. Toxicol., 43, 47-57.
Lotti, M. and Johnson, M. K. (1978). Neurotoxicity of organophosphorus
pesticides: predictions can be based on in vitro studies with hen and human
enzymes. Arch. Toxicol., 41, 215-221.
Lotti, M. and Johnson, M. K. (1980). Repeated small doses of a neurotoxic
organophosphate. Monitoring of neurotoxic esterase in brain and spinal
cord. Arch. Toxicol., 45, 263-271.
Lotti, M., Caroldi, S., Capodicasa, E., and Moretto A. (1991). Promotion of
organophosphorus-induced delayed polyneuropathy by phenylmethanesul-
fonyl fluoride. Toxicol. Appl. Pharmacol., 108, 234-241.
Magee, P. N. (1956). Toxic liver injury: the metabolism of dimethylnitrosamine.
Biochem. J . , 64, 676-682.
Magee, P. N. and Barnes, J. M. (1956). The production of malignant primary
hepatic tumours in the rat by feeding dimethylnitrosamine. Br. J . Cancer, 10,
114- 122.
Magee, P. N. and Farber, E. (1962). Toxic liver injury and carcinogenicity:
methylation of rat liver nucleic acids by dimethylnitrosamine in vivo.
Biochem. J . , 83, 114-124.
Magos, L. (1981). Metabolic factors in the distribution and half time of mercury
after exposure to different mercurials. In : ‘Industrial and Environmental
Xenobiotics’, Eds I. Gut, M. Cikrt, and G. L. Plaa. Springer-Verlag, Berlin,
pp. 1-15.
Magos, L. (1982). Mercury induced nephrotoxicity. In: ‘Nephrotoxicity, Assess-
ment and Pathogenesis,’ Eds P. H. Bach, F. W. Bonner, J. W. Bridges, and
E. A. Lock. John Wiley and Sons, Chichester, pp. 325-337.
Mahler, W. (1954). The ionisation of certain bis-phenols. J . Am. Chem. SOC., 76,
3920-3921.
Martin, C. N. and Garner, R. C. (1977). Aflatoxin B,-oxide generated by
chemical or enzymic oxidation of aflatoxin B, causes guanine substitution in
nucleic acids. Nuture (London), 267, 863-865.
Mattocks, A. R. (1972). Toxicity and metabolism of Seneccio Alkaloids. In:
‘Phytochemical Ecology’, Ed. J. B. Harbourne. Academic Press, New York,
pp. 179-200.
Neal, R. A . (1980). Metabolism and mechanisms of toxicity of compounds
W. N. Aldridge 81
containing thiono-sulphur. In : ‘The Scientific Basis of Toxicity Assessment’,

Ed. H. Witschi. Elsevier/North-Holland Biomedical Press, Amsterdam,
pp. 241-250.
Rojkind, M. and Martinez-Palomo, A. (1976). Increase in type I and type I11
collagens in human alcoholic liver cirrhosis. Proc. Natl. Acad. Sci., 73,
539-543.
Roughton, F. J. W. and Darling, R. C. (1944). The effect of carbon monoxide on
the oxygen dissociation curve. Am. J . Physiol., 141, 17-31.
Scatchard, G. (1949). The attractions of proteins for small molecules and ions.
Ann. N.Y. Acad. Sci., 51, 660-672.
Schmahl, D. (1979). Problems of dose-response studies in chemical car-
cinogenesis with special references to N-nitroso compounds. CRC Crit. Rev.
Toxicol., 6,257-281.
Seyer, J. M. (1980). Interstitial collagen polymorphism in rat liver with
CC1,-induced cirrhosis. Biochim. Biophys. Acta, 629, 490-498.
Shank, K. E. and Vetter, R. J. (1981). Model description of cadmium transport
in a mammalian system. In: ‘Cadmium in the Environment’, Ed. J. 0.
Nriagu. John Wiley and Son, New York, pp. 583-616.
Smith, L. L. and Wyatt, I. (1981). The accumulation of putrescine into slices of
rat lung and brain and its relationship to the accumulation of paraquat.
Biochem. Pharmacol., 30, 1053-1058.
Swann, P. F. and McLean, A. E. M. (1968). The effect of diet on the toxic and
carcinogenic action of dimethylnitrosamine. Biochem. J . , 107, 14-15P.
Swensen, D. H., Miller, E. C., and Miller, J. A. (1974). Aflatoxin B1-2,3-oxide:
evidence for its formation in rat liver in vivo and by human liver microsomes
in vitro. Biochem. Biophys. Res. Commun., 60, 1036-1043.
Tobias, R. S. (1966). a-Bonded organometallic cations in aqueous solutions and
crystals. Organomet. Chem. Rev., 1, 93-129.
Tohyama, C., Shaikh, Z . A., Nogawa, K., Kobayashi, E., and Houda, R. (1982).
Urinary metallothionein as a new index of renal dysfunction in ‘Itai-Itai’
disease patients and other Japanese women environmentally exposed to
cadmium. Arch. Toxicol., 50, 159-166.
Van Bladeren, P. J. (1988). Formation of toxic metabolites from drugs and other
xenobiotics by glutathione conjugation. Trends Pharmacol. Sci., 9,295-299.
Vandekar, M. (1980). Minimizing occupational exposure to pesticides: cholines-
terase determination and organophosphorus poisoning. Residue Rev., 75,
67-79.
Vandekar, M. and Heath, D. F. (1957). The reactivation of cholinesterase after
inhibition in vivo by some dimethyl phosphate esters. Biochem. J . , 67,
202-208.
Vander Mallie, R. J. and Garvey, J. S. (1979). Radioimmunoassay of metallo-
thioneins. J . Biol. Chem., 254, 8416-8421.
Wainio, W. W. and Greenlees, J. (1960). Complexes of cytochrome C oxidase
with cyanide and carbon monoxide. Arch. Biochem. Biophys., 90, 18-21.
Webb, M. (Ed.) (1979). ‘The Chemistry, Biochemistry and Biology of Cadmium’.
Elsevier/North Holland Biomedical Press, Amsterdam, pp. 1-465.
Weiland, E. and Neumann, H. G. (1978). Methaemoglobin formation and
binding to blood constituents as indicators for the formation, availability and
reactivity of activated metabolites derived from trans-4-stilbene and related
aromatic amines. Arch. Toxicol., 40, 17-35.
Wyman, J. (1964). Linked functions and reciprocal effects in haemoglobin: a
second look. Adv. Protein Chem., 19, 223-286.
CHAPTER 6
Animal Husbandry
A. B . G. LANSDOWN
1 Introduction
Advances in medical research and in predictive safety evaluation will
continue to be dependent upon experiments in living animals for the
foreseeable future, despite considerable efforts to develop alternative
studies. Although a vast range of species and strains of laboratory animal
is available to the experimental toxicologist, the greater proportion of
controlled, reliable studies conducted for legislative purposes have
employed a comparatively narrow range of species. As a consequence?
the biology and animal husbandry for such animals as the Sprague-
Dawley rat, New Zealand white rabbit, or the beagle dog is well known
and their basic requirements under laboratory conditions appreciated.
In most aspects of experimental toxicology where sub-human species
are employed as ‘models’ for studying human disease processes or in
predicting the human response to toxic chemicals in the environment, the
accent will be on good laboratory practice and reproducibility of results.
As will be readily appreciated, both parameters are entirely dependent
upon appropriate technician training, correct experimental procedure,
and management of animal facilities and husbandry. Experience has
shown that subtle variations in the conditions in animal holding rooms,
diets, caging, and bedding, and inconsistencies in the genotype and
microbiological status of the test animals are potentially responsible for
interlaboratory differences in study results.
Animal husbandry is an immense subject and whereas ideas on ‘what is
acceptable’ will vary widely according to the country and the legislative
requirements in force at the time, the experimentalist has the ultimate
responsibility for satisfying such criteria in his animals of genetic purity,
microbiological cleanliness, and dietary adequacy. In Great Britain, the
Universities Federation for Animal Welfare (UFAW) the Royal Society
and the Medical Research Council (MRC) have set out guidelines for
animal husbandry, their recommendations being updated as new infor-
82
A . B . G. Lansdown 83
mation comes to hand. Practical guidelines on the use of living animals in

research by the Biological Council (1984) are useful reading for the new
research worker.
Clearly, the species and strain of laboratory animal selected for a
particular project will vary according to the scientific rationale of the
work and experience gained in previous studies. Animal models and their
suitability in studying environmental toxicology, immunogenicity, car-
cinogenicity, reproductive toxicity, and teratogenicity will be discussed in
the relevant chapters elsewhere in this volume. It is the purpose of the
present review to discuss some specific and more general aspects of the
husbandry of laboratory animals as they relate to experimental research.
2 The Research Animal

Experimental toxicology and other forms of research employing labora-
tory animals embodies scientific and humanitarian considerations, aspects
of animal welfare, and legal obligations (Biological Council, 1984). In the
light of intensive research into the replacement of animals in predictive
studies, the experimentalist is faced with such questions as: is this test
animal suitable for the work I have in mind, are the experiments ethical,
and what is the likelihood of me achieving my desired intention?
(Remfrey , 1984). The International Council for Laboratory Animal
Science (ICLAS) (1978) has noted that too many scientific investigations
fail to provide sufficient information on the animals used and on the
husbandry and experimental techniques. Clearly, such information is
indispensable for the correct and optimal interpretation of the results ,
and will be a prerequisite for the repetition of the work by other
investigators, as is so often the case where the results of studies are
contentious .
The policy of using the so-called ‘defined’ animal for research purposes
is becoming more evident as legislative authorities recognise the impor-
tance of control in regulatory studies. As will be emphasised later, the
microbiologically clean and genetically defined laboratory animal, main-
tained in a controlled environment by trained technicians supervised by a
competent curator, is a very superior animal (Festing, 1981). Such species
compare favourably with the pure chemical demanded by most research
chemists!
A Legal Aspects
The Cruelty to Animals Act was introduced into Great Britain in 1876 to
control experiments on living animals which were calculated to cause pain
(Advisory Committee on Animal Experiments, 1981). That Act has been
in force, largely unchanged until the issue of the 1986 Animal (Scientific
Procedures) Act. Internationally, the USA, Canada, and most countries
of the European Economic Community, have legislation which regulates
84 Animal Husbandry
animal experimentation. These regulations vary in their details but in most

cases, control the premises in which the work is conducted, the training
and the competence of the scientists and technicians, and the source,
transportation, and husbandry of the experimental animals (Ray and
Scott, 1973). It is clear that present legislation governing the use of animals
in some countries is broad and extends to farm animals and domestic
species. Occasionally, as in the case of certain exotic species, special
legislation such as the Dangerous Wild Animals Act (1976) (Cooper, 1978)
is passed. Conceivably, such legislation would apply to most species of
non-human primate used experimentally.
In Great Britain, between the issue of the 1876 Act and the
recommendations of the 1986 Act, there had been a marked surge in the
use of animals in predictive safety evaluation of medicines and
proprietary products. To some, this excessive testing of pharmaceutical
and related products represents unnecessary use of animals (Remfrey ,
1976). This led to a strong desire to reduce, refine, and possibly replace
the use of animal species in some experiments.
This thinking has been partly achieved in the 1986 Act, which in
addition to safeguarding the use of experimental animals, seeks to
monitor the rationale of the experimental procedure, the technical
competence of the investigator, and the husbandry, source of supply and
transportation of the animals to be used. The Act identifies the ‘protected
animal’ as all living vertebrates (other than humans) and embraces all
experiments employing foetal, larval, or embryonic forms which have
attained specific levels of development (Lansdown, 1992). An experiment
conducted within a designated establishment will be a ‘regulated pro-
cedure’. For present purposes, it is of interest to note that the 1986 Act
requires that animals used for regulated procedures should be supplied by
designated breeding establishments approved by the Home Office.
In the United States, a vast amount of experimental toxicology is
conducted under the auspices of the Department of Health and Human
Resources, and the National Toxicology Program established in 1978 to
co-ordinate and strengthen research and testing of a wide range of
substances present in the environment, food, and consumer products.
They have developed an integrated array of tests tailored to study
harmful effects for each substance, whilst at the same time attempting to
validate the tests and improve their sensitivity and reproducibility.
In the United States, much current toxicology involving animals
derives from the Food and Drug Administration (FDA) who stipulate
that the source of supply, species, and strain of experimental animal will
be recorded in all protocols in accordance with their codes of Good
Laboratory Practice (US FDA, 1978a). The FDA administer the USA
Food, Drugs and Cosmetics Act, the Public Health Service Act, and
Medical Device Amendments of 1976. Experimental laboratories are
inspected regularly and the conditions in animal rooms, food storage, and
experimental equipment checked. Special procedures are required in
A . B. G . Lansdown 85
respect of quarantine, isolation, parasite control and vaccination, and

experimental procedures. Such details as the source of animals used,
identification, allocation to experimental groups, housing (including cage
size, bedding, and number of animals per cage), environmental control,
and diet are recorded (US FDA, 1978b).
B Animal Supply
In 1947 the MRC established the Laboratory Animals Bureau in Great
Britain to monitor and control the quality and supply of laboratory
animals (MRC, 1974). A census conducted at the time showed that
animals were supplied by:
(a) Breeders who sold their own animals direct to laboratories
(b) Dealers who obtained animals from diverse sources and then
resold them for research purposes.
Although this last situation may still obtain in some countries, accredita-
tion schemes are more widely adopted nowadays for scientific reasons as
well as for the health and safety of the staff concerned (especially those
working with dogs, cats, and exotic species).
The production and supply of laboratory animals under an accredita-
tion scheme aims to protect an experimentalist from zoonotic diseases
and safeguard his in-house breeding and experimental animals. Accredi-
tation schemes monitor the health status and genetic purity of laboratory
animals. The first aspect relates to the hygiene and cleanliness of
breeding stocks, weaning, and adult animals pending their issue to
laboratories. Genetic purity will be of immense importance in research
programmes employing highly inbred strains exhibiting enzymic defects
or morphological abnormalities as in the scoliotic mouse (Ky/Ky) or
Brattleboro mutant of the Long Evans rat (vasopressin deficient) (Valtin,
1967).
The various species/strains of laboratory animals available for research
studies include:
(1) Those bred under conventional conditions, e.g. rat, mouse, rabbit,
guinea pig, and cat
(2) Larger species bred under more expansive conditions and possibly
requiring less rigidly controlled conditions, e.g. dog, ferret, sheep,
and pig
(3) Exotic species obtained from the wild state or bred in captivity in
tropical countries in which they are indigenous, e.g. macaques,
baboons, and other monkeys.
Limited success has been achieved in breeding marmosets in captivity in
Great Britain and Europe but these are expensive and difficult to obtain.
The broad aim of any accreditation scheme will be to raise the quality
of laboratory animals. Two main schemes are available to suppliers,
namely a control scheme applicable to those species classified in the first
category above, and secondly, a scheme which will be more relevant to
86 Animal Husbandry
breeders and suppliers of animals under groups 2 and 3. The first of these
requires that establishments be approved by the Animals Inspectorate of
the Home Office; whereas the second will require approval by the Home
Office and/or the Ministry of Agriculture, Fisheries and Food. Particular
problems do arise in the importation of animals from overseas where
quarantine regulations apply (see below).
Nowadays, recognised suppliers of laboratory animals commonly
provide certification of microbiological cleanliness and genetical purity
for their animals, particularly in the case of mice, rats, rabbits, and
guinea pigs which are intensively reared under defined barrier maintained
conditions. Genetical purity will be monitored periodically by recognised
techniques of serum protein assay, skin grafting, immunological markers,
and breeding performance. True inbred animals will be isogenic and thus
identical at more than 99% of their genetic loci segregated in the original
colony. They will be histocompatable and will not reject skin grafts from
other animals of the same strain (Festing, 1987). They are essential in
immunological experiments.
Suppliers of larger laboratory species will normally provide quality
assurance data and certification of good health, vaccinations (distemper,
parvovirus, etc. for dogs, and leukopenia in cats), and deworming
schedules. Veterinary certification of fitness to travel is normally required
for sheep, pigs, and goats. Monkeys and animals imorted from overseas
and potential cariers of zoonotic infections can be quarantined as
required, and screened for infections where a scientist does not have the
expertise or facilities available.
C Transportation
The principle objective in the transportation of animals is that they
should arrive at their destination in a similar condition to that at their
departure (Clough and Townsend, 1987). They should be subjected to
minimal stress and provided with caging, boxes, etc. which allow
acceptable environmental conditions, sufficient space for movement, and
access to food and water. Consideration should be given to any special
needs that a species may have in the course of a journey. In the case of
larger animals, it may be necessary to obtain veterinary advice before
moving an animal. When an animal has been subjected to a scientific
procedure, veterinary certification of fitness for transportation is obliga-
tory in Great Britain.
The welfare of animals in transit is subject to continual discussion and
governmental legislation (News and Reports, 1986; Gibson, Paterson,
and McGonville, 1986). Although much debate surrounds the transporta-
tion of larger species-horses, cattle, pigs, and poultry, many of the
recommendations apply to smaller species. Animal transportation is
discussed in detail and guidelines provided by the British Veterinary
Association and UFAW. It is noteworthy that, in Great Britain at least,
probably as many as 95% of all laboratory animals are transported from
breeders and suppliers to laboratories in appropriately ventilated and

equipped vehicles, as a service.
3 Experimental Animal Facilities

Each species of laboratory animal has been derived from the wild state at
some time and as such will retain some biological features of its wild
ancestors. The laboratory environment is far removed from the wild state
but good laboratory design and husbandry should attempt to mimic
natural conditions as far as is practicable. Rats and mice, for example,
are crepuscular species and tend to feed, mate, and move about during
twilight or dark hours. Animals exposed to bright light develop retino-
pathy. On the other hand, larger species including dogs and monkeys are
more active and require suitable exercise facility.
A Environment
Experimental protocols these days should specify the conditions "1
temperature, lighting, heating, and humidity under which animals will be
maintained during breeding and experimental procedures. Recommenda-
tions for common laboratory species are summarised in Table 1. It will be
appreciated that deviations from these conditions may constitute a stress
situation and contribute to alterations in feeding, drinking, reproductive
behaviour, and responses to toxicological situations. Consistency in
environmental conditions is a prerequisite in obtaining consistent and
reproducible results.
Laboratory animals should be allowed sufficient oxygen and not be
Table 1 Enuironmental Recommendations for Common Species of

Laboratory Animals
Room
temperature Humidity
("C) (%) Lighting
Mouse 20-22 50-60 12 h light: 12 h dark

Rat 20-22 50-60 12 h light: 12 h dark
Hamster 21-22 50-60 14 h light: 10 h dark
Gerbil 20-24 50-60 14 h light: 10 h dark
Rabbit 16-20 50-60 14 h light: 10 h dark
Guinea pig 18-22 45-70 12-16 h light day-'
Dog 15-21 50-60 12 h light: 12 h dark
Cat 21-23 50-60 12 h light: 12 h dark
Pig 16-18 60-70 12 h light: 12 h dark
Non-human primate:
marmoset 20-25 40-60 12 h light: 12 h dark
macaque 24 40-60 12 h light: 12 h dark
baboon 24 40-60 12 h light: 12 h dark
~
* Recommendations apply to adult animals. Breeding animals may require special

environments.
88 AnirnaE Husbandry
exposed to toxic or excessive levels of carbon dioxide, ammonia (from

degenerating excreta), or other pollutants. Air in animal rooms should be
filtered to remove microbiological contamination. Ventilation is thus an
important aspect of an animal house, at least 15-20 complete changes of
air per hour are recommended for fully stocked small animal rooms
(Clough, 1984). Additionally, the air in the immediate proximity of
animals should be regularly replenished; small animals like rats and mice
spend much of their time grouped in darker corners of cages and may be
subjected to higher concentrations of carbon dioxide and ammonia.
B Caging and Accommodation

Accommodation and caging for laboratory animals varies greatly from
one laboratory to another. The European Convention for the Protection
of Vertebrate Animals used for Experimental and other Scientific
Purposes (Council of Europe, 1986) has drawn up guidelines on this
subject but its recommendations are not mandatory and are used with
discretion. More recently, guidelines compiled by the Royal Society and
UFAW (1987) should enable scientists to achieve desirable standards of
animal welfare and husbandry.
By the European Convention (Council of Europe, 1986), any animal
used or intended for use in a procedure shall be provided with
accommodation, an environment, freedom of movement, food, water,
and care appropriate to its well being. These will vary according to the
age and anticipated use of an animal and the level of microbiological risks
involved. Broadly, the size, shape, and fitting of pens, cages, etc. will be
designed to accommodate the physiological and behavioural needs of a
species. It should allow space for expected body growth. Recommended
cage sizes are summarised in Table 2.
For small mammals, cages are constructed of galvanised iron, wood,
Table 2 Guidelines on Housing Laboratory Animals in Conventional

Conditions
Cage Animal weight
Height Floor area per animal range
Mouse 12 cm 60-1OOsq. cm >30 g
Rat 18-20 cm 100-400 sq. cm 50->550 g
Guinea pig 20-23 cm 200-750 sq. cm 150->650 g
Hamster 15 cm 89-154 sq. cm 60->120 g
Rabbits 40-45 cm 2000-6000 sq. cm 2-6 kg
Cats 50-80 cm 0.5-0.75 sq. m. >3 kg
Dogs 1.5-2.0 m 4.5-8 sq. m. 5->35 kg
Non-human primate:
marmoset 80 cm 0.25 sq. m. 0.025-0.65 kg
macaques 100-200 cm 0.6-2.8 sq. m. 6->6 kg
*Figures quoted are for animals caged in groups except in the case of dogs, rabbits, and
macaques.
stainless steel, and various plastics. Whatever the material used, it is
imperative that it be durable, readily sterilised, and not harmful to the
animals. Catches, locks, and restraints will be adequate to contain the
animals without risk to limbs or tails. Similar precautions are adopted in
the selection of grids used as cage floors. It is not uncommon in long term
studies, that large rodents develop ulcerative foot lesions through contact
with galvanised or rough metal grids.
Larger animals will be confined in pens, these will normally have hard,
non-slip floors which are readily flushed and sterilised. Where metal grids
are used in indoor accommodation, these should not cause injury to an
animal’s feet.
C Bedding
A variety of bedding materials including sawdust, straw, sand, and
synthetic materials is available for laboratory animals. To limit micro-
biological contamination, the materials selected should be sterile and
provide insulation, but be without injurious effect for the animals and
technicians responsible for their husbandry.
Bedding materials may facilitate the spread of infections within an
animal house. Sterilisation by formalin fumigation, hot air exposure,
irradiation, and autoclaving is recommended (Spiegel, 1965). Under all
circumstances, soiled bedding, faecal material, and animal waste should
be incinerated.
D Water
Water is a further source of infection in an animal house and should be
regularly monitored for impurities. Good laboratory practice requires
that water used in drinking water bottles should be checked and if mains
sources are defective, other sources should be used. Water bottles will be
regularly sterilised and checked for leaks.
Chlorination is widely used to sterilise water supplies. It is simple and
effective, and readily controlled. Cats will accept water containing
0.2-1.0 mg 1-l free chlorine without health risk. Most species of labora-
tory animal survive well on mains water supplies, although high levels of
lead or other heavy metals may be detrimental to health.
4 Diet and Nutrition

Growth, maturation, and maintenance of normal body function in man
and laboratory animals depends upon the availability of sufficient supplies
of food, and nutritional sufficiency. Ill-health may result from generalised
food deprivation or inadequacies in any of essential carbohydrates, fats,
amino-acids and proteins, vitamins and minerals. The literature on this
subject is vast but it is clear that different species of laboratory animals
90 Animal Husbandry
Table 3 Composition of Diets for Common Species of Laboratory
Animals
Rat/Mouse Rabbit Dog
b m b m b m G/pig Cat Primate
Ash (96) 5.5 5.7 7.2 6.411.3 9.5 7 1.8 6.1

PROTEIN 22 16.4 16.4 13.623.4 18.4 17.4 10.3 18.5
FAT/OIL 3.7 2.9 1.9 2.7 8.2 8.5 3.4 4.5 3.9
CARBOHYDRATE 48 56 . 50 54 4.7 52 50 8.4 54
STARCH 41 38 22 29 40 43 27 ? 34
FIBRE 2.1 5.6 13.2 10.8 2.3 3 8.8 ? 4.3
CALCIUM 1.05 0.89 1.0 0.75 1.1 0.96 1 0.5 0.81
POTASSIUM 1.00 0.63 0.77 0.910.59 0.47 1.10 0.3 0.93
SODIUM 0.29 0.28 0.31 0.360.49 0.37 0.47 0.2 0.36
PHOSPHATE 0.70 0.80 0.71 0.600.84 0.71 0.78 0.3 0.71
MAGNESIUM 0.15 0.16 0.17 0.240.15 0.14 0.42 ? 0.28
CHLORIDE 0.40 0.49 0.45 0.590.54 0.45 0.73 0.2 0.56
MANGANESE (mg) 50 77 60 0.81
33 26 63 ? 57
COPPER 14 13 14 14.711 9 98 ? 11
IRON 175 156 185 115 163 140 202 ? 91
ZINC 60 70 54 74 83 69 71 ? 71
IODINE 0.25 0.28 0.16 0.162.8 2.2 1 ? 2
Vitamin A (IU kg- ') 12000 15000 14000 20000
27000 16000 29000 13500 42000
D3 (1U kg-') 1200 2400 1400 14005500 3200 1400 440 4000
c (mgkg-') - - - - - - 3000 - 3000
E 40 91 6.4 73 90 50 72 40 90
B1 5.5 13 9 13 6 5 23 2.5 14
B2 6 14 12 15 7 5 31 5.5 15
B6 6 16 10 12 7 6 24 2.2 35
Niacin 50 42 38 37 38 31 77 8.8 151
Pantothenic acid 17 22 17 22 26 23 29 5.5 25
Vitamin B12 0.015 0.083 0.03 0.031 0.047 0.033 0.036 0.022 0.017
Biotin 0.2 0.27 0.30 0.29 0.32 0.17 0.30 ? 0.77
Vitamin K3 0.35 150 15 15 4.8 3.8 ? ? 15
Choline 1100 1100 1130 1045 1500 1400 1340 880 1410
vary in their basic requirements, a feature which becomes evident when

one examines the commercial diets available (Table 3).
Food deprivation and nutritional inadequacy can have varying effects
according to the age and physiological state of an animal. For example,
protein deprivation in a pregnant animal is a cause of maternal ill-health
and may result in foetal wastage and the birth of smaller than normal
offspring having an increased frequency of congenital cardiovascular,
immunological, and neurological abnormality (Widdowson, 1965; Lans-
down, 1977). In long term feeding trials, a 25% reduction in food intake
will reduce the growth rate in rodents but tends to enhance their
longevity and resistance to spontaneous tumours (Tucker, 1979; Berg,
1960). In contrast, overfeeding can lead to obesity, inactivity, and
reduced life span.
On occasions, laboratory animal diets may be nutritionally adequate
but animals exhibit evidence of malnutrition. This may arise in the case
of infection, genetical change, or metabolic defect which render the
animals incapable of digesting and assimilating dietary constituents. Thus
A . B. G. Lansdown 91
mice infected with Coxsackie B viruses became subject to exocrine

pancreatic insufficiency and incapable of digesting dietary protein (Lans-
down, 1976; 1975). Alternatively, animals treated with lead or cadmium
compounds tend to become deficient in zinc, copper, or iron, the heavy
metal ions inhibiting essential enzyme systems and metabolic pathways
for the trace metals (Landsown, 1983).
With improving standards of animal health and husbandry, there are
noticeable improvements in the quality of animal foods available. These
days, considerable attention is paid to the purity of nutritional con-
stituents (especially materials obtained from natural resources) , freedom
from infective organisms, storage, and product stability or shelf-life. The
diets available are formulated to satisfy the nutritional requirements of a
species, and are mostly packed in sealed bags indicating the date and
shelf-life. Where perishable foods are necessary, these will be stored in
refrigerators or cold rooms (Royal Society-UFAW, 1987). Otherwise
animal foods will be stored in dry, well ventilated facilities, free from
vermin and other pests.
From an experimental view, it will be appreciated that deviations and
inconsistencies in animal diets, particularly for rodents, can be a cause of
irregularities in serum enzyme levels, haematological profiles? and liver
and kidney function tests as are performed as a vital part of toxicological
studies (Clapp, 1980).
A Breeding and Reproduction

The first thing that influences the growth and development of an animal is
its size at birth and the number of surviving young in a litter, features
that invariably relate to a mother’s nutritional state (Widdowson, 1965).
In species like the rat, mouse, and pig in which the gestational length is
not appreciably altered by litter size, birth weight is a measure of the
food available to each offspring (McCance, 1962). The small immature
neonate has a greater surface area in proportion to its body weight than
at later stages, and consequently loses more heat energy. Its food
requirement is thus proportionately greater than that of larger litter
mates, but frequently its ability to compete for available food is less. The
low survival rate of the runt piglet is well documented.
Dietary requirements for the various species of laboratory animals
during reproductive and breeding phases is a large subject and the reader
is referred to reviews and proceedings of specialised symposia (Albanese,
1973; Hammell et al. 1976; Jones, 1976). Hurley (1977) for example, has
conducted many studies to examine the importance of various dietary
constituents in rodent pregnancies with particular reference to their effect
on the foetus. Although her work relates exclusively to the rat, it is clear
that imbalances in any essential nutrient will be detrimental in the foetal
development of any animal, or humans (Lansdown, 1983). Vitamin
deficiencies and prenatal growth defects are well understood generally,
92 Animal Husbandry
congenital neurological, optic, and skeletal defects resulting from excess

vitamin A in early pregnancy, skeletal and cardiovascular deformities in
hypervitaminosis D, and foetal wastage and infertility associated with
vitamin E deficiency are recorded (Lansdown, 1983).
Cattle fed sweet clover containing the vitamin K antagonist, dicouma-
rol, developed symptoms of vitamin K deficiency and calves born to them
died with haemorrhage. In other species, hypovitaminosis K is a known
cause of foetal wastage, blindness, and mental retardation.
Nutritional inadequacy will have a detrimental influence on any animal
in pregnancy and during lactation, the effect depending upon the species
and the stage in the reproductive cycle. In addition to its effect on
maternal health, undernutrition will often retard DNA-synthesis and
RNA and protein replication in the offspring and may predispose it to
diseases later in life (Roeder, 1973). Other effects may include adverse
changes in germ cell formation, reproductive behaviour, conception,
gestational development , parturition, and post-natal development (Wid-
dowson, 1965). The lactating ability of a dam will relate to her age and
physiological state, but will be a reflection on her nutritional state (Dean,
1951). Restriction of food after weaning may impair body growth, but if
not excessive, an infant may accomplish normal growth and development
with improved nutrition later. Detailed nutritional requirements for
breeding laboratory animals are discussed by UFAW (1987).
B Maintenance
Whilst the nutritional requirements of most laboratory animals are
similar, reflecting basic similarities in metabolic pathways , the sources
from which these nutrients are obtained varies (Ford, 1987). Natural
diets relate closely with an animal’s eating habits and growth patterns in
the wild state. These have become adapted and refined under laboratory
conditions such that, with rodents at least, good correlations exist
between food consumption and growth rate. Each strain of laboratory
mouse or rat exhibits a well defined growth curve, which can serve as a
reliable guide to an animal’s dietary state and health status.
For larger species of laboratory animals, laboratory diets more closely
resemble natural rations. Thus monkeys will receive fruit, nuts, and
vegetables freshly as a normal part of their diet. These are supplemented
with mineral and vitamin mixes. In dog and cat husbandry, fish or meat
rations will be supplemented with commercial food products where
necessary. In all cases, it will be important that these commercial rations
are designed to provide all nutrients not available in natural foods.
The guinea pig is unusual in that it is incapable of synthesising ascorbic
acid as obtained from green vegetables, and without this vitamin they
die. Commercial rations provide about 3000 mg kg-’ vitamin C, thus
alleviating the need to provide green vegetables in the laboratory (Table
3 ) . The cat has an unusually high requirement for vitamin A compared
A . B . G . Lansdown 93
with most other species. New world monkeys including tamarins and
marmosets require dietary supplements of Vitamin D, .
In summary, the nutritional requirements for laboratory animals are
ideally tailored to the individual demands of the species and its age and
physiological state.
C Feeding Patterns and Dietary Manipulation

Rats, mice, rabbits, guinea pigs, and most small mammals will be allowed
diet ad libitum under laboratory conditions unless the experimental
protocol dictates otherwise. However, it is common practice with larger
animals to restrict feeding to once or twice daily, allowing food intake to
be controlled. Under certain experimental conditions, it may be neces-
sary to limit food intake or vary the composition of the diet. Thus
‘pair-feeding’ is a common practice in experimental toxicology. Where a
particular treatment results in a diminution in food consumption in
experimental animals, the food intake of control animals is reduced to
eliminate this variable. In other situations, it is quite usual to remove
food (but not water) from experimental animals for up to 24 hours before
operations or terminal sacrifice. In rats and mice, food withdrawal for up
to 24 hours is not excessively stressful.
It is usual in toxicological studies to vary the constitution of laboratory
animal diets, either by replacing a natural constituent with a synthetic
one, or by placing test substances in the food. Either way, the change
may impair the acceptability of the ration, and its palatability or nutritive
value. These variables are best explored in preliminary experiments. A
further possibility to be considered is that alterations in the diet may
influence the intestinal microflora, mucosal physiology of absorptive
capacity of the gut. Thus alcohol is high in calorific value, it is readily
absorbed in the rodent intestine, but it impairs the regenerative activity
of the intesinal mucosa, induces changes in enzyme systems and reduces
the absorptive capacity of the tissue in the long term (Lansdown, 1987).
5 Animal Maintenance and Breeding

A Rats and Mice
The vast proportion of toxicological experiments are conducted using rats
or mice. These species are relatively inexpensive to maintain, they are
economical on space and their biology and husbandry is well known.
Although at least 300 different strains of inbred or outbred rodents are
available for investigations, the majority of studies reported employ a
relatively narrow range of common strains such as the CBA, TO, Balb/c,
or C3H mouse, or the Sprague-Dawley, Wistar, or Fischer F344 rat.
These strains breed well under laboratory conditions.
Solid-walled plastic cages with stainless steel mesh lids are preferred to
94 Animal Husbandry
older forms of rodent cages since they are secure, readily sterilised, and
convenient to house on wall or free-standing racking. Current trends are
towards standardisation of rodent cages (see Royal Society-UFAW
guidelines, 1987). As a general rule, rodent cages will be cleaned at least
twice weekly, sterilised, and animals placed in fresh cages with clean
bedding as appropriate. Dead animals will be removed early as rodents
are prone to cannibalism. Loss of a dead animal through cannibalism can
frustrate toxicological experiments where post-mortems will be conducted
routinely.
Rodents are social animals and where possible should be housed in
groups of 2-5. However, it may be necessary to house singly in metabolic
or dietary consumption studies. In animal husbandry, the psychological
influence of this and other stresses are of interest (Lawlor, 1984). Some
rodent strains are appreciably more nervous than others, particularly
inbred strains. Irregularities in animal husbandry here may have marked
influences on behavioural patterns in experimental procedures (Levine,
1960; Wells, 1975).
In another context, rodent housing has become the object of ethogram
studies. Species exhibit stereotyped movements associated with survival
within defined surroundings (Barnett, 1963). Locomotion, respiration,
feeding, nesting, fighting, and reproductive activity have been monitored
as a means of selecting suitable cages. They may be useful in comparing
the behavioural characteristics of different strains of rats and mice under
defined sets of conditions.
In rodent breeding, it is common to place animals in cages equipped
with stainless steel grid bottoms. Oestrus cycles vary from 4 to 6 days in
rats and mice and, in laboratory breeding, it is usual to arrange breeding
groups of 3 female animals with one male. Male animals placed in close
proximity to the female animals but in separate cages for 24 hours, will
be introduced into the main group at approximately 1600 hours on the
first day and removed at 0900h on the second. Successful mating is
indicated by the presence of sperm or copulation plugs in the vaginal
orifice. Females will then be separated to breeding cages containing
nesting materials and maintained in isolation through pregnancy, through
lactation, to weaning. The grid bottomed cages used in mating prevent
the loss of the vaginal plugs. In our laboratory, infant rats and mice will
be separated from mothers at 3 to 4 weeks post-partum. It will be
important in rodent breeding to keep accurate records to enable the
selection of breeding stock and to identify downward trends in breeding
performance. (Weihe, 1987; Cunliffe-Beamer and Les, 1987).
B Rabbits
It is more usual to maintain rabbits singly under laboratory conditions
using specially designated rooms. Rabbits are unusually sensitive to
changes in environmental conditions, noise, lighting, husbandry, and
even technicians. When animals are maintained on a large scale, it is

important that their husbandry and maintenance be conducted quietly
and methodically. As for rodents, a wide variety of rabbit caging is
available commercially, usually constructed of metal. Recommended
sizes vary from 40-45 cm height and 2000-6000 sq. cm floor area as a
minimum for rabbits weighing from 2 to over 6 kilogrammes, (ca. 1000
sq. cm floor kg body weight-'). It is usual for animals to be housed on
wire mesh grids (16 mm mesh) allowing faecal material and urine to be
disposed of in trays containing sawdust, sand, or other absorbent
substance at regular intervals. Where animals are maintained in large
numbers, hygiene is greatly improved when banks of cages are equipped
with automatic watering and cleaning devices.
Rabbits are obligate vegetarians and as such their incisors are subject to
continual growth. This will require regular checks by trained technicians
who will trim their teeth from time to time to avoid excessive over-
growth. Nails will also require clipping periodically. A further problem
encountered in rabbit colonies is bezoar or hair ball formation. The
pelage is generally thicker than in most other laboratory species and hair
invariably contaminates food hoppers. When this is consumed, hair balls
develop in the oesophagus leading to obstruction and death.
As rabbits are temperamental animals, they may be difficult to breed in
the conventional laboratory even though the rabbit is a recognised
species for reproductive and teratological studies (Gibson, Staples, and
Newberne, 1966). Unlike rats and mice, rabbits ovulate at copulation,
the egg is well developed and post-fertilisation development is rapid
(Adams, 1970). Pregnancy is 31-32 days and litter sizes vary from 2-12
depending upon the age and strain of rabbit studied.
Conditions allowed for breeding rabbits are such as to allow a nesting
compartment, draft free environment, and freedom from disturbance.
After mating, the females are placed in cages with nest boxes containing
straw or nesting materials 3-4 days pre-term. Care is always necessary in
rabbit breeding that the young are not disturbed, and caging used is such
that the very immature young do not fall out of nest boxes or become
trampled upon by the mother. Newborn and young rabbits are also very
sensitive to hypothermia.
C Dogs
Dogs are extensively used in laboratory investigations these days and are
often selected as non-rodent species in regulatory toxicology studies.
Being a larger and more active species than most others used in
experimental work, dog pens are equipped with exercise areas. These
may be indoors or outdoors, depending on the situation of laboratories
and available space, but should exceed 4.5 sq. m per animal. Either way,
sleeping accommodation should be dry, clean, and suitably ventilated.
Commonly, dogs are housed singly but with an opportunity to see other
96 Animal Husbandry
animals. It is usual for dog areas to be delineated with metal bars and
with solid floors that enable regular flushing and cleansing. Straw and
woodshavings will be supplied as bedding material as an alternative to a
constructed bed of wood, plastic, or other material. Accommodation for
laboratory dogs in Great Britain and some other countries is legally
controlled (Institute for Laboratory Animal Resources, 1978; Canadian
Council for Animal Care, 1980-1984; Research Defense Society, 1979).
D Primate Species
Primate species used in laboratory studies present a number of special-
ised features in management not encountered with other species. They
are all regarded as exotic species and with few exceptions are bred and
obtained from sources in tropical or sub-tropical countries. As such, they
may be subject to zoonotic infections transmissible to man. Not surpris-
ingly, these animals are the object of numerous managerial publications
which detail safety measures and specific recommendations on husbandry
(Royal Society-UFAW, 1987; Medical Research Council, 1985; Medical
Research Council, 1976). Apart from ethical considerations, it is impor-
tant that an experimenter gains a good understanding of the biological
and behavioural needs of the species he wishes to use.
Current guidelines (Royal Society, 1987) insist that no monkey should
be housed in a cage which is less than twice an animals crown-rump
length in height. Monkeys are sociable species and usually benefit from
being housed so that they can see other animals. Except with marmosets,
it was not usual to house monkeys other than singly, in purpose designed
metal caging equipped with automatic water devices, grid floors to avoid
accumulation of faecal material, and a crush-back to facilitate husbandry,
anaesthesia, and dosing. Recently however, there have been moves in
Great Britain to persuade laboratories to provide exercise facilities and
multiple housing for large primates as a form of ‘environmental enrich-
ment’. Automatic cleaning of dropping trays is of obvious benefit as a
time-consuming and safety measure.
In most laboratories, work with primate species is normally undertaken
by technicians who are experienced in handling them and who are aware
of the safety requirements. Thus, in my laboratory, all staff entering
monkey rooms will wear safety clothing and observe the guidelines set
out by the MRC Simian Virus Committee (Vizoso, 1971-1972).
E Other Species
Cats, guinea pigs, pigs, and occasionally gerbils, hamsters, and sheep are
used as experimental animals. Of these species, cats only may require
special considerations in management and husbandry. Animals will be
purpose bred and of minimal disease status, maintained in a controlled
environment with solid floors, and with adult males and queens separ-
ated. They are social animals and may be housed in groups of 5-20 with
sufficient space to allow 0.5-0.75 sq. m per animal. Like dogs, cats do
need some facility to exercise as well as for clawing (Royal Society, 1987)
and climbing. Dirt trays are placed in indoor cat compounds. Cats do
need constant temperature under laboratory conditions, and floors
heated to 20°C are recommended. Technicians involved in cat husbandry
will be aware of the scrupulous cleanliness required at all times.
Guinea pigs are frequently used in immunological studies. They are not
usually bred in experimental laboratories because of their long gestation
and economical factors, but animals obtained from commercial suppliers
present few special difficulties in management or husbandry. They are
housed singly or in single sex groups, allowing up to 750 sq. cm per
animal depending upon its age and weight.
Pigs, sheep, and occasionally goats will be used in some experimental
studies for surgery or serum production. As farm animals, they may be
housed in the open or in indoor accommodation. In the latter case, grid
floors are preferred for ease of cleaning and hygiene. Areas of 2-7.5 sq.
m are allowed, depending upon an animals size. Good ventilation,
automatic watering, and exercise facilities are essential for indoor
accommodation. Regular weighing and food consumption monitoring
provides a good indication of an animal’s well-being, or otherwise.
6 Isolator Technology
In the past 25 years, isolators have become an established part of many
large experimental animal facilities. They have developed from simple
filtered-air animal boxes, fitted with long-sleeved rubber gloves for sterile
manipulation, to the more sophisticated forms capable of holding up to
200 animals. Larger varieties are available for large-scale animal breed-
ing. Isolator development has created an opportunity to produce large
numbers of germ-free or gnotobiotic animals, animals of a hypo-immune
status (the nude rats and mice), and to study the pathogenicity of
hazardous infections transmissible to man, or to other animals.
Isolators available these days are mostly of the flexible-film type
consisting of an envelope of plastic supported on a rigid metal frame.
Various isolators exist but they are essentially of two main types. There
are those designed for ease of operation, as in caesarian derivation of the
germ-free stock, microbiological sampling and isolations, and tumour
implantations. Secondly, there are the larger cuboidal or elongate
isolators which are preferred for animal breeding. This type will be fitted
with several pairs of long sleeved rubber gloves. Animal cages will be
mounted on racks to economise on isolator space. All modern isolators
are fitted with controlled air filtration, environmental monitoring devices,
internal illumination, and reliable power units. Air inlets function under
negative pressure, whereas exhausts will operate under positive flow and
be designed as to prevent a backflow of air should there be a power
failure.
Operation of isolators and the derivation of gnotobiotic animals is
described elsewhere (Trexler , 1987). Broadly, isolator technology re-
98 Animal Husbandry
quires that all materials used within the envelope will be germ-free
(except where specified organisms are introduced for study). This will
include the animals and all material involved in their husbandry and
experimental procedures. Thus food, bedding, instruments, and other
materials will be sterilised and prepacked before introduction to isolators
via dunk tanks or double door chambers. Hypochlorite or peracetic acid
are often used in the sterilisation process.
Commonly, gnotobiotic animals will be moved from one isolator to
another using ‘transporter isolators’ which couple to the double-door
portals on the main isolators. Sterile animal cages can then be passed
directly without exposure to the open environment.
Derivation of gnotobiotic stock is by caesarian section using sterile
procedures at near term. The pregnant uterus will be removed from the
mother under aseptic conditions and introduced into the isolator via a
sterile bath. Freshly isolated young will be fostered by a gnotobiotic dam
until weaning. A second method of deriving germ-free animals involves
the transfer of fertilised eggs or preimplantation embryos to a suitably
prepared recipient dam. Once a nucleus of breeding stock has been
established in the isolator, successive generations can be established in
other isolators or elsewhere.
Once a germ-free colony of animals has been established, it will be
essential to monitor the animals and the isolator environment periodically
for evidence of breakdown in the conditions. In event of contamination,
an isolator may require complete sterilisation and restocking with freshly
derived animals. In extreme cases, envelopes will be renewed. Most
isolators have a life of about 5 years before envelopes require renewal.
7 Quality Assurance
The need to maintain animals in good quality should be the aim of all
experimentalists. Not only will it be desirable to achieve microbiological
cleanliness, but quality control is essential in obtaining reproducible
results. Quality assurance is essential in the case of large animal breeders
aiming to supply animals of high standard, but it will be important in the
preservation of in-house stocks from microbiological and genetical
contamination. Contamination may occur through aerobic means, by
food and water, and through contact with other animals, or even humans.
Of particular concern will be organisms ‘imported’ with animals arriving
from overseas. In Great Britain at least, rigid quarantine requirements
are enforced. Genetic contamination is less of a problem for most
experimental toxicologists using highly outbred strains. However? to an
immunologist involved in transplant studies, and experimenters requiring
inbred strains, genetic ‘drift’ is a problem.
A Quarantine and Isolation

Quarantine, by definition, is the compulsory maintenance of animals
newly imported into a country in isolation for a minimum period of 6
months (Medical Research Council, 1976). In Great Britain, the accom-

modation is approved by the Ministry of Agriculture Fisheries and Food
acting under the Rabies Order (1974). Although quarantine applies to all
species, it is of greatest importance in the case of non-human primates.
The MRC Committee on Simian Viruses (1985) considers the isolation
period as a ‘conditioning phase’ during which animals will be observed by
a veterinarian. Rhesus monkeys, for example, should provide at least
three negative tuberculin tests and reactions for anergic infections.
To maintain the quality of animals in a laboratory, a curator may insist
that animals arriving in his laboratory, albeit from his own country, will
be isolated from in-house stocks. Some microbiological sampling is
essential. Quarantine rooms will be physically isolated from all other
animal holding rooms and, as far as possible, staff working there will not
move to other premises without a complete change of clothing. All
quarantine rooms should be subject to strict barrier conditions. Where a
laboratory does not have a suitable facility for quarantine, negative
pressure isolators are a convenient alternative.
B Microbiological Monitoring
In larger animal houses, regular microbiological screening of the air in
animal rooms, filters, bedding, and animals is a routine procedure. It will
form an essential part of an animal house engaged in breeding. This
monitoring will involve checks for viruses, bacteria, helminth infections,
and ectoparasites.
In the smaller laboratory, microbiological monitoring may be con-
ducted using sentinel animals which will be killed periodically and blood,
urine, faecal, and maybe throat swabs sampled. Larger animals will be
screened for infections when they arrive. Where worm burdens are
identified, these will be treated before animals are committed to
experimental rooms.
C Air Filters
In purpose designed animal rooms, quarantine rooms, and isolators,
ventilation will be through filters which are sufficient to exclude con-
tamination from the surrounding environment. These filters will be
changed regularly and checked microbiologically.
D Barrier Maintenance
In most modern laboratories, animals are bred and maintained under
barrier-managed conditions. These are designed as to allow high quality
specified pathogen free (SPF) animals to be available as recommended by
the MRC Laboratory Animal Centre guidelines (Townsend, 1969). The
form that the barrier takes will vary between laboratories but usually is
such as to limit the passage of personnel, and escaped rodents to
100 Animal Husbandry
contaminate the premises. Scrupulous cleanliness will obtain throughout
and rooms will be equipped with devices to exclude insects.
E Containment of Infection
However well an animal house is constructed, occasions will arise when
systems for preventing infection break down. Decontamination measures
in respect of Ectromelia, Shigella sp., Salmonella sp. etc., will vary, but
in event of a serious infection, all animals in a room will be killed and the
room fumigated. Formaldehyde used to be the fumigant of choice, but as
its toxicity has become widely publicised recently, less harmful agents like
Virkon (R&S. Biotech. Finedon, Northants) are preferred. Eaton (1987)
indicated that once the nature and gravity of an infection is known, the
curator is faced with the task of identifying the source and taking
appropriate preventive measures.
Latent infections will occasionally present difficulties in quality control.
These are a particular problem in laboratory monkeys such as Cynomol-
gus, Rhesus, and Vervet breeds. Animals may exhibit negligible serum
titres for Herpesvirus simiae (B-virus), hepatitis virus, or measles during
quarantining or under conventional laboratory conditions, but if the
animals are subjected to immunosuppressive treatments or even stress,
latent viral infections in spinal ganglia become active, and present a very
serious risk to handlers (Perkins and O’Donoghue, 1969; Vizoso,
1971-1972; Prior et al., 1964; World Health Organisation, 1971).
8 Experimental Procedures
A Dosing
Though it is possible to use any route or method of administration, the
use of injection is limited by the local tissue reactions provoked, and the
duration of such treatments may be limited. It is, however, possible to
carry out repeated and long term studies by subcutaneous injection if the
sites used are varied. This still depends, however, on the rapid clearance
of the material from the site of injection and usually will be confined to
water soluble chemicals not causing a local reaction.
Oral administration is by far the most common method for long term
studies. This may take the form of daily gastric intubation (gavage) or
incorporation of the material into the diet. It is possible to use daily
intubation even for life-time experiments and although it is time
consuming and expensive, where the procedure is carried out by suitably
experienced and sensitive operators there need be no unwarranted stress
to the animals or loss of experimental subjects due to technical errors. It
is, of course, essential that control animals are subject to the same
procedures. The method has the advantage of accuracy of dosage.
Mixture of the test chemical with the diet is the simplest and most
common method of administration, especially where large numbers of
A. B . G. Lansdown 101
animals and long periods of treatment are involved. It results in some loss
of accuracy of individual dosing but in most studies this is acceptable.
Individual housing with frequent monitoring of body weight and food
intake (at the extreme, daily) allows retrospective monitoring of the
dosage but this is seldom absolutely necessary.
When giving a chemical in the diet it is important to bear in mind the
changing ratio of food intake and body weight in the young animal. If a
fixed concentration of test material is fed to rats there may be a 3-4 fold
reduction in dosage (mg kg body weight-’) between post-weaning and
3-6 months later. The worst of these variations can be avoided by
appropriate modification of the dietary concentration.
In order to ensure adequate control of dosage during dietary ad-
ministration, analytical investigations are important to ensure that:
The dietary mixture is homogeneous
The dietary mixture is stable over the period required (e.g. there is no
settling of the test component or loss through volatility, and that there
is no chemical interaction between the diet and the test compound)
There is no selection by the animals, to avoid or preferentially select
the test material from the diet
Each batch of mixture prepared is of the required concentration.
The monitoring of the intake from the diet requires an accurate
measurement of food intake and is an essential part of any good
experimental design. As well as monitoring the dosage of the test
material, the recorded body weights and food intakes can be used as
guides to the condition of the animal. Changes of the food conversion
ratio whereby less weight is gained for the same food consumed may be a
useful measure of the toxic effect of a chemical.
In some instances, the presence of the test chemical reduces the
palatability of the diet. The test animals then fail to gain weight normally
because they eat less. In these circumstances it will be necessary to
conduct a ‘paired feeding assay’ in which the control animals are started
on trial one day later than the test animals, and are provided with
precisely the weight of diet consumed by the test animals the day before,
less, if necessary, the weight of the test material (usually negligible). It
may then be possible to compare the weight gained in relation to the diet
consumed. An adverse effect on palatability is not regarded as a toxic
effect.
The administration of the test substance in the drinking water (usually
provided ad libitum) adds to the complexity of the experiment in that
both food and water consumption need to be measured.
B Examination of Animals during Experimentation

Animals will be examined periodically throughout any experimental
procedure, and especially in toxicological studies. Records will include:
(a) Alterations in general appearance, physical and behavioural fa-
culties, feeding and water consumption
(b) Appearance of eyes (with or without ophthalmoscope)

(c) Skin and hair form
(d) Mouth, teeth, and throat (especially in larger species)
(e) Presence of palpable or overt masses (? tumours) and other lesions
(f) Evidence of infections, abscesses, sores, etc.
Additionally, experimental protocols and good laboratory practice will
normally specify sampling of blood, urine, and faeces to evaluate the
influence of test procedures. It is usual to collect samples at specific times
in the day to avoid changes due to diurnal rhythms
Blood sampling will depend upon the species involved. Thus for rats
and mice, bleeding from the tail vein, heart, or retro-orbital sinus is
usual. The posterior ear vein is the usual route in the rabbit. Cardiac
puncture is used in guinea pig bleeding. For larger species, cat, dog, pig,
goat etc., blood sampling for an experienced technician will usually be by
the jugular or femoral vein, or occasionally a more distal limb vein. The
marmoset is not an easy animal to bleed, but the femoral vein is most
suitable, with the animal being restrained in a body harness.
C Animal Identification and Randomisation

In good laboratory practice, it is usual to assign animals to experimental
groups by randomisation or by computerised methods. Such randomisa-
tion demands an unequivocal identification of each using colour coding,
tattooing, or by the use of neck or foot tags. It may be convenient in
large scale studies to use a colour coding scheme to mark individual diet
or treatment groups.
9 Records
In animal breeding and experimentation, a vast amount of numerical and
qualitative data will be accumulated. This will necessitate accurate
recording, assimilation, and storage pending report preparation. Most
larger laboratories and contract research companies will computerise this
information. Computer prints are admissible for legislative purposes.
In the foregoing, I have alluded to breeding and experimental
protocols. The former will be compiled to demonstrate:
(a) Progress and fluctuations in the breeding patterns of particular
strains and species
(b) Fecundity and the value of individual animals or pairs in a
breeding programme
(c) Rates of mortality and survival in various strains/species
(d) Seasonal variations in breeding performance.
This information is then used in the selection of breeding stock, forward
planning , evidence of genetic drift , breakdowns in environmental condi-
tions, etc. In a toxicological programme, computerised records and
statistical evaluations will be employed to identify changes in such
parameters as haematological profiles, body weights, food consumption,

and pathological lesions.
Certain records will be maintained by laboratories by legal obligation.
These will include:
Acquisition of animals, their age, sex, origin, and condition on
arrival
Record of animal health
Dates and forms of vaccination, treatment with anthelmintic drugs,
antibiotics, etc.
Fatalities and post-mortem observations
Emergencies-injuries to animals requiring veterinary supervision.
Discussion
The 1986 Animals (Scientific Procedures) Act in Great Britain provides a
realistic framework within which to use and protect all vertebrate animals
employed in regulated experiments. It sets out guidelines for the correct
treatment of various species and the responsibilities of those who handle
them. It is recognised that in a chapter of the present length, it is not
possible to discuss, other than briefly, the various aspects of husbandry,
nutrition, and quality assurance as expected under this new Act, the
reader is referred to publications by the UFAW, WHO (1971) and
Canadian Council on Animal Care (1980-1984) for more detailed
reading.
It is evident from the vast number of excellent publications issued
nowadays, that animal welfare, quality control, and monitoring of
infections and disease are essential parts of experimental toxicology and
animal research. The introduction of new technology in isolators,
quarantine facilities, and animal house design, coupled with higher levels
of technician training and expertise, enable us to perform our work with
improved accuracy and safety.
References
Adams, C. A. (1970). Ageing and reproduction in the female animal with
particular reference to the rabbit. J. Reprod. Fertil., Suppl., 12, 1.
Advisory Committee on Animal Experiments (1981). Report to the Secretary of
State on the framework of legislation to replace the Cruelty to Animals Act
1876. Home Office, HMSO, London.
Albanese, A. A. (1973). The effect of maternal nutrition on the development of
the offspring. Nutr. Rep. Int., 7 , 243.
Barnett, S. A. (1963). ‘A Study of Behaviour’, Methuen, London.
Berg, B. N. (1960). Nutrition and longevity in the rat. J . Nutr., 71, 242.
Biological Council (1984). Guidelines on the use of living animals in scientific
investigations. London.
Canadian Council for Animal Care (1980-1984). Guide to the care and use of
laboratory animals. Ottawa.
Clapp, M. J. L. (1980). The effect of diet on some parameters measured in
toxicological studies in the rat. Lab. Anim., 14, 253.
Clarke, H. E., Coates, M. E., Eva, J. K., Ford, D. J., Milner, C. K.,
O’Donoghue, P. N., Scott, R. P., and Ward, R. J. (1977). Dietary standards
of laboratory animals. In: ‘Report of the Laboratory Animals Centre, Diets
Advisory Committee’, Lab. Anim., 11, 1.
Clough, G. (1984). Environmental factors in relation to the comfort and well-
being of laboratory rats and mice. In: ‘Proceedings of Joint LASA-UFAW
Symposium: Standards in Laboratory Animal Management’. Pt. 1, p. 5.
Clough, G. and Townsend, G. H. (1987). In: ‘The Care and Management of
Laboratory Animals’. Ed. T. Poole, Longman Scientific and Technical, Bath,
p. 159.
Cooper, M. E. (1978). The Dangerous Wild Animals Act, 1976. Vet. Rec., 102,
475.
Council of Europe (1986). European Convention for the Protection of Vertebrate
Animals for Experimental and other Scientific Purposes. Strasbourg, HMSO,
London.
Cuncliffe-Beamer, Th. and Les, E. P. (1987). The laboratory mouse. In: ‘The
Care and Management of Laboratory Animals’. Ed. T. Poole, Longman
Scientific and Technical, Bath, p. 275.
Dean, R. F. A. (1951). The size of a baby at birth and the yield of breast milk.
In: ‘Studies on Undernutrition, Wuppertal, 1946-9’. MRC Special Report,
No. 275, 346, HMSO, London.
Eaton, P. (1987). Hygiene in the animal house. In: ‘The Care and Management
of Laboratory Animals’. Ed. T. Poole, Longman Scientific and Technical,
Bath, p. 275.
Festing, M. F. W. (1981). The ‘defined’ animal and the reduction of animal use.
In: ‘Animals in Research: New Perspectives in Animal Experimentation’.
Ed. D. Sperlinger, J. Wiley and Sons Ltd., Chichester, p. 285.
Festing, M. F. W. (1987). Introduction to animal genetics. In: ‘The Care and
Management of Laboratory Animals’, Ed. T. Poole, Longman Scientific and
Technical, Bath, p. 58.
Ford, D. J. (1987) Nutrition and feeding. In: ‘The Care and Management of
Laboratory Animals’, Ed. T. Poole, Longman Scientific and Technical, Bath,
p. 35.
Gibson, J. P., Staples, R. E., and Newberne, J. W. (1966). The use of the rabbit
in teratology studies. Toxicol. Appl. Pharmacol., 9, 398.
Gibson, T. E., Paterson, D. A., and McGonville, G. ( E d s ) (1986). ‘The Welfare
of Animals in Transit’, Proceedings of the Animal Welfare Foundation’s
Third Symposium. British Veterinary Association.
Hammell, D. L., Kratzer, D. D., Cromwell, G. L., and Hays, V. W. (1976).
Effect of protein malnutrition of the sow on reproductive performance and
on post-natal learning and performance in the offspring. J . Anim. Sci., 43,
589.
Hurley, L. S. (1977). Nutritional deficiencies and excesses. In: ‘The Handbook of
Teratology’, Eds. J. G. Wilson and F. C. Fraser, Plenum Press, New York,
1, 261.
Institute of Laboratory Animal Resources (1978). Guide for the use and care of
laboratory animals, US Department of Health, Education and Welfare ,
National Institutes of Health, Bethesda, Maryland.
International Council for Laboratory Animal Science (1978). ICLAS recommen-

dations for the specification of animals, the husbandry and techniques in
animal experimentation. ICLAS Bull., 42, 4.
Jones, D. G. (1976) The vulnerability of the developing brain to undernutrition.
Sci. Prog. Oxford, 63, 483.
Lansdown, A. B. G. (1975). Pathological changes in pregnant mice infected with
Coxsackievirus B3 and given dietary casein hydrolysate supplement. Br. J .
Exp. Pathol., 56, 373.
Lansdown, A. B. G . (1976) Pathological changes in the pancreas of mice
following infection with Coxsackie B viruses. Br. J . Exp. Pathol., 57, 331.
Lansdown, A. B. G. (1977) Histological observations on thymic development in
foetal and newborn mammals subject to interauterine growth retardation.
Biol. Neonat., 31, 252.
Lansdown, A. B. G. (1982). Teratogenicity and reduced fertility resulting from
factors present in food. In: ‘Toxic Hazards in Food’, Eds, D. M. Conning
and A. B. G. Lansdown, Croom-Helm, Beckenham, Kent, p. 73.
Lansdown, A. B. G. (1987). Alteration in the crypt cell population in the small
intestine as an early toxic response to sub-acute ethanol administration.
Arch. Toxicol., 59, 448.
Lansdown, A. B. G. (1992). The protected animal: an overview of the guidelines
controlling the health and well-being of laboratory animals in the United
Kingdom. Proceedings of the 8th Congress of the Federation of Asian
Veterinary Associations, Manila, Philippines.
Lawlor, M. (1984). Behavioural approaches to rodent management In: ‘Stand-
ards in Laboratory Animal Management’, UFAW, 1, 40,
Levine, S. (1960). Stimulation in infancy. Sci. A m . , May, 81.
McCance, R. A. (1962). Food, growth, and time. Lancet, 2, 621.
Medical Research Council (1974). The Accreditation and Recognition Schemes
for Suppliers of Laboratory Animals. Manual Series No 1.
Medical Research Council (1976). ‘Laboratory Non-Human Primates for Bio-
Medical Research in the United Kingdom’. MRC, London.
Medical Research Council (1985). The Management of Simians in Relation to
Infective Hazards to Staff, London.
News and Reports (1986). Welfare Problems in todays Noah’s Arks. Vet. Rec.,
119, 561.
Perkins, F. T. and O‘Donoghue, P. N. (1969). Hazards In Handling Simians,
Laboratory Animals Handbook, No. 4, Laboratory Animals Ltd., London.
Prior, J. E., Saur, R. M., and Fegley, H. C. (1964). Zoonoses associated with
laboratory monkeys. Lab. Anim. Care, 14, 48.
Ray, P. M. and Scott, W. M. (1973). Animal Welfare Legislation in the E. E. C.
Br. Vet. J . , 129, 194.
Remfrey, J. (1976). The control of animal experimentation. Das Tier im
Experiment. Ed. Herausgegeben Von Wolf H. Weihe, Verlag Hans Huber,
Bern.
Remfrey, J. (1984) Protecting experimental animals. Nature (London), 3l2, 191.
Research Defense Society (1979). Notes on the Law Relating to Experiments on
Animals in Great Britain, London. p. 15.
Roeder, L. M. (1973). Effect of level of nutrition on the rate of cell proliferation
and of RNA and protein synthesis in the rat. Nutr. Rep. Int. 7 , 271.
Royal Society-Universities Federation for Animal Welfare (1987). Laboratory
Animals and their use for Scientific Purposes. I. Housing and Care.
Wembley Press , London.
Spiegel, A. (1965). Sterilisation of food, bedding and other materials. Food
Cosmet. Toxicol., 3, 57.
Townsend, G. H. (1969). The grading of commercially available laboratory
animals. Vet. Rec., 85, 225.
Trexler, P. C. (1987). Animals of defined microbiological status. In: ‘The Care
and Management of Laboratory Animals’, Ed. T. Poole, Longman Scientific
and Technical, Bath, p. 85.
rodents. Int. J . Cancer, 23, 803.
Universities Federation for Animal Welfare (1980). Standards in Laboratory
Animal Care, Parts I and 11, UFAW, London.
Universities Federation for Animal Welfare (1987). ‘The Care and Management
of Laboratory Animals’, Ed. T. Poole, Longman Scientific and Technical,
Bath.
US Food and Drug Administration, (1978a). Laboratory Practice Regulations,
Department of Health, Education and Welfare. Washington, D.C.
US Food and Drug Administration (1978b). Compliance Program Guidance
Manual, Washington D.C., October 23rd.
Valtin, H. (1967). Hereditary hypothalmic diabetes insipidus in rats (Brattleboro
strain). A m . J . Med., 42, 814.
Vizoso, A. (1971-1972) Progress Report, 1967-1970. Annex to MRC Report,
London.
Weihe, W. H. (1987). The laboratory rat. In: ‘The Care and Management of
Laboratory Animals’. Ed. T. Poole, Longman Scientific and Technical, Bath.
Wells, P. A. (1975). The influence of early handling on the temporal sequence of
activity and exploratory behaviour of the rat. PhD Thesis, University of
London.
Widdowson, E. M. (1965). Factors affecting the growth rate of laboratory
animals. Food Cosmet. Toxicol., 3, 721.
World Health Organisation (1971). Health aspects of the supply and use of
non-human primates for biomedical purposes. WHO Technical Report
series, No. 470, Geneva.
CHAPTER 7
Inhala tion Toxic0logy

R. A. RILEY AND D. M. CONNING
1 Introduction
There are vast numbers of compounds in use throughout the world that
are capable of being inhaled into the respiratory system. The investiga-
tion of the adverse effects of such airborne materials constitutes
inhalation toxicology. This branch of toxicology is, more than any other,
dominated by the technology of exposure and by the special morphology
of the target organ (Leong, 1981; Salem, 1986). As a consequence, much
effort is needed to ensure that test exposures are consistent in physical
format and concentration, and that these parameters are sufficiently
controlled to ensure adequate penetration of the lung. The biological
consequences are very different where the test material is deposited at
the bronchiolar level, for example, rather than the alveolar. If penetra-
tion to the alveoli is achieved, absorption may occur, in which case the
toxicological evaluation is akin to any other study of systemic toxicity
with one notable, and sometimes important, distinction. This is that
absorption is effectively into the systemic circulation and not via the liver
as occurs with most (but not all) materials administered orally. Thus, the
major detoxifying organ may be effectively bypassed for a limited period
of time and this may have consequences not identifiable following
ingestion.
There is another aspect to inhalation toxicology that warrants special
attention. This is the close relationship with occupational hygiene. Many
of the techniques used to measure atmospheric concentrations are
common to both activities and, theoretically, it should be possible to
compare the human and animal reactions more closely for a given level of
exposure. In practice such comparisons are rarely achieved , usually
because the levels of, for example, industrial exposure are orders of
magnitude less than those generated in the laboratory.
This chapter deals with the technical requirements to ensure consistent
exposure of experimental animals to materials administered by inhala-
tion, the aim being usually to achieve alveolar penetration. We do not
dwell here on the range of pulmonary reactions that may occur.
107
108 Inhalation Toxicology
2 Physical Classification of Airborne Pollutants

The many thousands of compounds which have the potential to become
airborne, and therefore be inhaled, may be broadly classified according
to their physical properties, although obviously there will be some
overlap between groups. At the molecular level materials may be gases,
or substances in the gaseous state at normal temperature and pressure, or
they may be uapours, where the compound exists as a liquid or solid but
is translated at a steady rate into the surrounding environment. Particu-
late materials, either liquid or solid, when dispersed in air are termed
aerosols and are produced when the compound is reduced to a particle
size that allows it to become airborne. Atmospheres consisting of solid
particulate materials may be generated by grinding, crushing, impaction,
or degradation or may be naturally occurring, e.g. pollens. Fume, usually
an oxide, is produced as the result of heating the material to such a
temperature that a gas is evolved which then recondenses to solid
particles of a size less than 1micron. Fumes are generated, for example,
during welding processes.
Advances in technology have increased the number and amount of
chemicals in the atmosphere. It is estimated that there are over 50,000
different compounds in use today and this is increasing by around 1000
per annum. The health effects of inhalation of some of these chemicals
can be predicted to some extent by experimental investigation.
In order to simulate environmental conditions, a special technology has
evolved relating to the design and operation of inhalation chambers and
the generation and characterisation of aerosols and vapours, etc.
The costs associated with evaluation of toxicity by the inhalation route
are considerably higher than for studies using other routes of exposure,
essentially because of the cost of the specialised equipment. It is
estimated that an inhalation study costs two to three times that of a
comparable study by ingestion.
The procedures are those suitable for laboratories concerned with the
toxicity of chemicals which may contaminate industrial or agricultural
atmospheres or which may give rise to a community air pollution
problem. They may also find application in the study of the effects of
volatile or airborne drugs, of smokes, and of gases used in chemical
warfare.
3 Methodology of Inhalation Toxicology

The principle is to establish a dynamic environment in which the test
system, usually an experimental animal, is located in an airstream that
contains a fixed concentration of the test material in suitable format.
Static systems, in which the animal is placed in a container with a fixed
volume of air with the requisite concentration, are rarely used.
Where it is important to exclude exposure by ingestion, it is necessary
R. A . Riley and D.M. Conning 109
to arrange for individual nasal exposure only, to obviate grooming.

Where this is less important, groups of animals may be housed in a
chamber through which air is passed. The chamber must be of a design
that ensures uniformity of distribution of the test material in the air
stream, with the avoidance of turbulence which would allow local
variation in concentration. The air flow must be such as to ensure the
levels of water vapour emanating from the animals, of temperature, and
of oxygen are maintained within fairly narrow confines, depending on the
number of animals in the chamber. These constraints are more easily
controlled for gases and solid-state particles than for aerosols.
Inhalation chambers that are generally used are those of a basic
cuboidal design (Drew and Laskin, 1975; Hinners et al., 1968; Hemen-
way et al., 1982; Laskin and Drew, 1970; Leach et al., 1959; Hemenway
and MacAskill, 1982; Beethe et al., 1979; Doe and Tinston, 1981;
Macfarland, 1983). However, the ideal chamber to ensure uniform
distribution is essentially circular in design. It should be constructed of
materials which will not react chemically with the test material, should
allow good visibility for observation of the animals, and should be easily
decontaminated.
In studies involving small groups of animals a suitably modified
desiccator may conveniently be used. The animals, usually rats, mice, or
hamsters, are housed within the chamber supported on a metal grid
through which excreta may pass. Stainless-steel partitions can be inserted
to segregate the animals if required. Atmospheres are introduced to the
chamber down a central column under positive pressure. They then rise
through the areas where the animals are housed and finally exit through
holes positioned radially around the central input column. These output
holes also allow the insertion of probes for analysis of the test
atmosphere and the measurement of temperature and relative humidity.
Because these chambers operate under positive pressure they must, for
the safety of the investigators, sit within a ventilated space such as a fume
cupboard. According to good laboratory practice for inhalation toxicol-
ogy, the total space occupied by the animals in any chamber should not
exceed 5% of its internal capacity. Consequently, chambers of this design
limit the number of animals which may be exposed to a maximum weight
of 1kg.
Until recently, exposure chambers of cuboidal design capable of
housing larger groups of animals have been fitted with pyramidal top and
base units, and constructed of materials ranging from perspex to stainless
steel shell with glass windows. However, the chambers recently de-
veloped at BIBRA (Riley, 1986) are constructed entirely out of borosili-
cate glass (Figure 1). The wide range of glass components available make
construction of exposure chambers extremely flexible, both in terms of
chamber volume and in the addition of external components which may
be required for a particular study. Complete chambers are essentially
circular in shape and therefore provide excellent aerodynamic properties.
n N
A
-
I
t
2
'
! 3 -Input 1
view
<C
1
' Ports
-1Catch pans
*--Y-- --*---*
Extract
0
2sz Catch pan
Rubber /
stopper
Figure 1 Diagrammatic representation of a typical exposure chamber
Because they are constructed of glass they afford good all-round

visibility, they are inert to the atmospheres passing through them, and
they are easily cleaned. Atmospheres are introduced to the chambers
through a side-arm at the top, and this causes a swirling motion to ensure
complete uniformity of concentration throughout the chamber. After
passing through the chamber the atmospheres are exhausted under
negative pressure via in-line filtration systems. By careful off-balancing of
the exhaust and input air, a slight negative pressure of around 1ml water
gauge may be set and maintained; a water manometer connected to the
chamber gives a visible indication of this pressure. All joints between
components of the chamber are ground glass and consequently leaks
either inwards or outwards are minimised. Holes are provided around the
chamber to enable probes to be inserted to measure temperature, relative
humidity, and pressure and also to enable atmospheres to be drawn to a
suitable analytical system.
Animals are exposed in specially designed stainless-steel wire mesh
cages, each having the capacity to house 10 animals individually. The
cage units are supported on a metal frame and the chambers currently in
use each have the capacity to hold six of these units. Catch pans may be
inserted to prevent excreta falling onto animals positioned in lower tiers.
Where nasal exposure is required the individual animals are restrained in
R . A . Riley and D. M . Conning 111
tubes which allow only the nose or head to protrude. These tubes may be
plugged directly into the side of the exposure chamber.
The generation of test atmospheres of a calculated concentration is
determined by the introduction of known amounts of the test material
into a constant volume of air over a given period. For gases this is readily
achieved by using calibrated rotameters. The generation of vapours from
solvents may be approached by numerous methods , including bubbling
air through the liquid, diffusion of head space into the air stream, or by
atomisation of the liquid at a constant rate (Gage, 1970a and b). The
preferred method is atomisation, which can operate for lengthy periods
and avoids the problems that can arise if the test material contains
impurities and is unstable to continual aeration. The sample is introduced
into an atomiser from a syringe connected to an infusion pump. The
syringe size, injection rate, the air volume through the atomiser, and the
density and molecular weight of the material determine the actual
generated concentration.
Special difficulties attend the generation of atmospheres containing
liquid or solid aerosols. In addition to a known and constant concentra-
tion, the particle size distribution is important, as this will determine the
regions of the respiratory tract in which the aerosol is deposited, and thus
influence toxicity. Particles greater than 10 microns in aerodynamic
diameter tend to be trapped in the nasal cavity and main bronchi (Hatch
and Gross, 1964). They may cause local damage or may be absorbed
through the mucosa. Particles between 10 p m and 5 ,um achieve varying
degrees of penetration of the bronchial tree, the smallest likely to impact
in the terminal bronchioles. Particles less than 1 pm to 5 pm in diameter
will impact in the alveoli either by turbulence and velocity or by
Brownian movement. Particles smaller than 0.1 pm usually will not
impact but be carried out of the lung during exhalation.
The establishment of a known, constant concentration of a dust or mist
within a narrow size range presents technical difficulties. For solid
aerosols a generator has been developed in which the powder is
compressed into a pellet in a precision bore glass tube (Figure 2). When
installed in the generator the pellet is pushed at a pre-determined rate
into contact with a continually rotating knife edge. This action loosens
the top surface of the pellet and a negative pressure air stream removes
the particles. Before introduction into the animal chamber, the aerosol is
firstly passed through a cyclone which removes any particles greater than
30 pm and also assists in the separation of agglomerates which may have
formed during pelleting.
Atmospheric concentrations of liquid aerosols are somewhat easier to
prepare. Conventional atomisers are employed but these give a wide
range of droplet size. Again, as in the solid particle generation system,
coarse particles are removed by introduction into a cyclone (Gage, 1968).
The coarse particles are collected by impinging on the cyclone’s outer
surface whereas the finer ‘respirable’ particles remain in the air stream
which is directed into the exposure chamber.
ATOM I SER
( b r a s s ) 60rnrn long.
9 m m wide orifice 1 rnrn
between i n n e r & outer
BRASS COGS E & F

80 tooth. E being
connected permanently
FLOWMETER to pick u p tube C.
0 -50 l i t r e l r n i n diameter 50 rnm.
DUST P I C K UP T U B E C MOTOR
(s.s. 1 2 0 0 m m long 3 5 m m wide 4 0 m m h i g h
2 m m bore. 3 r n r n O.D.
FLOW METER
0-10 l i t r e l r n i n
DUST TUBE
KNIFE EDGE l O m r n O.D. 1 4 m m I.D.
(s.s. 1 l O r n m diarn. 2 0 0 r n m long (Glass)
height -body 6 mrn.
knife 2 m m .
ptfe p l u g
INFUSION P U M P R A M I
52 m m wide
i
I
Figure 2 Diagrammatical representation of dust feed mechanisms
It is excellent practice, and in most cases is now mandatory, to check

the concentration of the test atmosphere within the breathing zones of
the exposed animals. The monitoring of gases and vapours in general
present little difficulty-gas chromatography, infrared spectrometry, mass
spectrometry, and various specific analysers are readily available. How-
ever, it is preferable to check the analyser reading against a suitable
chemical method of analysis at least once per day. This will ensure that
the chosen instrument is standardised for measurement of the specific
atmosphere.
With aerosols the analysis is more complex as not only should the total
weight concentration be determined but more essentially the particle size
distribution, for reasons mentioned above. The total mass concentration
may be determined by drawing a measured volume through a membrane
filter and by analysing the amount collected either by direct weighing or
R. A. Riley and D. M . Conning 113
stage stage stage stage
:L
4 3 2 1
80 I
Stage 1 20
0
0
.5 1 2 5 10 20
Stage2 Particle diameter - microns
C al ib rat ion of cent r ipet e r for s p he r ical
particles of unit density
Stage3
Stage4
Section through a
centripeter stage
showing a i r flow lines
Figure 3 A typical cascade impactor
preferably by a specific chemical determination. The particle size

distribution should be measured using analysers which take into account
the aerodynamic properties of the material. For this reason it is desirable
to use a device such as a Cascade Impactor (Figure 3). This instrument
employs various stages in which jets are provided in decreasing diameters
(Hounam and Sherwood, 1965). As air is drawn through these stages at a
constant rate, the jet velocity increases. This causes particles of decreas-
ing size to be impacted on collection media positioned below each
descending stage in the instruments. Analysis of the amount collected is
again either by weight difference, or by a suitable chemical method,
When an atmosphere is introduced into a chamber, there is a delay
until the calculated concentration is reached. If good mixing occurs the
concentration c, at time t is given by the following expression, (Silver,
1946):
V c
n
where c is the calculated concentration, n is the air flow in litres per

minute, and V is the chamber volume in litres.
Eai
.-[r
t
0
.-
c
L
c
W
[r
t
0
0
2 6 10 14 18 22 26 30
Minutes into exposure

Figure 4 Means and standard deviations of nitrogen dioxide atmospheresfrom 40
sample points at one minute intervals
In simplified form the time taken to reach 99% of the desired

concentration is given by:
4.605 x V
t99 =
n
Figure 4 shows the rise to equilibrium of nitrogen dioxide atmospheres

in a typical BIBRA chamber. The mean values at 1minute intervals are
identical to the theoretical values calculated from the above equation,
and therefore an animal may be placed in any chamber position and will
receive the same concentration x time product of inhaled material as the
remainder of the exposed group.
4 Lung Clearance Mechanisms

The bronchiolar tree is equipped with ciliated cells that move the layer of
mucus, lining the lumen, continuously towards the epiglottis. Material
that is deposited on the bronchial and bronchiolar surface is cleared from
the lung by this means. Material deposited on the terminal bronchioles is
usually cleared within twenty-four hours.
R. A. Riley and D. M . Conning 115
Material that is deposited within the alveoli is cleared by macrophage

activity and transported to the local lymphatics where it is rapidly
conveyed to the bronchial lymph nodes. Inert particles may be cleared
from the alveoli in about five days. Where the particles are cytotoxic (e.g.
silica) clearance may take many days.
The rate of clearance may affect local toxicity quite considerably.
Numerous sensory endings are located at the surface of the respiratory
tract, from the nose to the alveolar level. They are easily stimulated by a
variety of chemicals and their stimulation is followed by a variety of
reflex responses. Such responses will result in bronchoconstriction which
will, in turn, further impede both inhalation and clearance. The
differential control of airway potency by such mechanisms is not fully
understood but where it is necessary to measure such effects, an estimate
of respiratory and expiratory resistance is required. In man and larger
animals this can be done with flow meters that measure the velocity of air
moving into and out of the lungs. For the measurement of respiratory
patterns in smaller mammals, a pressure sensor and plethysmograph are
required. Such a sensor (Figure 5) and body plethysmograph (Figure 6)
have been developed at BIBRA to measure the intensity of these reflex
responses. The sensor (UK Patent No. 2119932) will detect small
pressure changes in a plethysmograph which result from the inspiration
and expiration of the animal. These pressure changes cause a movement
of a permanent magnet located on a rubber tambour and changes in
magnetic flux density are detected by a fixed Hall Effect i.c. The output
signals, which are directly proportional to the tidal volume and respira-
tory rate of the animal, may be channelled to a suitable recording system
for evaluation.
' H a l l effect' I . C . mounted
on small P . C . B . directly
above magnet Brass bush to
guide bar magnet
8mm wide x 1.5mm thick

brass arm. Adjustable
Tambour rubber for height. 1" overall.
located by ' 0 ' - r i n g
Clamp n u t
Guide p i n to
keep arm upright
Main body
Figure 5 The BIBRA pressure sensor

Connection to pressure sensor
exposure chamber
--------
Figure 6 Body plethysmograph

R. A . Riley and D . M . Conning 117
In some circumstances, it may be necessary to more closely mimic

respiratory patterns adopted by man. Essentially, man is a facultative
mouth breather whereas most experimental animals, including primates,
are obligatory nose breathers. In the latter, much of the inhaled material
may be filtered out in the nasal cavity whereas in man the mechanism
may be bypassed. In this situation the dog may be a suitable animal.
Exposure can be by chamber or by a mask suitably equipped with a
mouth tube or by a means of obstructing the nostrils. One problem with
the dog is the well developed musculature surrounding the distal bronchi
and the residual numbers of intra-alveolar ducts. These allow much
greater distribution of material at the level of the alveolar spaces than
occurs in man. This is of greater importance with gaseous materials
where absorption may be enhanced. With particulate materials, including
aerosols, alveolar impaction is still the most important mechanism.
5 Conclusion
Irrespective of the species, the main aim of inhalation studies is to
generate and expose animals to consistent atmospheres over a defined
period. This requires a multi-disciplinary approach and in most instances
relates to the characteristics of the test article. Experiments of this nature
should be conducted with constant surveillance.
References
Beethe, R. L. et al. (1979). Evaluation of a recently designed multi-tiered
exposure chamber. Inhal. Toxicol. Res. Znst., LF-67.
Doe, J. E. and Tinston, D. (1981). Novel chambers for long term inhalation
studies. In: ‘Inhalation Toxicology and Technology’, Ed. B. K. J. Leong.
Ann Arbor, Michigan. Ann Arbor Science Publishers, pp. 77-88.
Drew, R. T. and Laskin, S. (1975). Environmental Inhalation Chambers. In:
‘Methods of Animal Experimentation’, Ed. W. I. Gray, Vol. 4. Academic
Press, New York. pp. 1-41.
Gage, J. C. (1968). Toxicity of paraquat and diquat aerosols generated by a size
selective cyclone: effect of particle size distribution. Br. J . Ind. Med., 16,
11-14.
Gage, J. C. (1970a). Experimental Inhalation Toxicology. In: ‘Methods in
Toxicology’, Ed. G. E. Paget, Blackwell Scientific Press, pp. 258-278.
Gage, J. C . (1970b). The sub-acute inhalation toxicity of 109 industrial chemicals.
Br. J . Ind. Med., 27, 1-18.
Hatch, T. F. and Gross, P. (1964). ‘Pulmonary Deposition and Retention of
Inhaled Aerosols’. Academic Press, New York.
Hemenway, D. R. et al. (1982). Inhalation toxicology chamber performance: a
quantitative model. Am. Ind. Hyg. Assoc. J . , 43, 120-127.
Hemenway, D. R. and MacAskill, S. M. (1982). Design, development and test
results of a horizontal flow inhalation facility. Am. Ind. Hyg. Assoc. J . , 43,
874-879.
Hinners, R. G. et al. (1986). Animal inhalation exposure chambers. Arch.

Environ. Health, 16, 194-206.
Hounam, R. F. and Sherwood, R. J. (1965). The cascade centripeter: a device for
determining the concentration and size distribution of aerosols. A m . Znd.
Hyg. ASSOC. J . , 26, 122-131.
Leach, L. J. et al. (1959). A multi-chamber exposure unit designed for chronic
inhalation studies. A m . Znd. Hyg. Assoc. J . , 20, 13-22.
Leong, B. K. J. (1980). ‘Proceedings of the Inhalation Toxicology and Technol-
ogy Symposium’. Kalamazoo, Michigan. Ann Arbor Scientific Publishers.
Macfarland, H. N. (1983). Designs and operational characteristics of inhalation
exposure equipment: a review. Fundam. Appl. Toxicol., 3 , 603-613.
Riley, R. A. (1986). A new approach to the construction of small inhalation
chambers: design and evaluation. A m . Ind. H y g . Assoc. J . , 47(3), 147-151,
Salem , H. (1986). ‘Inhalation Toxicology Research Methods, Applications and
Evaluation’. Marcel Dekker Inc., New York and Basel.
Silver, S. D. (1946). Constant flow gassing chambers: principles influencing
chamber design and operation. J . Lab. Clin. Med., 31, 1153-1161.
CHAPTER 8
Histopathology in Safety
Evaluation
J. G. EVANS AND W. H. BUTLER
1 Introduction
Histopathology has long been the most important of the investigations
undertaken for safety evaluation, because the morphological evidence of
a pathological process is the most consistent of the changes that can be
identified as the result of a toxic process (Butler, 1982). Consequently,
the practice of pathology in toxicological studies has become very
extensive and specialised. It is not practical in a chapter of this type to
cover the field in detail, so only the important principles are described
and exemplified by commonly encountered lesions. References for more
extensive reading are provided, should more specialised information be
required.
In essence, treatment may induce changes that would not otherwise
occur or result in increased severity or in the earlier occurrence of a
lesion that is normally found in the species or strain of animals being
used. This is of particular importance in studies designed to detect
chronic toxic effects and conducted over a substantial part of an animal’s
lifetime as the incidence of degenerative or neoplastic lesions is often
associated with ageing (Burek, 1978). The elucidation of such problems
requires careful attention to experimental detail and an awareness of the
disease activities likely to be encountered. The following account will
provide a basis for such expertise, particularly in respect of degenerative
or infective diseases. The issues with respect to neoplastic disease are
dealt with in Chapter 15.
2 The Role of Histopathology

Histopathological examinations of tissue form an important and time
consuming part of many safety evaluation studies. The examination is
dependent on the correct identification of lesions that may be present,
119
120 Histoputhology in Sufety Evuluution
and a proper appreciation of their significance. Therefore separation of

spontaneous from induced pathology is of major importance in the
majority of toxicological studies. It must be remembered, however, that
in certain instances treatment may not result in an additional, clearly
defined effect, but merely modify the severity or the extent of a disease
that occurs naturally.
Spontaneous diseases in laboratory animals may be divided into two
broad classes. The first are infectious diseases, resulting from viruses,
bacteria, richetsiae, or myoplasm; parasitic infestation may also be
included in this group. The second group are non-infectious diseases.
These are often age related and may be markedly influenced by genetic
factors and modified by dietary, hormonal, and animal husbandry
(Butler, 1982).
Infectious diseases are largely controlled, in modern animal units, by
good husbandry and management practice. Animals are now available
that are free from defined parasites and micro-organisms so that the
diseases that these agents cause do not form a major problem in
histological examination of tissue from toxicological studies (Clough,
1987). The range of infectious disease in laboratory animals has been
extensively documented (see bibliography), and will not be discussed in
detail. A number, however, which are of specific historical interest
still give rise to sporadic outbreaks of disease and these will be briefly
mentioned. The non-infectious diseases, in contrast, form an important
group for the pathologists practicing in toxicology. With this in mind
particular emphasis will be placed on non-neoplastic diseases as they are
commonly seen in the laboratory rat and mouse.
3 Examination of Animals and Tissue

Preparations
It cannot be emphasised too strongly that a proper systematic examina-
tion of animals at necropsy with an accurate description of all lesions is
an essential prequisite for a proper histopathological assessment. Failure
to do so at this stage may seriously compromise any conclusions that are
drawn from a study, or indeed may completely invalidate them. For this
reason staff performing these duties should be properly trained, and
supervised at the necropsy by the pathologist assigned to the study.
Selection of tissue is generally specified by the protocol and, unless
otherwise stated, covers samples of all tissues forming the animal’s body.
In particular, if electron microscopic examination is envisaged it is
important to design the protocol to ensure the optimum conditions for
such an examination. This may require the use of additional animals for
fixation by perfusion to obtain optimal results. Unless sufficient care is
taken over sampling and preservation of tissue for this purpose, any
conclusions may be strictly limited (Butler, 1982).
J . G. Evans and W . H . Butler 121
To prevent autolytic changes, tissues are generally ‘fixed’ to inhibit

degradative enzyme activity. Commonly used fixatives for both electron
and light microscopy are solutions of various aldehydes: formaldehyde
for light microscopy and glutaraldehyde for electron microscopy. The
fixatives themselves are made up as weak solutions (2-4%), buffered at
the pH and osmolality of body fluids (Hayat, 1981). To cut thin sections
of ‘fixed’ tissue it must be sufficiently rigid and so the tissue is dehydrated
through graded alcohols and then infiltrated and embedded in a suitable
material so that it can be cut. In the case of light microscopy this is
paraffin wax; and for electron microscopy an epoxy resin or araldite, or
possibly a mixture of the two, is used. Sections for light microscopy are
generally cut at 5 pm on a microtome fitted with a steel knife. For
electron microscopy, a diamond or glass knife is used and sections cut at
approximately 50 nm.
Sections in their natural state are colourless and show little contrast. In
order to differentiate the various cytological components, the tissue has
to be stained. In the case of electron microscopy the stains used are
heavy metals, such as lead or uranium salts, that bind differentially to the
various organelles and thus inhibit the passage of electrons to varying
degrees (Glauert, 1974). For light microscopy a wide variety of natural
and synthetic dyes are used. These are broadly classified as basic or acidic
dyes. The former are charged positively at the pH of the staining solution
and bind to acidic groups bearing negative charges, and the latter are
negatively charged and behave in the opposite manner. A typical basic
dye is toluidine blue which stains RNA and DNA. An example of an
acidic dye is acid fuchsin which stains the proteins of the cytoplasmic
matrix, mitochrondria, and smooth endoplasmic reticulum. Commonly
used dyes in routine histology are eosin (E) and haematoxylin (H). Cells
stained by H and E show a blue nucleus and a pink cytoplasm that may
contain clumps of blue or basophilic material (Bancroft and Stevens,
1977). In addition, a variety of stains have been developed to demonstr-
ate, in a more or less specific manner, various structural components of
the tissue. Examples of these are Van Gieson stain for connective tissue,
Feulgen’s method for DNA, and Periodic Schiff stain for glycol groups in
sugars. The last two methods fall more appropriately into the realm of
histochemistry and may require special methods of fixation. Histochemi-
cal methods also include the demonstration of enzyme activity in tissue
sections, The aim of enzyme histochemistry is to demonstrate the
presence of specific enzymes at their site of activity in life. To a certain
extent these two requirements, that of localisation and maintenance of
activity, are mutually exclusive. Some enzymes diffuse rapidly from their
initial site while others are particularly susceptible to the inhibitory
effects of fixation. For the latter, including majority of oxidative
enzymes, the tissue is rapidly frozen and sections cut in a cryostat at
-20°C. The section is then incubated to demonstrate the activity of the
enzyme prior to fixation. Others, such as the hydrolytic enzymes, may be
122 Histopathology in Safety Evaluation
subjected to controlled fixation in aldehydes at 4°C for a few hours,
following which the tissue is frozen and the sections cut.
The general principles of enzyme histochemistry are the same for all
enzymes and are typified by the method for glucose-6-phosphatase. This
enzyme catalyses the hydrolysis of glucose-6-phosphate in the following
way:
glucose-6-phosphatase+ glucose + phosphate
The method is dependent on the capture of the phosphate group by lead

ions that are included in the incubation medium. Lead phosphate
precipitates as a colourless deposit which is then treated with hydrogen
sulphide to form lead sulphide. This appears as a black or brown deposit
(Bancroft and Stevens, 1977).
Detailed accounts of histological and histochemical techniques may be
obtained from a number of texts referenced in the bibliography.
Immunocytochemical methods have become standard in routine
human diagnostic pathology using both immunofluorescent and immuno-
peroxidase techniques. The same techniques have been applied in safety
assessment but are not, at present, widely used (True, 1990).
Advances in molecular biology have led to the development of methods
that allow the in situ analysis of nucleic acids. Both messenger ribonucleic
acid and genomic deoxyribonucleic acid may be identified using these
techniques. Duplex strands of nucleic acid are separated by gentle heating
in the presence of labelled complimentary strands of nucleic acid. On
cooling, and in the right conditions, the labelled probe will hybridise with
the target sequence and can be visualised by the appropriate method. The
introduction of non-radiolabelled probes has meant that these techniques
may be used for routine diagnostic purposes. For example, biotin labelled
probes may be detected immunologically or by highly specific binding of
avidin or streptavidin. The latter may be labelled with a fluorochrome,
with colloidal gold, or with an enzyme so that they be seen with the light or
electron microscope (Burns and Magee, 1992).
The techniques of molecular biology that are applicable in diagnostic
histopathology are being continually upgraded, and to an extent simpli-
fied, and so may be used for investigative purposes in toxicological
pathology.
4 General Histopathological Terms

The structures of the developed body are derived from three germ layers
in the embryo: the ectoderm, mesoderm, and endoderm. The ectoderm
forms the skin and its appendage and much of the nervous system. The
mesoderm forms the supporting structures of the body-bone, muscle,
the connective tissues, and the vasculature and blood components. The
endoderm forms the gut and associated glands such as the liver, salivary
glands, and pancreas.
J. G. Evans and W. H . Butler 123
Clearly organs such as the liver, formed typically from endoderm, have
within their substance blood vessels and connective tissue that are of
mesodermal origin and thus the whole organ is of mixed embryonic
origin. In the case of the liver, the hepatocytes are arranged in a
glandular pattern around the vascular tree (Rappaport, Borowy,
Lougheed, and Lotto, 1954). The liver is the organ that has been most
extensively studied following both acute and chronic injury and many of
the concepts of cellular injury have been developed from the examination
of the reaction of hepatocytes (Farber and Fisher, 1979). It must be
emphasised that the response of cells to injury is limited and generally
recognised as a consequence of a disturbance in the ionic and water
balance within the cell. In addition, there may be the accumulation of
metabolic products such as free lipid or the breakdown products of
phospholipid membranes. These changes are often referred to as cellular
degenerations, and in the case of water balance are recognised in classical
pathology by varying degrees of cell swelling (La Via and Hill, 1971). In
the least severe of these, the cell loses its normal texture and has a faint
granular appearance. When the fluid accumulation is more pronounced,
discrete, fine droplets may be seen in the cell, while in the most severe
form extensive large vacuoles are formed displacing much of the normal
contents of the cell to the periphery.
It was once considered that these changes represented stages that would
eventually lead to the death of the cell, although this is now not generally
held. Indeed, cells exhibiting even the most advanced of these changes
are not inevitably committed to die, in fact, cell death may occur without
the cell passing through any of the stages associated with excess water
accumulation (Farber, 1982).
Dead or dying cells are recognised in living tissue by a series of
cytoplasmic and nuclear changes. Tissue recognised as such is said to be
necrotic and must be distinguished from the general autolytic changes
that occur after generalised body death. In the cytoplasm there is loss of
detail often associated with increased eosinophilia and shrinkage of the
cell. The nucleus may appear contracted and dense (pyknotic) or may be
broken into several pieces (karryorhexsis).
In addition to these direct manifestations of cellular injury, the body
may mount a series of tissue changes that are designed to limit the extent
of the injury or that may precede the stage of tissue repair. This
response, termed inflammation, is largely vascular and involves both the
blood vessels themselves and the cellular elements of the blood.
Inflammation has, by convention, been termed acute or chronic in
respect to the time for which the inflammatory change has been present.
In addition, certain histological features are associated with either an
acute or chronic reaction, although they are not totally exclusive. Thus
acute inflammation is largely vascular with the accumulation of fluid and
neutrophils at the site. Fluid accumulation is not a prominent feature of
chronic inflammation, the main histological feature being the presence of
mononuclear cells (macrophages and lymphocytes) and with the prolife-
124 Hktopathology in Safety Evaluation
ration of fibrous tissue and the formation of new blood vessels. Chronic
inflammation is also often associated with various stages of repair and
adaptation to tissue damage (Thomson, 1978).
The mechanism of the inflammatory reaction is complex and depend-
ent on the release of chemical mediators, both by damaged cells and by
the cells of the inflammatory response itself. The latter may indeed be
central to the maintenance of long-term recalcitrant chronic inflammatory
changes (Taussig , 1984).
5 Non-neoplastic Histopathological Changes in

Laboratory Animals
A wide variety of bacterial, virological, and parasitological diseases have
been described in laboratory animals. The majority of these are rarely
seen in practice as they have been eliminated by good animal husbandry
in well-maintained animal houses. This process begins at the phase of
animal production with breeding from specific, pathogen-free stock. This
is continued by barrier techniques in the animal house. In addition, many
of the diseases of the larger laboratory animals, such as the cat and dog,
are well controlled by vaccination policies and the common spontaneous
diseases of these species such as distemper, canine viral hepatitis,
leptospirosis, feline panleukopenia, and feline infectious peritonitis are
now more correctly in the realm of clinical veterinary pathology and
should be of little consequence in experimental toxicology. For the
majority of this section, therefore, the histopathological changes that may
confuse toxicological investigation will be confined to non-infectious
diseases. A number of infectious diseases, however, have been central to
the development of good laboratory practice and have in the past led to
conflicting findings. Among these are Tyzzer’s disease and mouse
hepatitis. The former is caused by the bacterium Bacillus piliformis, and
the latter occurs as a result of a virus infection. Both agents cause a focal
necrotising hepatitis and, although the former was first reported in the
mouse, it is known now to affect a wide range of species (Ganaway,
Allen, and Moore, 1971).
Other examples are Sendai virus, which in mice has been associated
with a necrotising bronchiolitis that may progress to a low-grade purulent
pneumonia; and Sialodacroadenitis, which is caused by another virus and
which results in necrosis within the salivary glands (Parker and Richter,
1982; Jacoby, Bhatt, and Jonas, 1979).
While the infectious diseases listed may cause problems from time to
time, the most confusing histopathological findings in laboratory animals
fall within the group of diseases generally described as constitutional and
are largely degenerative in nature and often coincident with old age. In
some instances these may affect almost 100% of animals and to be so
severe as to cause their death. Thus, recognition and diagnoses of these
diseases may be of considerable importance in for example, car-
J . G . Evans and W , H. Butler 125
cinogenicity studies as a judgement may be required as to the probable

cause of the animal’s death. In contrast, others may affect only a small
proportion of animals and are probably of little consequence in affecting
the overall mortality rate, although the incidence may be affected by
treatment.
Among the former is nephrosis, or chronic glomerulonephrosis, in the
aged rat. This is a severe degenerative disease of the kidney characterised
by glomerulonephrosis, the formation of hyaline casts, tubular atrophy,
and fibrosis (Gray, 1977). Associated with these structural changes are
severe metabolic disturbances, particularly of phosphorus and calcium,
Retention of phosphorus and loss of calcium through the kidney results in
hyperplasia of the parathyroid gland, which raises the calcium levels in
the serum and results in the deposition of calcium at various sites in the
body including the walls of blood vessels, the mucosa of the gastro-
intestinal tract, the lung, the heart, and the kidney itself. Other forms of
calcium deposits, not associated with chronic renal disease, may also be
encountered. One that may confuse toxicological interpretation is nephro-
calcinosis, which is characterised by deposition of accretions of calcium
in kidney tubules at the cortico-medullary junction. In some instances
only an occasional tubule may be affected and the amount of calcium
deposited is small. In other cases the deposit of calcium may be
extensive, resulting in necrosis with further deposits of calcium in the
dead tubular cells (dystrophilic calcification). The pathogenesis of this
lesion is poorly understood. The initial lesion may occur at a site distant
from the calcium deposit while the dietary balance of calcium, phos-
phorus, and magnesium, and the hormonal status of the animal play an
important contributory role (Glaister, 1986).
Cardiovascular disease is also common in old rats. This may be
associated with the chronic glomerulonephrosis previously described, and
result in extensive deposition of calcium salts in blood vessel walls and in
the heart. Areas of degeneration, necrosis, and fibrosis may also be seen
within the myocardium and this may eventually lead to myocardial
fibrosis. The latter may also occur in the absence of other degenerative
diseases and in this instance is predominantly in the immediate sub-
endocardia1 region and is characterised by the replacement of myocardial
fibres by an orderly arrangement of fibrous tissue. Again, the pathogene-
sis of the lesion is not understood. It has been suggested that the
myocardial necrosis results from a viral infection of anoxia, which induces
an abnormal response to physiological demand. Certain drugs, such as
isoproternol, can cause a similar pathological effect (Anvers and Cohen,
1979). Myxoid degeneration of the cardiac valves may also be seen in old
rats and an association between this condition and nephrosis has been
made.
Other parenchymatous organs in the rat are also susceptible to various
degenerative effects in old age. Thus, in the thyroid there may be varying
degrees of colloid loss and in some instances fibrosis. In addition, a
126 Hbtopathology in Safety Evaluation
diffuse and nodular hyperplasia of the C-cells is a feature that may be
confused with a benign tumour. The adrenal gland may also show areas
of hyperplasia and hypertrophy; the affected cells often being filled with
lipid (Burek, 1978). Testicular atrophy is occasionally seen, both in
young and old animals. There may be complete loss of germinal
epithelium or merely a relative decrease in cell numbers with the
formation of multinucleate giant cells from spermatids. Atrophy of the
ovary is a prominent feature in old female rats and may be of two
morphological forms: either the ovary appears shrunken and cystic with
loss of follicular tissue or there is overgrowth of granulosa-thecal cells
which may mimic a benign tumour of the granulosa cells (Glaister, 1986).
Mammary glands in old animals may occasionally show mild acinar
hyperplasia, sometimes with duct ectasia and sometimes with more
pronounced cystic development.
Paralysis of the hind limbs of ageing rats has been described in several
strains. Microscopically this is accompanied by demyelination and axon
degeneration and distention of the axon sheaths in the white matter of
the cord and the peripheral nerves. There is an associated atrophy of the
lumber and hind limb musculature. This condition is particularly promin-
ent in male ( W A G X BN)F1 rats over three years of age (Anvers and
Cohen, 1979).
Lymphoid tissue shows marked atrophic changes with age and reduc-
tion of germinal follicles is often associated with an abundance of
haemosiderin-containing macrophages. Some nodes may show cystic
dilatation of sinusoids and occasionally mesenteric lymph nodes may be
grossly distended. Splenic atrophy may mimic angioma in that loss of
lymphoid tissue leaves little other than the supporting structure. The
thymus also shows an age associated atrophy. This is characterised by a
loss of lymphocytes and a proliferation of the epithelial component. In
the most advanced form, little other than a fibro-vascular stroma
remains, often heavily infiltrated by adipose tissue.
Old mice may show similar degenerative conditions. In addition,
certain strains (e.g. strain A) are susceptible to the development of
deposits of amyloid in various tissues. Sites include the kidney, the
gastro-intestinal mucosa, and the heart. Amyloid itself is a complex
glycoprotein that has a characteristic periodicity when examined in the
electron microscope. Amyloid stains with Congo Red and shows charac-
teristic red-green birefringence under the polarising microsome. In the
kidney the deposit is most pronounced in the glomerular tuft, and in the
most severe form there may be extensive loss of protein through the
kidney with eventual death of the animal (Conner, Conner, Fox, and
Rogers, 1983).
The mouse shows marked strain and sex differences in histological
structure and disease patterns. Reference to the literature should be
made, as only a few specific examples will be mentioned here. In the
kidney of female animals, Bowman’s capsule has a flattened squamous
J . G. Evans and W. H. Butler 127
epithelium, whereas in the sexually mature male, the epithelium is

cuboidal in form. In mature males the terminal duct of the submaxillary
gland is formed of columnar cells filled with intensely eosinophilic
granules, while that of the female is formed of low-cuboidal cells lacking
granules. The adrenal gland of female mice contains more lipid than that
of the male, giving it a paler gross appearance. In the adrenal gland of
young mice there is a layer of cells contained between the cortex and
medulla and is called the X zone. This layer disappears rapidly in males
at the time of sexual maturity, and more slowly at the time of the first
pregnancy in females. If pregnancy does not intervene, the zone may
persist throughout the life of female animals (Dunn, 1965).
Small cystic structures are commonly found in the thyroid and pituitary
of old mice, while ectopic thymic tissue may be seen in the thyroid gland
of certain strains (e.g. BALB/c). The adrenal gland may show prolifera-
tion in the subcapsular region of spindle cells. These are most common in
female animals and may be increased by the presence of tumours.
Proliferation of lymphoreticular tissue is also prominent in aged mice and
great care is required in distinguishing benign proliferations from those
associated with malignancy.
The conditions described represent some of the common spontaneous
diseases of laboratory rats and mice. Many of them have underlying
genetic mechanisms, although it is the interaction of genetic and
environmental factors that determine their expression. It should be noted
that changes in dietary or husbandry procedures can markedly alter the
pattern of disease, irrespective of experimental treatment (Tucker, 1979).
Indeed, even if there is a treatment-related effect, this may operate
through an indirect rather than a direct mechanism. Thus, the induction
of strain may result in considerable changes in the hormonal status of the
animal which may markedly affect the pattern of disease seen.
References
Anvers, M. R. and Cohen, B. J. (1979). Lesions associated with aging. In ‘The
Laboratory Rat’ Eds H. J. Baker, J. R. Lindsey, and S. H. Weisbroth, Vol.
1, p. 377. Academic Press, New York.
Bancroft, J. D. and Stevens, A. ( E d s ) (1977). ‘Theory and Practice of
Histological Techniques’. Churchill-Livingstone, Edinburgh, London, and
New York.
Burek, J. D. (1978). ‘Pathology of Ageing Rats’. CRC Press, West Palm Beach.
Burns, J. and McGee, O’D. (1992). Tissue nucleic acid analysis. In ‘Oxford
Textbook of Pathology’ Eds O’D. McGee, P. G. Isaacson, and
N . G. Wright, pp. 2284-2288. Oxford University Press, Oxford, New York,
and Tokyo.
Butler, W. H. (1981). Histological control of spontaneous animal pathology. In
‘Animals in Toxicological Research’ Eds I. BartoSetc, A. Guaitani, and E.
Pacei, p. 71. Raven Press, New York.
128 Histopathology in Safety Evaluation
Clough, C. (1987). Quality in laboratory animals. In ‘Laboratory Animals’ Ed.

A. A. Tuffery, p. 79, A Wiley-Interscience publication, Chichester, New
York.
Conner, M. W., Conner, B. H., Fox, J. G. and Rogers, A. E. (1983).
Spontaneous amyloidosis in outbred CD-1 mice. Sum. Synth. Path. Res., 1,
67.
Dunn, T. B. (1965). Spontaneous lesions of mice. In ‘The Pathology of
Laboratory Animals’ Eds W. E. Ribelin and J. R. McCoy, p. 303. Charles C.
Thomas Publisher, Springfield, Illinois, USA.
Farber, E. and Fisher, M. M. (1979). ‘Toxic Injury of the Liver’. Part A. Marcel
Decker, New York and Basel.
Farber, J. L. (1982). Membrane injury and calcium homeostatis in the pathogen-
esis of coagulative necrosis. Lab. Invest., 47, 114.
Ganaway, J. R., Allen, A. M. and Moore, T. D. (1971). Tyzzer’s disease. A m . J .
Pathol., 64, 717.
Glaister, J. (1986). ‘Principles of Toxicological Pathology’. Taylor and Francis,
London and Philadelphia.
Glauert, A. M. (1974). ‘Practical Methods in Electron Microscopy’ Vol. 3 part 1.
Fixation, dehydration, and embedding of biological specimens. North
Holland Publishing Co., Amsterdam, Oxford.
Gray, J. E. (1977). Chronic progressive nephrosis in albino rat. CRC Crit. Rev.
Toxicol., 5 , 115.
Hayat, M. A. (1981). ‘Fixation for Electron Microsocopy’. Academic Press. New
York, London.
Jacoby, R. O., Batt, P. N. and Jonas, A, M. (1979). Viral diseases. In ‘The
Laboratory Rat’ Eds H. J. Baker, J. R. Lindsey, and S. H. Weisbroth Vol.
1, p. 272 Academic Press, New York.
LaVia, M. F. and Hill, R. B., Jr. (1975). ‘Principles of Pathobiology’. 2nd edn.
University Press, London, Oxford.
Parker, J. C. and Richter, C. D. (1982). Viral diseases of the respiratory system.
In ‘The Mouse in Biomedical Research’ Eds H. L. Foster, J. D. Small, and
J. G. Fox, Vol. 1, p, 28. Academic Press, New York.
Rappaport, A. M., Borowy, Z. J., Lougheed, W. M. and Lotto, W. N. (1954).
Subdivision of the hexagonal liver lobules into a structural and functional
units. Anat. Rec., 119, 11.
Taussig, M. J. (1984). ‘Processes in Pathology and Microbiology’. 2nd edn.
Blackwell Scientific Publications. Oxford, London.
Thomson, R. G . (1978). ‘General Veterinary Pathology’. W. B. Saunders Co.
Philadelphia, London, Toronto.
True, L. D. (1990). ‘Atlas of diagnostic immunohistopathology’. Ed. J. P.
Lippincott, Gower Medical, New York and London.
rodents. Int. J . Cancer, 23, 803.
J. G. Evans and W. H. Butler 129
Bibliography
Histological Methods
‘Manual of Histological Techniques’. J. D. Bancroft and H. C. Cook,
Churchill-Livingstone, Edinburgh, London, Melbourne, and New
York, 1984.
‘The Theory and Practice of Histological Techniques’. Eds. J. D.
Bancroft, and A. Stevens. Churchill-Livingstone, Edinburgh, Lon-
don, and New York, 1977.
General Pathology
‘An Introduction to General Pathology’. W. G. Spector, Churchill-
Livingstone, Edinburgh, London, New York 1977.
‘Processes in Pathology’. M. J. Taussig, Blackwell Scientific Publications,
Oxford, London, Edinburgh, 1979.
Veterinary and Laboratory Animal Pathology

‘Pathology of Laboratory Animals’, Vol. I and 11. eds K. Benirschhe, F.
M. Garner, and J. C. Jones. Springer-Verlag, New York, Heidel-
berg, Berlin, 1978.
‘Pathology of Ageing Rats’. J. D. Burek, CRC Press, Inc. 1978.
‘Pathology of Laboratory Animals’. Eds E. Cotchin, and F. J. C. Roe.
Blackwell Scientific Publications, Oxford and Edinburgh, 1967.
‘Laboratory Animal Medicine’. Eds J. G. Fox, B. J. Cohen, and F. M.
Loew. Academic Press Inc. (Harcourt Brace Jovanovich, Publishers)
New York, London. 1984.
‘Principles of Toxicological Pathology’. J. R. Glaister, Taylor and
Francis. London and Philadelphia, 1986.
‘Rat histopathology . A Glossary for use in toxicity and carcinogenicity
studies’. P. Greaves and J. M. Faccini, Elsevier Scientific, Amster-
dam, 1984.
‘Veterinary Pathology’. H. A. Smith and T. C. Jones, Lea and Feiberger,
Philadelphia.
CHAPTER 9
The Metabolism and

Disposition of Xenobiotics
S. D. GANGOLLI AND J. C . PHILLIPS
Introduction
Metabolism can refer to the sum of the chemical reactions that serve to
maintain life or to the chemical transformations effected by living systems
on foreign chemicals (Hayes, 1975). Pharmacokinetics is the discipline
which quantifies, as a function of time, the absorption, distribution,
metabolism, and excretion of chemicals in the body (Watanabe, Ramsey,
and Gehring, 1980). The WHO Scientific Group, in their recommenda-
tions on the procedures for investigating intentional and unintentional
food additives (WHO Scientific Group, 1967), recognised the importance
of metabolic and pharmacokinetic data in the safety evaluation of food
chemicals. This group recommended that, whenever possible, toxicity
studies should be conducted on an animal species most similar to man
with regard to its metabolic, biochemical, and toxicological characteristics
in relation to the test substance. Furthermore, the group considered that
at a relatively early stage there was a need to obtain information on the
metabolic fate of the test compound in experimental animals and in man.
A similar view was expressed by the Scientific Committee of the Food
Safety Council in their proposed Safety Decision Tree Scheme for food
safety assessment (Food Safety Council, 1978). The Committee recom-
mended that studies on metabolism and pharmacokinetics should be
undertaken immediately after acute toxicity has been determined, and
that information on the metabolic disposition of a compound should be
used both in the choice of animal model and in the definition of dose
regimes and other conditions relevant to the conduct of long-term animal
studies. The objectives of metabolic and pharmacokinetic studies as
defined by the Committee may be summarised as follows:
(1) To obtain information on the absorption, biotransformation, disposi-
tion, and excretion of a test substance after single or multiple doses
in experimental animals and in man.
130
S . D. Gangolli and J . C . Phillips 131
(2) To obtain a quantitative measure of these events at different dose

levels of the compound and to elucidate the conditions likely to stress
or overload the metabolic and excretory capacity of the experimental
animal.
In this chapter, we aim to provide a general review of metabolic and
excretory processes in animals and man. The sequence of the various
sections in this chapter is:
(1) Sites of absorption
(2) Distribution
(3) Metabolic reactions
(4) Sites of metabolism
(5) Bioactivation and detoxification processes
(6) Factors influencing metabolism and toxicity
In the following chapter, a brief review of the theoretical considera-
tions relevant to understanding the change in concentration of a
compound in blood and tissues following administration is given and
practical methods and techniques in metabolic studies discussed.
1 Sites of Absorption
The main sites of exposure and routes of entry of foreign chemicals into
the human body are the skin, lungs, and the gastro-intestinal tract.
Although there are special anatomical and physiological features charac-
terising and regulating the entry of substances from each of these sites,
there are certain common mechanisms applicable for the absorption of
chemicals across membranes and into the systemic circulation. These
‘uptake’ mechanisms fall broadly into two categories: (1) passive trans-
port, and (2) specialised transport systems.
A Passive Transport Systems

Two mechanisms mediate the passive transport of chemicals across
membranes. The first is simple diffusion and applies to very small,
hydrophilic, non-ionic compounds and gases capable of penetrating
through waterfilled pores (about 4 A radius) on the surface of mem-
branes. With passive transport, molecules pass across a concentration
gradient and the phenomenon is described by Fick’s first law of diffusion.
At dynamic equilibrium, the equation as applied to transport across a
single phase membrane (Davson, 1943) is:
KA(C, - Ci)
dD/dt =
X
where dD/dt is the rate of transport of a compound across a membrane, K
is a constant, A is the cross-sectional area of membrane exposed to the
compound, C, and Ci are the concentrations of the compound on the
132 The Metabolism and Disposition of Xenobiotics
outside and inside of the membrane respectively, and X is the thickness

of the membrane.
Examples of compounds that are absorbed by simple diffusion are the
gases (e.g. CO, C02, HCN, NO2) in the lung, the nerve gas, sarin, in the
skin, and urea in the gastro-intestinal tract.
The second mechanism involved in passive transport relates to the
ability of a compound to diffuse through both the aqueous and lipid
components of membranes. The degree of lipophilicity possessed by a
compound determines the facility with which it will be absorbed, and
therefore the relationship between the partition coefficient ( P ) of a
compound and its rate of permeation through membranes has been
extensively studied (see, for example review by Houston and Wood,
1980). When applied to intestinal transport, the absorption rate constant
of a compound ( K , ) is related to logP by:
log K , = - a(1og P ) 2 + b log P + c

where a, b, and c are the regression coefficients (Hansch and Clayton,
1973).
The measurement of partition coefficients by the somewhat tedious and
time-consuming two-solvent distribution technique (using n-octanol and
water) has been supplemented by the more elegant thin-layer chromatog-
raphy method developed by Biagi, Barbaro, Gamba, and Guerra (1969),
and the more recent high performance liquid chromatography method of
Unger, Cook, and Hollenberg (1978).
The lipophilicity of a compound is influenced by its degree of ionisation
in solution. This depends on the pK, of the compound and the pH of the
medium, the relationship being described by the Henderson-Hasselbalch
equations for acids and bases:
for an acid: pK, - pH = log (c,/ci)

for a base: pK, - pH = log(ci/c,)
where ci and c, are the concentrations of the ionised and unionised

compound , respectively.
Thus the degree of ionisation of weakly acidic organic compounds,
such as salicylic acid, benzoic acid, and the barbituric acids, is least in
the acidic conditions prevailing in the normal stomach and hence they are
readily absorbed from this site. Conversely, weakly basic compounds
(e.g. alkaloids, paracetamol, etc.) are more readily absorbed from the
small intestine. However, the lipophilicity, and hence the absorption, of
strongly ionised compounds can be increased by interactions with
counter-ions. Compounds capable of such ion-pair formation include
tetracyclins, quaternary ammonium compounds, and sulphonic acids.
S. D. Gangolli and J . C. Phillips 133
B Specialised Transport Systems

The two major processes classified as specialised transport are active
transport and carrier-mediated facilitated diffusion. Active transport
processes, involving uptake of compounds against a concentration
gradient using energy-dependent enzymic reactions, are largely respon-
sible for the absorption of essential components such as sugars, amino
acids, and nucleic acids. The translocation of these substances across
membranes involves their interaction with appropriate cell surface
receptors and the process, which requires the expenditure of energy, is
governed by the kinetics of enzymic reactions. These active transport
processes are rate limiting with respect to availability of substrate,
receptor sites, and ATP or other related sources of energy, and can be
inhibited by competitive substrates and metabolic inhibitors. Thus, for
example, the active transport of glucose is inhibited by the non-nutritive
analogue deoxyglucose, acting as a competitive substrate, and the uptake
of potassium by the sodium/potassium ATPase pump mechanism can be
reduced by the presence of rubidium (Tobin, Akera, Han, and Brody,
1974). Metabolic inhibitors, such as ouabain and tetrodotoxin can inhibit
active transport by interfering with Na/K ATPase or sodium uptake
(Ullrick, Capasso, Rumrich, and Sato, 1978; Kao, 1981).
Facilitated transport is the other specialised process in which carrier-
mediated mechanisms are involved. This process is not energy dependent
and does not operate against a concentration gradient. Examples of such
processes are the phlorizin-sensitive transport of sugars across mem-
branes and the transport of calcium ions by the calcium-binding protein,
calmodulin. The latter process can be inhibited by a number of organic
bases, such as the phenothiazines, acting as competitive substrates (Weiss
and Levin, 1978).
Other mechanisms of absorption include pinocytosis and phagocytosis,
whereby inert or high molecular weight substances are translocated intact
across membranes. Examples of such processes are the persorption of
macromolecular substances, such as degraded carrageenan from the small
intestinal lumen of the guinea pig (Sharratt, Grasso, Carpanini, and
Gangolli, 1970) and the uptake of horse-radish peroxidase from the rat
gut (Walker, Isselbacher, and Block, 1972).
Skin
Absorption of both lipophilic and hydrophilic compounds through the
intact skin is mediated primarily by passive diffusion mechanisms,
compounds being taken up principally by transepidermal diffusion.
However, diffusion via hair follicles and sebaceous glands can also occur.
The principles and methods for studying percutaneous absorption have
been reviewed by Barry (1983) and the ‘state-of-the-art’ in predicting skin
absorption recently summarised (Scott, Guy, and Hadgraft, 1990).
The Metabolism and Disposition of Xenobiotics
Lung
In the lung the mechanism for the absorption of inhaled gases is by
passive diffusion. Toxicants absorbed into the circulation from the lungs
include carbon monoxide, oxides of nitrogen and sulphur, and vapours of
industrial solvents such as chlorinated hydrocarbons. Particles of 1pm
diameter and below can penetrate the alveoli and gain access to the
circulation by phagocytosis.
Gastro-intestinal tract
Both passive and active carrier-mediated transport mechanisms effect the
absorption of compounds from the gastro-intestinal tract. Simple
diffusion allows the penetration of non-ionic hydrophilic substances such
as urea, and of anionic substances with hydration radii of less than 2.9A,
such as Br-, Cl- and NO;. Cations, on the other hand, require a
specialised carrier mediated transport system for absorption from the
gastro-intestinal tract. Lipophilic foreign compounds that are absorbed by
passive diffusion include drugs (barbiturates, paracetamol, etc.), food
additives (butylated hydroxyanisole and other antioxidants, benzoic acid,
and flavouring esters) and environmental chemicals (phthalic acid esters
and other industrial solvents and plasticisers, halogenated pesticides,
etc. ). Examples of compounds absorbed by active transport include
5-fluorouracil and some serine and threonine derivatives of nitrogen
mustards (Schanker and Jeffrey, 1961). In addition some dipeptides,
tripeptides, and oligopeptides are also actively absorbed from the small
intestine (Matthews, 1975; Das and Radhakrishnan, 1976).
The neonatal absorption of intact proteins and particulate material is
well documented. Although macromolecular absorption supposedly
ceases on the maturation of the epithelial cells, there is evidence showing
that adult animals continue to absorb biologically significant quantities of
large molecules and particulates (Le Fevre and Joel, 1977), probably by
pinocytosis. Examples of such compounds include colloidal silver, poly-
styrene latex, asbestos fibres, and starch particles.
2 Distribution
Following the absorption of a compound into the systemic circulation, it
may be distributed throughout the body to tissues, prior to metabolism
and excretion. Unless tissue distribution is very rapid, the blood can be
considered to be the ‘central compartment’ in a kinetic description of the
uptake, distribution, and excretion of a compound from the body (see the
following chapter). The rate of removal of a compound from the blood is
dependent on a number of factors, including blood flow through the
various organs and the extent of binding of the compound to tissue and/or
plasma proteins. The pharmacokinetic consequences of drug-protein
S . D . Gangolli and J . C. Phillips 135
interactions have been discussed (Jusko and Gretch, 1976). The plasma
concentration of a compound does not , therefore, necessarily reflect the
concentration in any particular tissue, and in many instances, tissues have
been found to have a specific affinity for a compound. This may explain
in part tissue specific injury, and frequently results in prolonged retention
of compounds in the body. Thus, paraquat accumulates preferentially in
the lungs, and at early times after administration, the concentration in
this organ increases as the plasma concentration falls (Smith et al., 1978).
Similarly, lipophilic compounds such as hexabromobiphenyl, DDT, and
dioxins accumulate in skin and adipose tissue, but not in muscle and liver
(Tuey and Matthews, 1981).
3 Metabolic Reactions-Bioactivation and

Detoxication Processes
The metabolism and biotransformation of foreign compounds in the body
may be effected by a variety of enzymic processes, which can be
categorised as follows:
(1) Hydrolysis
(2) Oxidation
(3) Reduction
(4) Conjugation
( 5 ) Miscellaneous (including synthesis)
Williams (1959) proposed that these five categories could be simplified
into two types of reactions, namely Phase I reactions (hydrolysis,
oxidation, and reduction) and Phase I1 reactions (conjugation and
synthesis). In general, Phase I reactions convert compounds into sub-
strates suitable for Phase I1 metabolism. Some workers consider the
further metabolism of glutathione conjugates to be a third phase of
metabolism (Phase 111).
Each of these categories of metabolic reactions will be described briefly
and their role in the bioactivation and detoxification of compounds
discussed.
A Hydrolysis
Hydrolases (EC. 3xxx) constitute a heterogeneous group of enzymes
generally of low specificity, capable of the hydrolysis of carboxylic acid
esters, amides and lactones, organophosphate and sulphate esters,
saccharides, glycosides and glucuronides, peptides, and related structures
of both endogeneous and exogeneous origin. Examples of some of the
exogenous compounds used as drugs, food chemicals, pesticides, and
industrial solvents capable of being hydrolysed in the body are shown in
Table 1. Recently it has been shown that there are multiple isozymes of
hepatic carboxyles terase (EC 3.1.1 .l),many of which have been character-
ised (see e.g. Hosokawa et al., 1990).
Table 1 Examples of some foreign chemicals capable of being metabo-

lized by enzymatic hydrolysis in the body
Use Chemical class Example
Carboxylic acid ester Acetylsalicylic acid

Pethidine, atropine
Carboxylic acid amide Phenacetin, penicillin,
chloramphenicol
G1ycoside Digitalis alkaloids
Food additive
Flavouring agents Carboxylic acid ester Ethyl acetate
Emulsifiers Fatty acid esters Acetylated monoglycerides
Preservatives Carboxylic acid ester Hydroxybenzoates, gallates
Sweetener Dipeptide Aspartame
Agrochemicals
Herbicides and Carboxylic acid esters Pyrethroids, carbamates
insecticides and amides Carbanilates,
Industrial Chemicals
Solvents and Carboxylic acid esters Dialkyl phthalates, adipates
plasticizers
The hydrolysis of various esters by serum albumin has also been

reported (Kuruno, Kondo, and Ikeda, 1983).
B Oxidation
Mixed -function Oxidase System
A group of enzymes which are associated principally with the membra-
nous components of the endoplasmic reticulum and collectively referred
to as the microsomal mixed function oxidase (MFO) system, constitute
the most important group of enzymes mediating the oxidation of both
endogenous and foreign compounds. A wide variety of foreign compounds
including drugs, food additives, and industrial chemicals are metabolised
by the MFO system and examples of some of the oxidation reactions and
substrates are shown in Table 2. Essential requirements for M F O activity
are molecular oxygen, NADPH, the flavoprotein reductase, and the
haemoprotein ‘cytochrome P-450’. The range of enzymic reactions, the
mechanisms involved, and other related topics pertaining to the M F O
system have been extensively reviewed [see, for examples, Jakoby, Bend,
and Caldwell (1982); Sat0 and Omura (1978); Testa and Jenner (1976);
Ullrich, Roots, Hildebrandt, Estabrook, and Conney (1977); and more
recently Schuster (1989)l. Although a detailed discussion of the M F O
system is beyond the scope of this chapter, certain important aspects of the
system will be mentioned briefly. These are:
(1) The essential components of the enzyme complex
(2) Characteristics of the substrate/cytochrome P-450 interaction

(3) The cytochrome P-450 superfaniily
(4) Inducers and inhibitors of enzymic activity.
Following the observation that MFO activity was present in the
microsomal fraction of tissue homogenates and that the enzyme activities
were greater in the smooth endoplasmic reticulum, various isolation
procedures were developed to investigate the nature of the complex. The
solubilisation, separation, and reconstitution procedures pioneered and
developed by Coon and his co-workers (Coon, Autor, Boyer, Lode, and
Strobel, 1973) have shown that three distinct components are essential for
microsomal MFO activity. These are: lipids (phosphatidylcholine),
NADPH-cytochrome P-450 (cytochrome c) reductase, and the terminal
oxidase cytochrome P-450. Historically, substrates of the MFO system
were divided broadly into two groups based on their effects on
the spectral properties of cytochrome P-450 preparations (Remmer,
Shenkman, Estabrook, et al., 1966). Type I substrates, exemplified by
hexobarbital, interact with the oxidised haemoprotein to give a spectral
peak at 385-390 nm and a trough at 418-427 nm. Type I1 substrates (e.g.
aniline) exhibit an interaction spectra with a peak at 425-435nm and
trough at 395-405 nm. Some substrates (e.g. phenacetin) also contain
chemical structures associated with both type I and type I1 compounds
and consequently interact with cytochrome P-450 to give a reverse type I
(or modified type 11) spectrum (i.e. peak at 420 nm and trough at
392 nm). This is a concentration-dependent effect, probably caused by
the interaction of the substrate with binding sites different from type I
sites. There is generally a good agreement between the spectral dissocia-
tion constant values (K,) for type I substrate-binding spectra and K ,
values for enzyme activity indicating that the former probably manifests a
true reflection of the enzyme substrate complex. On the other hand, the
correlation between K , and K , values for type I1 substrates is extremely
poor. The binding of substrates to cytochrome P-450, largely determined
by the lipophilicity of the compound, is accompanied by changes in
protein conformation and spin state of the iron (Shichi, Kumaki, and
Nebert, 1978). Thus, type I binding results in the transformation of a low
spin haem-iron to a high spin state, whereas type I1 and reverse type I
bindings result in the formation of a low spin cytochrome P-450
accompanied by the direct co-ordination of the substrate with the
haem-iron.
The ‘Cytochromes P-450’ are now recognised to comprise a superfamily
of haemoproteins with an identical prosthetic group, but different
apoprotein structures, which are responsible for the different substrate
specificities. Currently at least 27 families with up to eight subfamilies and
up to 23 individual forms have been identified (Nebert et a/., 1991; Soucek
and Gut, 1992). Many of these families are present in all mammals, and
although the greatest number of forms have been characterised in the rat,
human orthologues have been found for a substantial number of them.
c
Table 2 Enzymic Metabolism of Foreign Compounds w
00
Reaction Category
[Enzyme System] Specific biotransformation Chemicals
OXIDATION Aromatic hydroxylation Benzene, polycyclic aromatic hydrocarbons

[Microsomal mixed- Aromatic epoxidation acetanilide
function oxidase]
Aliphatic epoxidation: Olefins, vinyl chloride, allyl alcohol
Aliphatic hydroxylation: Alkanes, fatty acids
0-dealky lation : Codeine, phenacetin,
S-dealkylation: Methyl mercaptan, methylisothiourea
S-oxidation: Carbon disulphide, thioacetamide, parathion
N-oxidation Hydroxylamine, 2-acetylamino fluorene
N-hydroxylation Dimethylaniline, morphine, imipramine
de-sulphur ation: Parathion
de-halogenation: Halothane, DDT, dichloromethane
N-dealkylation: t-Amines, amphetamine, dialkylnitrosamines
OXIDATION
[Alcohol dehydrogenase] Alcohol+ Aldehyde: Ethanol, allyl alcohol, methoxyethanol
[Aldehyde dehydrogenase, Aldehyde + Acid Formaldehyde, tolualdehyde
Aldehyde oxidase
Xanthine oxidase]
[Monoamine oxidase, Amines- Aldehyde: Tryptamine, tyramine, benzylamine
Microsomal mixed-
function amine oxidase]
REDUCTION
[Alcohol dehydrogenase] Carbonyl reduction Cyclohexanone
[Aromatic aldehyde- Benzaldehyde
ketone reductase p -Nitroacetophenone
(NADPH-dependen t)] Chloral hydrate
[Microsomal mixed- Polycyclic arene oxide Benzo( a)pyrene-4,5-oxide 9
function epoxide reduction P
reductase]
[Microsomal and bacterial Azo reduction to primary Prontosil, butter-yellow and other azo-colours
azo reductase] amines azobenzene
[Microsomal and bacterial Nitro aromatic compounds Nitrobenzene chloramphenicol

nitro-reduct asel amines niridazole, nitrazepam 4
[xanthine oxidase]
[Microsomal N- N-oxide + tertiary amine Nicotine-N-oxide, indacine-N-oxide

oxidoreductase]
[Microsomal dehalogenase] Aryl and alkylhalide DDT, trichlorofluoromethane

dehalogenation
HYDROLYSIS
[Carboxylic acid Ester hydrolysis Ethyl acetate, malathion, propyl gallate,
esterase] acetylsalicylic acid
[Amidases] Amide hydrolysis Penicillin dimethoate

[Microsomal epoxide Epoxide + Diol Styrene oxide, benzo(a) pyrene-4,5-oxide
hydrolase]
[Glycosidases-
glucuronidase Glucuronide hydrolysis Phenolphthalein glucuronide
gly cosidases] Glycoside hydrolysis Cycasin, amygdalin, thioglycosides
[Aryl sulphatase] Aryl sulphate hydrolysis Steriod sulphates
Table 2 (continued)
2
Reaction Category
[Enzyme System] Specific biotransformation Chemicals s
is
e
CONJUGATION z
[UDP-glucuronyl
transferase] Glucuronic acid conjugation
3.
ta
3
with:
aryl and alkyl alcohols Phenol, propane-1, 2-diol, tJ
h
7-hydroxycoumarin, sterols .a"'

primary and secondary amines Aniline, sulphonamides, cyproheptadine
N-hydroxy arylamines
2
sulphydric compounds Thiophenols $
carboxylic acids Phenylacetic acid, other arylacetic acids, %
carbamic acids sd
30
[Acyl-CoA: amino acid Amino acid (e.g. glycine) Benzoic acid, salicylic acid, cinnamic acid z
0
N-acyl transferases] conjugation with 2.
carboxylic acids Q
[Sulphotransferases] Sulphate conjugation Phenol, 2-acetylarninofluorene,
with: phenols, oestradiol and other sterols pl
aromatic amines, alkanols
[Acetyl CoA: arylamine Acetylation Isonicotinic acid hydrazide , sulphanilamide

k
00
0
transferases] of aromatic amines s
[Methyl transferases] 0- and N-Methylation reaction Catechol, imidazole, histamine,
[Phosphotransferases] Phosphate conjugation with Phenol
phenols
[Glutathione-S- Glutathione conjugation with: Phenanthrene-9,lO-oxide,styrene-7,&oxide,

transferases] arene and alkene oxides and benzo( a)pyrene-4,5-oxide
epoxides
[-glutamyl trans Aryl and alkyl halides, Chloracetonitrite 1,2-dichloro-4-nitrobenzene
peptidase + nitrates methyl bromide
cysteinyl glycinase
+ N-acetylase]
The early approaches to characterisation of cytochrome P-450, namely

affinity and ion-exchange chromatography (see, e.g. Gibson and Schenk-
man, 1978; Ryan et al., 1979) have now largely been replaced by gene
isolation and directed expression in suitable cell lines (see Gonzalez, 1988,
1990). Since each P-450 gene nearly always produces a single protein, the
current nomenclature for the gene, transcript and product are the same.
Thus in rat and human, italicised CYPlAl indicates the gene and
non-italicised CYPlA 1 both the corresponding mRNA and protein.
A wide variety of compounds of differing chemical structures have been
found to induce the MFO system, with at least six different classes of
inducers now recognised (Soucek & Gut, 1992). These are barbiturates
(e.g. phenobarbitone), polycyclic aromatic hydrocarbons (e.g. 3-methyl-
cholanthrene), steroids (e.g. dexamethasone), simple aliphatic compounds
(e.g. acetone), hypolipidaemic drugs (e.g. chlofibrate), and macrolide
antibiotics (e.g. rifampicin). Possible mechanisms involved in the induc-
tion process include transcriptional activation, mRNA stabilisation and
protein stabilisation (Gonzalez, 1988; 1990; Kimura et al., 1986; Song et
al., 1989). The cellular expression of polycyclic hydrocarbon-induci ble
P-450s is mediated by a cytosolic receptor, the Ah receptor, which
interacts with the inducer, is translocated to the nucleus, and stimulates
translation by interaction with nuclear DNA. Receptor mechanisms may
also be involved in the induction of other subfamilies, although no
receptor for phenobarbitone has been found (Murray and Reidy, 1990).
Many inhibitors of drug metabolism have been identified, although few
are specific for particular subfamilies (Testa and Jenner, 1981; Murray,
1987). Inhibitors may exert their effect by interfering with substrate
binding or electron transport and others may require metabolic activation
for their activity. The major families/subfamilies of cytochrome P-450,
with some known inducers and inhibitors are summarised in Table 3.
Mono-oxygenase activity is also found in other liver cell organelles. The
outer membrane of mitochondria contain cytochrome P-450, but little
cytochrome c-reductase; reduction being mediated by NADH-
cytochrome b5-reductase (Raw, 1978). Nuclear mono-oxygenase activity
is thought to be responsible for the toxification of many genotoxic
compounds including 2-acetylaminofluorene and aflatoxin B1 (Kawajiri,
Yonekawa, Hara, and Tagashira, 1979; Guengerich, 1979). Activity at
this site is inducible, and results in qualitatively similar patterns of
metabolism to that of the microsomal enzyme.
Other Oxidative Reactions

These reactions include the NAD-linked dehydrogenases responsible for
the oxidation of ethyl alcohol and other primary and secondary alkanols
to aldehydes (alcohol dehydrogenase) and the oxidation of aldehydes to
acids by aldehyde dehydrogenases. Other enzyme systems known to
mediate these reactions include catalase for methanol oxidation, and
Table 3 The cytochrome P-450 superfamily-some inducer, inhibitors,

and substrates
Subfamily/
protein Inducera Inhi bi torsb Substrates
IA
1Al 3MC; TCDD; BNF; ARO aNF; 9-HE 7-Et hoxresorufin
1A2 ISO; 3MC; TCDD; PNF aNF; 9-HE Caffeine; acetanilide
2A
2A 1 PB; BNF; 3MC; PCN Coumarin, testosterone
2A2 Non-inducible
2B
2B1 PB; ethanol; DEX SB; SKF-525A Pentoxyresorufin,
2B2 PB, ARO; PCN; DEX - Benzphetamine
2R4 PB -
2D
2D1 Non-induci ble Ajamalicine; Debrisoquine
quinidine
2E
2E1 Ethanol, DMSO Diallyl sulphide Dimethylnitrosamine
Isoniazid 4 Nitrophenol
3A
3A 1 PCN; PB; DEX Erythromycin Testosterone
3A2 PB - Testosterone
3A4 PB; DEX 17a-Ethynyl- Nifedipine
estradiol
3A6 RI F -
4A
4A I CLOF Terminal Lauric acid
acet yli nic
fatty acids
a3MC = 3-methylcholanthrene; TCDD = 2,3,7,8-tetrachlorobenzodioxin; PNF = P-

naphthoflavone; IS0 = isosafrole; PB = phenobarbitone; ARO = Aroclor 1254;
DMSO = dimethylsulphoxide; PCN = pregnenolone-16a-carbonitrile; DEX = dex-
amethasone; RIF = rifampicin; CLOF = clofibrate.
b~-N= F a-naphthoflavone; 9-HE = 9-hydroxyellipticine;SB = secobarbital.
Data from Murray and Reidy (1990); Soucek and Gut (1992).
aldehyde oxidase and xanthine oxidase for the oxidation of aldehydes to

acids (McMahon, 1971).
Other oxidation reactions of importance in the metabolism of foreign
compounds include the oxidative deamination of amines by either the
monoamine oxidase system or by a microsomal FAD-containing mono-
oxygenase enzyme (Ziegler, 1980). The oxidation of nitrogen in organic
molecules has recently been reviewed (Gorrod and Damani, 1985). A
number of novel oxidative reactions have also been identified. These
include the co-oxidation of polycyclic aromatic hydrocarbons and aroma-
tic amines by prostaglandin hydroperoxidase (Eling , Boyd, Reed, Mason,
and Sivarajah, 1983) and various peroxidative activities mediated by
cytochrome P-450 (Testa, 1987).
C Reduction
The two main classes of chemicals that are enzymically reduced in the
body are those containing either carbonyl groups or nitrogen functional
groups. The oxido-reductive interconversion of primary and secondary
alcohols to aldehydes and ketones, respectively, is mediated by cytosolic
NAD-linked alcohol dehydrogenase. At physiological pH, the reaction
equilibrium favours reduction. However, although ingested ketones, such
as cyclohexanone and hexobarbitone are reduced to the corresponding
secondary alcohol, reactive aldehydes are usually further oxidised to
acids. Examples of aldehydes reduced in uivo, however, include phthal-
aldehydic acid (Shiobara, 1977) and 2[ 1'-phenoxy-2'chloro]acetaldehyde
(Osterloh, Karakaya, and Carter, 1980). Additionally, cytosol from a
number of tissues contains at least three NADPH-dependent aldehyde
reductases (M.W. range 34,000-36,000) capable of the reduction of many
aromatic ketones and aldehydes (Wermuth, Munch, and Von Wartburg,
1977; Sawada, Hara, Kato, and Nakayama, 1979). An interesting feature
of these enzymes is their stereo-selectivity. Thus, for example, R -
warfarin is reduced almost exclusively to R,S-warfarin alcohol in humans,
only a trace of the R,R-alcohol being formed.
Aromatic nitro- and azo-compounds and tertiary amine N-oxides are
three important groups of nitrogen containing compounds metabolised by
reduction. Aromatic nitro compounds are reduced to amines, via nitroso
and hydroxylamine derivatives, by mammalian hepatic microsomal
NADPH-linked nitro-reductase activity. This enzyme activity, inducible
by phenobarbitone and DDT treatment, and inhibited by SKF-525A,
suggests the obligatory involvement of cytochrome P-450. A number of
other proteins possess nitro-reductase activity. These include NADPH-
cytochrome c reductase, DT-diaphorase, xanthine oxidase, aldehyde
oxidase, and lipoyl dehydrogenase (Wang, Behrens, Ichikawa, and Bryan
1974; Hewick, 1982). Aldehyde oxidase, for example, is a cytosolic
molybdenum-containing enzyme, which catalyses the reduction of com-
pounds such as nitrofurazone, N-nitrosodiphenylamine, and various
sulphoxides (Tatsumi, Yamada, and Kitamura, 1983).
An hepatic azo-reductase enzyme capable of metabolising aromatic azo
compounds to their component amines, probably via an intermediate
hydrazo-derivative, is found in both microsomal and cytosol fractions.
The properties of the microsomal enzyme, such as inhibition by oxygen
and inducibility by typical inducers of cytochrome P-450,suggests that
this enzyme is associated with the cytochrome P-450 dependent MFO
system. The cytosolic enzyme from rat liver has been purified and shown
S. D . Gangolli and J . C. Phillips 145
to be a flavoprotein of molecular weight 52,000, containing two molecules

of FAD per enzyme molecule (Huang, Miwa, and Lu, 1979). The
enzyme activity is inhibited by dicoumarol and markedly induced by
3-methylcholanthrene. NADPH and NADH can serve as co-factors. The
electron transport properties, spectral characteristics and mechanism of
inhibition by dicoumarol suggests that this enzyme is identical with
DT-diaphorase (Huang, Miwa, and Lu, 1979).
N-Oxide reduction to tertiary amines may be carried out enzymatically
or non-enzymatically, depending on the substrate, site of metabolism,
and the animal species. An hepatic microsomal NADPH-linked cyto-
chrome P-450 dependent enzyme, inducible by phenobarbital, catalyses
the reduction of indicine N-oxide to indicine (Powis and Wincentsen,
1980). In contrast, a cytosolic enzyme, probably a flavoprotein, requiring
NADPH in preference to NADH, mediates the reduction of nicotine
1'-N-oxide to nicotine. Haemoglobin non-enzymatically catalyses the
reduction of a number of N-oxide compounds including those of
trimethylamine , nicotinamide , imipramine , and dimethylamino-
azobenzene, to their corresponding amines (Bickel, 1969). In the process,
haemoglobin is oxidised to methaemoglobin. Other nitrogen containing
compounds known to be biologically reduced include aliphatic nitro,
nitroso, hydroxylamine, oxime, and hydroxamic acids.
Reduction of carbon-carbon double bonds has not been widely noted,
however the saturation of the a,P-unsaturated ketone LY140091 by a
cystolic enzyme was recently reported (Lindstrom and Whitaker, 1984).
D Conjugation
These reactions constitute a diverse group of enzyme-mediated processes
whereby foreign compounds or their metabolites are linked to en-
dogenous compounds to form more polar water-soluble products
capable of being readily excreted from the body in the bile or the urine,
depending on overall molecular weight. The major conjugation reactions
lead to the formation of glucuronides, mercapturic acids, sulphates, and
amino acid and acetyl derivatives. Less common conjugation reactions
include glycosylation, methylation, fatty acid conjugation, and
phosphorylation.
Glucuronide conjugates
UDP-glucuronyltransferase (UGT) catalyses the transfer of glucuronic
acid from UDP-glucuronic acid (UDPGA) to endogenous or foreign
substrates to form O(ether)-, O(ester)- N - , S-, and less commonly
C-glucuronides. Exogenous substrates include alcohols, phenols, car-
boxylic acids, hydroxylamines, aromatic amines, and sulphydryl com-
pounds. In the conjugation reactions involving these functional groups,
the C - 1 carbon atom of the glucuronic acid moiety of UDPGA is
activated to facilitate nucleophilic attack by the oxygen, nitrogen, carbon,

or sulphur atom of the substrate. The glucuronide derivative thus fromed
has the 0-configuration, and UDP is the leaving group.
The UGTs are a group of isoenzymes of 50-60 kDa localised primarily
in the hepatic endoplasmic reticulum and nuclear envelope (Burchell and
Coughtrie, 1992). In human liver, nine enzymes in two sub-families have
been identified, based on sequence identity. The four isoenzymes in
sub-family 1 catalyse the glucuronidation of phenols and bilirubins, but
generally not steroids or bile acids, and the five isoenzymes of sub-family 2
catalyse the glucuronidation of steroids, bile acids, and some xenobiotics.
UGT enzyme activity can be enhanced in uitro by the addition of
detergents such as Triton X-100 and Lubrol or organic solvents (e.g.
n-hexane, chloroform, and ether). Treatment of animals in vivo with
phenobarbital, 3-methylcholanthrene, DDT, chlorobiphenyls, or TCDD
leads to the induction of hepatic microsomal glucuronyl-transferase
activity. On the other hand, inhibition is effected by treatment with
hydrazines, harmol, and by the major metabolite of hydantoin, 5-
(p-hydroxyphenyl)-5-phenylhydrantoin (Mulder, 1974; Batt, Ziegler, and
Siest, 1975). In vitro inhibition of enzyme activity is effected by the
addition of UDP-galactose, UDP-xylose, galactose, galactosamine (prob-
ably at the level of UDP-glucose dehydrogenase) and by piperonyl
butoxide (Hanninen and Marniemi, 1970; Lucier, McDaniel, and
Matthews, 1971).
Mercapturic acid conjugates

The formation of mercapturic acids constitutes the end product of a series
of biochemical reactions initiated by the conjugation of electrophilic,
hydrophobic substrates with the tripeptide glutathione. Substrates for
this conjugation reaction, which is mediated by glutathione S-transfer-
ases, (Habig, Pabst, and Jakoby, 1974) include alkyl and aryl halides,
alkylmethane-sulphonates, dialkylphosphoric acid triesters, epoxides,
alkenes, organic thiocyanate, and triazines. The sequence of enzymic
reactions leading to the formation of mercapturic acids is exemplified by
the metabolism of benzyl chloride to benzylmercapturic acid (Figure 1).
Glutathione S-transferase (GST: EC 2.5.1.18) is a family of enzymes of
overlapping substrate specificity (Vos and Van Bladeren, 1990), each
enzyme species being a dimer differing in sub-unit composition. The
isoenzymes are mainly present in cytosol, although a microsomal form has
been identified. Elution chromatography of the transferases from rat liver
cytosol on carboxymethylcelluloses has revealed up to 12 different
isozymes (Stockman, Beckett, and Hayes, 1985; Mannervik, 1987), which
along with the enzymes from mouse and human liver cytosol are assigned
to one of three classes (a, p and n ) based on structural and catalytic
properties (see Table 4).
Individual isoenzymes of glutathione S-transferase are inducible by a
0 C t i 2 C 1 + GSH 1,~ C H 2 - S - C y s -I t l u
+ HCl
e C H 2 - S - C y st - ~CH2--S-Cys-NH2
0 CH,-S-Cys -NH -Acet yl
Benzylmercapturic acid
Figure 1 Conjugation of benzyl chloride with glutathione. The enzymes catalysing
the reactions are: (1) Glutathione S-transferme, ( 2 ) y-Glutamyl transferme, (3)
Peptidase, (4) Acetyl-CoA acetyl transferase.
variety of chemicals (Mannervik, Alin, Guthenberg, Jensen, and War-

holm, 1989, principally by translational activation of the corresponding
genes (Vos and Van Bladeren, 1990). Induction by phenobarbitone,
3-methylcholanthrene, and benzo(a)pyrene has been reported by Clifton
and Kaplowitz (1978), and polychlorinated biphenyls and TCDD have
also been shown to induce this enzyme activity in rat liver cytosol
(Kohli, Mukhtar, Bend, Albro, and McKinney, 1979; Baars, Jansen, and
Table 4 Classification of rat cytosolic

glu ta thione S-t runsferases
Class
Alpha (a) M u (PI Pi (71)

1.1 3.3 7.7
1.2 3.4
2.2 4.4
8.8 3.6
4.6
6.6
Isoenzyme 5.5 not yet classified.

Mouse isoenzymes-class Alpha N4.4
clas Mu N1.l; C1.l; ( 2 . 2 ; D1.l
class Pi N3.3.
Data from Vos and Van Bladeren (1990).
Breimer, 1978). Other compounds known to be inducers of GST activities

in hepatic and extrahepatic tissues include the antioxidants butylated
hydroxyanisole (BHA) and ethoxyquin (Benson, Batzinger, Ou, et al.,
1978) and trans-stilbene oxide (Guthenberg, Morgenstern, DePierre, and
Mannervik, 1980). A wide range of inhibitors of GSTs are also known
(Mannervik and Danielson, 1988) many of which have differential
effects on the various isoenzymes. Interestingly, the herbicides 2,4-D and
2,5,4-T, which are strongly inhibitory to p enzymes, appear to act by a
mechanism not asociated with binding to the active site (Dierickx, 1983).
In vivo, inhibition may also result from the depletion of GSH by treatment
with compounds such as diethylmaleate, and by thiol blocking agents (e.g.
p-hydroxymercuribenzoate) and alkylating agents (Askelof, Guthenberg,
Jakobson, and Mannervik, 1975). Other compounds possessing a carbon-
yl (or similar) group with an electrophilic character which are capable of
reacting non-enzymatically with glutathione and other thiols, include the
2-methylmercaptotriazineherbicide, Cyanatryn (Bedford, Crawford, and
Hutson, 1975).
Sulphate conjugates
A variety of sulphotransferases present in mammalian liver cytosol
catalyse the formation of 0-sulphate esters of simple phenols, other
aromatic alcohols, and steroids (androstenolone and estrone) . Addition-
ally, arylamines are metabolised by sulphate conjugation to sulphamates.
The donor of the sulphate group is 3'-phosphoadenosine-5'-
phosphosulphate (PAPS) formed from ATP and sulphate ions by the
action of three interlinked enzymic reactions, namely ATP-sulphurylase,
ADP-sulphurylase, and APS-kinase. Thus, the depletion of the in-
tracellular ATP pool by mitochondrial inhibitors, e.g. rotenone, can
reduce the availability of the co-factor (PAPS) and thereby inhibit the
phosphotransferase activities. Other inhibitors of these enzymes include
pentachlorophenol and 2,6-dichloro-4-nitrophenolwhich act as 'dead
end' competitive substrates. The sulphotransferase have been reviewed in
detail by Jakoby, Sekura, Lyon, Marcus, and Wang (1980).
Other Conjugation Reactions :Amino Acid Conjugates

Glycine is the amino acid most commonly used for the conjugation of
aromatic acids in mammals. In the typical sequence of biochemical
processes involved in the conjugation of benzoic acid, for example, the
first step is the esterification of the acid with CoA in the inner
mitochondrial matrix, to form benzoyl CoA. This reaction,
mediated by butyryl-CoA synthetase and requiring ATP, proceeds via
the formation of the intermediate AMP derivative (C6H5CO-AMP)
which then reacts with CoA to produce benzoyl CoA. Subsequently the
acyl CoA combines with glycine, catalysed by benzoyl transferase to form
S. D. Gangolli and J. C. Phillips 149
hippuric acid. ‘Benzoyl transferase’ and ‘phenylacyl transferase’ both of

which are acyl-CoA: amino acid N-acyl transferases have been isolated
from bovine-liver mitochondria. The former enzyme is specific for
benzoyl-CoA, salicyl-CoA and certain short linear and branch-chain fatty
acyl-CoA’s, whereas the latter specifically utilises the CoA esters of
phenylacetic acid and indole-3-acetic acid. These related enzymes,
separated by SDS disc-electrophoresis and by other procedures, have
been shown to be single polypeptide chains (molecular weight approxi-
mately 33kDa). In addition, there appears to be a third enzyme in rat
liver specific for bile acid-CoA derivatives which is capable of forming
bile acid conjugates with both glycine and taurine (Killenberg and
Jordan , 1978).
Taurine and glutamine conjugates have been shown to be formed with
certain arylacetic acids, including 1- and 2-naphthylacetic acids, hydratro-
pic acid, and 3-phenoxybenzoic acid in mammals (Idle, Millburn, and
Williams, 1978). Other amino acids participating in conjugation reactions
include histidine, serine, and alanine in bats, ornithine in reptiles and
certain birds, and arginine in Arachnida. In addition, conjugation with
peptides (other than GSH) has been reported (e.g. with phenothiazine in
calves: Waring and Mitchell, 1985).
Acetyl conjugates
Liver cytosol contains a group of enzymes, the acetyl CoA: N -
acetyltransferases, which catalyse the acetylation of aliphatic and aroma-
tic amines including drugs such as isoniazid, dapsone, and the sul-
phonamides. Animal studies have shown a heterogeneity and poly-
morphism in the N-acetyltransferases with respect to tissue distribution,
substrate specificity and pH-activity characteristics. One enzyme, found
primarily in the liver and gut, catalyses the acetylation of a wide range of
compounds and is considered to be the enzyme responsible for genetic
polymorphism. Another acetyltransferase found in extrahepatic sites is
active towards p -aminobenzoic acid and has a limited substrate
specificity.
In marked contrast to the other conjugation reactions, acetylation
decreases the hydrophilicity of the compound.
Methy 1 conjugates
Enzymic methylation of 0-, N - and S - atoms in a wide range of
substrates is effected by the corresponding methyltransferases using
S-adenosylmethionine as the methyl donor.
Soluble (i. e. cytosolic) catechol- , hydroxyindole-, and di-iodotyrosine-
0-methyl transferases have been identified as separate enzymes. In
contrast , phenol-0-methyltransferase is a microsomal enzyme. Substrates
for the N-methylation reaction include primary, secondary, and tertiary
amines and azaheterocyclics, and for S-methylation include diethyl

dithiocarbamate and carbon disulphide.
Phosphate conjugates
Rat liver mitochondria contain an ATP-dependent enzyme capable of
phosphorylating 1-aminopropan-2-01. Other examples of enzyme
mediated phosphorylation reactions are the formation of 1-naphthyl di-
hydrogen phosphate from 1-naphthol and of di-(2-amino-l-naphthyl)
hydrogen phosphate from 2-naphthylamine (Binning, Darby , Heenan,
and Smith, 1967; Boyland, Kinder, and Manson, 1961; Troll, Tessler,
and Nelson, 1963). However, little is known of the enzymes mediating
these reactions and the extent to which they are involved in the
metabolism of foreign compounds.
Novel conjugates
Many novel conjugate reactions, including those with endogenous
chemicals, such as cholesterol, fatty acids, and phospholipids have been
reported (see reviews by Paulson et al., 1986 and Testa, 1987).
4 Sites of Metabolism
Before discussing the biotransformation and disposition of foreign
compound with respect to specific tissues, the organisation and functions
of various intracellular organelles involved in metabolic processes at the
cellular level will be outlined. The liver cell is used as an illustrative
example, and the various sub-cellular components are shown in the
electron micrograph (Figure 2). A detailed discussion of the methods for
the isolation and purification of these organelles is beyond the scope of
this chapter, and many texts describing the relevant methods for
disruption and fractionation (by centrifugation) have appeared (see, for
example, Birnie, 1972; Snell and Mullock, 1987). A brief outline of the
structure and function of the various organelles follows.
A Intracellular Organelles
Nucleus
The various components of the cell nucleus are contained within a double
membrane. The outer one is part of the cell’s endoplasmic reticulum, and
carries a number of ribosomes. The inner membrane, which is separated
from the outer by the perinuclear space, is lined by the lamina densa, and
the whole structure contains a number of pores. Within the membranes is
the finely dispersed euchromatin, the fibrous hetero- and nucleolar-
associated chromatin, the nucleolus, and a number of other bodies of
Figure 2
granular appearance. (Bouteille, Laval, and Dupuy-Coin, 1974). A wide

range of enzyme activities have been found associated with the nucleus,
apart from those involved in RNA and DNA synthesis and processing,
These include methylases, acetylases, proteases, kinases, deacetylases,
steroid dehydrogenase, cytochrome oxidase, glycosyl transferases,
ATPase , carboxylesterases , phosphatases , mixed-function oxidases, and
5’-nucleotidase (Olson and Busch, 1974).
Mitochondria
The mitochondria, of which there are about 400 in a liver cell, contain
the enzymes responsible for oxidative phosphorylation and related energy
transfer processes, and are the site of production of high-energy
phosphates (e.g. ATP) in the cell. These rod-shaped bodies, with sizes up
to 5 p by 0.7 ,u, are bounded by a double membrane. The outer, limiting
membrane encloses a folded, inner membrane which gives rise to the
cristae. Thus, the mitochondrion consists of four compartments in which
enzymes can be located, namely the outer membrane, the inter-membrane
space, the inner membrane including the cristae, and the matrix enclosed
by the inner membrane’s surface. Enzyme activities associated with the
outer membrane include monoamine oxidase , NADH-cytochrome b5
reductase, acyl transferases, phosphotransferases and kinases, and cyto-
chrome P-450-dependent monooxygenases. NADH-linked de-
hydrogenases, succinate dehydrogenase, fatty acyl transferases, and
cytochrome P-450are associated with the inner membrane and citric acid
Table 5 Tissue localisation of some enzymes mediating xenobiotic tran:formation reactions
Tissue Hydrolysis Oxidation Reduction Conjugation
Skin Carboxylic acid esterase MFO system e.g. AHH; Ketosteroid reductase UDP-GT
EH ethoxycoumarin Sulphotransferase
0-deethylase ; Progesterone desaturase COMT
Steroid dehydrogenase,
MA0
Lung EH MFO system e.g. AHH, N-oxide reductase UDP-GT %

N-and 0-demethylases Phenol sulphotrans- s
0-
N-oxidase and hydroxylase ferase 0
Gastro- Non-specific carboxylic MFO system e.g. AHH, ‘N-hydroxy-AAF reductase’ UDP-GT
intestinal acid esterase N- and 0-deethylase GSH-T
tract biphenyl hydroxylase N-acetyl trans- b
Small ferase 8.
intestine Sulphotransferase
Glycine conjugase 82
3
Microflora Carboxylic acid esterase N- and 0-dealkylase Nitro-reductase Methyl and acetyl %
Sulphate ester hydrolase Dehalogenase Azo-reductase transferase s6
Amide hydrolase Deaminase N-oxide reductase 20
Glucuronidase Epoxide reductase %
Aldehyde reductase 0
Glucosidase 2.
Sulphamatase Aromatase 2
Liver Carboxylic acid esterase MFO system e.g. aromatic Azo-reductase UDP-GT; GSH-T; 9
Amidase and aliphatic hydroxylase; Nitro-reductase Sulphotransferase P
EH N-, 0-and S-dealkylases: N-oxide reductase Acetyl and methyl c,
epoxidase; sulphoxidase; Ketoreductase transferases 2
N-hydroxylase, etc. 2
ADH and Ald DH s
E,
Catalase, a
MA0 EL
5
Kidney Glucuronidase MFO system e.g. aromatic UDP-GT; GSH-T 9
Aryl sulphatase and aliphatic hydroxylase Sulphotransferase 2
Carboxylic acid esterase N-dealkylase; Fatty acid Amino-acid conjugase ~ k
Amidase hydroxylase Acetyl transferase t,
EH ADH
Adrenals MFO system e.g. steroid UDP-GT

hy droxylase GSH-T
Sulphotransferase
Testes Glucuronidase MFO system e.g. steroid Aromatase UDP-GT

Aryl sulphatase hydroxylase GSH-T
Carboxylic acid esterase Sulphotransferases
Abbreviations: EH-Epoxide hydratase; MFO-Mixed function oxidase; UDP-GT-Uridine diphosphoglucuronsyltransferase GSH-T-Glutathione-S-

transferase; COMT-Catechol-O-methyl transferase; ADH-alcohol dehydrogenase Ald.DH-Aldehyde dehydrogenase; MA%
Monoamine oxidase; AAF-N-acetylaminofluorene
cycle enzymes, transaminases , kinases, and NADP-linked de-
hydrogenases are among the enzymes found in the matrix. Adenylate
kinase and nucleoside diphosphokinase have been found in the inter-
membrane space. Other enzyme activities found in mitochondria include
esterases and superoxide dismutase. An extensive discussion of the
nature and distribution of enzymes in the mitochondrion can be found in
the review of Ernster and Kylenstierna (1969).
Microsomes
The microsomes are a heterogeneous mixture of membrane vesicles
derived from the endoplasmic reticulum (e.r.). The separation of
microsomal preparations into rough (ribosome-containing) and smooth
(ribosome-free) fractions by density gradient centrifugation procedures,
has led to the study of enzyme topology in these membranes. Rough e.r.,
which is composed of thin lamellae made up of two opposing membranes,
often occurring in more or less parallel stacks, contains a high concentra-
tion of electron transport proteins, glucose-6-phosphatase, and ATPase
activity. Smooth e x . is made up of a maze of fine tubules, the
membranes being in intimate contact with the mitochondria (Claude,
1969). It contains high levels of cytochromes b5 and P-450 (De Pierre and
Erster, 1977; Bergman and Dallner, 1976) and is generally more active in
metabolising xenobiotics than the rough e.r. (Holtzman, Gram, Gigon,
and Gillette, 1968; Gram, Schroeder, Davis, Reagan, and Guarino,
1971). Other enzyme activities associated with microsomes include
monoamine oxidase , NAD(P)H-cytochrome c reductase , nucleoside
diphosphatase, esterase, aldolase, and glutamine synthetase (Amar-
Costesec, Beaufay, Feytmans, Thines-Sepoux, and Berthet, 1969).
Peroxisornes
Peroxisomes are widely distributed in animal cells, and consist of a single
limiting membrane enclosing an electron-dense granular matrix and a
crystalloid core. They can be differentiated from lysozomes by the
absence of other matrix features such as vacuoles, lipid droplets, or
membranous structures. Catalase is the principal marker enzyme for
peroxisomes, however other oxidases are present including D-amino acid
oxidase, urate and glyoxylate oxidase. Other enzmes found in peroxi-
somes include carnitine acyl transferase, NAD(P)-dependent de-
hydrogenases, thiolase, enoyl-CoA hydratase and fatty acyl CoA synthe-
tase (Reddy and Lalwani, 1983).
Lysozomes
Lysozomes are present in the cytoplasm of the majority of mammalian
cells, with the exception of the erythrocytes. They appear as ‘dense
bodies’ in electron micrographs, consisting of a fine, granular matrix

enclosed by a single membrane. The matrix often contains other
structures such as osmiophilic droplets, vesicles and membrane-like
material. Lysozomes contain a wide range of hydrolytic enzymes includ-
ing acid phosphatase, DNAase and RNAase, esterases, aryl sulphatases,
glycosidases, and lipases (for a detailed discussion of lysozomal enzymes,
see Tappel, 1969).
Plasma membranes
The plasma membrane of a liver cell is that part of the surface of the
hepatocyte which is in contact with other cells. The fraction as prepared
by discontinuous density gradient centrifugation (Lansing, Belkhode,
Lynch, and Lieberman, 1967), contains desmosome-rich trilamellar
structures and circular vesicles. As well as 5’-nucleotidase, the widely
accepted marker enzyme for plasma membranes, these structures also
contain acid and alkaline phosphatase activity, phosphodiesterase and
nucleoside triphosphate pyrophosphohydrolase. In addition, the plasma
membrane contains a large number of receptor sites for the active uptake
of essential nutrients, and hormones, including those for amino acids,
sugars, fats, glucagon, and prolactin.
The main mammalian tissues involved in the metabolism of foreign
compounds are shown in Table 5, along with representative examples of
the major enyme categories known to be present at these sites.
Although enzymes such as the mixed-function oxidases and conjugates
are generally widely distributed in most tissues, certain important
biochemical and morphological features distinguish and characterise the
range of biotransformation reactions effected in each of the organs. A
brief description of these features, as they relate to the biotransformation
and disposition of foreign compounds, follows.
B Skin
This tissue constitutes approximately 10% of the normal body weight of
mammals, and is thus one of the largest organs of the body. It consists of
two distinct components; the outer and thinner epidermis and the
thicker, underlying dermis. In addition, there are several structures
within the skin, such as sweat glands, sebaceous glands, and hair follicles.
The skin is involved in several important physiological and biochemical
functions, including the regulation of body temperature, of water loss and
retention, of sebum production via androgens, and of carbohydrate and
lipid metabolism via insulin. The skin is also one of the most active
tissues in the body for protein synthesis (Freedberg, 1972). In addition,
the skin is capable of activating or inactivating several steroidal hor-
mones, such as the androgen dehydroepiandrosterone, which is con-
verted first to testosterone and subsequently to 5a-dihydrotestosterone.
Steroid metabolism in the skin has been reviewed comprehensively by

Hsia (1971) and Rongone (1977).
The presence of haem-proteins in the skin has been known for some
time, although cytochrome P-450 was first demonstrated in rat skin
microsomes less than 20 years ago (Bickers, Kappas, and Alvares, 1974).
The amount of microsomal protein in weanling rat skin is small
(3-5 mg g-' wet tissues) compared with the liver (20-25 mg g-' wet
tissue) (Bickers and Kappas, 1980) and MFO enzyme activities are also
low (Willemsens, Vanden Bossche, and Lavrijsen, 1989).The activities are
inducible by polycyclic hydrocarbons, PCBs, TCDD, and other com-
pounds known to be inducers of P4501A1 in the liver. The most
extensively studied xenobiotic metabolism is that of the microsomal
MFO-associated AHH activity, reported to be present predominantly in
the epidermal layer (Thompson and Slaga, 1976). Other MFO activities
found in the skin include aniline hydroxylase and 7-ethoxycoumarin
deethylase (Pannatier, Jenner, Testa, and Etter, 1978; Pohl, Philpot, and
Fouts, 1976). Due to technical difficulties in the preparation of cutaneous
tissues for metabolic studies, very few definitive experiments have been
carried out to delineate the substrate specificities of the enzyme activities
associated with the MFO system in the skin.
C Lung
The mammalian lung contains over 40 different cell types, although
only four are unique to the lung (Sorokin, 1970). These four types
of pulmonary epithelial cells, concentrated in the broncho-alveolar
region and known to contain enzyme systems capable of metabolising
foreign chemicals are: (1) non-ciliated bronchiolar cells (Clara cells) ;
(2) squamous alveolar cells (membranous pneumocytes: type I cells);
(3) great alveolar cells (granular pneumocytes type I1 cells); and
(4) pulmonary alveolar macrophages. The Clara cells, dome-shaped in
form and containing lipid droplets in their apices, have an abundant
complement of smooth endoplasmic reticulum. Type I cells have a
flattened shape and attenuated cytoplasm with limited intracellular
organelles, whereas type I1 cells are cuboidal in shape, and rich in rough
endoplasmic reticulum, Golgi apparatus, lysosomes, and lamellar bodies.
Type I1 cells eventually mature into type I cells over a period of
approximately 3 months. Although type I cells are only half as numerous
as type I1 cells, they cover about 25 times as much of the alveolar surface
as the latter cell type. The pulmonary alveolar macrophages, present in
the alveolar spaces, are particularly rich in lysosomes containing
hydrolases.
The main locations for the enzymes of the MFO complex in this organ
are the type I1 pneumocytes and the Clara cells. The wide variety of
substrates metabolised in the lung has been reviewed by Gram (1980).
Whereas phenobarbitone treatment has no significant inductive effect on
pulmonary MFO enzymes (Uehleke, 1968), polycyclic aromatic hydro-

carbons, naphthoflavone and phenothiazine derivatives have a marked
inducing capability. Several P450s have been isolated from lung tissue,
including forms with similar properties to those of forms 4Al and 4A2
(Murray and Reidy, 1990).
D Gastro-intestinal Tract
Small intestine
The gastro-intestinal tract was originally thought to act only in the
absorption of foreign compounds, but it is now recognised that the
epithelium is an important site for the metabolism of ingested com-
pounds. Mixed-function oxidase enzyme activity, which is present along
the length of the intestine, is most active in the duodenum and decreases
towards the caecum. Activity also varies at different levels of the mucosa
cells, with the lowest activity found in the crypts and the highest in the
villous tip cells (Hoensch, Woo, and Schmid, 1975). Although the specific
activities of MFO enzymes in the small intestine are low compared to the
liver (see review by Houston and Wood, 1980) the extensive surface area
provided by the mucosal cells and their rapid turnover renders this tissue
quantitatively important. In contrast to the relatively low uninduced
Phase I (oxidative) reaction capability, this organ shows high glucur-
onyltransferase activity. The mucosal cells respond to most of the model
inducers of the MFO system (Wollenberg and Ullrich, 1977).
Microflora
The gastro-intestinal tract in mammals contains a wide variety of
micro-organisms. Over 400 species of bacteria have been identified and
the main groups found in man, namely the bacteridaceae, propionobac-
teridaceae, lactobacillaceae, enterobacteriaceae, micrococcaceae, and
yeasts are anaerobic organisms. There are large interspecies differences
in the numbers and types of micro-organisms present in the gut and also
in their distribution along the gastrointestinal tract (Rowland and
Walker, 1983; Savage, 1977).
In view of the anaerobic character of the resident microflora and the
low redox potential prevailing in the caecum and colon, reductive
reactions constitute the major biotransformations effected on foreign
chemicals. These reactions include the reduction of nitrate, nitro-, and
azo-compounds (see Table 5). The other major class of metabolic
reactions carried out by the gut microflora is the enzymic hydrolysis of
esters, amides, peptides, and glucuronides. The gastro-intestinal hydroly-
sis of glucuronides excreted via the bile and the subsequent intestinal
reabsorption of the metabolite formed leads to entero-hepatic circulation
and the possibility of second-pass hepatic metabolism. Entero-hepatic
circulation has been found to be of significance in the metabolism of
stilboestrol, butylated hydroxytoluene , and various food flavours, includ-

ing linalool.
A number of factors have been found to modify the enzymic profile of
gut microflora. For example, treatment of rats with cyclamate for
prolonged periods leads to adaptive changes in the microflora resulting in
an increase in the metabolism of this compound by the induction of gut
flora sulphamatase activity (Renwick, 1986). On the other hand, chronic
exposure of rats to metronidazole leads to the inhibition of its metabo-
lism by the gut flora (Eakins, Conroy, Searle, Slater, and Willson, 1976).
The metabolic capabilities of microflora in various species has been
reviewed by Scheline (1973; 1980) and by Rowland, Mallett and Wise
(1985).
E Liver
A major and possibly the most intensively studied site for the metabolism
of foreign compounds is the liver. The relative uniformity in the
morphology of the tissue, the abundance of metabolising enzyme
activities and the ease of preparation for assay has, in part, been
responsible for the wide ranging biochemical and histological investiga-
tions on this organ. The brief description of the morphology of the liver is
intended to provide an appreciation of the tissue distribution of enzyme
activities.
There is a growing body of evidence that the original description of the
liver in terms of a hexagonal configuration of liver cells proposed by
Kiernan in 1833, does not equate with the current concepts of the
functional unit of the liver, namely the acinus, first described by
Rappaport and his co-workers. This unit is seen as a parenchymal mass
consisting of a terminal portal tract, hepatic arteriole, bile ductule, and
lymph vessels, contained between two or more central veins. Thus, the
classical description of midzonal periportal, focal, and centrilobular
regions of the liver correspond to zones 1, 2, and 3 respectively in the
Rappaport description. (For a more detailed discussion of this scheme,
see Rappaport, 1969). Recent histochemical and electron microscopic
studies have demonstrated that enzyme distribution within the liver
lobule is not uniform. The activity of respiratory enzymes is particularly
high in zone 1, whereas NADP-dependent MFO enzyme activities are
concentrated in zone 3. Furthermore, immunohistochemical and micro-
spectrophotometric studies have shown that phenobarbital-inducible
cytochrome P-450 is predominantly in zone 3 whereas the 3-
methylcholanthrene-inducible cytochrome P-448 is more evenly distrib-
uted throughout the liver lobule. (Baron, Redick, Kapke, and Guen-
gerich, 1981).
The MFO enzyme system is most generally studied in the microsomal
fraction, although cytochrome P-450 has also been detected in hepatic,
mitochondrial, and nuclear fractions by ESR and microspectrophoto-
metry. The mitochondrial haemoprotein and associated AHH activity are

inducible by 3-methylcholanthrene, and the reaction was more rapid in
the presence of NADH than with NADPH. The specific content of
cytochrome P-450 in the rat hepatic nuclear fraction is about one-ninth of
that in the microsomes, and the NADPH-cytochrome c reductase specific
activity about 45%. (Sikstrom, Lanoix, and Bergeron, 1976). Rigorous
purification procedures have established that these MFO activities are
genuinely associated with hepatic mitochondrial and nuclear fractions and
not due to adventitious microsomal contamination.
F Kidney
The kidney contains many of the enzyme systems capable of metabolising
foreign compounds found in the liver, but the complex morphology and
biochemical functions of the former makes it difficult to compare the
tissue distribution and role of metabolising enzymes in the two organs.
The two major anatomical regions of the kidney are the outer cortex and
the medulla, the former being the major component. The cortex is
composed mainly of proximal tubule cells and also of glomeruli, distal
tubules, collecting ducts, blood vessels, and connective tissue. The
endoplasmic reticulum in the tubule cells is structurally similar to that
found in liver parenchymal cells, but constitutes only a small fraction of
the total tubule volume. The luminal surface of the tubular cells is lined
with a brush border consisting of numerous invaginated villi. The basal
surface is characterised by frequent infolding, within which lie rod-shaped
mitochondria. The medulla includes the papilla containing the loops of
Henle, the collecting ducts, the descending and ascending vasa recta and
their associated capillary plexus and interstitial tissue. (For a more
detailed discussion of kidney structure and function, see Brenner and
Rector, 1976.)
The renal cytochrome P-450 system, located mainly in the cortex, has
many similarities but also certain differences to the hepatic system. These
differences reside in the absorption maximum of the CO-cytochrome
P-450 interaction spectrum, substrate specificities, and inducibility. The
kidney cytochrome P-450 system, like that in the liver, is also involved in
the metabolism of endogeneous substrates, having an essential role in
fatty acid hydroxylation, prostaglandin synthesis and vitamin D
metabolism.
The kidney cortex contains at least two forms of cytochrome P-450,
one inducible by treatment with polycyclic aromatic hydrocarbons and
the other inducible by dietary lauric acid. The latter form is relatively
specific for fatty acid m-hydroxylation. The renal MFO system metabo-
lises substrates at rates generally lower than those found with the liver
enzyme and the array of substrates metabolised is also much restricted.
(Jones, Orrhenius, and Jakobson, 1980).
G Other Tissues
Other organs of importance in the metabolism of foreign compounds
include the adrenals, gonads, and placenta.
The adrenal cortex, in common with other steroid-producing tissues?
contains an abundance of cytochrome P-450, located both in the
microsomes and in the mitochondria. The former is required for the 17a-
and 21-hydroxylation of steroids, whereas the latter mediates the llp-
and 18-hydroxylation and the side-chain cleavage of cholesterol. The
mitochondria1 steroid hydroxylation system has been resolved into three
components-an NADPH-dependent flavoprotein (adrenodoxin reduc-
tase), a non-haem iron protein (adrenodoxin), and cytochrome P-450.
There is little information on the chemical inducibility of adrenal
cytochrome P-450; one report claims that phenobarbitone and 3-
methylcholanthrene treatment are without effect (Feuer, Sosa-Lucero,
Lumb, and Moddel, 1971).
Both the testes and the ovaries contain relatively low levels of MFO
enzymes. However, the close proximity of these enzymes to the site of
germ cell production confer particular toxicological relevance to their
metabolic capabilities. The placenta contains a wide range of enzymes
mediating the metabolism of foreign compounds, including the MFO
system, conjugases, reductases, and hydrolases (for reviews, see Juchau,
1980; Heinrichs and Juchau, 1980).
5 Bioactivation and Detoxification Processes

The metabolic biotransformation of foreign compounds leading either to
the generation of biological reactive metabolites (and thus to toxicity),
or to the formation of normal physiological constituents or biologically
inert excretory products, is depicted schematically in Figure 3.
In this scheme, five main types of metabolic events and the consequent
biological effects may be identified. These are as follows:
Type 1: Compound ‘X’, toxic per se is metabolised .via Pathway A to
yield a non-toxic end product.
Type 2: Compound ‘X’, non-toxic per se is metabolised by a Pathway B
reaction process to form a biologically reactive, and therefore,
toxic metabolite.
Type 3: Compound ‘X’, non toxic per se, is metabolised by a Pathway B
reaction process to a reactive (toxic) product which in turn is
metabolised by a Pathway A process to yield a non-toxic
product.
Type 4: Compound ‘X’, non-toxic per se, is metabolised as in type 3
reaction. The non-toxic product at step 2 is further metabolised
by a type B reaction to yield a toxic species.
Type 5: Compound ‘X’?non-toxic per se, can be metabolised either by a
type A reaction or by a type B reaction. Thus, in contrast to the
FOREIGN Fo
7COMPOUND P
X
METABOLIC
PATHWAY A 4
[Detoxifying Reactions] 9
i * I1 Biologically
Reactive Metabolite
( 1 Physico-Che m ica I/covalent

interaction with
(a) plasma membranes
Biologically Endogenous (b) lntracellular
inert products Components macromolecular
co nst it uents
1
Excreted
1
Normal metabolism
e.g. proteins, DNA, etc.
(2) Com petit
substrates in vital
biochemical reactions
NON-TOXIC TOXIC
Figure 3 Schematic representation of the metabolism of a foreign compound to toxic or non-toxic products and the interplay between
detoxifying and bioactivation processes in determining the eventual biological fate of a compound.
Table 6 Some examples of biotransformation reactions on foreign compounds leading to toxic or non-toxic metabolites
2
Compound Pathways Products
X A
so2 Sulphite oxidase Sulphate
Phenol UDP-GT 0-glucuronide
Ethylene oxide EH Ethylene glycol
B
Chloroform MFO Phosgene
Di-(2-ethylhexy1)phthalate non-specific esterase Mono-ester
Red 2G azo reductase Aniline
Ethylene dibromide GSH-T Thiiranium ion
Phenacetin UDP-GT or
Sulphotransferase II MFO
-glucuronide or sulphate 11
quinone?
-
g
UDP-GT: Uridine diphosphoglucuronosyltransferase; GSH-T: Glutathione-S-transferase; EH: Epoxide hydratase; MFO: Mixed-function oxidase; ADH:
Alcohol dehydrogenase.
164 The Metabohm and Disposition of Xenobiotics
situation with types 3 and 4 reactions where the metabolic

processes operate in tandem, in type 5 reactions the two
metabolic processes act simultaneously and therefore the biolog-
ical effect is determined by the relative metabolic contribution of
two processes.
It should be noted that Pathways A and B may each consist of a
number of enzymic reactions acting on the compound or its metabolites.
Examples of some of the foreign compounds metabolised by these
various types of reactions are shown in Table 6. The picture that emerges
from a survey of the complex and dynamic interplay of enzymic reactions
participating in the metabolic disposition of foreign compounds, is that,
in the main, enzymic reactions are toxicologically ‘neutral’, in that they
cannot be categorised exclusively as bioactivating or detoxifying pro-
cesses. Thus, for example, the MFO enzyme system can participate in
both bioactivation and detoxification reactions and conjugation reactions,
generally considered to be a detoxification step, can yield toxic end
products, such as glutathione or sulphate conjugates (for reviews, see Van
Bladeren, Bruggeman, Jongen, Scheffer, and Temmink, 1987; Mulder,
Meerman, and van der Goorbergh, 1986).
Two further factors have a determining influence on the metabolism
and biological fate of a compound. The first is the influence of the
kinetics of the enzymic processes involved, and the second is the
relationship of the various organs and intracellular organelles involved in
the bioactivation and detoxification of the compound, with respect to the
target site. Examples of the latter include the interactive metabolic
transformation of nitrobenzene by the liver and the gastro-intestinal
microflora (Levin and Dent, 1982) and of 1,3-hexachlorobutadiene by the
liver and kidney (Nash, King, Locke, and Green, 1984). Other important
determinants in the toxicology of a compound are the route of exposure
of the species to the compound and the pharmacokinetics following
absorption. The various factors known to influence pharmacokinetics and
metabolism of compounds will be discussed in the next section.
6 Factors Influencing Metabolism and

Pharrnacokinetics
Differences in the biological fate of foreign compounds observed both
between and also within animal species can be ascribed to the following:
(1) Differences in rates of absorption and transfer across cell
membranes.
(2) Differences in rates of metabolism.
(3) Differences in routes of metabolism.
A number of factors contribute to these differences in xenobiotic
metabolism. They include the age, sex, and strain of the animal model,
the nutritional status of the animal, and other dietary factors. An
appreciation of the influence of these various factors in modifying the

metabolism and pharmacokinetics of a compound in an animal model is
essential for the understanding of the underlying toxic mechanisms. A
brief description of these factors follows.
Rates of Absorption
In an earlier section, a brief description of the influence of molecular
size, lipophilicity, and degree of ionisation on the absorption of foreign
chemicals from various sites of exposure in the body was presented. In
addition to these physico-chemical properties, other factors of sig-
nificance in this context are the physical form of the compound, the dose
level administered and the frequency of exposure to the foreign
compound.
Whereas inert particulate material is unlikely to be absorbed into the
body, except for the minute amounts persorbed across the intestinal
lumen by pinocytosis, the uptake from the skin of absorbable compounds
can be accelerated by various means. These include the presence of
solvents, such as dimethylsulphoxide, and increased hydration of the
stratum corneum by a vehicle.
Both the dosage and frequency of exposure to a foreign compound can
affect its absorption, particularly when absorption is carrier mediated and
therefore potentially saturable.
Species differences in the skin absorption rates of chemicals have been
observed. Thus, whereas cutaneous permeability in the guinea pig, pig,
and monkey is similar to man, the skin of the rat and rabbit is much
more, and that of the cat much less permeable, compared with man
(Scala, McOsker, and Reller 1968; Coulston and Serrone, 1969; Wester
and Maibach, 1977).
Entero hepatic Circulation

Differences in extent of enterohepatic circulation (EHC), which is the
reabsorption of a compound excreted into the bile from the gut and
returned to the liver via the portal blood, are also known to affect the
pharmacokinetics and biological effect of xenobiotics. Since one conse-
quence of EHC is the maintenance of a disproportionately high con-
centration of compound in liver and gut for a longer period than would
occur in its absence, exposure of the animal to potentially toxic
compounds is increased. Thus, for example, species differences in the
toxicity of indomethacin to the gastric mucosa are directly related to the
extent of E H C (Duggan, Hooke, Noll, and Kwan, 1975).
Rates of Metabolism : Inter-species Differences

There is extensive literature on the quantitative differences observed in
the metabolic rates of chemicals in various animal species (see review by
Table 7 Species diferences in the rates of metabolism of foreign

compounds
Species Circulating half-life (min)
Hex0barbi tal Antipyrine Aniline
Man 60
Old World Monkey 40-50
Dog 0
Rodent (Rat, Mouse) 0-5
Hathway, Vol. 3, 4, 5 , and 6). Some examples of species differences in

the rate of metabolism of various drugs are shown in Table 7. It is clear
that the mouse is more efficient than man in the metabolism of
hexobarbital and antipyrine, for example, and that man converts quinic
acid to benzoic acid more rapidly than the old world monkey or the dog.
Other examples of species related differences in the rates of metabolism
of chemicals include the 2-hydroxylation of biphenyl in the mouse (high)
and in the rat (low), and the glucuronidation of monobutylphthalate in
the hamster (high) and in the rat (low).
Rates of metabolism : Intra-species Differences

The metabolism of foreign compounds within a particular species may
be modified by a number of factors, including the age, sex, strain,
nutritional status and diet of the animals. A brief note on each of these
factors follows:
Age. The activities of enzymes associated with the hepatic MFO complex
are generally extremely low in the developing foetus and newborn
animal. Liver size and the specific activities of the MFO enzymes
progressively increase with age until puberty, as do the enzymes involved
in conjugation reactions (Dutton, 1978). However, phenol sul-
photransferase activity is present in foetal liver at an earlier stage than
glucuronyltransferase activity.
The human foetus has appreciably higher levels of MFO enzyme
activities compared with its animal counterpart, but less than in human
adults. However, the glucuronidation ability in the human foetus is low.
Hence, factors leading to the increased synthesis of bilirubin without a
concomitant increase in UDP glucuronyltransferase activity can result in
jaundice in premature human babies. In contrast to glucuronidation,
sulphation of some drugs (e.g paracetamol) is more rapid in the foetus

than in the adult (Levy, Khanna, Soda, Tsuzuki, and Stern 1975).
Sex. In general, male rats and humans, but not male mice, tend to
metabolise compounds more rapidly than females. Thus, for example,
glyceryl guaiacolate ether, a centrally acting muscle relaxant, has a
shorter circulating half-life in male rats compared to female animals
because of the higher 0-demethylase activity in the males (Giri, 1973).
Similarly, the synthetic androgen, trengestone, is metabolised more
readily to the 16a-hydroxy-derivative by male rat liver slices than by
female preparations and, although a sex-difference was not seen with
human liver slices, marked differences in the excretion of unchanged
compound after oral administration to humans have been reported
(Breuer, Kime, and Knuppen, 1973). The exhalation of acetaldehyde
following the i.p. administration of ethanol in C57BL/6J mice is
approximately 5-6 times greater in male animals than in females
(Redmond and Cohen, 1972) and castration of the male mice abolishes
this difference. Increased xenobiotic metabolising activity in males is also
exemplified by the relative resistance of male rats to the toxicity of the
organophosphorus insecticide, parathion.
Strain Differences. The literature is replete with examples of intra-species
variation in the metabolism and disposition of foreign chemicals being
ascribable to genetically determined strain differences (for reviews, see
for example Hathway, Vol. 1-6; Nebert, Robinson, Niwa, Kumaki, and
Poland, 1975). The induction of arylhydrocarbon hydroxylase (AHH)
activity by 3-methylcholanthrene or P-naphthoflavone occurring in genet-
ically responsive (C57BL/6J and C3H/He) strains of mice but not in the
DBA/2 strain is due to an autosomal dominant trait capable of being
segregated in heterozygous offspring. The presence or the absence of
A H H induction has been found to correlate with the hepatic N -
hydroxylation of 2-acetylaminofluorene in C57BL/6N and DBA/2N
strains (Thorgeirrson, Felton, and Nebert, 1975). Other examples of
strain differences in metabolism include the 7-hydroxylation of coumarin
in mice (Lush and Andrew, 1978), the 4- and 6- hydroxylation of
debrisoquine in inbred strains of rat (Al-Dabbagh, Idle, and Smith, 1981)
and the N-acetylation of p-aminobenzoic acid and other arylamines in
rats and mice (Tannen and Webber, 1979). In humans, examples of racial
metabolic differences include the acetylation of drugs such as dapsone
and sulphamethazine and the hydroxylation of debrisoquine. A recent
example of genetic polymorphism in drug oxidation relates to the
anti-angina compound perhexiline (Gould, Amoah, and Parke 1986;
Cooper, Evans, and Whibley, 1984). For this drug, impaired oxidation
results in increased toxicity.
Dietary Constituents and Nutritional Status. There has been a growing

recognition of the importance of both macro- and micro-components in
the diet and of nutritional status in influencing the metabolism and
pharmacokinetics of foreign compounds since the early observations by

Mueller and Miller (1950) and Brown, Miller, and Miller (1954) showing
that dietary riboflavin modified hepatic azoreductase activity towards the
carcinogenic azo compound 4-dimethylaminoazobenzene in the rat and
mouse. Subsequent reports have shown that the biotransformation of
drugs can be significantly affected by the intake of carbohydrates, fat,
protein, vitamins, and minerals. Many reviews on the extensive and
growing literature on this subject have been published including those by
Campbell and Hayes (1974); Parke and Ioannides (1981); and Hayes and
Campbell (1980).
The high intake of glucose and other carbohydrates in experimental
animals and in man has been found to decrease haem synthesis and the
activities of enzymes associated with the hepatic MFO system. Conse-
quences of excess carbohydrate intake include prolonged phenobarbital
hypnosis (Strother , Throckmorton, and Herzer 1971) and depressed
hepatic dimethylnitrosamine demethylase activity (Venkatesan, Arcos,
and Argus, 1970).
Animals maintained on diets totally deficient in lipids have markedly
depressed hepatic MFO enzyme activities. The progressive addition of
saturated and unsaturated fats to the diet leads to the restoration of these
enzyme activities to normal levels. In particular a high intake of
unsaturated fats, e.g. herring oil rich in linoleic acid, has been found to
‘permit’ the maximum induction of hepatic cytochrome P-450 and the
MFO system (Marshall and McLean, 1971). Cholesterol has also been
found to exert a stimulatory effect on liver MFO enzymes (Lambert and
Wills, 1977).
Protein deficiency has been shown to decrease the content of phos-
phatidylcholine in cytochrome P-450 and to decrease the activities of
NADPH cytochrome P-450 reductase and of numerous MFO enzymes
(Campbell and Hayes, 1976). On the other hand, the excessive intake of
protein increases hepatic MFO activity in the rat and leads to a significant
reduction in the circulating half-life of antipyrine and theophylline in man
(Kappas, Anderson, Conney, and Alvares, 1976).
The inadequate intake of micronutrients such as ascorbic acid, ribofla-
vin , tocopherol , calcium, copper , magnesium, selenium , and zinc has
been found generally to decrease the activities of the MFO enzyme
system and other xenobiotic metabolising enzymes (see, for example,
Gibson and Skett, 1986). This may be as a result of an inhibition in the
synthesis of the enzyme or essential co-factors. Other elements such as
chromium, manganese, nickel, cadmium, and cobalt have been found to
decrease cytochrome P-450 content by inhibiting the 8-ALA synthetase
activity and inducing haem oxygenase activity.
Food deprivation has been found to alter markedly the hepatic MFO
system. The effect is sex dependent in that whereas in the male rat some
of the enzymes capable of being stimulated by androgens are depressed,
in female animals a slight increase in enzymic activity is observed (Kato
Table 8 Species diferences in the routes of metabolism of ,foreign

compounds
Principal Metabolic
Compound Reaction Species Reference
Phase I reactions
Coumarin 7-hydroxylation Man, Baboon
3-hydroxylation Rat, Guinea-
Pig >
Rabbit
BHT Oxidation of -OH Rat, Man
Oxidation of t-butyl Man
2-AAF N-oxidation Rat, Mouse, 3
Dog
C-hydroxylation Guinea-pig,
Lemming
Cyclohexylamine C-Hydroxylation Rat
Deamination Man
Phase I1 reactions
4-Chlorophenylacetic Glycine conjugation Rat 5
acid Glutamine conjugation Man
Phenol Sulphate conjugation Cat 6
Glucuronidation Pig
Sulphation/glucuronidation Rat
Ref. (1) Cohen (1979)

(2) Daniel, Gage, and Jones (1968)
(3) Lotlikar, Enomoto, Miller, and Miller, (1967)
(4) Renwick (1986)
(5) James, Smith, and Williams (1972)
(6) Capel, French, Milburn, Smith, and Williams (1972)
and Gillette, 1965; Gram, Guarino, Schroeder, Davis, Reagan, and

Gillette, 1970).
Routes of Metabolism-Species Differences

Qualitative differences in the pathways involved in the metabolism of
foreign compounds in various animal species have been found, and a
number of examples of species differences in either phase 1 or phase 2
metabolic processes are shown in Table 8. These differences have been
related to the varying susceptibility of the species to the biological effect
of the particular compound. Although many attempts have been made to
predict which route of metabolism would be favoured in a species or
genus derived from structural considerations, no clear predictive frame-
work has emerged (see, for example, the studies on conjugation of
arylacetic acids by Caldwell, 1981).
A number of enzymes involved in the metabolism of foreign corn-
170 9 The Metabolism and Disposition of Xenobiotics
pounds have been found exclusively in man and other primates. These
include the enzymes involved in the N’-glucuronidation of methoxysul-
phonamides, and the 0-methylation of 3,5-di-iodo-4-hydroxybenzoic
acid.
Concluding Remarks
This review has attempted to present a general perspective of some of the
important factors and sites in the body involved in the absorption,
metabolism, and biological disposition of foreign compounds.
The rationale for this scheme of presentation was to enable the reader
to glean an appreciation and recognition of the inter-relationship of the
various facets involved in determining the metabolic fate of foreign
compounds in experimental animals. Such an understanding could
usefully form the basis for the interpretation of metabolic data and the
application of the information in the design and conduct of animal
toxicity studies.
It must be reiterated that each of the factors mentioned constitutes a
complex topic in its own right. Active and vigorous research is still
proceeding in order to explain the role of these various factors in
xenobiotic metabolism. At this stage, the state-of-the-art, while illumina-
ting important aspects of the subject, awaits further development to
enable a realistic assessment to be made of the biological effects of
chemicals to man.
References
Al-Dabbagh, S. G., Idle, J. R., and Smith, R. L. (1981). Animal modelling of
human polymorphic drug oxidation-the metabolism of debrisoquine and
phenacetin in rat inbred strains. J. Pharm. Pharmacol., 33, 161-164.
Amar-Costesec, A. , Beaufay, H., Feytmans E. , Thines-Sempoux, D. , and
Berthet J. (1969). Subfractionation of rat liver microsomes. In: ‘Microsomes
and Drug Oxidations’, Eds J. R. Gillette, A. H. Conney, G. J. Cosmides, R.
W. Estabrook, J. R. Fouts, and G. J. Mannering. Academic Press NY., pp.
41-58.
Askelof, P., Guthenberg, C. , Jakobson, I. , and Mannervik, B. (1975). Purifica-
tion and characterisation of two glutathione S-aryltransferase activities from
rat liver. Biochem. J . , 147, 513-522.
Baars, A. J., Jansen, M., and Breimer, D. D. (1978). The influence of
phenobarbital, 3-methylcholanthrene and 2,3,7,8-tetrachlorodibenzo-p-
dioxin on glutathione S-transferase activity of rat liver cytosol. Biochem.
Pharmacol., 27, 2487-2494.
Baron, J., Redick, J. A., Kapke, G . F., and Guengerich, F. P. (1981). An
immunohisto-chemical study on the localisation and distribution of
phenobarbital- and 3-methylcholanthrene-inducible cytochrome P-450within
the livers of untreated rats. J . Biol. Chem., 256, 5931-5937.
Barry, B. W.(1983). Dermatological Formulations-Percutaneous Absorption
Drugs and the Pharmaceutical Sciences, Vol. 18. Marcel Dekker Inc., New
York.
Batt, A-M., Ziegler, J-M., and Siest, G. (1975). Competitive inhibition of

glucuronidation by p-hydroxyphenyl hydantoin. Biochem. Pharmacol., 24,
152- 154.
Bedford, C . T., Crawford, M. J., and Hutson, D. H. (1975). Sulphoxidation of
cyanatryn, a 2-mercapto-sym-triazine herbicide, by rat liver microsomes.
Chemosphere, 4, 311-316.
Benson, A. M., Batzinger, R. P., Ou, S.-Y. L., Bueding, E . , Cha, Y.-N., and
Talalay , P. (1978). Elevation of hepatic glutathione S-transferase activities
and protection against mutagenic metabolites of benzo(a)pyrene by dietary
antioxidants. Cancer Res., 38, 4486-4495.
Bergman, A. and Dallner, G. (1976). Properties of a rat liver smooth microsomal
subfraction not aggregated by Mg2+. Life Sci., 18, 1083-1090.
Biagi, G. L., Barbaro, A. M., Gamba, M. F., and Guerra, M. C. (1969).
Partition data of penicillins determined by means of reversed-phase thin-
layer chromatography. J. Chromatogr., 41, 371-379.
Bickel, M. H. (1969). The pharmacology and biochemistry of N-oxides.
Pharmacol. Rev., 211, 325-354.
Bickers, D . R. and Kappas, A. (1980). The skin as a site of chemical metabolism.
In: ‘Extrahepatic Metabolism of Drugs and Other Foreign Compounds’. Ed.
T. E. Gram, Ch. 8, MTP Press Ltd, London U.K. pp. 295-318.
Bickers, D. R., Kappas, A., and Alvares, D. P. (1974). Differences in
inducibility of hepatic and cutaneous drug metabolising enzymes and
cytochrome P-450 by polychlorinated biphenyls and 1,1,l-trichloro-2,2-bis(p-
chloropheny1)ethane (DDT). J. Pharmacol. Exp. Ther., 188, 300-309.
Binning, A,, Darby, F. J., Heenan, M. P., and Smith, J. N. (1967). The
conjugation of phenols with phosphate in grass grubs and flies. Biochem. J.,
103, 42-48.
Birnie, G. D. (1972). ‘Subcellular components. Preparation and fractionation’.
2nd Edn, Butterworths/University Park Press.
Bouteille, M., Laval, M., and Dupuy-Coin, A. M. (1974). Localisation of nuclear
functions as revealed by ultrastructural autoradiography and cytochemistry .
In: ‘The Cell Nucleus’. Vol I, Ch 1. Ed. H. Busch, Academic Press, New
York, pp. 3-71.
Boyland, E., Kinder, C. H., and Manson, D. (1961). The biochemistry of
aromatic amines. 8, Synthesis and detection oi’ di-(2-amino-l-naphthyl)
hydrogen phosphate, a metabolite of 2-naphthylamine in dogs. Biochem. J.,
78, 175-179.
Brenner, B. M. and Rector, F. C. (1976). ‘The Kidney’. W. B. Saunders,
Philadelphia, USA.
Breuer, H., Kime, D. E., and Knuppen, R. (1973). Metabolism of 6-chloro-9P-
lO-pregna-l,4,6-triene-3,20-dione in rat, rabbit, monkey, and man. Acta
Endocrinol. (Copenhagen), 74, 127-143.
Brown, R. R., Miller, J. A., and Miller, E. C. (1954). The metabolism of
methylated azo-dyes. IV Dietary factors enhancing demethylation in uitro. J .
Biol. Chem., 209, 211-222.
Burchell, B.and Coughtrie, M. W. H. (1992). UDP-glucuronosyl-transferases. In:
‘Pharmacogenetics of Drug Metabolism’. Ed. W. Kalow. Pergamon Press
Inc., New York, pp. 195-225.
Caldwell, J. (1981). Current status of attempts to predict species differences in
drug metabolism. Drug Metab. Rev., 12, 221-237.
Campbell, T. C. and Hayes, J. R. (1974). Role of nutrition in the drug
metabolising enzyme system. Pharmacol. Rev., 26, 171-197.

Campbell, T. C. and Hayes, J. R. (1976). The effect of quantity and quality of
dietary protein on drug metabolism. Fed. Proc. Fed. A m . Soc. Exp. Biol.,
35, 2470-2474.
Capel, I. D., French, M. R., Millburn, P., Smith, R. L., and Williams, R. T.
(1972). The fate of [14C]phenolin various species. Xenobiotica, 2, 25-34.
Claude, A (1969). Microsomes, endoplasmic reticulum and interactions of
cytoplasmic membranes. In: ‘Microsomes and Drug Oxidations’. Eds J R.
Gillette, A. H. Conney, G. J. Cosmides, R. W. Estabrook, J. R. Fouts, and
G. J. Mannering. Academic Press, NY, pp. 1-39.
Clifton, G. and Kaplowitz, N. (1978). Effect of dietary phenobarbital, 3,4-
benzo(a)pyrene and 3-methylcholanthrene on hepatic, intestinal, and renal
glutathione S-transferase in the rat. Biochem. Pharmacol., 27, 1284-1287.
Cohen, A. J. (1979). Critical review of the toxicology of coumarin with special
reference to interspecies differences in metabolism and hepatotoxic response
and their significance to man. Food Cosmet. Toxicol., 17, 277-289.
Coon, M. J., Autor, A. P., Boyer, R. F., Lode, E. T., and Strobel, H. (1973).
In: ‘Oxidases and Related Redox Systems’. Eds T. E. King, H. S. Mason,
and M. Morrison. University Park Press, Baltimore, p. 529.
Cooper, R. G., Evans, D. A. P., and Whibley, E. J. (1984) Polymorphic
hydroxylation of perhexiline maleate in man. J . Med. Genet., 21, 27-33.
Coulston, F. and Serrone, D. M. (1969). The comparative approach to the role of
non-human primates in evaluation of drug toxicity in man. Ann. N . Y . Acad.
Sci., 162, 681-704.
Daniel, J. W., Gage, J. C., and Jones, D. I. (1968). The metabolism of
3,5-di-tert-butyl-4-hydroxytoluene in the rat and in man. Biochem. J . , 106,
783-790.
Das, M. and Radhakrishnan, A. N. (1976). Role of peptidase and peptide
transport in the intestinal absorption of proteins. World Rev. Nutr. Diet, 24,
58-87.
Davson, H, (1943). In: ‘Permeability of Natural Membranes’. Eds H. Davson
and J. F. Danielli. Cambridge University Press, p. 11.
DePierre, J. W. and Ernster, L. (1977). Enzyme topology of intracellular
membranes. Annu. Rev. Biochem., 46, 201-262.
Dierickx, P. J. (1983). Interaction of chlorophenoxyalkyl acid herbicides with rat
liver glutathione S-transferases. Food Chem. Toxicol., 21, 575-579.
Duggan, D. E., Hooke, K. F., Noll, R. M., and Kwan, K. C. (1975).
Enterohepatic circulation of indomethacin and its role in intestinal irritation
Dutton, G. J. (1978) Developmental aspects of drug conjugation, with special
references to glucuronidation Annu. Rev. Pharm. Toxicol., 18, 17-36.
Eakins, M. N., Conroy, P. J., Searle, A. J., Slater, T. F., and Willson, R. L.
(1976). Metronidazole (Flagyl) a radio-sensitizer of possible clinical use in
cancer chemotherapy: some biochemical and pharmacological consideration.
Eling, T., Boyd, J., Reed, G., Mason, R., and Sivarajah, K. (1983). Xenobiotic
metabolism by prostaglandin endoperoxide synthetase. Drug Metab. Rev.,
14, 1023-1053.
Ernster, L. and Kylenstierna, B. (1969). Mitochondria1 membranes-structure
composition and function. In : ‘Mitochondria-Structure and Function’, Eds
L. Ernster and Z . Drahota, FEBS Symposium 17, pp. 5-31.
Feuer, G., Sosa-Lucero, J. C., Lumb, G., and Moddel, G. (1971). Failure of
various drugs to induce drug-metabolising enzymes in extrahepatic tissues of
the rat. Toxicol. Appl. Pharmacol., 19, 579-589.
Food Safety Council (1978). Proposed System for Food Safety Assessment. Food.
Cosmet. Toxicol., 16 (Suppl 2), 1-36.
Freedberg, I. M. (1972). Pathways and controls of epithelial protein synthesis. J .
Invest. Dermatol., 59, 56-65.
Gibson, G. G. and Schenkman, J. B. (1978). Purification and properties of
cytochrome P-450 obtained from liver microsomes of untreated rats by lauric
acid affinity chromatography. J . Biol. Chem., 253, 5957-5963.
Gibson, G. G. and Skett, P. (1986). In: ‘Introduction to Drug Metabolism’.
Chapman and Hall, Ch. 5.
Giri, S. N. (1973). The pharmacological action and 0-demethylation of glyceryl
guaiacolate ether in male and female rats. Toxicol. Appl. Pharmacol., 24,
513-518.
Gonzalez, F. J. (1988). The molecular biology of P450s. Pharmacol. Rev., 40,
243-288.
Gonzalez, F. J. (1990). Molecular genetics of the P450 superfamily. Pharmacol.
Ther., 45, 1-38.
Gorrod, J. W. and Damani, L. A. (1985) ‘Biological Oxidation of Nitrogen in
Organic Molecules’. Ellis Honvord Ltd., Chichester, UK.
Gould, B. J., Amoah, A. G. B., and Parke, D. V. (1986). Stereoselective
pharmacokinetics of perhexiline. Xenobiotica, 16, 491-502.
Gram, T. E. (1980). The metabolism of xenobiotics by mammalian lung. In:
‘Extrahepatic Metabolism of Drugs and Other Foreign Compounds’. Ed. T.
E. Gram. Ch. 3, MTP Press Ltd., London. pp. 159-209.
Gram, T. E. (1982). ‘Extrahepatic Metabolism of Drugs and Other Foreign
Compounds’. MTP Press Ltd., London.
Gram, T. E., Schroeder, D. H., Davis, D. C., Reagan, R. L., and Guarino, A.
M. (1971). Further studies on the submicrosomal distribution of drug
metabolising components in liver. Localisation in fractions of smooth
microsomes. Biochem. Pharmacol., 20, 2885-2893.
Gram, T. E., Guarino, A. M., Schroeder, D. H., Davis, D. C., Reagan, R. L.,
and Gillette, J. R. (1970). The effect of starvation on the kinetics of drug
oxidation by hepatic microsomal enzymes from male and female rats. J .
Pharmacol. Exp. Ther., 175, 12-21.
Guengerich, F. P. (1979). Similarities of nuclear and microsomal cytochromes
P-450 in the in vitro activation of aflatoxin B,. Biochem. Pharmacol., 28,
2883-2890.
Guthenberg, C., Morgenstern, R., DePierre, J. W . , and Mannervik, B. (1980).
Induction of glutathione S-transferases A , B and C in rat liver cytosol by
trans-stilbene oxide. Biochem. Biophys. Acta, 631, 1-10.
Habig, W. H., Pabst, M. J., and Jakoby, W. B. (1974). Glutathione S-
transferases; the first enzymic step in mercapturic acid formation. J. Biol.
Chem., 249, 7130-7139.
Hanninen, 0. and Marniemi, J. (1970). Inhibition of glucuronide synthesis by
physiological metabolites in liver slices. FEBS Lett., 6, 177-181.
Hansch, C. and Clayton, J. M. (1973). Lipophilic character and biological activity
of drugs. 11, The parabolic case. J . Pharm. Sci., 62, 1-21.
Hathway, D. E. (1970-1981). Specialist Periodical Reports, Vol. 1-6 ‘Foreign
Compound Metabolism in Mammals’. The Chemical Society. London.
Hayes, J. R. and Campbell, T. C. (1980). Nutrition as a modifier of chemical

carcinogenesis. In : ‘Carcinogenesis’ Vol. 15, ‘Modifiers of Chemical Car-
cinogenesis’ Ed. J. J. Slaga, Raven Press, New York, Chapter 11, pp.
207-241.
Hayes, W. J. (1975). ‘Toxicology of Pesticides’. Williams & Wilkins Co.
Baltimore.
Heinrichs, W. L. and Juchau, M. R. (1980). Extrahepatic drug metabolism: The
Gonads, In: ‘Extrahepatic Metabolism of Drugs and other Foreign Com-
pounds’ Ed. T. E. Gram, Ch 9. MTP Press Ltd., London. pp. 319-332.
Hewick, D. S. (1982). Reductive metabolism of nitrogen containing functional
groups. In : ‘Enzymic Basis of Detoxification: Metabolism of Functional
Groups’. Edr W. B. Jackoby, J. R. Bend, and J. Caldwell. Academic Press,
New York. pp. 151-170.
Hoensch, H., Woo, C. H., and Schmid, R. (1975). Cytochrome P-450 and drug
metabolism in intestinal villous and crypt cells of rats: Effect of dietary iron.
Biochem. Biophys. Res. Commun., 65, 399-405.
Holtzman, J. L., Gram, T. E., Gigon, P. L., and Gillette, J. R. (1968). The
distribution of the components of mixed-function oxidase between the rough
and the smooth endoplasmic reticulum of liver cells. Biochem J . , 110,
407-412.
Hosokawa, M., Maki, T., and Satoh, T. (1990). Characterisation of molecular
species of liver microsomal carboxylesterases of several animal species and
humans. Arch. Biochem. Biophys., 277, 219-227.
Houston, J. B. and Wood, S . G. (1980). Gastro-intestinal absorption of drugs
and other xenobiotics. In: ‘Progress in Drug Metabolism’, Vol. 4. Eds J. W.
Bridges and L. F. Chasseaud. J. Wiley and Sons, pp. 57-129.
Hsia, S. L. (1971). Steroid metabolism in human skin. Mod. Trends Dermatol., 4,
69-88.
Huang, M-T., Miwa, G. T., and Lu, A. Y. H. (1979). Rat liver cytosolic azo-
reductase. Purification and characterisation. J . Biol. Chem., 254, 3930-3934.
Idle, J. R., Millburn, P., and Williams, R. T. (1978). Taurine conjugates as
metabolites of arylacetic acids in the ferret. Xenobiotica, 8, 253-264.
Jakoby, W. B., Bend J. R., and Caldwell, J. (1982). ‘Metabolic Basis of
Detoxification: Metabolism of Functional Groups’. Academic Press, New
York.
Jakoby, W. B., Sekura, R. D., Lyon, E. S . , Marcus, C . J., and Wang, J-L.
(1980). Sulphotransferases. In: ‘Enzymic Basis of Detoxification’, Ed. W. B.
Jackoby. Academic Press, New York. Vol. 11, pp. 199-228.
James, M. O., Smith, R. L., and Williams, R. T. (1972). The conjugation of
4-chloro- and 4-nitro-phenylacetic acid in man, monkey and rat. Xenobiotica,
2, 499-506.
Jones, D. P., Orrhenius, S . , and Jakobson, S. W. (1980). Cytochrome P. 450-
linked mono-oxygenase systems in the kidney. In : ‘Extrahepatic Metabolism
of Drugs and Foreign Compounds’, Ed. T. E. Gram, Ch. 2, MTP Press Ltd.,
London. pp. 123-158.
Juchau, M. R. (1980). Extrahepatic drug biotransformation in the placenta. In:
‘Extrahepatic Metabolism of Drugs and Foreign Compounds’. Ed. T. E.
Gram, Ch. 4, MTP Press Ltd. London. pp. 211-238.
Jusko, W. J. and Gretch, M. (1976). Plasma and tissue protein binding of drugs
in pharmacokinetics. Drug Metabol. Rev., 5, 43-140.
Kao, Y. (1981). Tetrodotoxin, Saxitoxin, Chiriquitoxin. New persepctives on
ionic channels. Fed. Proc., Fed. A m . SOC.Exp. Biol., 40, 30-35.

Kappas, A., Anderson, K. E., Conney, A. H., and Alvares, P. (1976). Influence
of dietary protein and carbohydrate on antipyrene and theophylline metabo-
lism in man. Clin. Pharmacol. Ther., 20, 643-653.
Kato, R. and Gillette, J. R. (1965). Sex differences in the effects of abnormal
physiological states on the metabolism of drugs by rat liver microsomes. J .
Pharmacol. Exp. Ther., 150, 285-291.
Kawajiri, K., Yonekawa, H., Hara, E., and Tagashira. Y. (1979). Activation of
2-acetylaminofluorene in the nuclei of rat liver. Cancer Res., 39, 1089-1093.
Killenberg, P. G. and Jordan, J. T. (1978). Purification and characterisation of
bile acid-CoA: amino acid N-acetyltransferase from rat liver. J . Biol. Chem.,
253, 1005-1010.
Kimura, S., Gonzalez, F. J., and Nebert, D. W. (1986). Tissue-specific expression
of the mouse dioxin-inducible PI-450 and P,-450 genes: differential tran-
scriptional activation and mRNA stability in liver and extrahepatic tissues.
Mol. Cell Biol., 6, 1471-1477.
Kohli, K. K., Mukhtar, H., Bend, J. R., Albro, P. W . , and McKinney, J. D..
(1979). Biochemical effects of pure isomers of hexachlorobiphenyl-Hepatic
microsomal epoxide hydrase and cytosol glutathione-S-transferase activities
in the rat. Biochem. Pharmacol., 28, 1444-1446.
Kuruno, Y., Kondo, T. and Ikeda, K. (1983). Esterase-like activity of human
serum albumin: enantioselectivity in the burst phase of reaction with
p-nitrophenyl-methoxyphenyl acetate. Arch. Biochem. Biophys., 227, 339-
341.
Lambert, L. and Wills, E. D. (1977). The effect of dietary and lipid peroxide
sterols and oxidised sterols on cytochrome P-450 and oxidative demethyla-
tion in the endoplasmic reticulum. Biochem. Pharmacol., 26, 1417-1421.
Lansing, A. I., Belkhode, M. L., Lynch, W. E., and Lieberman, I. (1967).
Enzymes of plasma membrane of liver. J . Biol. Chem., 242, 1772-1775.
LeFevre, M. E. and Joel, D. D. (1977). Mini review. Intestinal absorption of
particulate matter. Life Sci., 21, 1403-1408.
Levin, A. A. and Dent, J. G. (1982). Comparison of the metabolism of
nitrobenzene by hepatic microsomes and caecal microflora from Fischer F344
rats in vitro and the relative importance of each in vivo. Drug Metab.
Dispos., 10, 450-454.
Levy, G., Khanna, N. N., Soda, D. M., Tsuzuki, O . , and Stern, L. (1975).
Pharmacokinetics of acetaminophen in the human neonate. Formation of
acetaminophen glucuronide and sulfate in relation to plasma bilirubin
concentration and D-glucaric acid excretion. Pediatrics, 55, 818-825.
Lindstrom, T. D. and Whitaker, G. W. (1984). Saturation of a,P-unsaturated
ketone: a novel xenobiotic biotransformation in mammals. Xenobiotica, 14,
503-508.
Lotlikar, P. D., Enomoto, M., Miller, J. A., and Miller C. (1967). Species
variation in the N - and ring hydroxylation of 2-acetylaminofluorene and
effects of 3-methylcholanthrene treatment. Proc. SOC.Exp. Biol. Med., 125,
341-346.
Lucier, G. W., McDaniel, 0. S . , and Matthews, H. B. (1971). Microsomal rat
liver UDP Glucuronyltransferase: Effects of piperonyl butoxide and other
factors on enzyme activity. Arch. Biochem. Biophys., 145, 520-530.
Lush, I. E. and Andrews, K. M. (1978). Genetic variation between mice in the
metabolism of coumarin and its derivatives. Genet. R e x , 31, 177-186.
176 9 The Metabolism and Disposition of Xenobiotics
Mannervik, B. (1987). Glutathione transferase: In: ‘Drug Metabolism-from

Molecules to Man’ Eds D. J. Benford, J. W. Bridges, and G. G. Gibson.
Taylor and Francis Ltd., London, pp. 30-39.
Mannervik, B., Alin, P., Guthenberg, C., Jensson, H., and Warholm. M. (1985).
In: ‘Microsomes and Drug Oxidation’ Eds A. R. Boobis, J. Caldwell, F. De
Matteis, and C. Elcombe. Taylor and Francis, London, p. 221.
Mannervik, B. and Danielson, U. H. (1988). Glutathione transferases. Structure
and catalytic activity. CRC Crit. Rev. Biochem., 23, 283-336.
Marshall, W. J. and McLean, A. E. M. (1971). A requirement for dietary lipids
for induction of cytochrome P-450 by phenobarbitone in rat liver microsomal
fraction. Biochern. J . , 122, 569-573.
Matthews, D. M. (1975) Intestinal absorption of peptides. Physiol. Rev., 55,
537-608.
McMahon, R. E. (1971). Enzymic oxidation and reduction of alcohols, aldehydes
and ketones. Handb. Exp. Pharmacol., 28, 500-517.
Mueller, G. C. and Miller, J. A. (1950). The reductive cleavage of 4-dimethyl-
aminoazobenzene by rat liver: Reactivation of carbon dioxide treated
homogenates by riboflavin-adenine dinucleotide J . Biol. Chem., 185, 145-
154.
Mulder, G. J. (1974). On non-specific inhibition of rat liver microsomal
UDP-glucuronyltransferase by some drugs. Biochem. Pharmacol., 23, 1283-
1291.
Mulder, G. J., Meerman, J. H. H. and van der Goorbergh A. M. (1986). In:
‘Xenobiotic Conjugation Chemistry’ Eds. G. D. Paulson, J. Caldwell, D. H.
Hutson, and J. J. Menn. Am. Chem. SOC., Washington, p. 282.
Murray, M. (1987). Mechanisms of the inhibition of cytochrome P-450-mediated
drug oxidation by therapeutic agents. Drug Metah. Rev., 18, 55-8 1.
Murray, M. and Reidy, G. F. (1990). Selectivity in the inhibition of mammalian
cytochromes P-450 by chemical agents. Pharmacol. Rev., 42, 85-101.
Nash, J. A., King, L. J., Locke, E. A., and Green, T. (1984). The metabolism
and disposition of hexachloro-1,3-butadiene in the rat and its relevance to
nephrotoxicity . Toxicol. Appl. Pharmacol., 73, 124-137.
Nebert, D. W., Robinson, J. R., Niwa, A., Kumaki, K., and Poland, A. P.
(1975). Genetic expression of aryl hydrocarbon hydroxylase activity in the
mouse. Cell Physiol., 85, 393-414.
Nebert, D. W., Nelson, D. R., Coon, M. J., Estabrook, R. W., Feyereisen, R., et
al. (1991). The P450 swperfamily: Update on new sequences, gene mapping
and recommended nomenclature. D N A Cell Biol., 10, 1-1 4.
Olson, M. 0. J. and Busch, H. (1974). Nuclear proteins. In: ‘The Cell
Nucleus’, Vol. 111, Ch. 6, Ed. H. Busch Academic Press, New York, pp.
211-268.
Osterloh, J. D., Karakaya, A., and Carter, D. E. (1980). Isolation and
identification of the polar metabolites of chloropheniramine in the dog. Drug
Metab. Dispos., 8, 12-14.
Pannatier, A,, Jenner, P., Testa, B., and Etter, J. C. (1978). The skin as a
drug-metabolising organ. Drug Metab. Rev., 8, 319-343.
Parke, D. V. and Ioannides, C. (1981). The role of nutrition in toxicology. Annu.
Rev. Nutr., 1, 207-234.
Paulson, G. D., Caldwell, J., Hutson, D. H., and Menn, J. J. (1986). Xenobiotic
Conjugation Chemistry, ACS Symposium Series 299, Washington DC, USA.
Pohl, R. J., Philpot, R. N., and Fouts, J. R. (1976). Cytochrome P-450 content
and mixed function oxidase activity in microsomes isolated from mouse skin.
Drug Metab. Dispos., 4, 442-450.
Powis, G., and Wincentsen, L. (1980). Pyridine nucleotide co-factor require-
ments of indicine-N-oxide reduction by hepatic microsomal cytochrome
P-450. Biochem. Pharmacol., 29, 347-351.
Rappaport, A. M. (1969). Anatomic considerations: In: ‘Diseases of the Liver’.
Ed. L. Schiff. 3rd Edn., J. B. Lippincott Co, Philadelphia. pp. 1-49.
Raw, I. (1978). Cytochrome P-450 in the liver mitochondria1 outer membrane of
20-methylcholanthrene or Aroclor treated rabbits. Biochem. Biophys. Res.
Commun., 81, 1294-1297.
Reddy, J. K. and Lalwani, N. D. (1983). Carcinogenesis by hepatic peroxisome
proliferators: Evaluation of the risk of hypolipidemic drugs and industrial
plasticizers to humans. Crit. Rev. Toxicol., 12, 1-58.
Redmond, G. P. and Cohen, G. (1972). Sex difference in acetaldehyde
exhalation following ethanol administration in C57BL mice. Nature
(London), 236, 117.
Remmer, H., Schenkman, J., Estabrook, R. W., Sasame, H., Gillette, J.,
Narasimhula, S . , Cooper D. Y., and Rosenthal, 0. (1966). Drug interaction
with hepatic microsomal cytochrome. Mol. Pharmacol., 2, 187-190.
Renwick, A. G. (1986). The metabolism of intense sweetners. Xenobiotica, 16,
1057-1071.
Rongone, E. L. (1977). Cutaneous metabolism. In: Adv. Mod. Toxicol. Vol. 4:
‘Dermatotoxicology and Pharmacology’. Eds R. N. Margulli, and H. J.
Marbach. John Wiley and Sons, New York.
Rowland, I. R. and Walker, R. (1983). The gastrointestinal tract in food
toxicology In: ‘Toxic Hazards in Food’. Eds D. M. Conning and A. B. G.
Lansdown. Ch. 6. Croom Helm, London and Canberra. pp. 183-274.
Rowland, I. R., Mallett, A. K., and Wise, A. (1985). The effect of diet on the
mammalian gut flora and its metabolic activity. CRC Crit. Rev. Toxicol., 16,
31- 103.
Ryan, D. E., Thomas, P. E., Korzeniowski, D., and Levin, W. (1979).
Separation and characterisation of highly purified forms of liver microsomal
cytochrome P-450 from rats treated with polychlorinated biphenyls, pheno-
barbital, and 3-methylcholanthrene. J . Biol. Chem., 254, 1365-1374.
Sato, R. and Omura, T. (1978). ‘Cytochrome P-450’. Academic Press, New
York.
Savage, D. C. (1977). Interactions between the host and its microbes. In:
‘Microbial Ecology of the Gut’. Eds R. T. J. Clark and T. Bauchop.
Academic Press, London. pp. 277-310.
Sawada, H., Hara, A., Kato, F. and Nakayama, T . (1979). Purification and
properties of reductases for aromatic aldehydes and ketones from guinea-pig
liver. J. Biochem. (Tokyo), 86, 871-881.
Scala, J., McOsker, D. E., and Reller, H. H. (1968). The percutaneous
absorption of ionic surfactants. J. Invest. Dermatol., 50, 371-379.
Schanker, L. S. and Jeffrey, J. J. (1961). Active transport of foreign pyrimidines
across the intestinal epithelium. Nature (London), 190, 727-728.
Scheline, R. R. (1973). Metabolism of foreign compounds by gastrointestinal
micro-organisms. Pharmacol. Rev., 25, 45 1-523.
Scheline, R. R., (1986). I n : ‘Extrahepatic Metabolism of drugs and Other
Foreign Compounds’. Ed. T. E. Gram, MTP Press Ltd., London, pp.

551-580.
Schuster, I. (1989). Cytochrome P-450 Biochemistry and Biophysics. Taylor and
Francis, London.
Scott, R. C., Guy, R. H., and Hadgraft, J. (1990). Prediction of Percutaneous
Penetration. Methods, Measurement, Modelling. IBC Technical Services
Ltd., London.
Sharratt, M., Grasso, P., Carpanini, F. M. B., and Gangolli, S. D. (1970).
Carrageenan ulceration as a model for human ulcerative colitis. Lancet, 2,
932.
Shichi, H., Kumaki, K., and Nebert, D. W. (1978). Circular dichroism studies on
the binding of type I substrates and reverse type I compounds to rabbit liver
microsomal cytochrome P-450. Chem. Biol. Interact., 20, 133- 148.
Shiobara, Y., (1977)., The effect of carboxyl substituents on the metabolism of
aromatic aldehydes. Xenobiotica, 7, 457-468.
Sikstrom, R., Lanoix, J., and Bergeron, J. J. (1976). An enzymic analysis of a
nuclear envelope fraction. Biochem. Biophys. Acta, 448, 88-102.
Smith, L. L. et al. (1978). In: ‘Industrial and Environmental Xenobiotics’. Eds J.
R. Fouts and I. Gut. Excerpta Medica, Amsterdam, pp. 135-140.
Snell, K. and Mullock, B. (1987) ‘Biochemical Toxicology, a Practical Approach’
IRL Press, Oxford.
Song, B.-J., Veech, R. L., Park, S. S., Gelboin, H. V., and Gonzalez, F. J. (1989).
Induction of rat hepatic N-nitrosodimethylamine demethylase is due to
protein stabilization. J . B i d . Chern., 264, 3568-3572.
Sorokin, S. P. (1970). The cells of the lung. In: ‘Morphology of Experimental
Respiratory Carcinogenesis’. Eds P. Nettesheim, M. R. Hanna, and J. W.
Deatherage U.S. Atomic Energy Commission.
Soucek, P. and Gut, I. (1 992). Cytochromes P-450 in rats: structures, functions,
properties, and relevant human forms. Xenobiotica, 22, 83-1 03.
Stockman, P. K . , Beckett, G. J., and Hayes, J. R. (1985). Identification of a
basic hybrid glutathione S-transferase from human liver. Biochem. J., 227,
457-465.
Strother, A., Throckmorton, J. K . , and Herzer, C. (1971). The influence of high
sugar consumption by mice on the duration of action of barbiturates and in
vitro metabolism of barbiturates, aniline and p-nitroanisole. J . Pharmacol.
Exp. Med., 179, 490-498.
Tannen, R. H. and Webber, W. W. (1979). Rodent models of the human
isoniazid-acetylator polymorphism. Drug Metab. Dispos., 7 , 274-279.
Tappel, A. L. (1969). Lysozomal enzymes and other components. In: ‘Lysozomes
in Biology and Pathology’, 2. Eds J. T. Dingle and H. B. Fell. Frontiers of
Biology 14B, Ch. 9, North Holland Publishing Co., Amsterdam, pp.
209-244.
Tatsumi, K . , Yamada, H., and Kitamura, S. (1983) Reductive metabolism of
N-nitrosodiphenylamine to the corresponding hydrazine derivative. Arch.
Biochem. Biophys., 226, 174-181.
Testa, B. (1987). Recently discovered routes of metabolism. In: ‘Drug
Metabolism-from Molecules to Man’. Eds D. J. Benford, J. W. Bridges,
and G. G. Gibson. Taylor and Francis, London pp. 563-580.
Testa, B. and Jenner, P. (1976). ‘Drug Metabolism: Chemical and Biochemical
Aspects’. Marcel Dekker Inc., New York.
Testa, B. and Jenner, P. (1981). Inhibitors of cytochrome P-450’s and their

mechanism of action. Drug Metub. Rev., 12, 1-1 17.
Thompson, S. and Slaga, T. J. (1976). Mouse epidermal aryl hydrocarbon
hydroxylase. J . Invest. Dermatol., 66, 108-111.
Thorgeirrson, S. S . , Felton, J. S. and Nebert, D. W. (1975). Genetic differences
in the aromatic hydrocarbon inducible N-hydroxylation of 2-
acetylaminofluorene and acetaminophen-produced hepatotoxicity in mice.
Mol. Pharmacol., 11, 159-165.
Tobin, T., Akera, T., Han, C. S . , and Brody, T. M. (1974). Lithium and
rubidium interactions with sodium- and potassium-dependent adenosine
triphosphatase. A molecular basis for the pharmacological actions of these
ions. Mol. Pharmacol., 10, 501-508.
Troll, W., Tessler, A. N., and Nelson, N. (1963) Bis(2-amino-1-naphthyl)
phosphate, a metabolite of P-naphthylamine in human urine. J . Urol.
Nephrol., 89, 626-627.
Tuey, D. B. and Matthews, H. B. (1981). Distribution and excretion of
2,2’ ,4,4’,5,5’-hexabromobiphenyl in rats and man: Pharmacokinetic model
predictions. Toxicol. Appl. Pharmacol., 53, 420-43 1.
Uehleke, H. (1968). Extrahepatic microsomal drug metabolism. In : ‘Sensitization
to Drugs’. Proc. Tenth European SOC. for the Study of Drug Toxicity
Excerpta Media Intn. Congress Series 181, Oxford pp. 94-100.
Ullrich, V., Roots, A . , Hildebrandt, A., Estabrook, R. W., and Conney, A. H.
(1977). ‘Microsomes and Drug Oxidation’. Pergamon Press, Oxford.
Ullrick, K. J., Capasso, G., Rumrich, G., and Sato, K. (1978). Effect of
p-chloromercuribenzoate (pCMB), ouabain, 4-acetamido-4’-isocyanoto-
stilbene-2,2’-disulphonic acid (SITS) on proximal tubular transport proc-
esses. In: ‘Membrane Toxicity’ Eds M. W. Miller and A. E. Shamoo.
Plenum Press, NY. pp. 3-10.
Unger, S. H., Cook, J. R. and Hollenberg, J. S. (1978). Simple procedure for
determining octanol-aqueous partition, distribution and ionisation
coefficients by reversed-phase high-pressure liquid chromatography. J .
Pharm. Sci., 67, 1364-1367.
Van Bladeren, P. J., Bruggeman, I. M., Jongen, W. M. F., Scheffer, A. G., and
Temmink, J. H. M. (1987). The role of conjugating enzymes in toxic
metabolite formation. In: ‘Drug Metabolism-From Molecules to Man’. Eds
D. J. Benford, J. W. Bridges, and G. G. Gibson. Taylor and Francis,
London, pp. 151-170.
Venkatesan, N., Arcos, J. C., and Argus, M. F. (1970). Amino acid induction
and carbohydrate repression of dimethylnitrosamine demethylase in rat liver.
Cancer Res., 30, 2563-2567.
Vos, R. M . E. and Van Bladeren, P. J . (1990). Glutathione S-transferases in
relation to their role in the biotransformation of xenobiotics. Chem. Biol.
Interact., 75, 241-265.
Walker, W. A., Isselbacher, K. J., and Bloch, K. J. (1972). Intestinal uptake of
macromolecules: Effect of oral immunisation. Science, 177, 608-610.
Wang, C. Y . , Behrens, B. C., Ichikawa, M . , and Bryan, G. T. (1974).
Nitroreduction of 5-nitrofuran derivatives by rat liver xanthine oxidase and
reduced nicotinamide adenine dinucleotide phosphate-cytochrome c reduc-
tase. Biochem. Pharmacol., 23, 3395-3404.
Waring, R. H. and Mitchell, S. C. (1985). The metabolism of phenothiazine in
180 The Metabolism and Disposition of’ Xenobiotics
the neonatal calf: identification of drug-polypeptide conjugates from urine.

Xenobiotica, 15, 459-468.
Watanabe, P. G., Ramsey, J. C., and Gehring P. J. (1980) Pharmacokinetics and
metabolism of industrial chemicals. In: ‘Progress in Drug Metabolism’,
Vol. 5. Eds J. W. Bridges and L. F. Chasseaud. J. Wiley and Sons, pp.
311-343.
Weiss, B. and Levin, R. M. (1978). Mechanism for selectively inhibiting the
activation of cyclic nucleotide phosphodiesterase and adenylate cyclase by
antipsychotic agents. Advan. Cyclic Nucleotide Res., 9, 285-303.
Wermuth, B . , Munch, J. D. B., and von Wartburg, J. P. (1977). Purification and
properties of NADPH-dependent aldehyde reductase from human liver. J.
Biol. Chem., 252, 3821-3828.
Wester, R. C. and Maibach, H. I. (1977). Percutaneous absorption in man and
animals: a perspective. In: ‘Cutaneous Toxicity’, Eds V. A. Drill and P.
Lazar. Academic Press, New York.
WHO Scientific Group (1967). Procedures for Investigating Intentional and
Unintentional Food Additives. WHO Tech Rep. Ser., 348, WHO, Geneva.
Williams, R. T. (1959) ‘Detoxification Mechanisms’, 2nd Edn. Chapman and Hall
Ltd., London.
Willemsens, G., Vanden Bossche, H., and Lavrijsen, K. (1989). Cytochrome
P-450-dependent metabolism of retinonic acid in epidermal microsomes from
neonatal rats. Effects of azole antifungals. In: ‘Cytochrome P450: Biochem-
istry and Biophysics’ Ed. I. Schuster. Taylor and Francis, London, pp.
767-770.
Wollenberg D. and Ullrich, V. (1977). Characterisation of the drug mono-
oxygenase system in mouse small intestine. In: ‘Microsomes and Drug
Oxidation’. Eds V. Ullrich, et al., Pergamon Press, New York. pp. 675-679.
Ziegler, D. M. (1980). Microsomal flavin-containing mono-oxygenases. In:
‘Enzymatic Basis of Detoxification’ Vol. I. Ed. W. B. Jakoby. Academic
Press, New York. Vol. I, pp. 201-230.
CHAPTER 10
Theory and Practice in

Metabolic Studies
J. C. PHILLIPS AND S. D. GANGOLLI
Introduction
In the design and conduct of experimental studies on the metabolic fate
and biological disposition of a compound, two important aspects merit
consideration. These are the underlying theoretical principles and the
relevant practical techniques. A brief description of the salient features of
these two aspects follows:
1 Theoretical Considerations
The biological effect that a foreign compound has on a living organism is
a function of the concentration of that compound at the target site and
the duration of exposure. Pharmacokinetics, which is the discipline that
quantifies with respect to time, the uptake, distribution, metabolism, and
elimination of compounds, therefore provides important information with
which to assess the potential hazard of the compound. Understanding the
pharmacokinetic characteristics of a compound over a wide dose range,
after single and repeated exposure, and between species, greatly facilitates
the extrapolation of toxicity data from various test species to man. The
importance of pharmacokinetic studies in toxicology has been the subject
of many recent reviews (see for example, Gehring, Watanabe, and Blau,
1976; Withey, 1977; Watanabe, Ramsey, and Gehring, 1980).
A detailed discussion of the analysis of pharmacokinetics data and the
assumptions inherent in the various mathematical models is beyond the
scope of this chapter and the reader is referred to the many standard
texts (e.g. Goldstein, Aronow, and Kalman, 1974; Niazi, 1979; Gibaldi
and Perrier, 1975). However, some of the basic concepts will be outlined,
and the various pharmacokinetic parameters that can be derived will be
discussed.
181
182 Theory and Practice in Metabolic Studies
The most common approach to pharmacokinetic characterisation is to

depict the body as a system of compartment^^, although it must be
remembered that these compartments do not necessarily correspond to
anatomical or physiological entities. The one-compartment model de-
scribes the body as a single, homogeneous unit and is most useful for the
analysis of blood concentration-time data or urinary excretion-time data
when the administered compound is rapidly distributed in the body. If, as
is frequently the case, the administered compound takes some appreci-
able time to distribute within the body, a two-compartment (or multi-
compartment) model must be used. A two-compartment model consists
of a ‘central compartment’, which contains blood plasma and (in the
absence of data to the contrary) the metabolically active tissues such as
the liver and kidneys, connected to a ‘peripheral compartment’. The
choice of model depends on a number of factors, including the number of
tissues or fluids being sampled and the frequency of sampling.
Many pharmacokinetic processes , for example the rate of absorption,
elimination, or biotransformation of a compound, are satisfactorily de-
scribed by first-order (or linear) kinetics. Under these conditions the rate
of the process is proportional to the concentration of the compound. In
its simplest form, ‘linear’ kinetics obeys the equation:
-dc
-- - kc
dt
where c is the concentration of compound at the time t, and k is the rate
constant.
If, however, one of the pharmacokinetic processes is saturable, as in
the case of carrier-mediated uptake or enzyme-mediated metabolism, the
system may obey non-linear Michaelis-Menten kinetics. In this situation
-dc
-- -
V,C
dt K,+c
where V, and K , are the maximum rate and Michaelis constant of the
process.
At very low concentrations of compound, as is frequently the case in
‘real-life’ situations, c is much less than K,, so that first-order kinetics
apply. In contrast, where c is very large, as in maximum tolerated dose
experiments, the kinetics of the system approximate to zero-order.
A One-compartment Model
If a plot of the logarithm of $he plasma concentration of a compound,
given by intravenous injection, against time can be fitted to a single
straight line, the body is acting as a one-compartment system (Figure 1).
The slope of the line is equal to -ke1/2.303, where k,, is the apparent
J . C. Phillips and S. D. Gangolli 183
Time
Figure 1 Blood concentration-time curve for a compound after i.v. injection into
a one-compartment system. Elimination is apparent first-order with a rate constant
kel
first-order elimination rate constant. k,, is also equal to 0.693/t1,, where

t1,2 is the elimination half-life. The elimination half-life is independent of
the concentration of the compound and is constant for the whole period
of elimination of the compound (if t1,2 varies with dose, it is likely that
either the system is obeying zero-order kinetics, and/or that absorption is
prolonged).
The volume of distribution of the compound (Vd) is a measure of the
apparent space within the body available to contain the compound.
Thus:
v
d = Amount of drug in body/c (3)
where c = concentration in the fluid measured. For this example,
v
d = i.v. dose (D)/concentration at time zero (co)
co is the intercept of the plasma concentration-time curve on the

concentration axis. An alternative method for obtaining V, is to use the
relationship:
where A UC,, cc. is the area under the plasma concentration-time curve,
from time zero to infinity. The area can be estimated by a variety of
methods, including the ‘cut and weight’ method, the trapezoidal rule, or
the interpolation method (Corney and Heath, 1970). This method is
applicable to estimations for non-i.v. administered compounds, and also
to some multi-compartment linear system.
The rate constant of urinary excretion of a compound (kex)is the
difference between the overall elimination rate constant (k,,) and the sum
of the individual rate constants for elimination by all extra-renal
pathways (e.g. biotransformation, biliary excretion). A semi-log plot of
urinary excretion rate of unmetabolised compound against time should
yield a straight line with a slope of -ke1/2.303 and an intercept of
K,, vd co. It is important to remember that the urinary excretion rates
determined experimentally are not instantaneous rates, but averages over
a finite time period. However, the average rate closely approximates to
the instantaneous rate at the mid-point of the collection period, providing
this is no longer than the half-life of the compound. An alternative
method for calculating k,, and the problems of extra-renal elimination
and further biotransformation of primary metabolites are discussed by
Levy and Gibaldi (1975). For some drugs, k,, is proportional to kidney
function as reflected in creatinine or inulin clearance.
The kinetics of urinary excretion may also be characterised by renal
clearance value (CZR).This is the rate of urinary excretion of the
compound divided by the plasma concentration. Thus:
In practice, this is determined by dividing the urinary excretion rate at

the mid-point of the collection period by the plasma concentration at this
time. However, a more satisfactory approach is to plot urinary excretion
rate against plasma concentration at the collection mid-point, the slope of
the resultant straight line being CZ,.
As by definition,
(M,/dt) = k,,A
and A = Vdc from Equation (3), substitution in Equation ( 5 ) shows that
renal clearance also equals k,, vd.
The clearance ratio of a compound is defined as the ratio of its renal
clearance to that of inulin. Since inulin clearance is a measurement of
glomerular filtration rate, a clearance ratio of greater than one indicates
excretion, in part, by renal tubular secretion. If tubular reabsorption is
appreciable and/or glomerular filtration is incomplete (as in the case of a
drug bound to protein), the clearance ratio will be less than unity.
Whole body clearance (Cl,) is analogous to renal clearance and defined
as:
Clb = k,, Vd (7)
J . C. Phillips and S . D . Gangolli 185
Whole body clearance is the sum of the individual clearances by the

various tissues of the body (for example liver, lung, etc.).
Another way of defining clearance, which is independent of any
physiological model, is in terms of blood flow through an organ (Q) and
the concentration of the compound in the blood entering and leaving the
organ. Thus the rate of elimination of the compound from the circulation
by an organ at a steady state, is the difference between the amount
entering the organ (Qcin)and the amount leaving (QcOut). From Equation
(9,
QCin - Qcout
‘‘organ =
Ci n
or CZorgan= Q Eurgan, where Eorganis called the extraction ratio.

The extent of absorption of a compound from the gastro-intestinal tract
can be determined from plasma concentration-time curves following oral
and i.v. administration of the same dose of compound. Thus:
A UC, co (Oral)
Apparent availability =
A UC, co (i.v)
The fraction absorbed ( F ) expressed as a percentage is referred to as the

‘bioavailability’; it is only equal to the apparent availability if the volume
of distribution and elimination rate constant of the compound are
independent of the route of administration.
The rate of absorption of a compound can also be determined from
plasma concentration-time curves. A typical curve, showing an initial lag
phase (Figure 2) can be described by the equation:
where c is the concentration, F i s the fraction of the administered dose D

absorbed, Vd, is the apparent volume of distribution, t is the time and kab
and kel are the rate constants for absorption and elimination respectively.
If kab>kel, a plot of log c against t will eventually have a slope of
-ke1/2.303. Extrapolating this line to t = O and using the method of
residuals, the plot of log residual against t has a slope of -kab/2.303. If
kab<kel, the reverse situation obtains, with the slope of the
concentration-time plot equal to -kab/2.303 and the slope of the
residuals plot equal to -ke1/2.303. This is sometimes called a ‘flip-flop’
model. The details of the mathematical models and alternative methods
for calculating absorption rates are described by Levy and Gibaldi (1975)
as are the limitations of this approach, which applies only if the
compound is absorbed and eliminated by first order processes.
Time
Figure 2 Absorption and elimination of a compound involving two consecutive
first-order processes. The absorption rate constant kabis determined by the method
of residuals
The pharmacokinetics of a compound following repeated, rather than

single, administration is clearly of great importance in toxicological
studies. If the plasma tIl2 is small compared with the interval between
doses, the bulk of the compound may be eliminated during this period
and, thus, the situation after multiple doses will be very similar to that
after a single dose. However, if the tll, is long, the compound will
accumulate in the body as each new dose is added to the residue of the
previous dose, eventually reaching a steady state (plateau). The mathe-
matical treatment of repetitive dose kinetics has been discussed by many
authors, including Dost (1968), Levy and Gibaldi (1975), and Wagner,
Northam, Alway, and Carpenter (1965).
The change in concentration of compound in plasma following re-
peated i.v. administration is described by the equation
where D , V, and k,, are as in Equation (9) above, c is the concentration

in plasma of the compound during the dosing interval, n is the number of
dose, z the time interval between doses and t the time since the last dose.
The average plasma concentration at the plateau (tr)is given by:
where F = the bioavailability (Equation 8).

This equation is valid both for i.v. administration (where F = 1) and
the administration by other routes ( F d 1). The average concentration
C, is not directly related to either the minimum (c,) min or the
maximum (c,) max concentration achieved at steady state. These are
given by:
As the time taken for the plateau concentration to be reached is often

quite long, one practical application of this type of kinetic analysis is the
calculation of a 'loading dose' of compound that will rapidly produce any
desired plateau plasma concentration with any dosing interval and
maintenance dose. The average body burden of compound (X,)at the
steady state is given by:
B Multi-compartment Models
A compound that is injected intravenously into the body usually takes
some time to distribute and therefore the plasma concentration-time
curve is not mono-exponential. If the curve can be resolved into two
linear components (biexponential) it may be described by the equation:
and justify the representation of the body as an open two-compartment

linear system. It is usually assumed that the compound is eliminated
exclusively from the central compartment , and although elimination from
both compartments, or exclusively from the peripheral compartment, is
possible, the three models are not distinguishable on the basis of simple
plasma or urine elimination data. During the initial distributive phase,
the rate of decline of the plasma concentration of the compound is
greater than in the post-distributive phase. If the two curves are resolved
by 'feathering', the slopes of the two lines are -a/2.303 and -/3/2.303
respectively and the intercepts on the concentration axis are A and B
(Figure 3).
Time
Figure 3 Blood concentration of a compound following i.v. injection in a
two-compartmentsystem. A and a are determined by tfeathering’ the curve
It should be noted that the timing of blood sampling dictates the

number of compartments to which the data can be best fitted. For
example, the kinetics of calcium metabolism in man and the rat can be
fitted to a two compartment model if blood is sampled between 2 and 72
hours following an i.v. dose of 45CaC12or to a four-compartment model if
sampled between 0 and 72 hours (Bronner and Lemaire, 1969). The rate
constants for the movement of compound between the central and
peripheral compartment (K12 and K21) and for elimination from the
central compartment ( K l o )are given by:
AP + B a
K21 = K 1 2= a + P - KZ1- K l o and K I o= aP/K21 (14)
A+B
(Mayersohn and Gibaldi, 1971).

It should be noted that the elimination rate constant ( K l o ) is not the
same as the terminal half-life, as in the one-compartment model.
Determination of these rate constants allows an assessment to be made of
the effect of factors such as age, sex, and genetic influences on the
disposition of a compound. A knowledge of K12allows the calculation of
the amount of compound in the peripheral compartment (A,) after i.v.
J . C. Phillips and S. D . Gangolli 189
administration of dose D , using the equation:
Although this may be useful in rationalising pharmacological effects with

tissue levels of compound it should be remembered that A , may not
accurately reflect the actual amount of compound in a particular
‘peripheral’ tissue. The apparent volume of distribution of a compound in
the two-compartment model is given by:
This equation applies regardless of route of administration, but can be

simplified to:
D
v,=
[;+;IP
for i.v. administration only.
By analogy with the single-compartment situation, the fraction of an
oral dose absorbed is given by:
(AUC,,, ) oral x p oral

F=
( A U C , , , ) i.v. x [j i.u.
Whole body clearance in this model is defined as V’P and is equal to

the clearance from the central compartment; it can be calculated from
Equation (16). As in the one-compartment model, average steady state
concentrations after repetitive dosing are given by:
providing elimination is from the central compartment [k,, in Equation

(11) is equivalent to K,, in Equation (18)].
C Non-linear Pharmacokinetics
As mentioned earlier, many biological processes involving enzymes or
carrier-systems are saturable and are thus best described by Michaelis-
Menten equations. There are two limiting cases in in vivo situations; K ,
much larger than c and c much greater than K,. In the former situation,
Equation (2) reduces to:
which has the same form as the first order rate equation, with V,/K,
equivalent to k,,. In the latter case, Equation (2) reduces to
where the rate is constant and independent of c. The biotransformation

of ethanol (Lundquist and Wolthers, 1958) and salicylate (Levy, 1965)
approach the conditions described in Equation 20.
For a compound eliminated by a single capacity-limited process, the
plasma concentration-time curve in the post-absorptive, post-distributive
phase can be used to estimate the in vivo apparent K , and Vm values.
Plotting the rate of change of c(Ac/At) as a function of c at the mid point
of the sampling interval using a linearised version of the Michaelis-
Menten equation:
yields a straight line with a slope of l/Vm and an intercept of Km/Vm.The

change in tissue concentration of a compound and urinary excretion of
chemical and metabolites can also be investigated in non-linear systems
(see for example, Gibaldi and Perrier, 1975).
2 Practical Considerations in Metabolic Studies

The following practical considerations need to be taken into account in
the design and conduct of metabolic studies:
(1) The chemistry of the test compound
(2) The choice of animal model
(3) The choice of appropriate experimental techniques.
A Test Compound
The importance of establishing the chemical identity and purity of the
test compound, and the nature and levels of accompanying contaminants
prior to embarking on metabolic studies, is self evident. Ideally,
metabolic studies should be carried out on chemically defined com-
pounds, free of impurities. As analytical methodology and preparative
techniques improve, this objective is being achieved. However,
difficulties arise in the case of natural products and certain categories of
synthetic compounds which are manufactured to comply with technical

specifications, tailored specifically to meet a technological end use.
Analytical studies on synthetic materials, such as azo-dyes, flavouring
agents, emulsifiers, and industrial solvents, frequently show them to be
complex mixtures of parent compound and subsidiary constituents. Thus ,
metabolic studies may either be undertaken using the purified major
component, as in the recent study with the azo-dye amaranth (Phillips,
Bex, Mendis, Walters, and Gaunt, 1987), or with materials repre-
sentative of the commercially used products such as the studies with the
azo-dye Brown HT (Phillips, Mendis, and Gaunt, 1987) and emulsifier
YN (Phillips, Gaunt, and Gangolli, 1975).
In the case of naturally occurring products, the problems associated
with obtaining a pure compound or even material of standard composi-
tion are formidable. Many of these substances, such as gums, resins,
waxes, and essential oils, are not only complex, poorly defined mixtures,
but also subject to variations in chemical composition due to geographi-
cal, seasonal, and soil factors.
In addition to data on the chemical composition of a test compound, it
is essential to obtain, prior to embarking on metabolic studies, informa-
tion on the physico-chemical properties of the compound. The informa-
tion of particular relevance in this context is that relating to the stability
and reactivity of the test compound. This should include data on the
compound’s volatility, its likelihood to undergo spontaneous degradation
or polymerisation, and its likely reactivity with the vehicle used for
administration or with dietary constituents.
An essential consideration in conducting metabolic studies is the
availability of appropriate methods for the isolation and determination of
the parent compound and its metabolites in biological samples. As
human contact with foreign compounds generally occurs at low levels of
exposure, and metabolic studies are carried out at comparable dose
levels, it is important to establish that the methodologies for ‘clean-up7
procedures and analytical determinations are adequately sensitive and
specific for the purpose. These investigations may be carried out on the
compound per se, or on radiolabelled or stable isotope-labelled
preparations.
B Choice of Animal Model

Mention was made in the previous chapter of some of the factors
responsible for the variability in the absorption , metabolism, and
disposition of foreign compounds in experimental animals and that these
differences could be attributed in part to the species, strain, sex, and age
of the animal. Clearly the influence of these factors needs to be
considered in the choice of an animal model for metabolic studies and,
subsequently, for the conduct of toxicological studies to generate data for
the evaluation of the safety of the compound in man. However, the
seemingly simple problem of choosing an animal model capable of
metabolising a compound in manner similar to man is fraught with

difficulties. (For a discussion of the criteria for species similarities, see
Ruelius, 1975). Variability in the metabolism and pharmacokinetics of
compounds in humans can be as wide as in experimental animals. In
addition to marked differences due to age and sex, genetic polymorphism
is an important determinant for metabolic differences in man. Thus, for
example , the differences in the hydroxylation of the alicyclic antihyper-
tensive drug, debrisoquine, in humans, and in the oxidation of other
compounds including bufuralol, nortriptyline , phenacetin, and phenfor-
min (Davis and Boobis, 1983) have been ascribed to genetic polymor-
phism. Since the ideal situation, that of having identified an entirely
comparable and practical animal model, is unlikely to be achieved, a
working compromise based on a sound understanding of the similarities
and differences in the metabolism and disposition of a compound in man
and in an experimental animal forms the basis for the selection of a
suitable animal species. Difficulties in extrapolating experimental data
obtained in animals to man have been discussed recently by Garattini
(1985).
C Experimental Investigations in Metabolic Studies

The experimental animal species commonly used in metabolic studies are
rodents (rats, mice, guinea-pigs, and hamsters) , canines (dog) , felines
(cat and civet), rabbits, and non-human primates.
The in uivo and in uitro experimental procedures generally used for the
studies include:
(1) Whole animal studies
(2) Isolated perfused organ technique
(3) Cells in culture
(4) Preparations of tissues and subcellular fractions.
The experimental protocol for a particular metabolic study depends on
the nature of the compound and the specific information required. A
brief outline description of the various experimental procedures follows.
Whole Animal Studies

Investigations in the intact animal may be categorised as either non-
invasive procedures or terminal studies, i.e. studies where the animal is
killed following treatment for examination of the internal organs.
Non-invasive procedures are widely used for investigating metabolism
and pharmacokinetics of drugs and food additives in man and also in
large experimental animals (e.g. cat, dog). Briefly, the procedure
involves the collection of serial samples of blood, urine, faeces, saliva, or
respired air following the administration of the test compound, and the
subsequent analysis of the samples for the parent compound and/or
metabolites. Thus the pharmacokinetics of antipyrine, hexabarbitone,
and phenylbutazone in man can be determined by the sequential analysis
of blood samples. In the case of antipyrine, the salivary levels of the drug
have been found to reflect accurately the blood levels, and this method
has been widely used for assessing hepatic microsomal mixed function
oxidase activity in humans (Fraser, Mucklow, Murray, and Davis, 1976).
The ‘14C02breath test’ has been used for measuring the demethylation
of 14C-labelled antipyrine in man (Hepner and Vessell, 1975). This
method has also found application in the metabolism of 14C-labelled
glycodiazine , diazepam, and caffeine and other compounds metabolised
by enzymic demethylation, deacetylation, and decarboxylation reactions
(see review by Bircher and Preisig, 1981). A variation of this procedure is
the use of 13C-labelled compounds, whereby the amount of 13C02 and
other stable isotope-labelled metabolites are measured in the exhaled air
and in body fluids by mass spectrometry or Fourier transform NMR
spectroscopy. This method has been used in investigations on the
metabolism of chloroform in man (Fry, Taylor, and Hathway, 1972) and
the use of stable isotopes in pharmacological research has been reviewed
(Baillie, 1981; Hawkins, 1977).
In ‘terminal’ in vivo studies, the analysis of respired air, body fluids
(including bile), and excreta is accompanied by the examination of
internal organs for residues of the test compound, its metabolites, and
products formed by the incorporation or interaction of the metabolites
with endogenous components. Radiolabelled compounds are widely used
in these studies (see Table 1 for list of radionuclides frequently used in
whole-animal studies) and the vast literature on foreign compound
metabolism is based largely on this method. Recently, attention has
focused on the kinetics of the formation, disposition, and removal of
products formed by the covalent binding of electrophilic metabolites of
test compounds with endogenous macromolecules such as DNA, RNA,
and protein (Pohl and Branchflower, 1981). Examples of such studies
include investigations into the turnover of hepatic DNA and protein and
their adducts following exposure of rats and mice to vinyl chloride and
vinylidene chloride (Reitz, Watanabe, McKenna, Quast , and Gehring
1980).
Whole-body autoradiographic procedures provide another method for
demonstrating the tissue distribution of radiolabelled compounds in
experimental animals. The techniques employed have been described in
detail by Rogers (1967), Ullberg and Larsson (1981), and more recently
by Benard, Burgat, and Rico (1985). The general method consists of
administering to experimental animals the radiolabelled test compound,
and then at timed intervals killing and rapidly freezing the animals. Serial
sections of the carcass are cut using a freezing microtome, and the
sections removed using adhesive transparent tape. The sections adhering
to the tapes are dried while frozen and the radioactivity visualised by
exposure to X-ray film. At the end of the exposure period, the sections
may be fixed and stained. This technique provides a rapid survey of
radioactivity in many organs and has the advantage that loss or
Table 1 Some radionuclides used in whole animal

metabolic studies
Element Radioisotope Particle emission * Use t
Hydrogen 3H p ; 18; 12.3yr
Carbon l'C p'; 970; 20min
l4C p ; 156; 5730yr
Fluorine 18F EC, p'; 640; 1.8 hr
Sodium "Na EC, p'; 540; 2.6yr
24Na p ; 1390; 15hr
Phosphorus 32P p: 1710; 14.3d
33P p: 248; 25d
Sulphur 35s p: 167; 87d
Chlorine 36 c
1 p: 709; 307,000yr
Calcium 45Ca p: 254; 164d
47Ca p: 690,2000; 4.7d
Chromium 51Cr EC; 28d
Iron 55Fe EC; 2.7yr
59Fe p: 270,460; 45d
Zinc (j5Zn EC, p'; 325; 243d
Bromine 82Br p ; 440; 36 hr
Cadmium "'Cd EC; 462d
1251
Iodine EC; 60d
1291
p ; 150; 1.6 x 107yr
1311
p: 335,608; 8.ld
Gold 198A~ p ; 960; 2.7d
Mercury '03Hg p: 212; 47d
* Principle emission; Em,, (KeV); tl,, EC = electron capture
t A = Audoradiography; B = Whole Body
translocation of soluble compounds is kept to a minimum. The whole-
body autoradiography method has been widely used in the mouse and
rat; the availability of suitable freezing microtomes being the constraining
factor in using larger animals. The disadvantages of the technique are
that large amounts of radiolabelled compound are required, that the time
factor involved in developing and processing the X-ray films precludes
the generation of rapid results, and the results are at best semi-
quantitative, giving little information on the nature of the radiolabelled
material in tissues. Although the standard method cannot be used for
volatile compounds, since the autoradiographs are obtained from dried
sections, low temperature autoradiography, in which cut sections of the
frozen (-80°C) carcass are exposed to X-ray film has been successfully
used for inhalation anaesthetics (Cohen and Hood, 1969) and volatile
nitrosamines (Brittebo and Tjalve, 1982).
Isolated Perfused Organ Technique

This procedure has been used for metabolic studies on the liver, kidney,
lung and small intestines. The experimental techniques have been
described in detail (Colowick and Kaplan, 1981; Ross, 1972).
In studies on the perfused liver, the animal species commonly

employed is the rat. The organ can be completely isolated from the
animal or left in situ in the killed animal. Metabolic studies using this
technique have been carried out on a wide range of compounds including
phenacetin, paracetamol, suphanilamide, and isoniazid. This technique
has the advantage that the intact architecture of the liver is retained and
the flow rate and the composition of the perfusion medium can be
controlled. Furthermore, serial samples of the perfusate and bile can be
obtained. Additionally, test compounds and inhibitors can be introduced
into the perfusion medium at concentrations that would be toxic in the
intact animal. Disadvantages of the isolated liver system include the
difficulty of obtaining identical liver preparations at the same time, and
the relatively short period of time over which these preparations are
viable (about 6-8 hours).
The isolated perfused kidney method has been used for the study of
the transport and metabolism of endogenous compounds, such as choline
and vitamin D, and of foreign chemicals, e.g. isoproterenol, salicylates,
and paracetamol. The rat kidney is the one most extensively used,
although studies with dog and rabbit kidneys have also been reported.
The use of the isolated perfused kidney in drug disposition studies has
been reviewed by Bekersky (1983).
The use of the isolated perfused lung, generally obtained from small
laboratory animals, (e.g. rats, guinea-pigs, and rabbits), provides a
versatile technique for biochemical and metabolic studies. The test
compound can be readily introduced directly into the pulmonary artery
or into the trachea by inhalation, and serial blood samples can be
obtained during the course of the investigation. The principal limitation
of this technique is that experiments are restricted to a duration of about
4 hours.
The importance of the small intestine as a major site of xenobiotic
metabolism has led to the application of the perfusion technique to this
organ. The procedure is conducted either by the en bloc removal of the
entire small intestine from the animal followed by vascular perfusion with
heparinised or defibrinated blood in a recycling system or the autoperfu-
sion of a segment of the small intestine in situ. The former approach
requires considerable surgical skill and the use of glucocorticoid and
norepinephrine in the perfusate to compensate for denervation. The
animal generally used is the rat.
One of the principal advantages common to all of the isolated perfused
organ systems, is the ability to study the metabolism of the test
compound in isolation from the contributory role of other sites in the
body.
Isolated Cells in Culture

The isolated cell suspension and cell culture systems have been exten-
sively used in recent years for the study of xenobiotic metabolism. The
mammalian cell types commonly used are hepatocytes, kidney cells,

pulmonary cells, enterocytes, and nerve cells. Descriptions of the
methods for the preparation of cultures and application of the techniques
are given fully in Colowick and Kaplan (1981b).
Inevitably, isolated hepatocytes constitute the most widely used cell
type. These cells retain in culture their ability to carry out the wide range
of phase I and phase I1 metabolic transformations. Furthermore, the
levels of cofactors required for these reactions are intracellularly gen-
erated and maintained in a manner analogous to the situation that
obtains in the intact animal. Thus, sequential phase I and phase I1
biotransforrnation reactions on a substrate can be carried out by freshly
isolated hepatocytes at rates comparable to the whole organ. The
application of this technique for the study of the metabolism of a wide
variety of chemicals including biphenyl, norbenzphetamine and amino-
pyrine has been extensively reviewed (Fry, 1982; Fry and Bridges, 1977;
Fry, 1983).
Suspensions of intact, fully functional, isolated kidney cells have been
found to have widespread utility in biochemical research (e.g. Bach,
Kettley, Ahmed, and Dixit, 1986) and to a lesser extent in metabolic
studies. The most complete data are available for the renal metabolism of
paracetamol (Jones, Sundby, Ormstad, and Orrenius, 1979); however,
phase I metabolism of benzo( a)pyrene, 7-ethoxycoumarin, and biphenyl
has been investigated, as well as the phase I1 metabolism of 4-
methylumbelliferone and benzoic acid (Fry, Wiebkin, Kao, Jones,
Gwynn, and Bridges, 1975). Isolated kidney cells offer an ideal model for
the study of the relationship of cellular transport to metabolic processes.
Thus, kidney cell preparations active in amino acid uptake can be used
for the study of GSH synthesis and utilisation, e.g. in the paracetamol
conjugation reaction (Mold&, Anderson, Norling, and Ormstad, 1980).
The development of procedures for the isolation of specialised cell
types from the lung and the maintenance of these cells in culture has
provided the means for the investigation and elucidation of cell-specific
xenobiotic metabolism. Thus, for example, alveolar type I1 cells and
non-ciliated bronchiolar epithelial cells (Clara cells) have been separated
by centrifugal elutriation from rabbit and rat lung and the cytochrome
P-450 dependent MFO system investigated (Philpot , Anderson, and
Eling, 1977; Jones and Fouts, 1980).
Suspensions of enterocytes obtained from the small intestines of rat
and guinea-pig and, maintained in culture, have been used to study the
metabolism of benzo(a)pyrene, ethylmorphine, biphenyl, and other
chemicals (Pinkus, 1981). Although the overall levels of enzyme activities
in the enterocytes are low, most of the enzymes associated with the
hepatic MFO system have been found in the gut epithelium. The main
disadvantage of the isolated enterocyte system is the limited viability of
the preparations.
The mammalian nerve cell preparation best suited for culturing
purposes is that obtained from embryonic tissue, although recently Smith
and McInnes (1986) have succeeded in maintaining cells from adult

mouse dorsal root ganglia in culture. Two types of culture can be used:
the first consists of ‘mixed’ cultures containing neurons and a variety of
glial cells, ependymal cells, and fibroblasts, whereas the second approach
attempts to separate the different cell types before culturing in order to
achieve pure (mostly neural or glial) cultures. The former procedure
allows the normal physiological interactions between the different cell
types to be studied, and its uses have been reviewed recently by Spencer,
Crain, Bornstein, Peterson, and Van de Water (1986); the latter
procedure enables the interaction between a compound and a specific cell
type to be investigated. There appears to be little published information
on metabolic studies of foreign compounds by these preparations.
Testicular cells in culture have been used extensively to study the
mechanisms of toxicity of various toxins, including phthalate esters and
glycol ethers (Gray and Beamond, 1984; Gray, MOSS, Creasy, and
Gangolli, 1985). Individual cells types including Sertoli cells and sper-
matocytes have been used, and more recently co-culture systems have
been developed in which, for example, Sertoli and germ cells are cultured
together to provide a more realistic model of the effect of toxins in uiuu
with respect to cell-cell interactions. The development of compartmental
apparatus may further advance understanding of cell-cell interactions ,
and permit investigations into the role of metabolic activation in this
system.
Whole Tissue and Intracellular Homogenates

Whole tissue homogenates and subcellular fractions obtained by
differential centrifugation procedures constitute the most widely used
methods for studies on the metabolism of foreign compounds. The organ
of pre-eminence in these studies is the liver, and the literature on
methodology, enzyme kinetics, cofactor requirements, and effects of
inducers and inhibitors on metabolic reactions in respect of these systems
is extensive. Excellent symposia proceedings and reviews on this subject
appear regularly (e.g. ‘Microsome and Drug Oxidation’ meetings).
Detailed description of the preparation procedures and enzyme assay
methods are available (Brodie and Gillette, 1971; Snell and Mullock,
1987). The use of these in uitru preparations and the many refinements
introduced, such as purified enzyme fractions, have provided an invalu-
able tool for elucidating the biochemical mechanisms responsible for the
transformation of foreign chemicals. However, the limitations of the
system need to be taken into account in arriving at a realistic assessment
of the results. Metabolic changes in in uitru systems are subject to
variation due to preparation procedures, ionic strength and pH of the
medium, levels of cofactors and nutrients, and to end-product inhibition
or activation. In addition, the results of in uitro enzyme induction and
inhibition studies require to be interpreted with caution. Thus, for
example , whereas in uitro 3,3,3-trichloropropylene oxide and cyclo-

hexene oxide inhibit epoxide hydratase activity, in the intact animal these
compounds act as inducers of this enzyme, whilst reducing liver GSH
content (Oesch, 1974; Oesch, Jerina, Daly, and Rice, 1973).
Each of the various techniques and methods briefly described in this
section play an important part in providing an insight into the mechan-
isms mediating the metabolic transformation of compounds in the body.
Each of the techniques has advantages and disadvantages, a r d an
appreciation of these is essential for their intelligent use in generating
relevant metabolic data.
References
Baillie, T. A. (1981). The use of stable isotopes in pharmacological research.
Pharmacol. Rev., 33, 81-132.
Bach, P. H., Ketley, C. P., Ahmed, I., and Dixit, M. (1986). The mechanism of
target cell injury by nephrotoxins. Food Chem. Toxicol., 24, 775-779.
Bekersky, I. (1983). Use of the isolated perfused kidney as a tool in drug
disposition studies. Drug Metab. Rev., 14, 931-960.
Benard, P., Burgat, V., and Rico, A. G. (1985). Application of whole-body
autoradiography in toxicology. Crit. Rev. Toxicol., 15, 181-215.
Bircher, J. and Presig, R. (1981). Exhalation of isotopic CO,. In: ‘Methods in
Enzymology’, Vol. 77. Ed. W. B. Jakoby. Academic Press Inc. (London)
Ltd., pp. 3-9,
Brittebo, E. B. and Tjalve, H. (1982). Tissue specificity of N-nitrosodibutylamine
metabolism in Sprague-Dawley rats. Chem. Biol. Interact., 38, 231-242.
Brodie, B. B. and Gillette, J. R. (1971). In: ‘Handbook of Experimental
Pharmacology’, Vol. 28., ‘Concepts in Biochemical Pharmacology’, Pt, 2.
Springer Verlag, Berlin.
Bronner, F. and Lemaire, R. (1969). Comparison of calcium kinetics in man and
the rat. Calcif. Tissue Res., 3, 238-248.
Cohen, E. N. and Hood, N. (1969). Application of low temperature autoradiog-
raphy to studies of the uptake and metabolism of voltaile anaesthetics in the
mouse. I. Chloroform. Anaesthesiology, 30, 306.
Corney, P. L. and Heath, D. F. (1970). A simple way of estimating turnover
rates from specific activity-time curves. J . Appl. Physiol., 28, 672-674.
Colowick, S. P. and Kaplan, N. 0. (1981a). Organ perfusion: In: ‘Methods in
Enzymology’, Vol. 77, Pt. B. Ed. W. B. Jakoby. Academic Press Inc.
(London) Ltd. pp. 81-129.
Colowick, S. P. and Kaplan, N. 0. (1981b). Cells. I n : ‘Methods in Enzymology’,
Vol. 77, Pt C. Ed. W. B. Jakoby. Academic Press Inc. (London) Ltd., pp.
130-168.
Davis, D. S. and Boobis, A. R. (1983). Drug metabolism and the safety
evaluation of drugs and chemicals. In: ‘Animals in Scientific Research: An
Effective Substitute for Man?’ Ed. P. Turner. Macmillan Press, London, pp.
69-80.
Dost, F. H. (1968). ‘Grundlagen der Pharmakokinetics’. Thieme, Stuttgart.
J . C. Phillips and S. D . Gangolli 199
Fraser, H. S., Mucklow, J. C., Murray, S . , and Davis, D. S. (1976). Assessment

of antipyrine kinetics by measurement in saliva. Br. J . Clin. Pharmacol., 3,
321-325.
Fry, B. J., Taylor, T., and Hathway, D. E. (1972). Pulmonary elimination of
chloroform and its metabolites in man. Arch. Int. Pharmacodyn. Ther., 196,
98-1 11.
Fry, J. R. (1982). In: ‘Reviews on Drug Metabolism and Drug Interactions’ Eds
A. H. Beckett and J. W. Gorrod. Freund Publishing House, Tel Aviv.
Fry, J. R. (1983). A review of the value of isolated hepatocyte systems in
xenobiotic metabolism and toxicity studies. In : ‘Animals in Scientific
Research: An Effective Substitute for Man’, Ed. P. Turner. Macmillan Press,
pp. 81-90.
Fry, J. R. and Bridges, J. W. (1977). The metabolism of xenobiotics in cell
suspensions and cell cultures. In: ‘Progress in Drug Metabolism’, Vol. 2. Eds
J. W. Bridges and L. F. Chasseaud. John Wiley and Sons, pp. 71-118.
Fry, J. R., Weibkin, P., Kao, J., Jones, C. A., Gwynn, J., and Bridges, J. W.
(1978). A comparison of drug metabolising capability in isolated viable rat
hepatocytes and renal tubule fragments. Xenobiotica, 8, 113-120.
Garattini, S. (1985). Toxic effects of chemicals: difficulties in extrapolating data
from animals to man. Crit. Rev. Toxicol., 16, 1-29.
Gehring, P. J., Watanabe, P. G., and Blau, G. E. (1976). Pharmacokinetic
studies in evaluation of the toxicological and environmental hazards of
chemicals. In: ‘New Concepts of Safety Evaluation’, Eds M. A. Mehlman,
R. E. Shapiro, and H. Blumenthal. Hemisphere Publishing Corp.,
Washington DC, pp. 195-270.
Gibaldi, M. and Perrier, D. (1975). Pharmacokinetics. In: ‘Drugs and the
Pharmaceutical Sciences’, Vol. 1. Ed. J. Swarbrick, Marcel Dekker, New
York.
Goldstein, A., Aronow, L., and Kalman, S. M. (1974). ‘Principles of Drug
Action: The Basis of Pharmacology’, 2nd Edn., John Wiley and Sons, New
York.
Gray, T. J. B. and Beamand, J. A. (1984). Effect of some phthalate esters and
other testicular toxins on primary cultures of testicular cells. Food Chem.
Toxicol., 22, 123-131.
Gray, T. J. B., Moss, E. J., Creasy, D. M., and Gangolli, S. D. (1985). Studies
on the toxicity of some glycol ethers and alkoxyacetic acids in primary
testicular cell cultures. Toxicol. Appl. Pharmacol., 79, 490-501.
Hawkins, D. R. (1977). The role of stable isotopes in drug metabolism. In:
‘Progress in Drug Metabolism’, Vol. 2., Eds J. W. Bridges and L. F.
Chasseaud, John Wiley and Sons, Chichester, pp. 163-218.
Hepner, G. W. and Vesell, E. S. (1975). Quantitative assessment of hepatic
function by breath analysis after oral administration of (14C)-aminopyrine
Ann. Intern. Med., 83, 632-638.
Jones, D. P., Sundby, G. B., Ormstad, K . , and Orrenius, S. (1979). Use of
isolated kidney cells for study of drug metabolism. Biochem. Pharmacol., 28,
929-935.
Jones, K. G. and Fouts, J. R. (1980). Pharmacologkt, 22, 277.
Levy, G. (1965). Pharmacokinetics of salicylate elimination in man. J. Pharm.
Sci., 54, 959-967.
Levy, G. and Gibaldi, M. (1975). Pharmacokinetics. In: ‘Concepts in Biochemi-
cal Pharmacology’, Pt 3. Eds J. R. Gillette and J. R. Mitchell, Ch. 59,

Springer Verlag, Berlin. pp. 1-34.
Lundquist, F. and Wolthers, H. (1958). The kinetics of alcohol elimination in
man. Acta Pharmacol. ( K b H ) , 14, 265-289.
Mayersohn, M. and Gibaldi, M. (1971). Mathematical models in phar-
macokinetics. Solution of the two compartment open model. Am. J. Pharm.
Educ., 35, 19-28.
Moldeus, P., Anderson, B., Norling, A., and Ormstad, K. (1980). Effect of
chronic ethanol administration on drug metabolism in isolated hepatocytes
with emphasis on paracetamol activation. Biochem. Pharmacol., 29, 1741-
1745.
Niazi, S. (1979). ‘Textbook of Biopharmaceutics and Chemical Phar-
macokinetics’. Appleton, Century, Croft, New York.
Oesch, F. (1974). Purification and specificity of a human microsomal epoxide
hydratase. Biochem. J., 139, 77-88.
Oesch, F., Jerina, D. M., Daly, J. W., and Rice, J. M. (1973). Induction
activation and inhibition of epoxide hydrase: An anomolous prevention of
chlorobenzene-induced hepatotoxicity by an inhibitor of epoxide hydrase.
Chem. Biol. Interact., 6, 189-202.
Phillips, J. C., Bex, C. S., Mendis, D., Walters, D. G., and Gangolli, S. D.
(1987). The metabolic disposition of ‘‘C-labelled amaranth in the rat, mouse,
and guinea-pig. Food Chem. Toxicol., 25, 947-954.
Phillips, J. C . , Gaunt, I. F., and Gangolli, S. D. (1975). Studies on the metabolic
fate of 32P-labelledemulsifier YN in the mouse, guinea-pig, and ferret. Food
Comet. Toxicol., 13, 23-30.
Phillips, J. C., Mendis, D., and Gaunt, I. F. (1987). The metabolic disposition of
14C-labelled Brown HT in the rat, mouse, and guinea pig. Food Chem.
Toxicol., 25, 1013-1019.
Philpot, R . M., Anderson, M. W. and Eling, T. E. (1977). In: ‘Metabolic
Functions o f the Lung’. Eds Y. S. Bakhle and J. R. Vane. Marcel Dekker,
Inc., New York, p. 123.
Pinkus, L. M. (1981). Separation and use of enterocytes. In: ‘Methods in
Enzymology’, Vol. 77. Ed. W. B. Jakoby. Academic Press Inc. (London)
Ltd., pp. 154-161.
Pohl, L. R. and Branchflower, R. V. (1981). Covalent binding of electrophilic
metabolites to macromolecules. In: ‘Methods in Enzymology’, Vol. 77. Ed.
W. B. Jakoby. Academic Press Inc. (London) Ltd., pp. 43-49.
Reitz, R. H., Watanabe, P. G., McKenna, M. J., Quast, J. F., and Gehring, P.
J. (1980). Effects of vinylidene chloride on DNA synthesis and DNA repair
in the rat and mouse: a comparative study with dimethylnitrosamine.
Toxicol. Appl. Pharmacol., 52, 357-370.
Rogers, A. W. (1967). ‘Techniques of autoradiography’. Elsevier Publishing Co. ,
Amsterdam.
Ross, B. D. (1972). ‘Perfusion technique in Biochemistry’. Oxford Univ. Press
(Clarendon), London.
Ruelius, H. W. (1975). Interaction of drug metabolism with the other elements of
drug safety evaluation. Drug Metab. Rev., 4, 115-133.
Smith, R. A. and McInnes, I. B. (1986). Phase contrast and electron microscopi-
cal observations of adult mouse DRG cells maintained in primary culture. J.
Anat., 145, 1.
J . C. Phillips and S . D . Gangolli 20 1
Snell, K. and Mullock, B. (1987). ‘Biochemical Toxicology: a Practical Ap-

proach’, IRL Press, Oxford.
Spencer, P. S., Crain, S. M., Bornstein, M. B., Peterson, E. R., and Van de
Water, T. (1986). Chemical neurotoxicity: detection and analysis in or-
ganotypic cultures of sensory and motor systems. Food Chem. Toxicol., 24,
539-544.
Ullberg, S. and Larsson, B. (1981). Whole-body autoradiography. In: ‘Progress in
Drug Metabolism’. Vol. 6, Eds J. W. Bridges and L. F. Chasseaud. J. Wiley
and Sons, London, pp. 249-289.
Wagner, J. G., Northam, J. I . , Alway, C. D., and Carpenter, 0. S. (1965).
Blood levels of drug at the equilibrium state after multiple dosing. Nature
(London), 207, 1301-1302.
Watanabe, P. G., Ramsey, J. C. and Gehring, P. J. (1980). Pharmacokinetics
and metabolism of industrial chemicals. In: ‘Progress in Drug Metabolism’,
Vol. 5, Eds J. W. Bridges and L. F. Chasseaud. J. Wiley and Sons, London,
pp. 311-343.
Withey, J. R. (1977). Pharmacokinetic Principles. Proc. 1st Internat. Cong.
Toxicol.
CHAPTER 11
Zmmunotoxicology -
Conceptua1 Problems
H. E. AMOS
1 Introduction
Recognition that chemicals interacting with the immune system in a
‘non’ immunologically specific manner might constitute a potential hazard
to health, has not been universally embraced by toxicologists involved in
risk assessment. The difficulty has been to convincingly establish a cause
and effect relationship. Untoward responses resulting from transient
perturbations in immune responsiveness may involve a considerable
latent period before they become manifest. Moreover, interpretation of
immunological assay systems used in safety evaluation are themselves
open to criticism. Many assays depict in vitro phenomena, which cannot
be directly related to a clinical state. For example, a reduced in vitro
proliferative response of lymphocytes to the mitogen, phytohaemag-
glutinin, is frequently equated with a clinical state of immune suppres-
sion. This is quite unjustified unless account is taken of the reserve
capacity of the immunological system.
‘There is no doubt, however, that cells involved in the immune
response can be a target for toxic damage which will then compromise
the functional integrity of the response. It will either become down-
graded, leading to degrees of suppression, or upgraded, producing
autoallergic reactions and hypersensitivity responses. The perception that
modulation of the immune response by ‘immunotoxicants’ under
controlled conditions can be a therapeutic tool has been widely acclaimed.
Their value has been proven in the control of tissue rejection and
improvements in delivery systems particularly the humanisation of murine
monoclonal antibodies, offers considerable optimism for the future.
2 Inadvertant Immunomodulation
Of major interest to immunotoxicologists are those ‘immunotoxicants’
which inadvertantly participate in the immunological system. They may
202
H . E. Amos 203
include drugs, environmental chemicals and impurities in products

manufactured by recombinant DNA, and other biotechnologies
(D’Agnolo, 1983). In fact, the range of products is such that it is virtually
impossible to avoid contact with at least some of them. The question
must be asked therefore:
‘What is the clinical significance of inadvertant exposure to
immunotoxicants?’
It is well established that experimentally induced immunosuppression in
animals leads to an increased tumour incidence and in some cases
increased metastatic spread (Pimm and Baldwin, 1978; Baldwin, 1973).
The mechanism, however, is complicated, it is not a straightforward
reduction in immuno surveillance brought about by the ‘immunotoxicant’
targeting T lymphocytes. Studies with chemical carcinogens, for example,
have shown that suppression can be manipulated so as to remove T
suppressor cells which will in fact increase tumour immunity (North et
al., 1982).
The clinical effect of an ‘immunotoxicant’ which downgrades the
immunological system will depend upon which cell function has been
depleted. This was elegantly demonstrated by Spreafico et al. (1984) who
defined the immunological profile in mice treated with doxorabicin
(DXR) and its close chemical analogue daunrabicin (DNR). They
showed that DXR given up to the LDlo level had no effect on
macrophage function but with DNR, the multifunction macrophage was
severely depressed, affecting phagacytosis, antigen expression, and non-
specific direct cytotoxicity . Thus closely related chemicals subserving the
same therapeutic effect can differ in their toxicity towards a particular cell
type.
Spreafico et al. (1984) attempted to relate the susceptibility of
immunologically important cells to different immunosuppressive agents
and showed, for example, the sequence for corticosteroids to be:
Thymocytes (TH) > Precursor cytotoxic lymphocyte (CTLp)
> cytotoxic lymphocytes > T suppressor cells(Ts) > B cells
and for cyclophosphamide the ranking is:
Ts > TH > B > CTLp
The clinical significance of this diversity is that the overall response of the
immunological system to challenge by an ‘immunotoxicant’ is the
algebraic sum of various positive and negative cell responses. For each
agent, therefore, which specifically targets immunologically important
cell types, a fingerprint can be established which can then be used to
predict the clinical consequence. Unfortunately the clinical consequence
can be predicted only in the animal species which donated the cells.
Studies with the alkyl tin series had already shown that species variation
in cell responsiveness could be expected (Seinen et al., 1977). In these
204 ImmunotoxicoLogy-Conceptua L Problems
experiments thymocytes were the main target and neonatal rats the most
sensitive species, there was virtually no toxic effect on thymocytes from
man.
Three main points emerge from these quoted examples which have
implications for toxicologists involved in safety evaluation and risk
assessment.
(a) The degree and type of immunological imbalance is dependent upon
the immunopharmacological profile of chemicals.
(b) Unless a profile is defined it is difficult to predict with any degree of
confidence the likely clinical outcome.
(c) Profiles tend to be species specific.
The implications clearly relate to how can toxicologists effectively
screen for ‘immunotoxicants’, given that each one has a distinct target
cell profile? Before considering proposals which have been put forward
to answer the question, the issue of clinical significance needs comment.
Detectable changes at a clinical level are only significant to the
toxicologist if they are untoward and pose a hazard to health. There are a
few examples of chemicals down regulating the immunological system in
man following inadvertant exposure and these are associated with major
toxicological accidents. Two of which, in particular, are illustrative.
A Polybrominated Biphenyls in Michigan (PBBs)

In 1973 a commercial preparation containing polybrominated biphenyls
was mistakenly used instead of magnesium oxide in the preparation of a
feed supplement for lactating cows. Subsequently large numbers of cows
and farmworkers were exposed to PBBs and sustained toxic injury,
including an immunotoxic response (Carter, 1976).
PBBs are fat soluble and stored in the thymus, liver, brain, and
adipose tissue where they persist for long periods of time. A longitudinal
study of Michigan farm workers exposed to PBBs was undertaken firstly
in 1976-77 and again in 1981-83 (Bekesi et al., 1978; 1983). Changes in
the immunological profiles of long term survivors were monitored. The
main findings were decreased numbers of T cells, an increase in null cells,
and a decreased proliferative response of lymphocytes to in vitro
stimulation with mitogens. These changes, detectable during the first
survey, were still present at the time of the second analysis.
Plasma levels of PBBs ranged from 0.6 to 70 p.p.m. and in 25 percent of
the subjects followed there was no decrease in levels between the two
surveys. No correlation was found between serum levels and the
prevalence of clinical symptoms or immunological findings, but the PBB
content of white blood cells did directly correlate with immune disfunc-
tion (Robez et al., 1982).
It would appear, therefore, that continued exposure of immunologi-
cally functional cells to the immunotoxicant contributed to the persist-
H . E. Amos 205
ence of the immunological defect. What remains still to be unequivocally

proven is whether the degree of suppression resulting from the persistent
defect will eventually lead to an increased cancer risk. The work has been
extensively reviewed by Bekesi et al. (1987).
B Accidents Involving Exposure to Polychlorinated

Dibenzo-p-dioxins (TCDD)
Polychlorinated dibenzo-p-dioxins and polychlorinated dibenzofurans are
two series of tricyclic aromatic compounds with similar physical, chemical,
and biological properties. They are present as contaminants in different
commercial products such as phenoxy herbicides, polychlorinated bi-
phenyls, diphenyl ether herbicides, and hexachlorophene. They have also
been reported to occur in fly ash from industrial incinerated waste.
There have been at least seven major accidents involving TCDD over
the past fifteen years, but the one involving inadvertant exposure at
Seveso in Italy offered the best opportunity to study the immunological
defect. The official conclusion to the studies was ‘that no significant
difference in immuno function between exposed and control children
could be found’ (Seveso, 1979). This finding could not be predicted on
the basis of animal model studies.
It was demonstrated in the early 1970s that the immunological system
was a particularly sensitive target for TCDD in laboratory animals (Vos
and Moore, 1974). The cells involved were those participating in
differentiation and expression of cell mediated immunity and T cell
dependent antibody production, which are located within the thymus
gland. It has been suggested that the action of TCDD on thymus
epithelial cells, which are involved in T cell maturation, is through the
expression of the Ah receptor (Ah designing the genetic locus for aryl
hydrocarbon responsiveness) (Poland and Glover, 1980).
The Ah receptor can be detected in different tissues. Rats have a high
concentration of Ah receptor activity expressed in the thymus, whereas in
mice the concentration in the thymus is about 1/4 of that expressed in the
liver (Mason and Okey, 1982). Moreover, some strains of mice
[DBA/2J(D2)] are Ah-receptor deficient and resistant to the toxic effect
of TCDD at concentrations which cause marked toxicity in sensitive
strains [C57BL/6 (B6)]. Thus, the Ah locus segregates with TCDD
induced thymus atrophy in mice (Vecchi et al., 1983).
The recognition that genetic constraints can play a role in producing
inter and intra species differences in response to challenge with ‘im-
munotoxicants’ might be one explanation why the official report of the
Seveso incident concluded that there was no demonstrable immunological
defect in man.
These two illustrative studies do not, as yet, prove that exposure to
PBB, or TCDD, cause a hazard to health which is mediated through the
immunological system. Time, undoubtedly, will be the final arbiter, but
206 Immunotoxicology-Conceptual Problems
faced with a scenario that TCDD, for example, is an active tumour

promoter in animals (Pitot et al., 1980; Poland and Glover, 1982) and
since the T cytotoxic effector cell is a major cell type, participating in
immune surveillance against neoplasms (Herberman and Ortaldo, 198l),
modulation of T lymphocyte maturation by the Ah receptor could be a
mechanism associated with a potential carcinogenic risk of TCDD. It is,
therefore, prudent for toxicologists to devise methods for identifying
‘immunotoxicants’ in safety evaluation assessments.
3 Relevance to Man of Immunotoxicological

Models in Animals
The basis of any safety evaluation study is to model an ‘effect’ in animals
then make judgements on the relevance of the data to man. There is,
therefore, an element of empiricism in safety assessment studies, but the
empirical approach has proved its value in that man has been protected
from most major poisons. It is not quite so certain, however, whether
confidence in animal model data can be extended to the field of
immunotoxicology .
As stated in an earlier section, each immunotoxicant will have a
particular cell profile and it is the summation of the ‘effects’ which
determine the ultimate clinical response. Many of the early workers in
immunotoxicology failed to appreciate this fact, and drew clinical
conclusions on the basis of a single assay system. Thus, many claims for
specific xenobiotics causing an immuno depressed state were premature,
Sharma and Zeeman (1980) illustrated this point rather well. They
studied the effect of D D T on antibody production in mice and showed
that they could either increase or decrease the production by altering
conditions of the model system. The main variables influencing the result
were dose and dosing schedules, nature of the antigen and time between
antigen challenge, and the effect studied.
Even immunotoxicity data generated in man is not without question.
The Bekesi studies on PBBs quoted earlier (Bekesi et al., 1978) have
been challenged by Silva et al. (1979). These investigations were unable
to find any differences in the numbers or function of various classes of
lymphocytes in high or low PBB exposed individuals vemw a control
group.
Immunotoxicologists are now well aware of these pitfalls and efforts
are being made to address the central issue which is to determine how
loss of immune function, as measured in vitro, equates to a clinical effect
in vivo. Much of the credit for this approach must go to Dean and
co-workers (Dean et al., 1984). They have assembled data obtained over
a number of years with a variety of chemicals tested at three different
dose levels in host resistant model systems. They evaluated the lack of
independence of two variables-in vitro immune function and altered
host susceptibility, using Spearman’s rank correlation co-efficient .
H . E. Amos 207
They were able to generate data in which significant correlation was

obtained between depressed cytolysis by natural killer cells (NK) and
increased susceptibility to in vivo challenge with B16F10 melanoma cells
( R h o = 0.758). Similarly, they also demonstrated a correlation between
loss of cytotoxic lymphocyte (CTL) activity and increased susceptibility to
challenge with PYB6 sarcoma cells.
These data go a long way towards answering the question-what
degree of immune suppression is necessary to induce directly overt toxic
effects? From Dean and co-workers’ data it was possible to deduce that
the functional reserve for the CTL model was approximately <lo-50
percent, whereas with the NK system it was more like 60-70 percent. It is
only when this reserve has been exhausted that the susceptibility to
tumour challenge is noted.
Functional assay of this type, in which it is possible to judge the clinical
implications of changes in the immunological systems, must be the basis
for developing any immunotoxicity safety evaluation screen. The data
base for recommending the introduction of immunotoxicity testing into
routine screens for risk assessment is, however, not yet adequate. A fact
fortunately recognised by most regulating agencies.
4 Hypersensitivity as an Immunotoxic Response

Difficulties associated with screening chemicals for their direct effect on
cells of the immunological system are relatively minor when compared to
screening chemicals for immunogenicity. This is due to our inability to
control in assay systems the many variables which govern the im-
munological response. The demonstration that a chemical will induce
antibody formation in experimental animals is not by itself indicative of a
toxic response or a prediction that the compound will cause hypersen-
sitivity tissue damage in man.
Practically all chemicals can be made to stimulate antibodies in model
systems, given the right conditions. But small molecular weight chemicals
are not complete antigens in their own right. They are referred to as
haptens.
The hapten hypothesis is the assumption that drugs become covalently
bound to a macromolecular carrier and are then able to perform the
functions of antigens and induce an allergic state. It is now well
recognised that the ability of low molecular weight organic chemicals to
stimulate an antibody response is a direct function of their reactivity with
nucleophilic groups on proteins or other macromolecules.
An aspect of research, therefore, into immunotoxic drug reactions, has
concentrated on the formation of complete immunogenic units from a
composition of a hapten and carrier molecule as this is the one variable
which can to some extent be controlled. Once a complete antigen is
formed, however, host variables will dictate the final clinical response
and these are not sufficiently characterised to be taken into account in

safety evaluation systems, but it should be noted that one major regulatory
authority, the Ministry of Health and Welfare in Japan, does demand that
all drugs be tested for their irnmunogenicity in experimental systems.
5 Antigen Formation with Haptens

There are a number of possible mechanisms by which haptens can form
covalent linkages with macromolecules under physiological conditions.
The main ones are as follows.
A Direct Protein Reactivity

Compounds with functional groups capable of covalently linking to
nucleophilic side chains of proteins in neutral conditions, without prior
biotransformation or chemical modification , clearly have immunogenic
possibilities. Drug designers are well aware of this and steps are taken to
limit direct protein reactivity by careful molecular manipulation.
B Structural Alteration to Instability

Degradation
The beta lactam antibiotics, such as penicillin, induce antibodies in the
majority of subjects taking the drug. It has been shown that the
antibodies are not directed towards penicillin itself , but to various
breakdown products. The major antigenic determinant, the penicilloyl
conjugate, can be formed in a number of ways, one of which is through
the penicillin rearrangement product, penicillenic acid , a highly reactive
anhydride capable of rapidly conjugating free lysine residues in proteins
(Parker et al., 1962).
The ease by which penicillin undergoes rearrangement to form a
variety of antigenic determinants is not shared by the majority of drugs,
yet most of our information about immunotoxic drug reaction has been
gained by studying penicillin!
Poly m erisa tion

Some small molecular weight compounds undergo polymerisation in
solution to produce structures of considerable size and complexity. The
net effect of such a process is to convert a non-immunogenic hapten into
a molecule capable of directly stimulating antibodies. It is not known to
what extent polymer formation contributes to the antibody response
generated against drugs, but the potential should be recognised.
H . E. Amos 209
6 Metabolic Conversion and Enzyme Interactions

It is difficult to determine with any degree of certainty exactly what
happens to a drug after it gets inside the body. Enzymic mechanisms exist
which are aimed at converting the parent drug into polar metabolites
which facilitate elimination. Some metabolic intermediates can be highly
protein reactive but the evidence that such complexes stimulate anti-
bodies is sparse. The following two examples, however, support the
contention that antigens can be formed from metabolic products.
A Practolol
The cardioselective B-receptor blocking agent, practolol, induced in a
small number of patients a syndrome defined as ‘oculomucutaneous’. As
the name implies the main sites of tissue damage were confined to the
eyes and serosal surfaces, including the peritoneum and skin.
Epidemiological studies identified two factors which suggested that the
tissue damage might not be due to a direct toxic effect of the drug but
might implicate an immunological element in the pathogenesis. These
were:
(a) the syndrome was not dose related
(b) the incidence was low.
Although attempts to identify an antibody with specificity for the
practolol hapten were unsuccessful it was possible to show antibody
activity against a practolol product generated in an in-vitro mixed
function oxygenase system using rat and hamster liver microsomes
(Amos, Lake, and Atkinson, 1977; Amos, Lake, and Artis, 1978). The
clinical significance of the antibody in relation to the pathogenesis of the
syndrome was not established, but its demonstration did indicate that
metabolic products can stimulate a specific immunological response.
B Halothane
Halothane is a fluorinated hydrocarbon used as a volatile anaesthetic and
there is a clinical association between repeated use of the drug and
hepatic damage. The possibility that a hypersensitivity mechanism may
be involved in the pathogenesis arose following a number of epidemiolo-
gical observations.
(a) Accelerated onset of hepatic changes after more than one halothane
exposure, (Inman and Mushin, 1974).
(b) Unexplained post-operative pyrexia following exposure, (Trey Lip-
worth, and Chatman, 1968).
(c) Frequent past history of allergy.
The first observation that a minor metabolite may be involved in the
production of hepatic necrosis was the finding that a product of halothane
metabolism, via the major oxidative pathway was capable of rendering

rabbit hepatocytes antigenic (Neuberger et al., 1980). It is not clear from
the work whether the metabolite is essential for the antibody specificity
or whether it functions to alter self determinants, but either way the
immunogenic role of the metabolite must be acknowledged.
7 Alteration of Drugs by Chemical Interaction

The proposition that environmental chemicals may convert non protein
reactive compounds into ones which can covalently link, has not been
fully appreciated. At present there is one example which illustrates the
possibility. The topical antimicrobial agent, chlorhexidine, is a bis-
biguanide structurally unsuited for creating covalent linkages. There is
evidence, however, implicating chlorhexidine in IgE antibody formation
and clinical anaphylaxis (Ohtoshi et al., 1986). Compliance with the
hapten hypothesis would demand that chlorhexidine must covalently link
to proteins to stimulate the IgE response. Experimental studies showed
that chlorhexidine could be converted into a protein reactive species by
chlorine or hypochlorite. The amount of chlorine required to convert the
-NH- residues of chlorhexidine into the corresponding chloramide , NC1,
was less than 1 p.p.m. in water and the conversion was very rapid in
neutral conditions (Rose, personal communication). The converted
N-chloro derivative bound avidly to model proteins and the conjugates
were shown to induce IgE antibodies in experimental animal models
(Layton, Stanworth, and Amos, 1986; 1987).
The points to note in this example are that the chlorine interaction was
very rapid and required very little free chlorine; the level of 1 p.p.m. is
reached in the outlet source of many domestic water supplies. Protein
conjugates formed by N-chloro conversion of the hapten were im-
munogenic and stimulated IgE responses. If such a mechanism was
reponsible for chlorhexidine antigen formation it is possible to ask might
it have a more general role in immunotoxicology? Further work will tell.
8 Pseudo-allergic Reactions
A fundamental requirement for a hypersensitivity reaction is an antigen
reacting with specific antibody or with specifically allergised cells. The
consequences of such interactions alter effector pathways which can lead
to tissue damage or, as in the case of anaphylaxis, produce potentially life
threatening conditions.
The term ‘pseudo-allergic’ refers to reactions which mimic signs and
symptoms of hypersensitivity disorders, but without involving im-
munological mechanisms. Thus the inducing compound does not behave
as an antigen.
One of the most important pseudo-allergic reactions is with compounds
which induce signs of clinical anaphylaxis, such as synacthen (Jaques and
H . E . Amos 21 1
Brugger, 1969), and intravenous anaesthetic drugs (Lorenz et al., 1969).

An important target of these compounds is the complement system,
particularly direct fixation of C3. Fixation of this major complement
component by chemicals, without involving C1, C4, or C2, will initiate
fixation of the end terminal sequence of the complement cascade and
generate biologically active fragments called anaphylatoxins. The frag-
ments have the property of reacting with mast cells and liberating
vasoactive amines which bring about the anaphylactic process unless
tested in a cohort of genetically high responders to the specific antigen
(Gleichmann et al., 1989).
The implication for the immunotoxicologist is that pseudo-allergic
response can respond to dose-response relationships, whereas true
immunological reactions are not dose-related in the accepted manner,
9 Conclusion
The identification of chemicals which target the immunotoxicological
system and alter its function is becoming a task for toxicologists involved
in safety evaluation. Systems are being devised and validated which could
eventually be incorporated into routine toxicology screens, but the
fundamental question whether transient modulation producing either a
downgrading or an upgrading of the immunological response relates to a
serious health hazard, such as an increased tumour incidence, is still
unresolved. Definitive data are lacking but it must be stated that the
evidence which is accumulating would make it irresponsible for toxicolo-
gists to ignore the possibility of a correlation.
References
Amos, H. E., Lake, B. G., and Atkinson, H. A. C. (1977). Allergic drug
reactions: An in vitro model using a mixed function oxidase complex to
demonstrate antibodies with specificity for a practolol metabolite. Clin.
Allergy, 7 , 423.
Amos, H. E., Lake, B. G., and Artis, J. (1978). Possible role of antibody specific
for a practolol metabolite in the pathogenesis of oculomucocutaneous
syndrome. Br. Med. J . , 1, 402-404.
Baldwin, R. W. (1973). Immunological aspects of carcinogenesis. A d v . Cancer
Res., 18, 1.
Bekesi, J. G., Holland, J. F., Anderson, H. A., Fischbein, A. S . , Rom, W . ,
Wolff, M. S . , and Selikoff, I. J. (1978). Lymphocyte function of Michigan
dairy farmers exposed to polybrominated biphenyls. Science, 199, 1207-
1209.
Bekesi, J. G., Roboz, J. P., Solomon, S., Fischbein, A. S . , and Selikoff, I . J.
(1983). Altered immune function in Michigan residents exposed to polybrom-
inated biphenyls. In: ‘Immunotoxicology’, Eds G. G. Gibson, R. Hubbard,
and D. V. Parke; pp. 181-191. Academic Press, New York.
Bekesi, 1. G., Roboz, J . P., Fischbein, A. S., and Selikoff, I. J. (1987). Clinical
212 Immunotoxicology-conceptual Problems
immunology studies in individuals exposed to environmental chemicals. In:

‘Immunotoxicology’, Eds A. Berlin, J. Dean, M. H. Draper, E. M. B. Smith,
and F. Spreafico. pp. 347-367. Martinus Nijhoff.
Carter, L. T. (1976). Michigan’s PBB incident: chemical mix up leads to disaster.
Science, 192, 240-243.
D’Agnolo, G. (1983). The control of drugs obtained by recombinant DNA and
other biotechnologies. In: ‘Current Problems in Drug Toxicity’. Eds G.
Zbinden, F. Cohadon, J. Y. Detaille, and G. G. Mazue. John Libby
Eurotext. Paris and London. pp. 241-247.
Dean, J. H., Lloyd, D., Lauer, L. S., House, R. V., Ward, E. C., and Murrary,
M. J. (1984). Experience with validation of Methodology for Immunotoxicity
Assessment in Rodents. In: ‘Immunotoxicology’. Eds A. Berlin, J. Dean, M.
H. Draper, E. M. B. Smith, and F. Spreafico. Martinus Nijhoff. pp. 135-158.
Gleichman, E., Kimber, I., and Purchase, I. F. H. (1989). Immunotoxicology:
suppressive and stimulatory effects of drugs and environmental chemicals on
the immune system. Arch. Toxicol., 63, 257-273.
Herberman, R. A. and Ortaldo, J. R. (1981). Natural killer cells; their role in
defence against disease. Science, 214, 24-30.
Inman, W. H.W. and Mushin, W. W. (1974). Jaundice after repeated exposure
to halothane: An analysis of reports to the committee on safety of medicines.
Br. Med. J . , 1, 5.
Jaques, R. and Brugger, M. (1969). Synthetic polypeptides related to cor-
ticotrophin acting as histamine liberators. Pharmacology, 2 , 361.
Layton, G. T., Stanworth, D. R. and Amos, H. E. (1986). Factors influencing
the immunogenicity of the haptenic drug chlorhexidine in mice. 11. The role
of the carrier and adjuvants in the induction of IgE and IgG anti-hapten
responses. Immunology, 59, 459-465.
Layton, G. T.,Stanworth, D. R., and Amos, H.E. (1987). Factors influencing
the immunogenicity of the haptenic drug chlorhexidine in mice. I. Molecular
requirements for the induction of IgE and IgG anti-hapten antibodies. Mol.
Immunol., 24, 133-141.
Lorenz, W.,Doenicke, A., Halbach, S., Krumey, I., and Werle, E. (1969).
Histaminfreisetzung and magensaftsekretion bei narkosen mit propanid
(Epantol). Klin. Wochenschr., 47, 154.
Mason, M. E. and Okey, A. B. (1982). Cytosolic and nuclear binding of
2,3,7,9-tetrachlorodibenzo-p-dioxinto the Ab receptor in extra hepatic
tissues of rats and mice. Eur. J . Biochem., l23, 209-215.
Neuberger, J. M., Mieli-Vergani, G., Tredger, M., Davis, M., and Williams, R.
(1980). Oxidative metabolism of halothane in the immuno-pathogenesis of
the associated liver damage (abstract). Gut, 21, 467.
North, W. J., Dye, E. S . , Mills, C. D. , and Chandler, J. P. (1982).Modulation of
antitumour immunity; immunological approaches. Springer Seminars in
Immunobiology, 512. 193-200.
Ohtoshi, T.,Yamanuchi, N., Tadokoro, K., Miyachi, S., Suzuki, S . , Miyamoto,
T., and Muranaka, M. (1986). IgE antibody-mediated shock reaction
caused by topical application of chlorhexidine. Clin. Allergy, 16, 155-162.
Parker, C. W., de Weck, A. L., Kern, M., and Eisen, H. N. (1962). The
preparation and some properties of acid derivatives relevant to penicillin
hypersensitivity. J . Exp. Med., 115, 803-819.
Pimm, M. V. and Baldwin, R. W. (1978). Immunology and immunotherapy of
H . E. Amos 213
experimental and clinical metastases. In: ‘Secondary Spread of Cancer’. Ed.

R. W. Baldwin. Academic Press, London, pp. 163-209.
Pitot, H. C., Goldsworth, T., Campbell, H. A., and Poland, A. (1980).
Quantitative evaluation of the promotion by 2,3,7,8-tetrachlorodibenzo-p-
dioxin of hepatocarcinogenesis from diethylnitrososamines. Cancer Res., 40,
3616-3620.
Poland, A. and Glover, E. (1980). 2,3,7,8-tetrachlorodibenzo-p-dioxin segrega-
tion of toxicity with the Ah locus. Mol. Pharmacul., 17, 86-94.
Poland, A. and Glover, E. (1982). Tumour promotion by TCDD in skin of
HRS/J hairless mice. Nature (London), 300,271-273.
Robez, J., Greaves, J., Holland, J. F., and Bekesi, J. G. (1982). Determination
of polybrominated biphenyls in serum by negative chemical ionization mass
spectrometry. Anal. Chem., 54, 1104-1108.
Rose, F. (1984). Personal communication.
Seinen, W., Vos, J. G., Van Spenje, I., Snock, M., Brando, R., and Hooykaas,
H. (1977).Toxicity of organotin compounds. 11. Comparative in vivo and in
uitro studies with various organotin and organolead compounds in different
animal species with special emphasis on lymphocyte cyctotoxicants. Toxicul.
Appl. Pharmacol., 42, 197-212.
Seveso, (1979)The escape of toxic substances at the ICEMSA establishment on
10 July 1976,and the consequent dangers to health and the environment due
to industrial activity. A translation by the Health and Safety Executive of the
offical report of the Parliamentary Commission of Enquiry, by permission of
the Parliament of the Republic of Italy.
Sharma, R. P. and Zeeman, M. G. (1980). Immunologic alterations by
environmental chemicals: relevance of studying mechanisms versus effects. J.
Immunopharmacol., 2, 285-307.
Silva, J., Kaufmann, C. A., Simon, D. G., Landrigan, P. J., Humphrey, H. E.
B., Heath, C. W., Wilcox, K. R., van Amburg, G., Kaslow, R. A., Ringel,
A., and Hoff, K. (1979). Lymphocyte function in humans exposed to
polybrominated biphenyls. J. Reticulendothel. SUC.,26, 341-346.
Spreafico, F., Merendino, A,, Braceschi, L., and Sozzani, S. (1987). Immuno-
depressive drugs as prototype immunotoxicants. In : ‘Immunotoxicology’.
Eds A. Berlin, J. Dean, M. H. Draper, E. M. B. Smith, and F. Spreafico.
Martinus Nijhoff. pp. 192-207.
Trey, C., Lipworth, L., and Chalmers, T. C. (1986). Fulminant hepatic failure.
New Engl. J. Med., 279, 798.
Vecchi, A., Sireni, M. A., Canearata, M., Recchia, M., and Garatini, S. (1983).
Immunosuppressive effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin in strains
of mice with different susceptibility to induction of aryl hydrocarbon
hydroxylase. Toxicol. Appl. Pharmacol., 68, 434-441.
Vos, J. G. and Moore, J. A. (1974).Suppression of cellular immunity in rats and
mice by maternal treatment with 2,3,7,8-tetrachlorodibenzo-p-dioxin. Int.
Arch. Allergy Appl. Immunol., 47, 777-794.
CHAPTER 12
Perspectiv es-the Evaluation

of Reproductive Toxicity and
Teratogenicity
A. B. G. LANSDOWN
1 Introduction
The quality of an infant at birth and its capacity to undergo further
development and maturation is an expression of the viability and
fertilising capability of germ cells, successful conception, and subsequent
intrauterine growth. Peri- and post-natal development will usually
depend upon the growth potential of the individual and its ability to cope
with environmental circumstances. Toxicologists have become aware that
these environmental circumstances may act at any stage from germ cell
production in the male or female of the species to produce infertility,
genetic damage, or failure in reproductive potential, through conception
to intrauterine development and peri-natal stages, leading to structural,
functional, and behavioural changes in the offspring. Intrauterine de-
velopment is a highly vulnerable stage in life when growth in the offspring
depends on its genotype and its response to environmental factors.
Life before birth has become a highly emotive subject in most modern
societies. Declining birth rates, local increases in the birth of children
with structural and functional abnormalities, and causes of peri-natal
death have presented considerable epidemiological concern in recent
years, and clinical research has been directed towards developing
methods such as ultrasonography and biochemical tests to identify
‘at-risk’ pregnancies, as well as in identifying genetical or environmental
causes of foetal distress. Thus, in a recent study of foetal and neonatal
mortality in Curacao and the Netherlands Antilles, an analysis revealed
an overall death rate of 34.2 per 1000 total births (Wildschutt et al.,
1987). In 28 of 223 consecutive foetal and neonatal deaths, malformation
was the principle cause of fatality; others included hypertension, antepar-
tum haemorrhage, and asphyxiation. Among the environmental causes of
214
malformation, infection, complications associated with intrauterine

growth retardation, and maternal metabolic disorders were identified.
The role of infectious agents is well known and numerous excellent
reviews on this subject are available (Dudgeon, 1970; Brown and
Karunas, 1972; Mims, 1968; Overall, 1972; Overall and Glasgow, 1970).
Low birthweight and its post-natal consequences for survival and
subsequent growth are widespread international problems and a subject
for intense research among epidemiologists, clinicians, and experimental
toxicologists (Coid and Ramsden, 1973; Lansdown and Coid, 1974;
Lansdown, 1977; Levin et al., 1975; Edwards, 1969; Lechtig et al., 1975).
Such environmental causes as dietary inadequacy, maternal exposure to
sub-toxic thresholds of chemicals in the environment, or genetical factors
have been identified. Of human pregnancies recognised by a missed
menstrual period, about 15% end in miscarriage, although this level may
be as high as 50% if one includes pregnancies diagnosed by means of the
more sensitive hormonal assays. To the Darwinist, this pre-natal loss
involving progeny subject to retarded growth, genetical damage, or
unspecified complications, is a form of natural selection. The implications
of this concept in limiting the dissemination of such life-threatening
diseases as insulin-dependent diabetes, blood group disorders, anaemias,
and other conditions through the community are considerable (Valdheim
1986; 1985).
To the clinician and the toxicologist, germ cell production conception
and pre- and peri-natal growth in the offspring present different chal-
lenges. In the former case, the past 25 years have seen a surge of
interest in the development of in uitro fertilisation and ovum transplanta-
tion in an attempt to overcome the problem of reproductive failure, and
the use of ultrasound, serum a-foetoprotein, and amniocentesis in the
diagnosis of congenital disorders. Toxicologists in contrast, directed by
the lessons learned in the thalidomide episode in the late 1950s, and
subsequent catastrophes such as the Minamata situation and the inadver-
tent use of Agent Orange in the Vietnam war, have sought to identify
chemicals and environmental hazards which are potentially harmful in the
broad reproductive process in man or other species. Interestingly, the
techniques developed as an integral part of clinical in uitru fertilisation
programmes have proved particularly useful in studying the responses of
pre-conception embryos to potentially toxic agents.
It is the purpose of the present review to focus on tests available in the
toxicological evaluation of substances to impair processes in germ cell
production, reproductive activity, and pre-natal development. As will be
seen with several examples selected, many present attitudes and legisla-
tive requirements have been fashioned by real-life experiences.
2 Historical
Man has exhibited a strong interest in reproductive disorders or abnor-
malities in the development of members of his own race and in various
216 Evaluation of Reproductive Toxicity and Teratogenicity
species of animals since earliest times. Many illustrations exist in wall

paintings and ancient manuscripts of individuals with bizarre structural
deformities, dating from the times of the ancient Babylonian and
Egyptian dynasties. Even through the Middle Ages and until compara-
tively recent times, reproductive inadequacy or deformity in the neonate
was attributed to the action of demons or curses, or even to acts of God.
Perhaps the first person to apply embryological principles to the study
of congenital abnormalities was the eminent physician William Harvey
(1651), who is reputed to have noticed that harelip, as occurs in some
deformed infants, resembles a condition seen normally at an earlier stage
in development. To Harvey, this harelip condition represented a state of
‘arrested growth’, a hypothesis that has been employed subsequently to
explain the development of such abnormalities as cleft palate, ectopia
cordis, and gastrorachischisis.
In the 18th century, Etienne Geoffrey St. Hilaire (1832) and his son
performed a series of classical studies using chick embryos. They
demonstrated that developmental abnormalities such as spina bifida or
anencephaly could be induced by shaking the eggs or partly coating the
shells with an impermeable material. Later, they extended their work to
study patterns of human developmental abnormality. They coined the
term ‘teratology’ to describe the study of monstrosities and grotesque
developmental anomalies.
Important work showing that developmental anomalies may be in-
duced by environmental agents dates from the early 1940s when pre-natal
infection was recognised as a cause of structural and behavioural
abnormality. Thus, Gregg and workers in Australia demonstrated a
correlation between rubella infection in early pregnancy and the birth of
infants with congenital heart disease, deafness, cataract, and other
deformities, known collectively as the ‘rubella syndrome’ (Gregg, 1941;
Swan et al., 1943). Subsequently, a wide range of viruses and other
infections have been identified as potential causes of reproductive failure,
infertility, or congenital deformity in humans or sub-human species
(Lansdown, 1978).
Perhaps the most important single event, creating an awareness that
foetal development can be influenced by xenobiotic agents administered
to a pregnant mother at sensitive stages in pregnancy, was the identifica-
tion of the hypnotic drug thalidomide as a cause of limb and spinal
abnormalities (Lenz, 1961; McBride, 1961).
Even now, there is controversy as to how this growth impairment
occurs but it is clear that the foetus is most vulnerable during the period
of maximum cell division and organogenesis. Whilst this concept is
generally true in respect of teratogens that pass transplacentally to
accumulate in foetal tissues, work with a broader range of chemical
teratogens has established that, depending upon the nature of a teratogen
and its site of action-accumulation-metabolic breakdown, foetal de-
velopment may be adversely influenced at any stage from conception to
birth. It is evident also from recent studies that nutritional inadequacy,
drugs, or exposure of pregnant mothers to adverse environmental

conditions in late pregnancy may influence post-natal development and
maturation (Dobbing and Sands, 1971; Dobbing, 1974). At these later
stages, organogenic processes in the foetus have reached an advanced
state. However, organs such as the heart, lungs, and brain are in the
essential phase of ‘functional development’. The foetal brain is particu-
larly susceptible to environmental changes at this late stage in gestation
and Dobbing and his colleagues have demonstrated, using suitable animal
models, that nutritional deprivation or factors leading to foetal growth
retardation during the stage known as the ‘brain growth spurt’ can
precipitate irreversible damage and be manifest post-natally as mental
retardation (Dobbing and Sands, 1971). The prospect of catch-up growth
has been investigated in human babies and in a variety of animal models,
and the science of behavioural teratology has emerged.
It would be unproductive to list the wide range of events occurring
over the past 25 years from which present day thinking and the design of
tests to identify environmental agents capable of impairing reproductive
processes derive. It is purposeful to note that the emergence of
reproductive toxicology has been greatly influenced by real-life problems
as diverse as infertility and reproductive failure in men following
occupational exposure to lead or chronic alcoholism, and foetal abnor-
mality and pre-natal death resulting from the consumption of bread
contaminated with mercurial fungicides. As new information comes to
hand, so testing procedures and precautionary measures will be revised
accordingly (Lansdown , 1987).
3 Principles of Reproductive Toxicity and

Teratogenesis
The reproductive process involves germ cell production and maturation
in the parental generation, fertilisation, conception, and intrauterine
growth of the offspring, parturition, and development after birth. In each
case, normal development depends upon the viability of the genotype
and appropriate genetical instructions for morphogenic events. Constitu-
ent cells in the embryo/foetus will undergo a programmed sequence of
cyto-differentiation, metabolic and functional changes, migration, and
even death, in achieving normal organogenic events. Within a species
these development patterns may show small variations, but for a
particular strain or species they will fall within the ‘expected range’.
Small and subtle deviations in development are documented for humans
and most strains of laboratory animals commonly employed in reproduc-
tive studies.
Reproductive toxicology embodies several basic toxicological principles
but also involves some specialised approaches and expertise. The period
of intrauterine growth represents a highly vulnerable stage in develop-
ment when the offspring are particularly sensitive to the toxic effects of
exogenous agents, nutritional changes in the mother, and adverse
environmental physical factors. Maternal illness, inappropriate choice of
drugs, hypoxia, and hyperthermia have been implicated variously in
reproductive failure, defective foetal development, and reduced post-
natal survival (Edwards, 1969; Lechtig et al., 1975). Epidemiological
evidence suggests that up to 80% of human conceptions abort and that at
least 6% of live births are complicated by some structural or functional
defect.
Estimates of the frequency of these abnormalities vary greatly on
account of the wide racial, socio-economic, and geographic variation in
population studies, Statistics are also complicated by difficulties arising
in the definition of ‘what constitutes an abnormality’. The identification
of minor deviations in human development is a further problem. In early
foetal mortality and stillbirth, it is quite likely that an undiagnosed
deformity is present, possibly reflecting a mutagenic or functional defect.
Defects in reproductive performance may be due to genetical or
environmentally induced damage in germ cell production, alteration in
physiological factors, disease processes in male or female partner, or
through immunological factors. In controlled laboratory experiments
conducted under defined conditions, most of these variables may be
eliminated such that the true influence of an environmental compound
acting on a particular stage in the reproductive cycle can be assessed
accurately.
Six general principles apply in teratological studies, each being
supported by abundant clinical and experimental evidence (Table 1)
(Wilson, 1959). As a generalisation, the susceptibility of foetal tissues to
deformation or damage relates largely to the nature of the exogenous
influence and its pharmacological effect on tissues differentiating at the
Table 1 General principles of teratogenicity

The susceptibility of a conceptus to a teratogenic agent depends upon its
genotype and the manner in which this interacts with the environmental
factor.
The susceptibility of a conceptus to a teratogenic agent relates closely to its
development stage at the time of exposure.
Teratogenic influences act upon sensitive cells and tissues of the embryo or
foetus by specific mechanisms to induce deviant development.
Deviant development may be manifest as growth retardation, structural
deformity, functional disorder, or by mortality.
A teratogenic agent will induce abnormalities in sensitive tissues, the effect
being related to the nature of the agent and the route by which exposure
occurs.
Manifestations of deviant development increase in degree as the dose level
of the teratogenic agent increases. They range from a no-effect threshold to
a totally lethal effect.
A. B. G. Lansdown 219
time of exposure. In the foetus, each tissue system seems to exhibit a

stage of critical sensitivity outside which it exhibits little or no response.
For example, in the rat foetus exposed to trypan blue, dye accumulated
in the lysosomes of the yolk sac endoderm resulting in nutritional
insufficiency in the foetus and deformities in those organs developing at
that time. At later stages, when the function of the yolk sac placenta is
superceded by the chorioallantoic placenta, trypan blue exerted minimal
effect on foetal development (Beck and Lloyd, 1963). Stockard (1921)
recognised these critical phases of development which he indicated could
occur at any stage from conception to birth and involve functional or
morphological parameters. In the present discussion, it is appropriate to
make a distinction between the terms ‘embryo’ and foetus. For present
purposes, embryonic development is defined as that period during which
organ development occurs and when a defect in cell function results in
malformation. In the foetal stage the principle organ systems are
differentiated and development failure is a potential cause of pathological
damage or functional disorder.
The genetic make-up of a conceptus is the setting in which teratogene-
sis occurs; genetic and extrinsic factors interact to varying degrees to
initiate developmental abnormality. Responses range from embryo-letha-
lity to no-effect, depending upon the teratogen and species of animal
involved. Experience with such drugs as thalidomide, salicylate, and
anti-cancer agents has shown that the dose causing maternal toxicity is an
unreliable guide to the level required to induce foetal deformity.
Thalidomide exhibits a very low toxicity threshold in mothers, but is
highly teratogenic to the foetus at low concentrations. In contrast, the
teratogenic dose of sodium salicylate, in rats at least, closely approaches
that causing maternal toxicity (Lansdown, 1970). Many examples exist to
illustrate ways in which the mother exerts a protective effect against
teratogenic influences. This is particularly true in the case of nutritional
deficiencies. In the case of drugs, compounds are possibly detoxified in
the maternal liver. The placenta may act as a barrier for some toxic
agents and infections (influenza virus). Occasionally, as in the case of
cadmium and lead ions, which are strongly teratogenic in rodents if they
enter the circulation, substances ingested with the diet are poorly
absorbed from the maternal gastrointestinal tract.
It may well be that for many teratogens the concentration absorbed
into the maternal circulation correlates well with that seen in the foetus
(Gillette, 1977). However, since drug metabolising enzyme systems are
poorly developed or not present at all in the foetal liver, toxic compounds
tend to concentrate in the tissues. This would account in part for the
greater toxicity seen in foetal tissues compared to those of the mother.
Examples also exist to show that the teratogenic potential of a compound
in a particular species relates closely to the rate at which it is
metabolised. Alcohol is a well known human teratogen but is appreciably
less effective in the rat, probably on account of the very much faster rate
of metabolism in that species (Kennedy and Persaud, 1979). Variations in

the maternal metabolism of teratogens are attributable to differences in
plasma-protein binding, tissue distribution patterns, and excretion (Smith
and Caldwell, 1977). In the laboratory situation, these factors will
normally be examined in simple preliminary studies well before terato-
genicity and reproductive trials are planned.
4 Epidemiological and Experimental Aspects

Strongest evidence that a particular chemical or environmental factor is
likely to prove injurious to human reproduction is provided by large scale
population studies, such as those conducted in Japan and Iraq in relation
to accidental mercury poisoning (Takeuchi and Matsumoto, 1969; Bakir
et al., 1973). Chronic maternal alcoholism (Jones et al., 1973), rubella
virus infection (Swan et al., 1943; Gregg, 1941), ionising radiation
(Hiraoka, 1961), and exposure to drugs like thalidomide (Lenz, 1961;
McBride, 1961), are further examples of teratogenicity and reproductive
failure where irrefutable clinical evidence exists.
Occasionally, clinical studies suggesting that a particular agent or
condition is responsible for increased numbers of still-births , congenital
deformities, and reproductive failures are equivocal. One such example
concerned the potential teratogenic risks posed by trace levels of
anaesthetic gases to pregnant nurses and other personnel working in
hospital operating theatres (Corbett et al., 1974; Lansdown, Pope et al.,
1976). These and other workers failed to identify a clear cause-effect
relationship or to establish threshold levels of halothane , nitrous oxide,
or other gas necessary to cause toxic symptoms. Subsequent studies in
which pregnant rats were exposed to high sub-anaesthetic levels of single
or combinations of more than one gas failed to confirm teratogenicity
(Lansdown, Pope, et al., 1976; Pope, Halsey, et al., 1975; Pope, Halsey,
et al., 1978). Pregnant animals were exposed at periods of gestation at
which maximum teratogenicity occurs with other teratogens. Where
foetal growth retardation was evident, this was attributed to reduced food
consumption through anaesthetic-induced drowsiness. Although it would
be unwise to extrapolate these observations in terms of a no-effect in
humans without further wide-ranging trials, possibly employing other
species, it is conceivable that the problems identified in the human cases
resulted from a complex interaction of many factors, i.e. stress, tiredness,
etc. Nevertheless, as a precaution a large proportion of hospital operating
theatres in the United Kingdom are fitted with scavenging devices to
control air quality and to remove trace levels of anaesthetic gases.
Many examples exist of clinical evidence that a substance or environ-
mental condition causes reproductive distress or teratogenicity where the
results have been confirmed experimentally. Thalidomide, salicylates,
hypoxia, hyperthermia, and various nutritional imbalances are examples.
Occasionally, however, reproductive or teratogenic risks identified
through accurate clinical observations have not been reproduced under

experimental conditions. Perhaps the best known of these is the rubella
syndrome. A large number of studies have been conducted using a wide
variety of laboratory animals, including sub-human primates. Although
some congenital deformities have been produced resembling those
produced by rubella infection in humans, viraemia has not been recorded
on many occasions and the true rubella syndrome of congenital heart
disease, abortion, cataract, and other changes has not been reproduced
(Sever, Meyer, et al., 1966).
In the main, experiments designed to demonstrate teratogenicity or
reproductive toxicity these days tend to be prospective rather than
retrospective in outlook. It is hoped that, as a consequence of the
experience gained in the last 25 years, the vast proportion of hazards
presently in the environment or in permitted drug lists will be ap-
preciated, and recommendations made accordingly. In the years ahead,
through nationally and internationally controlled safety evaluation proce-
dures, it is expected that most hazards among new drugs, food additives,
industrial processes, and agricultural preparations will be identified
before the human risk is encountered.
5 Mechanisms of Reproductive Toxicity and

Foetal Abnormality
Agents in the environment may adversely influence reproduction in
humans and other species by their toxic action at any time, from germ
cell production and maturation in the parental generation, to post-natal
growth and development in the offspring. This toxicity may involve a
specific stage in the reproductive cycle or be of a more general type. For
simplicity, it is preferable to consider the action of the various agents
according to whether they act directly or otherwise on germ cell
production and reproductive performance in male or female parent, or
alternatively are toxic in some way to post-conceptional development of
the progeny. It is recognised that normal development in the progeny
either pre-natally or post-natally depends upon the viability of the
genotype and the way in which this interacts with its environment.
A Germ Cell Production and General Reproductive

Performance
A wide range of exogeneous agents including viral infections, metal ions,
alcohol, anti-cancer and anti-viral drugs, pesticides, and food additives
are known to impair spermatogenesis, leading to reduced reproductive
capacity and maybe sterility in the long term. On occasions, defects in
spermatozoa1 morphology have been linked with metabolic, functional,
and structural abnormalities in the surviving progeny. It is also evident
from clinical observations with certain drugs that induced hormonal
changes in the male or female partner may impair libido, reproduction

cycles, or germ cell production, leading to reproductive failure (Lans-
down, 1983; Dorrington and Gore-Langford, 1981; Nicholl, 1980).
Reproductive activity may be depressed in the event of ill-health. This
frequently leads to altered physiological and endocrinological activity.
It occurs with infections like influenza, cytomegalovirus, and several
enterovirus infections, conditions which are acknowledged causes of
impaired foetal development.
B Teratogenesis
Teratogenesis in the present context is used in its narrowest sense to
define the action of any xenobiotic chemical or environmental condition
on structural or functional development in the embryo/foetus. This may
occur by toxic action in susceptible tissue or organ systems, or through a
detrimental effect upon the health and nutritional state of the pregnant
Table 2 Levels of action of environmental agents in cells of the

developing embryo1foetus
ENVIRONMENTAL
AGENT
cell e n v i r o n m e n t
1Nutritional
and
Physical defects
Deformities i n
Structure and
Function
a. Mutagenesis
b. Chromosomal dam
c. Mitotic inhibition
d. I m p a i r e d DNA-
synthesis
Defects i n enzymic p a t h w a y s
and biomolecular syntheses
V
Abnormalities i n i n t e r c e l l u l a r communication,
cell recognition, i n d u c t i o n a n d m i g r a t i o n
mother. By this latter mechanism, foetal growth is disturbed through

imbalances in the intrauterine environment or the availability of
nutrients essential for optimal growth. The mother is unable to satisfy the
physiological demands of her offspring.
In view of the extreme diversity of environmental agents capable of
impairing foetal growth in one or more species, it is clearly impracticable
to attempt formulating a unifying hypothesis to explain the ways in which
development abnormalities arise. For present purposes, a simplified
scheme is proposed where specific levels of toxic action are identified. In
each case, the text will be illustrated by experimental studies employing
specific chemical factors or genetically determined conditions. The
scheme envisaged (Table 2) recognises the cell as the main functional unit
and identifies four principle routes of action (Saxen, 1976). This
represents a considerable simplification of the mechanisms for abnormal
development suggested by Wilson (1972), who considered up to ten
possible ways in which sensitive tissues might respond to toxic insults. In
the present discussion, it is recognised that cell sensitivity varies
according to generative phases of the tissue and the gestational age of the
individual.
The four-level model for teratogenesis allows for the fact that certain
teratogens including metal ions, hypervitaminosis A, and some drugs
influence tissues to different degrees according to their development
state, functional condition, and activity. It offers a broad framework
within which to discuss mechanisms of deviant development , avoiding
many of the controversies that have arisen over what constitutes a
primary route and that which is clearly of secondary importance.
The Intracellular Compartment

Genetical or chromosomal damage is an expression of changes resulting
from exposure to toxic agents capable of altering DNA structure. On
occasions, this may be difficult to distinguish from responses due to such
factors as ageing, irradiation, or ultraviolet light, all of which are capable
of mutagenicity (Kolata, 1980). Chromosomal damage is more frequently
seen in patients with cancer or congenital defects but the reverse is not
invariably true. The incidence of defects in the newborn attributable to
chromosomal damage is an indication of clastogenicity minus the rate of
intrauterine death.
In the human race, the relative contribution of genetical factors to
congenital defects is not known. Approximately 5% of malformations are
due to somatic mutations, 10% to chromosomal damage, and 5% to
identifiable causes in the environment (Wilson, 1977). The so-called
‘inborn errors of metabolism’ are included in this group. The mutation is
reflected often in terms of an enzymic deficiency affecting a specific
metabolic pathway.
The contribution of genetic damage to teratogenesis has been discussed
on many occasions (Wilson, 1977) but often the data for a particular
group of chemicals are incomplete and the interpretations questionable.
Studies 'conducted with a range of well known organochlorine and
organophosphorus pesticides are an illustration (Table 3) (Sternberg,
1979). In summary, the mutagenic potential of these agents seems to be a
good indication of their carcinogenic potential in the tests available but
not of their teratogenic activity. DDT, for example, caused chromosomal
damage in mice, and in two cell culture systems it induced dominant
lethal mutations in rats and mice, but it was not teratogenic in either
species. Strangely, DDT actually prolonged reproductive life in rats by
up to five months, and protected against the teratogenic effects of
salicylates, Benlate, and chloridine. Multigeneration studies in rats have
demonstrated that aldrin, carbaryl, dieldrin, and malathion may lead to
Table 3 Mutagenic potential of a range of organochlorine and organo-

phosphorus pesticides
Insecticide Toxicity
Mutagenicity Teratogenicity Fertility
DDT
Aldrin
Dieldrin
Endrin
Kepone
Chlordane
Heptachlor
Mirex
Lindane
Dichlorvos
Malathion
Parathion
Carbaryl
Diazinon
Captan
Nitrosocarbaryl
Key-1. Chromosomal damage in vivo or in vitro
2. Unscheduled DNA synthesis
3. AMESTEST
4. DNA base pair substitution test
5. Dominant lethal assay (rats or mice)
6. Heritable translocation test
7. Chromosomal damage in human peripheral blood culture
8. Chromosomal damage in Drosophila sp. salivary gland
a. Embryotoxicity
b. Teratogenicity
* * Oestrus changes in dogs, rats and rabbits with or without
pregnancy failure
* * * Reduced fertility
(Reproduced by permission from Sternberg (1979) Pharm. Therap., 6, 147)
infertility when tested over three generations (Smith and Caldwell, 1977),
although the experimental evidence has not been altogether convincing in
every case.
More consistent evidence of teratogenicity is seen with those agents
which impair DNA synthesis and mitotic activity at periods of high
organogenesis. Ionising radiation is a well known cause of mutation in
human patients and an acknowledged cause of congenital abnormalities
(Yamazaki ef al., 1954; Brent, 1972; Wilson, 1964a). Whole-body
radiation leads to the widespread production of free radicals and is also
regarded as a cause of mutation (Sternberg, 1979). Mitotic arrest, cell
death , and impaired cell proliferation occur with milder teratogenic
influences like rubella infection, folate antagonists, cortisone , and agents
which inhibit the mitotic spindle (the stathmokinetic agents). Impaired
cell division in any tissue is a potential cause of asynchronous growth,
hypoplasia, and possibly agenesis, depending upon the susceptibility of
the tissue at the time of exposure.
Neural crest cells are particularly sensitive to the effects of purine and
pyrimidine antimetabolites. These agents will induce exencephaly , an-
cephalocoele, and spina bifida if they are administered to the pregnant
mother at the stage of neural fold formation in the embryo.
Defective regulatory mechanisms are a further teratogenic manifesta-
tion of intracellular toxicity. Anomalies in cell migration , proliferation ,
and intercellular communication occur. Naeye and Blanc (1965) con-
cluded that hypoplastic development in the offspring of mothers infected
with rubella early in pregnancy was due to a viral inhibition of cell
proliferation. The organs had, in fact, failed to achieve the ‘critical mass’
for the particular gestational age and subsequent events were delayed or
absent.
Vitamin A is an amphiphilic compound which readily penetrates the
cell membrane to cause mitochondria1 swelling, release of lysosomal
enzymes, and inhibition of DNA synthesis. Impaired cell division and
migration occurring at critical stages in the development of the cephalic
mesoderm leads to derangements in the overlying neuroectoderm and the
formation of exencephaly and spina bifida (Marin-Padilla, 1966). Cell
death, mitotic arrest , and defective migration patterns probably underlie
the micromelias, cleft palate , and cardiovascular deformities associated
with cortisone, cadmium ion, salicylates, and possibly thalidomide
(Saxen, 1970; Poswillo, 1975; Andrew and Zimmerman, 1971).
The several genetical deformities operating at the cellular level seem to
involve deficiencies in cytoplasmic enzyme systems. Phenylketonuria,
lysosomal alpha-L-iduronidase deficiency (Hurler’s Syndrome) , and dia-
betes mellitus are examples. As a consequence of the various enzyme
deficiencies , some metabolites reach toxic levels, whereas others exhibit
defective or alternative patterns of biodegradation. The defects are
manifest as foetal mortality or as structural, functional, or behavioural
abnormalities which are appreciated in the neonatal period.
The Cell Membrane

Foetal development involves a complex pattern of cell aggregation,
morphogenetic movements, and cell migration, all of which depend to
some degree on the biochemical characteristics of the cell membrane, and
the intercellular passage of essential nutrients and inducer substances.
Much of this subject is very unclear and speculative. However, some
interesting genetic and non-genetic models exist to illustrate how damage
induced in the cell surface can adversely affect morphogenesis and
functional development.
The vertebrate limb, for example , involves extensive cellular interac-
tion in development. The early limb bud exists as a finger-like projection
of mesenchyme ensheathed in ectodermal tissue. The apical ridge of this
ectoderm exhibits a profound influence on the morphogenetic patterns of
skeletal structures within the mesenchyme (Cooke and Summerbell,
1980; Saunders and Fallon, 1966). Extirpation, reversal, or injury to this
ridge of ectoderm has been shown to cause marked abnormalities in the
limb bud (Wolpert, 1976). Although a large number of teratological
agents have been shown to affect adversely the development of the limb
bud, the mechanisms are probably not specific for that tissue, but form
part of a more complicated condition.
Cell death is an important event in the morphogenesis of a number of
structures including the limb bud, kidney, and palate. This means that
any teratogen selectively altering morphogenic patterns of cell death is a
potential cause of structural abnormality in that tissue. Janus Green B,
for example, has been shown to inhibit the degeneration of the
interdigital tissues in the development of the limb bud, syndactyly being
the resulting deformity (Menkes and Deleanu, 1964). Involutionary
changes characterise the development of the thymus and thyroid from
branchial pouch primordia, and also the derivation of the pituitary gland
from Rathke’s pouch of stomodeal origin. Defects in the degeneration of
these primitive tissues result in a persistence of the embryonic tissues into
post-natal life as ‘embryonic rests’ or hamartomata. Such changes are
occasionally seen as part of the normal background pathology in
laboratory rodents (Lansdown and Grasso, 1971; Lansdown, 1982) but
occur rarely in other species (Rest, 1987).
Morphogenetic cell death has an important implication in the normal
development of the mammalian palate. Palatal primordia develop in the
lateral mesenchyme of the buccal region and migrate towards the midline
where fusion occurs following the degeneration of the intervening
epithelial tissues. Defects in this degenerative process underlie the
formation of cleft palate commonly associated with cortisone treatment in
foetal mice. Cleft palate may also be attributable to defects in mucopoly-
saccharide synthesis and cell division (Ross and Walker, 1967). Although
hypervitaminosis A is also a cause of cleft palate in various species, the
biochemical mechanism is probably complex, involving changes in the
synthetic capacity of mesenchymal cells and their interaction with tissues

of ectodermal origin (Marin-Padilla, 1966; Kochhar, 1968).
The Synthesis and Composition of the Extracellular Matrix

Many foetal tissues are characterised by an intracellular matrix of specific
chemical composition. Probably the best known example is the ground
substance of cartilage and bone, which is frequently deformed or absent
in foetuses from mothers exposed to teratogenic agents in early or
mid-gestation. Skeletal abnormalities are seen where exogeneous agents
or genetical determinants adversely influence the synthesis and degrada-
tion of mucopolysaccharides of cartilage matrix. Achondroplasia,
chondrodystrophy , and chondrodysplasia are all well known conditions
affecting bone development in humans, cattle, and laboratory animals
(Duffell, Lansdown, and Richardson, 1985).
Teratogens may impair the biosynthesis of chondroitin sulphate and
lead to defective cartilage formation and ossification patterns. An
example is seen with sodium salicylate, which has been shown to inhibit
the sulphation of the polysaccharide moiety in early cartilage develop-
ment (Lansdown, 1970). Deviant cartilage formation with hypertrophy of
chondocytes was shown to impair subsequent ossification patterns.
Damage in the skull and spinal regions lead to manifestations of
exencephaly and spina bifida in affected foetuses. Micromelias were also
common observations.
Tetracycline, hypervitaminosis A, and cortisone exhibit a pronounced
effect on skeletal development in many species (Saxen and Kaitilia, 1972;
Morris, 1972; Lansdown, 1983b). The mechanism is probably complex.
However, tetracycline shows a particular affinity for calcium ions in intact
foetuses and in cultured tissues. Impaired skeletal development in this
case is attributed to a state of ‘non-availability’ of essential micronutri-
ents. Other examples exist where ionic ‘competition’ is a cause of
structural deformity. Collagen and elastic tissue are other examples of
intercellular tissues which may be abnormal in teratogen exposed
foetuses. Abnormal collagen formation is seen in cattle with the
hereditary disease dermatosparaxis. Affected individuals exhibit defects
in procollagen synthesis (Lenaers et al., 1974). Lathyrogenic agents, such
as beta-aminopropionitrile, inhibit procollagen peptidase activity, and
abnormalities involving vascular and skeletal tissues are seen experimen-
tally. Nutritional copper deficiency is a further well known condition
affecting collagen development. It is probable that copper acts in some
way in the cross-linking of collagen fibres.
The Cellular Environment

In accordance with Wilson’s (1959) six principles of teratogenicity, foetal
development is determined by the close interaction between the foetal
genotype and its environment. This environment relates not only to the
intrauterine medium, which is intimately linked to the state of health of
the mother, but also the intrafoetal conditions. This latter aspect reflects
foetal nutrition, intercellular ionic balances, and physiological conditions.
Maternal metabolic diseases, malnutrition, and abnormal environmen-
tal conditions (hypoxia, hyperthermia, etc.) are well known causes of
distress in human pregnancies and they have been examined experimen-
tally in several animal models (Edwards, 1969). Often, conditions such as
these are expressed at the foetal level in terms of nutritional insufficiency
(Hurley and Cosens, 1974). This may arise through defects in utero-
placental circulation or through the inability of the mother to metabolise
nutrients essential for normal foetal growth. The influence of maternal
ill-health in pregnancy is well documented in the medical press.
A further aspect of maternal ill-health which is well known is that of
viral infection in pregnancy, especially with influenza, rubella, and herpes
viruses. With the influenza virus at least, where infection of the conceptus
has not been confirmed, foetal deformities and mortality are probably
largely due to fever (hyperthermia) in the mother (Edwards, 1969). The
work has been conducted in a number of experimental animal models,
the structural defects seen closely resembling those seen in infected
mothers. Metabolic imbalances attributable to disease processes are a
further possible explanation.
Coxsackievirus infection in mice leads to a profound exocrine pan-
creatic insufficiency and atrophy within two days of infection, and as a
consequence mothers become incapable of digesting dietary protein
(Lansdown, 1975b). Labile reserves of protein accumulated in the
maternal liver become depleted with the result that foetal development is
retarded and mortality increased (Lansdown, 1976). The level of foetal
effect seen correlated well with the stage in pregnancy at which infection
occurred (Lansdown, 1975a). Subsequent studies demonstrated that this
foetal distress could be prevented using immunological means or by
feeding the animals with diets containing protein in a digested form (i.e.
casein hydrolysate) . Foetal development, as determined by measuring
a-foetoprotein and albumin levels, was within normal limits for the
species and gestational age (Lansdown, Coid, and Ramsden, 1975).
Nutritional causes of congenital deformity are well documented with
reference to the serious problems encountered in developing countries.
Protein and/or deficiency in specific amino acids is associated with a
range of defects involving the development of skeletal, vascular, im-
munological, and nervous systems. Usually, these deformities form part
of a more general pattern of growth retardation with reduced cell
production (Naeye, Blanc, and Paul, 1973). Deficiencies in essential
vitamins, minerals, fats, and amino acids occurring through starvation,
metabolic deficiencies, or through the action of toxic compounds in the
diet are discussed elsewhere (Lansdown, 1983b; Hurley, 1977).
6 Legislative Aspects of Reproductive Toxicity

Testing
Reproductive toxicity is defined as that part of general toxicology dealing
with the adverse influence of exogeneous agents upon the reproductive
process (ECETOC, 1983). Teratogenicity represents a specific module
within this framework and relates only to the effects of test chemicals on
embryonic/foetal growth and their ability to induce malformations
recognisable at birth or at later stages.
Early reproductive toxicity studies were conducted only on food
additives, pesticides, and other preparations for which exposure could be
expected over several generytions. Studies would also have been con-
ducted with hormonal preparations or products which might have been
administered to women of child-bearing age. The test most commonly
used before 1960 was the two-litter test where animals were dosed from
before mating and through two complete cycles of mating, pregnancy,
and lactation. Toxicity was assessed on the basis of reduced levels of
conception, litter size, and post-natal viability. Surprisingly, thalidomide
induced changes in all three parameters but was not suspected of being
teratogenic. Retrospectively, it seems likely that its teratogenic properties
were not identified in the two-litter test on account of its exceedingly low
toxicity. In due course, Lenz (1961) and McBride (1961) clearly
recognised the association between thalidomide therapy in women in the
first trimester and the birth of babies with limb and spinal defects. Since
that time, there has been a dramatic reappraisal of the safety evaluation
of substances for reproductive toxicity.
The publication ‘Guidelines for Reproductive Studies for the Safety
Evaluation of Drugs for Human Use’ by the Food and Drug Administra-
tion of the United States (FDA, 1966) is the basis of most predictive
safety evaluation studies conducted these days. The recommendations are
for single generation studies for the evaluation of drugs and chemicals
likely to be ingested by humans over short periods, and for substances
that are voided from the body within a short time. In contrast, for agents
to which people might be exposed over a prolonged period, such as drugs
used in chronic treatments, pesticides, and food additives, and where
ingestion might extend over several generations, multigeneration studies
would be conducted.
A Single Generation Studies

These tests recommended by the Food and Drug Administration (1966)
are perhaps better known as the ‘Three Segment Studies’ on account of
the specification for separate fertility studies with general reproductive
performance (segment 1), embryotoxicity and teratogenicity (segment 2),
and peri- and post-natal dosing (segment 3) modules.
Experiments conducted under segment 1 provide an overall insight of
the potential risks accompanying a test compound without identifying

either its target tissue or mechanism of action. Rats and mice are
preferred for this study and they will be dosed for a sufficient time before
mating to reveal adverse effects on gametogenesis, and then through
mating, gestation, and post-natally until the end of lactation and
weaning. Half the female animals will be killed and examined in
mid-gestation to allow an investigator to assess the influence of his test
compound on conception and foetal survival. Comparison of the number
of corpora lutea in each ovary and the conception rate in each uterine
horn provides an assessment of pre-implantation loss. It may be that, if a
non-pregnancy situation is diagnosed at this stage, a study will be
terminated and repeated using lower dose levels. Supplementary studies
may prove useful in determining reasons for the non-pregnancy.
Segment 2 studies are more appropriately termed the teratogenicity
phase, since they seek to identify the influence of a test compound upon
post-conception development; they also cover the broader aspects of
embryo-lethality and altered growth patterns. Rats, mice, and rabbits are
used, pregnant animals being dosed throughout the period of optimal
organogenesis. Experience with a wide range of known teratogens has
shown that highest rates of malformations occur following dosing during
this period. Dosing at earlier stages may result in increased embryonic
loss. Experiments are terminated shortly before delivery and parturition,
thus avoiding the loss and cannibalism of deformed or dead offspring.
External examination of foetuses by hand lens, with or without transverse
sectioning to detect visceral abnormalities, is mandatory. Most teratolog-
ists routinely conduct skeletal examinations in foetuses stained with
alizarin red dye. On occasions, foetal tissues may be examined histologi-
cally for subtle abnormalities.
Segment 3 of the single generation studies is designed specifically to
evaluate the influence of a test compound on later stages of foetal
development, parturition, and early post-natal life. Again the rat, mouse,
and rabbit are species of choice. Animals are dosed during the latter third
of gestation and through lactation and weaning. The principal advantage
of this peri- and post-natal dosing study is that an investigator is able to
determine the influence of his test compound on functional development
of foetal organs. The specifications allow an appreciation of changes that
would not be detected under the recommendations in segments 1 or 2.
Tests for behaviour or organ function will be conducted in neonates at
the discretion of investigating scientists.
Where a test compound administered to a pregnant animal is excreted
in the milk, segment 3 studies should provide evidence of foetal toxicity
arising through transplacental exposure , and also of adverse effect
through consuming ‘contaminated milk’. The tests will demonstrate the
influence of a test compound upon lactation and maternal behaviour.
The three segment study is criticised on account of its inadequacy in
showing up the effects of test compounds on fertility and reproductive
A . B. G. Lansdown 23 1
performance (Palmer, 1981). Regulatory authorities in Great Britain and

in the E E C recommend that there may sometimes be an advantage in
conducting additional male and female fertility studies in segment 1, with
treated males being mated with untreated females and uice uersa.
However, experience with a large number of compounds has demon-
strated that this measure is warranted in less than 1% of cases. The
modification may allow a greater appreciation of toxicity than the
guidelines set by the Food and Drug Administration permit.
B Multigeneration Studies
Multigeneration studies are not conducted with some compounds in the
light of their chemical configuration and intended usage. They will be
performed, however? where prolonged exposure in the human environ-
ment is envisaged (DHSS, 1982). Multigeneration studies may extend
over three generations of test animals but this is time consuming and
costly. These days it is more common for the tests to be conducted over
two generations unless the intended use of the substance predicts that
further tests are desirable.
In the evaluation of multigeneration studies, the offspring of successive
generations will be examined for normality and maturation with re-
productive potential. For scientific and economic reasons, mice and rats
are normally selected for this type of work. A proportion of the offspring
from each generation will be sacrificed (usually 50%) and examined
macroscopically and histologically for abnormalities. The tests are
modified by the various regulatory authorities but these modifications
generally apply to the number of animals required in each test group and
in the specifications for sacrifice and evaluation of toxic changes.
C Additional Studies
Guidelines published by the various regulatory authorities for reproduc-
tive toxicity and teratogenicity are specific in their requirements; that is,
which species shall be used, the period of dosing, and the means of
evaluation of toxic change. It will be appreciated, however, that the costs
involved in conducting the entire spectrum of studies in intact animals
can be prohibitively expensive. As a consequence, authorities such as
ECETOC (1983) encourage the development of short term assays such
as:
(a) Male fertility and spermatozoa1 morphology tests
(b) Female fertility and the influence of test substances on oestrus
cycles
(c) In vitro tests for embryotoxicity
(d) Dominant lethal assay tests for mutagenicity.
Specific recommendations are not available yet for these additional tests?
but there is evidence that many research groups are using them on
account of the smaller number of animals required. The tests are also
preferred since they can provide specific information in a shorter time
than is possible where the more conventional tests are employed.
ECETOC (1983) consider that more emphasis should be placed upon
developing and validating the short term assays in reproductive toxicol-
ogy. It is hoped that, in due course, the tests required by the various
national and international regulatory authorities will be harmonised.
7 Experimental Considerations
A Species
Many strains and species of laboratory animal have been used in
reproductive studies, but the majority of regulatory studies have been
conducted in the popular strains of mouse, rat, and rabbit. More rarely,
dogs, ferrets, and pigs have been employed as non-rodent species.
Species selection has been largely based on the convenience of the
animals under laboratory conditions, a detailed knowledge of their
reproductive biology, and their sensitivity to known human teratogens
like thalidomide. Although monkey species are phylogenically closer to
humans, they exhibit a number of important metabolic and behavioural
differences such that rarely have they been found more suitable than
rabbits and rats. The latter species are considerably less expensive to use
and do not require the highly specialised level of management and
husbandry.
It is common practice in reproductive studies to administer test
compounds intragastrically or in the diet to mimic the anticipated
pattern of human exposure. Interspecies variations exist in the gastro-
intestinal absorption of exogeneous materials and food substances ,
principally on account of differences in gastric acidity, intestinal flora,
metabolising enzymes in the intestinal epithelium, and in the characteris-
tics of the mucosal surfaces. Preliminary studies are profitably conducted
to examine the pattern of intestinal absorption of test compounds before
conducting more definitive tests, since it is well known with many proven
teratogens that a close relationship exists between the level of compound
in the maternal circulation and the foetal response. The rate at which a
test compound is metabolised has a marked bearing on the levels of
teratogenicity seen. Alcohol, for example, is metabolised more slowly in
the human than in the rat and is appreciably more toxic. Imipramine is
eliminated more slowly from the body in rabbits and rats than in humans
(Harper, Palmer, and Davies, 1965).
The reproductive biology and foetal development are well known in rats
and mice, making them choice species for many types of toxicity study
(Wilson, 1964b). They were used in the early two-litter tests which
preceded the present reproductive toxicity tests. However, their value as
a sole test species is limited on account of their low sensitivity to
thalidomide. Nowadays, the rabbit is preferred as the main species for
A. B. G. Lansdown 233
most teratological tests, with rats and mice being employed as second
species.
Many experimental studies have been conducted in mice but their
value in teratological studies is limited in view of their known suscep-
tibility to spontaneous cleft palate (Loevy, 1962). This feature has been
used to advantage in demonstrating the interaction between genetical
factors and environmental factors in abnormality (Pinsky and DiGeorge,
1965; Biddle and Fraser, 1976).
The sensitivity of the rabbit foetus to the deforming influence of
thalidomide is well appreciated (Giroud, Tuchmann-Duplessis, and
Mercier-Parot, 1962), but other features that make this an attractive
species for reproductive toxicology include its spontaneous ovulation
pattern and comparatively large size (Gibson, Staples, and Newberne,
1966). Although the rate of spontaneous abnormalities in the rabbit is
slightly higher than in rodents, this is not considered to be a serious
disadvantage. In each study, there will be a statistical evaluation of the
number of congenital abnormalities in test and control groups (Stadler,
Kessidjan, and Perraud, 1983).
The ferret has been recommended as a suitable species in teratological
studies. It is a small carnivore and may be suitable where a non-rodent
species is sought (Beck, 1975). However, it seems that few studies so far
have been conducted using this species and there is still a paucity of
background data on its general biology and susceptibility to known
teratogens. On the other hand, the hamster has featured widely in
experimental studies and its sensitivity to several known human terato-
gens is appreciated (Ferm, 1965). Although it is a convenient laboratory
animal and has a comparatively short gestation (16 days), it is notoriously
difficult to dose by intravenous injection.
B Dose of Test Compound and Route of Administration

The magnitude of the dose and the route of administration of a test
compound to animals in reproductive toxicity experiments can have a
profound bearing on the outcome and levels of foetal abnormality.
Experience has shown that these two factors do not operate in isolation,
but can interact in the production of congenital abnormalities (Fraser,
1964; Mellin, 1964).
Experimental teratology and reproductive toxicity differ from most
pharmacological experiments in that test compounds are usually adminis-
tered at doses which are sufficiently high to produce an observable effect.
Occasionally, the doses administered are unrealistically high and cannot
mimic anticipated human exposure levels. In one such experiment, mice
were given 1900-2500 mg kg-' sodium chloride, a dose which could be
expected to cause extensive ionic and metabolic imbalances in maternal
and foetal tissues; but foetal abnormalities were reported (Nishimura and
Miyamoto, 1969). Other experiments have sought to demonstrate the
teratogenicity of caffeine in rats, but the doses administered were

equivalent to that contained in 38-72 cups of coffee (Le Chat, Borlee, et
al., 1980; Palm, Arnold, et al., 1978).
In regulatory studies, test compounds will normally be administered to
test animals by the route that is most relevant in human exposure.
Several examples exist to demonstrate that, whereas some substances are
without obvious risk under normal conditions of exposure, they may be
teratogenic under unusual circumstances. Cadmium and lead ions are not
teratogenic in hamsters or rats if given by oral route, but become so if
they are injected subcutaneously or intravenously (Ferm, 1969; McClain
and Becker, 1975). The relevance of this observation in terms of human
risk is unclear. In the case of lead ions, there is a reported ten-fold
difference in the intestinal absorption in rats and in humans (Moore,
Hughes, and Goldberg, 1979). Humans have been reported to absorb
about 10% of an ingested dose of lead (Rabinowitz, Wetherill, and
Kopple, 1976), but in their 1977 survey The Scandinavian Council for
Environmental Information Workshop concluded that ". . . lead is not a
teratogen in humans in the classical sense."
In conclusion, it must be emphasised that the biological properties of a
test substance cannot be equated with clinical risk. It is the circumstances
of exposure to the substance that are of paramount importance, that is
the route, level of dose, and the time of exposure.
C The Period of Dosing

Environmental agents may impair the reproductive activity of a species or
alter the developmental pattern of its progeny at any stage, from germ
cell formation in the parental generation through to post-natal growth
and maturation in the offspring. In the foetus there is a good correlation
between the type of abnormalities seen and the period in gestation at
which they are exposed to the deforming influence. For present purposes,
the reproductive cycle is considered in three main phases:
(a) Gametogenesis
(b) Development in the embryo/foetus
(c) Post-natal growth and maturation.
The various regulatory authorities take these factors into account in
making their specifications (Johnson and Everitt , 1984; Tuchmann-
Duplessis, 1984).
Experimental studies have demonstrated that each embryonic/foetal
tissue exhibits its own peak of sensitivity to deforming influences
(genetical or environmental). In most species, the period of high
teratogenicity tends to be during early or mid-gestation when or-
ganogenesis is optimal. In the mouse foetus, cleft palate was regularly
induced following administration of cortisone to pregnant mothers at
13-14 days (Biddle and Fraser, 1976). Exposure of rats to trypan blue
dye, sodium aurothiomalate, or cadmium ions produced maximum
Table 4 Range of reproductive efsects and manifestations of teratogenicity as a function of the time of exposure to toxic
agents during the reproductive cycle
State of Exposure Response
Male Female Conceptus
Before mating Reduced fertility, loss Reproductive failure , -

of libido, abnormal abnormal oestrus
sperm characteristics, cycles, hormonal
spermatocyte changes, defects in
mutation ovum transport,
mutation
During pregnancy Generalised maternal Preconception loss,
illness, metabolic, stillbirth, intrauterine
hormonal, and nutritional death, structural and
defects, circulatory and functional abnormalities,
other functional changes, biochemical changes, growth
toxaemia retardation, transplacental
carcinogenesis , somatic
mutations, post-natal distress.
Peri- and post-natal Failure in lactation, Peri-natal mortality, functional

period toxic metabolites in milk, and developmental abnormality,
prolonged gestation, mental and behavioural deformity,
deficiencies in maternal carcinogenic changes with
behaviour, weaning, etc. mutations, post-natal toxicity
through substances present in
milk
236 Evaluation of Reproductive Toxicity and Terutogenicity
teratogenicity during the period 8-10 days of gestation when foetal

nutrition is largely by way of the yolk sac placenta. In the human,
maximum sensitivity to thalidomide is seen at about 37-50 days (Lenz,
1961; Lenz and Knapp, 1962). Experimental evidence shows that, in
animals treated with teratogens at different stages in the reproductive
cycle, the responses seen are specific for the stage of exposure (Table 4).
D Experimental Observations
Identification of defects in reproductive performance or in foetal growth
constitutes a vital part of any predictive safety evaluation study. The
quality of the results relates directly to the sensitivity of the methods used
and to the experience of the investigator.
For comparative purposes, the influence of environmental agents on
the reproductive performance of a test species or the development of its
offspring may be assessed by the following indices:
(a) Pregnancy (i.e. the number of pregnancies recorded as a per-
centage of those exposed)
(b) Conception rate
(c) Foetal survival
(d) Malformation
(e) Parturition
(f) Post-natal survival (1, 10,20,50 days, etc.).
These indices are valuable in comparing the toxicity of different agents
(administered at comparable pharmacological doses), different dose
levels to construct a dose-response relationship, and test models.
The pregnancy index will be a measure of the influence of a test
compound upon germ cell production, reproductive behaviour , and
conception. Defects in the male animal may be attributable to impaired
spermatogenesis, sperm motility, and fertilising ability (capacitance),
hormonal changes, or lowered reproductive behaviour. An analysis of
spermatozoa1 morphology and fertilising ability may be conducted under
in vitro conditions. Other tests envisaged will include semen viscosity,
acidity, and chemical constitution. In the rabbit, sperm samples can be
readily obtained using the artificial uterus technique. In small species,
sperm samples are obtained from epididymal preparations.
Reproductive activity in female animals shows sensitivity to hormonal
changes as well as disturbances in general health. Drug-induced oestrus
changes are readily examined by vaginal smear techniques and assays of
circulating hormone levels. In vitro fertilisation is a further potentially
useful test in assessing the viability of the ovum.
A wide range of methods is available for determining foetal develop-
ment in teratological studies. Selection of procedure will normally be
limited by equipment available and economic factors. Commonly, an
investigator will count resorption sites, dead embryos/foetuses, and
stillbirths. Live progeny delivered by Caesarian section at near term are
examined routinely by hand lens and gross abnormalities in the shape and
proportions of the head and limbs will be noted. Transverse sections
(3-5 mm) allow an assessment of visceral changes (Wilson, 1959).
Skeletal abnormalities commonly occur in teratological experiments
where test compounds exhibit a direct influence on the development of
cartilage or bone, as well as those which induce a generalised impairment
in foetal growth. Dawson’s alizarin red-S technique is widely used in
staining the skeletons of foetuses treated with teratogens (Dawson,
1926). Addition of counterstains, such as methyl green, methylene blue,
and methyl blue, has proved useful in demonstrating cartilagenous
anomalies in whole foetal mounts. Teratological effects on skeletal
development may be assessed by the following:
(a) The number and extent of ossifications
(b) Observation of absent or defective ossifications
(c) Defects in the configuration of ossifications, e.g. wavy ribs,
supplementary ribs or digits.
In mechanistic studies, sections taken by Wilson’s (1959) method may
be preserved in formalin or Bouin’s fixative and preserved for histo-
pathological examination. Alternatively, as in Ullberg’s (1958) studies,
histological preparation with subsequent autoradiography may be
employed to identify sites of drug deposition/metabolism in foetuses
from mothers treated with radiolabelled compounds. This specialised
type of study will not be conducted routinely in many laboratories.
In peri- and post-natal experiments, toxic changes are sought in the
neonates and young offspring. Methods available for this include timing
of specific developmental events (e.g. eye opening, hair growth, etc.)
and behavioural pharmacology. This last parameter has become a very
specialised area and applies very much to the functional development of
the brain. Different laboratories have developed their own particular
expertise and such methods as the righting reflex, open-field exploration,
movement, and response to external stimuli are adopted (Barlow and
Sullivan, 1975). In many respects these studies resemble those conducted
routinely in pharmacological laboratories.
8 Statistical Evaluation
Reproductive and teratological studies generate a vast amount of
quantitative data. The onus will be upon the statistician to evaluate the
significance in differences in such parameters as:
(a) Reduced fertility
(b) Reproductivity
(c) Conception rates
(d) Intrauterine and post-natal growth rates
(e) Changes in the sex ratio of the offspring
(f) Increases in the incidence of structural and functional defects in
the offspring
(g) Viability
(h) Weaning index
(i) Patterns of post-natal growth in the offspring.
In order to obtain results that are statistically sound, it is important that
sufficient numbers of animals are assigned to test and control groups
using a table of random numbers. This procedure will be adopted also in
culling litters to a constant size for longer term observations. Randomisa-
tion will also be used in selecting animals for mating in multigeneration
studies.
In segment 1 tests, ten animals will be allocated per treatment group
for each step of the experiment; that is ten for examination in
mid-gestation and ten for autopsy at weaning. Experience has shown that,
at best, only agents having a potent effect on reproductive processes can
be detected (Palmer, 1981). It is necessary to obtain a 40-50% difference
between treated and control groups to achieve significance in Fischer’s
Exact Test.
Palmer (1974) considered that, on the basis of several hundred tests,
analysis of litters rather than individual foetuses provided the only valid
sample unit for statistical purposes and that litter values are not normally
distributed. Thus non-parametric modes of analysis are preferred , but
they are not infallable and may be used only as a guide for making a
judgement, and not the sole criterion for decision.
In assessing the significance of major malformations, a statistical
analysis may have limited application. It is likely that group sizes of at
least 60 animals-that is three times the number required by the Food
and Drug Administration-would be necessary to establish the statistical
significance of a 100-fold increase in the malformation rate in the event of
a 0.1% incidence of ‘background’ malformations.
9 General Observations
In more than 25 years since the thalidomide tragedy, toxicological
studies have assumed greater importance than at any time previously,
and research and development costs have increased disproportionately.
Legislative authorities throughout the world have sought to introduce , or
improve where necessary, screens to identify those environmental chemi-
cals liable to impair reproductive processes or development of the
offspring, long before a human is exposed by intention or accidentally.
As new information has come to hand, through experimental toxicology
or as a consequence of epidemiological study , legislative requirements
have become more stringent and far reaching. As in other disciplines of
toxicology, it will be the responsibility of the experimentalists to modify
their tests to take into account the nature of the substance they are
testing and the manner in which a human being may be exposed.
The three segment study for reproductive toxicology and teratology
will provide a lot of information, but it is complex and time consuming. It
may be preferable to conduct simpler tests to provide answers to specific

questions (Palmer, 1974). Alternatively, one could use the simple tests in
the early stages of toxicity evaluation and then progress to more
definitive trials. At all times, the toxicologist should be aware of the
limitations of any test he conducts, and if necessary apply further tests to
identify toxic effects.
Having completed necessary toxicological screens, the toxicologist will
attempt to extrapolate his observations in terms of human risk. This will
always be a contentious subject, but in reproductive toxicology at least,
experience has shown that the studies described above using the rat,
mouse, or rabbit as test species are normally adequate in revealing those
environmental agents likely to be injurious in human reproduction or
pregnancy. Although in vitro tests for studying the influence of environ-
mental agents on embryonic or foetal growth are widely conducted in
many larger laboratories at present, they are not currently accepted by
most major regulatory authorities as an alternative to studies conducted
in intact animals. It is likely, however, that if a test compound does
provide evidence of risk in preliminary in vitro studies a company may
decide against carrying out further studies at that stage for economic
reasons.
Acknowledgement
I am grateful to Mrs Anne Murray for her assistance in preparing this
chapter.
References
Andrew, F. D. and Zimmerman, E. F. (1971). Teratology, 4, 31.
Bakir, F., Damluji, S. R., Amin-Zaki, L., and Murtaiha, M . (1973). Science,
181, 230.
Barlow, S. and Sullivan, F. M. (1975). In: ‘Teratology: Trends and Applications’.
Eds C. L. Berry and D. E. Poswillo. Springer-Verlag, New York, p. 103.
Beck, F. (1975). In: ‘New Approaches to the Evaluation of Abnormal Develop-
ment’. Eds D. Neubert and H. J. Merker. Thieme, Stuttgart, p. 8.
Beck, F. and Lloyd, J. B . (1963). J . Embryol. Exp. Morphol., 11, 175.
Biddle, F. G. and Fraser, F. C. (1976). Genetics, 84, 743.
Brent, R. L. (1972). Davis’ Gynec. Obstr., 2, 1.
Brown, G. C. and Karunas, R. S . (1972). A m . J . Epidemiol., 95, 207.
Carter, C. 0. (1969). Br. Med. Bull, 25, 52.
Coid, C. R. and Ramsden, D. B. (1973). Nature (London), 241, 460.
Cooke, J. and Summerbell, D. (1980). Nature (London), 287, 697.
Corbett, T. H., Cornell, R. G., Endres, J. L., and Leiding, B. S. (1974).
Anaesthesiology, 41, 341.
Dawson, A . B. (1926). Stain Technol., 1, 123.
Department of Health and Social Security (1982). ‘Guidelines for the Testing of
Chemicals for Safety’, Report No. 27, p. 33. London.
Dobbing, J. (1974). Pediatrician, 53, 2.

Dobbing, J. and Sands, J. (1971). Biol. Neonat., 19, 363.
Dorrington, J. and Gore-Landford, R. E. (1981). Nature (London), 290, 600.
Dudgeon, J. A. (1970). In: ‘Clinical Virology: The Evaluation and Management
of Human Viral Infections’, Eds Debre and Celers. Saunders, Philadelphia,
p. 792.
Duffell, S. J., Lansdown, A. B. G., and Richardson, C. (1985). Vet. Rec., 117,
571.
ECETOC (1983). ‘Identification and Assessment of the Effects of Chemicals on
Reproduction and Development’, Monograph No. 5, Brussels.
Edwards, M. J. (1969). Teratology, 2, 313.
Food and Drug Administration (1966). ‘Guidelines for Reproductive Studies for
Safety Evaluation of Drugs for Human Use’, Washington.
Ferm, V. H. (1965). Lab Invest., 14, 1500.
Ferm, V. H. (1969). Experientia, 25, 56.
Fraser, F. C. (1964). In: ‘Proceedings of the Second International Conference on
Congenital Malformations’, Ed. M. Fishbein, International Medical Con-
gress, New York, p. 277.
Gibson, J. P., Staples, R. E., and Newberne, J. W. (1966). Toxicol. Appl.
Pharmacol., 9, 398.
Gillette, J. R. (1977). In: ‘Handbook of Teratology’, Eds J. G. Wilson and F. C.
Fraser. Plenum, New York, 3, p. 35.
Giroud, A., Tuchmann-Duplessis, M., and Mercier-Parot, L. (1962). C. R. SOC.
Biol., 156, 765.
Gregg, N. M. (1941). Trans. Ophthal. SOC.Austr., 3, 35.
Harper, K. H . , Palmer, A. K., and Davies, R. E. (1965). Arzneim-Forsch., 15,
1218.
Harvey, W. (1651). In: ‘Exercitiones et Generatione Animalium Typis du
Gardiansis’ , Pulleyn, Coemetaria Paulino , Amsterdam.
Hiraoka, S. (1961). Acta Anat. Nippon, 36, 161.
Hurely, L. S. and Cosens, G. (1974). In: ‘Trace Element Metabolism in
Animals’, Eds W. G. Hoekstra, J. W. Suttie, H. E. Ganther, and W. Mertz.
University Park Press, Baltimore, 2, p. 516.
Johnson, M. and Everitt, B. (1984). ‘Essential Reproduction’. Blackwell Scien-
tific Publications, Oxford.
Jones, K. L., Smith, D. W., Streissguth, A. P., and Mirianthopoulos, N. C.
(1973). Lancet, 1, 1076.
Kennedy, L. S. and Persaud, T. V. N. (1979). In: ‘Advances in the Study of Birth
Defects’; Ed. T. V. N. Persaud. MTP Press, Lancaster, 2, p. 223.
Knill-Jones, R. P., Moir, D. O., and Rodrigues, L. V. (1972) Lancet 1, 1326.
Kochar, D. M. (1968). Teratology, 1, 299.
Kolata, G. B. (1980). Science, 208, 1240.
Lansdown, A. B. G. (1970). Food Cosmet. Toxicol., 8, 647.
Lansdown, A. B. G. (1975a). Br. J . Exp. Pathol., 56, 119.
Lansdown, A. B. G. (1975b). Br. J . Exp. Pathol., 56, 373.
Lansdown, A. B. G. (1976). Teratology, 13, 291.
Lansdown, A. B. G. (1977). Biol. Neonat., 31, 252.
Lansdown, A. B. G. (1978). In: ‘Advances in the Study of Birth Defects’, Ed. T.
V. N. Persaud, MTP Press, Lancaster, p. 292.
Lansdown, A. B. G. (1982). Vet. Rec., 110, 429.
Lansdown, A. B. G. (1983a). Trends Pharmacol. Sci.

Lansdown, A. B. G. (1983b). In: ‘Toxic Hazards in Food’, Eds D. M. Conning
and A. B. G. Lansdown. Croom-Helm, London, p. 73.
Lansdown, A. B. G. (1987). In: ‘The Future of Predictive Safety Evaluation’,
Eds A. N. Worden, D. V. Parke, and J. Marks. MTP Press, Lancaster, 2, p.
77.
Lansdown, A. B. G. and Coid, C. R. (1974). Br. J . Exp. Pathol., 55, 101.
Lansdown, A. B. G., Coid, C. R., and Ramsden, D. B. (1975). Nature
(London), 254, 599.
Lansdown, A. B. G. and Grasso, P. (1971). J . Comp. Pathol., 81, 141.
Lansdown, A. B. G., Pope, V. D. B., Halsey, M. J., and Bateman, P. E. (1976).
Teratology, 13, 299.
Le Chat, M. F., Borlee, I., Bouckaert, A., and Mission, C. (1980). Science, 207,
1296.
Lechtig, A.,Delghado, H., Lasky, R. E., Klein, R. E., Engle, R. L., Yarburgh,
C., and Habicht, J. P. (1975). A m . J. Dis. Child. 129, 434.
Lenaers, A., Ansay, M., Nusgens, B. V., and La Piere, C. M. (1971). Eur. J .
Biochem., 23, 533.
Lenz, W. (1961). Diskussionsbemerkung Tagung der Rhein Westfah Kin-
derartzeverinigung, Dusseldorf, 11th November.
Lenz, W. and Knapp, K. (1962). Deutsche, Med. Wochenschr., 7, 253.
Levin, D. L., Stanger, P., Kitterman, J. A., and Heyman, M. A. (1975).
Circulation, 52, 500.
Loevy, H. (1962). Anat. Rec., 142,375.
McBride, W. (1961). Lancet, 2, 1358.
McClain, R. M. and Becker, B. A. (1975). Toxicol. Appl. Pharmacol., 31, 72.
Marin-Padilla, M. (1966). J. Embryol. Exp. MorphoZ., 15, 261.
Mellin, G. N. (1964). A m . J. Obstet., 90, 1169.
Menkes, B. and Deleanu, M. (1964). Rev. Roum. Embryol. Cytol. Ser.
Embryol., 1, 69.
Mims, C. A. (1968) Prog. Med. Virol., 10, 194.
Moore, M. R., Hughes, M. A., and Goldberg, D. J. (1979). Znt. Arch. Occup.
Health, 44, 81.
Morris, G. M. (1972). J . Anat., 113, 241.
Naeye, R. L. and Blanc, W. (1965). J . A m . Med. ASSOC., 194, 1277.
Naeye, R. L., Blanc, W., and Paul, C. (1973). Pediatrician, 52, 492.
Nicholl, C. S . (1980). Fed. Proc., 39, 2563.
Nishimura, H., and Miyamoto, S . (1969). Acta Anat., 74, 121.
Overall, J. C. (1972). Am. Heart J., 84, 823.
Overall, J. C. and Glasgow, L. A. (1970). J . Pediat., 77, 315.
Palm, P. E., Arnold, E. P., Rachwell, R. C., Leyczek, J. C., Teague, K. W., and
Kensler, C. J. (1978). Toxicol. Appl. Pharmacol., 44, 1.
Palmer, A. K. (1974). Teratology., 10, 301.
Pinsky, L. and Di George, A. M. (1965). Science, 147, 203.
Pope, W. D. B., Halsey, M. J., Lansdown, A. B. G . , Simmons, A., and
Bateman, P. E. (1978). AnaesthesioZogy, 48, 11.
Pope, W. D. B., Halsey, M. J., Lansdown, A. B. G . , and Bateman, P. E. (1975).
Acta Anaesthesiol. Belg., 26, 169.
Poswillo, D. E. (1975), Br. Med. Bull., 31, 101.
Rabinowitz, M. B., Wetherill, G. W., and Kopple, J. D. (1976). J. Clin. Invest.,

58, 260.
Rest, J. R., (1987). Vet. Rec., 31, 426.
Ross, L. M. and Walker, B. (1967). A m . J. Anat., l21, 509.
St. Hilaire, E. G. (1832). In: ‘Histoire generale et particuliere des animales de
l’organisation chez l’homme et les animaux’ Eds J. B. Baliere et fils, Paris.
Saunders, J. W. and Fallon, J. F., (1966). In: ‘Major Problems in Developmental
Biology’, Ed. M. Locke, Academic Press, New York, p. 289.
Saxen, L. (1970). Znt. J. Gynec. Obstet., 8, 784.
Saxen, L. (1976). J. Embryol. Exp. Morphol., 36, 1.
Saxen, L. and Kaitilia, I. (1972). A d v . Exp. Med. Biol., 27, 205.
Scandinavian Council for Environmental Information (1977). ‘Workshop on the
utility of an evaluated data bank in teratology and its use for a state-of-the-
art report on adverse effects of lead on reproduction and developments’,
Nyashamn, Sweden.
Sever, J. L., Meier, G. W., Windle, W. F., Schiff, G. M., Monif, G. R., and
Fabiyi, A. (1966). J. Zrzfect. Dis., 116, 21.
Smith, R. L., and Caldwell, J. (1977). In: ‘Drug Metabolism from Microbe to
Man’, Edrs D. Parke and R. L. Smith, Taylor and Francis, London, p. 331.
Stadler, J., Kessidjan, M.-J., and Perraud, J. (1983). Food Cosmet. Toxicol., 21,
631.
Sternberg, S. S. (1979). Pharmacol. Therap., 6, 147.
Stockard, C. R. (1921). Am. J. Anat., 28, 115.
Swan, C. P., Tostevin, A. L., Moore, B., Mayo, H., and Black, G. H. B.,
(1943). Med. J. Austr., 2, 201.
Takeuchi, T. and Matsumoto, H. (1969). In: ‘Methods for Teratological Studies
in Experimental Animals and Man’, Eds H. Nishimura and J. R. Miller,
Igaku Shoin, Tokyo, p. 280.
Tuchmann-Duplessis, H. (1984). Pharmacol. Therap., 26, 273.
Ullberg, S. (1958). Proc. 2nd Int. Conf. on Peaceful Use of Atomic Energy,
Sweden, 24, 248.
Valdheim, C . M. (1985). Diabetes, 34, 21A.
Valdheim, C. M. (1986). N . Engl. J. Med., 315, 1314.
Wildschutt, H. I. J., Tutein Nolthenius-Puyaert, M. C . B. J. E., Wiedijk, V.,
Treffers, P. E., and Huber, J. (1987). Br. Med. J., 295, 894.
Wilson, J. G. (1959). J. Chronic Dis., 10, 11.
Wilson, J. G. (1964a). J. Cell. Comp. Physiol., 43, 11.
Wilson, J. G., (1964b). J . Pharmacol. Exp. Therap., 144, 429.
Wilson, J. G. (1972). Am. J . Anat., 136, 129.
Wilson, J. G. (1977). In: ‘Handbook of Teratology’., Eds J. G. Wilson and F. C.
Fraser, Plenum, New York, 1, p. 47.
Wolpert, L. (1976). Br. Med. Bull., 32, 65.
Yamazaki, J., Wright, S . , and Wright, P. (1954). Am. J. Dis. Child., 87, 448.
CHAPTER 13
Genetic Toxicology
DIANA ANDERSON
1 Introduction
Genetic toxicology is a branch of general toxicology which addresses the
problems of toxicity to the DNA. It is a discipline which has arisen from
the early studies of Muller (1927) and Auerbach (1947) in which
irradiation and chemicals were shown to induce mutation in Drosophila,
the fruit fly. When it was shown that mutations could be induced in
mammals, the possibility arose that some of the hereditary diseases in
man might be induced by environmental agents. Such materials include
not only man-made chemicals, but natural carcinogens and mutagens
found in fungi and plants and produced by cooking processes (Anderson
and Purchase, 1983; Felton and Knize, 1991; IPCS, 1992; Rowland et al.,
1984; Sugimura, 1978; Wakabayashi et al., 1993). The impact of these
agents on human health of present and future generations could be of
consequence if the extent of exposure is sufficient to permit expression of
genotoxic properites in somatic or germ cells. Over the last generation
many substances shown to be mutagenic were also shown to be
carcinogenic, and early correlations as high as 90% between mutagenicity
and carcinogenicity were claimed (McCann et al., 1975a; 1976; Purchase,
1978; Sugimura, 1977). Such correlations stimulated a new interest in the
somatic theory of cancer of the 1950s and has resulted in intensive study of
the genetic effects of chemical substances.
It has been estimated that there is a genetic element in at least 10% of
all human pathological conditions; thus genetic changes, if produced in
man, could be of serious consequence. Abnormalities could arise as a
result of either gene or chromosome mutation (Carter, 1977). McKusick
(1975) has segregated ‘traits’ inherited at the level of the gene as
dominant or recessive, and as autosomal or sex-linked. A dominant gene
is immediately expressed in the next generation (e.g. Huntingdon’s
chorea and achondroplasia). A recessive mutation may take many
generations to be expressed (e.g. phenylketonuria) except for sex-linked
recessives (e.g. haemophilia). Many constitutional and degenerative
243
244 Genetic Toxicology
diseases such as epilepsy or schizophrenia could be caused by ir-

regularities of gene expression or arise from multiple genes.
Chromosome abnormalities may arise either by errors in the distribu-
tion of chromosomes leading to abnormalities of chromosome numbers,
such as non-disjunction [Down’s syndrome (mongolism) and Klinefelter’s
and Turner’s syndromes], where the effect is seen in the next generation;
or chromosome breakage (ataxia telangiectasia and Fanconi’s anaemia)
which arise from recessive autosomal gene changes. Some chromosome
changes are thought to give rise to early embryonic loss and thus have
little genetic impact on society. Deleterious mutations that are com-
patible with survival may be a heavy burden to society if the affected
person requires medical or institutional care (mongolism, for example,
which occurs in 1 in 1000 births).
It is not possible to give a quantitative estimate of the contribution of
chemical agents to the incidence of genetic disease, but that they could
constitute an aetiological factor in such disease must be accepted. Hence,
in toxicological assessments, there is a need to define by the appropriate
tests the mutagenic activity of chemical substances.
This is a complex and difficult task, and no single method gives
conclusive information about genetic risk. A variety of test systems has
been put forward, and some governmental agencies have produced
guidelines for recommended methods (Department of Health and Social
Security (DHSS), UK, 1981; Department of Health (DOH),1989; Official
Journal of the EEC, 1979; and the OECD, 1983).
Structurally DNA, the target of mutagenic action, is identical in all
living organisms but there are differences in the organisation of the
genetic material in different species so that some organisms may be
considered to constitute better predictive models for man. There are
many interactions between chemicals and organisms which may deter-
mine whether genetic damage is expressed-unlike ionising irradiation,
which is immediately active in terms of mutagenic potential. Irradiation
generates extremely short-lived free radicals, some of which are close to
or within the genetic target molecules, whereas there are many factors
that influence whether a chemical compound may reach or react with
such targets. Among such factors are the chemical structure of the
compound, the duration of treatment, the route of administration,
absorption, distribution, and excretion, the macromolecular binding, the
metabolic transformation (which may all be dependent on species, strain,
and sex), the membrane barriers, and the presence of pertinent defence
mechanisms.
Other factors determine whether the damage is expressed as genetic
damage. Such factors include the innate susceptibility of the cell (i.e.
genetic repair capacity), the numbers of susceptible cells, the type of
genetic target, the selection processes involved, and the mode of
inheritance (Matter, 1976).
Many compounds that require testing may not bind covalently with
Diana Anderson 245
biologically important molecules unless they are first metabolised to

highly reactive forms. Others disturb other functions and cause a genetic
effect indirectly. Thus bases and nucleosides can themselves be mutage-
nic, an effect thought to arise as a result of unbalanced DNA precursor
pools (Anderson et al., 1981; Meuth, 1984; Clode et al., 1986; Clode and
Anderson et al., 1988a and b). Clearly the opportunities that theoretically
exist for a given compound to attack genetic material are widespread, as
are the obstacles and defences against such an attack being effective. As a
consequence, the methods available for assessing such effects are numer-
ous and in many respects very specialised.
2 Mutagenicity Tests
The mutagenic activity of chemical substances has been studied in a
number of test systems including micro-organisms? plants, insects,
mammalian, and human cells using in vitro and in vivo techniques. The
mutagenic effect detectable in one system may not be found in another or
even in different tissues of the same system.
The main sub-mammalian test systems currently available are those
which reflect the present state of knowledge and may be improved with
the progress of experimental work in this field. Not all of the available
tests are described here. Three categories are used: (a) tests for gene
mutations; (b) tests for chromosome damage including aneuploidy/non-
disjunction; (c) tests for DNA repair. The tests and cells and organisms
used in them are listed in Table 1.
A Gene Mutation Tests

In molecular terms, gene mutations consist of the substitution, deletion
or insertion of one or more nucleotide base-pairs in DNA. Mutations due
to substitution are known as base-pair substitutions and those due to
deletions or insertions as frameshifts. When a base-pair substitution
mutation occurs, a wrong base is inserted, which then pairs with its
natural partner during replication (adenine with thymine, cytosine with
guanine) so that a new pair of incorrect bases is inserted into the DNA.
When a frameshift mutation occurs, if there is base-pair loss, the
messenger DNA, which reads the DNA in triplet codons, incorporates
the first base-pair from the next triplet codon. Thus the subsequent code
of the DNA becomes scrambled until a ‘nonsense’ or terminating codon
is reached, A similar process happens with a base-pair gain.
Gene (or point) mutations cannot be detected by cytological methods
but can be distinguished as variants of characteristics controlled by
specific gene loci or as recessive lethal conditions (generally linked to sex
chromosomes). Mutations of specific sites can arise in either a ‘backward’
(from mutant to normal ‘wild’ type) or a ‘forward’ direction (from a wild
type to mutant).
Table 1 Tests, cells, and organisms used in mutagenicity tests

Mutagenicity Tests
I
Gene mut'ation tests

I
Chromosome damage tests DNA re'pair tests
1
Microbial tests
I
Microbial tests Microbial tests
Bacteria Nondisjunction in fungi Spot tests for
Salmonella typhimurium Chromosome loss inhibition zones
Histidine independence Sacchuromyces cerevisiae Salmonellu tyhimurium
(reverse mutation) Aspergillus niduluns BacilIlns subtilis
8-Azaguanine resistance Sordariu hreviocollis Escherichia coli
Arabinose resistance Succuromjws cerevis iuc.
(forward mutation) Plants Mitotic recombination
Escherichiu coli Vicia jaba Gene conversion
Tryptophan independence Allium cepa Sacchuromyces cerevisiue
(reverse mutation) Tradescantiu Schizosacchuromyces
5-Methyltryptophan Zea Mays pombe
Streptomycin Tricitum vulgare Unscheduled DNA
(forward mutation) Hordeum vulgure synthesis
Prophage reduction Lycopersium exulent um in mammalian cells
Fluctation test Nicotiania tobacum in vitro and in vivo
Pisum sutivum synthesis
Fungi Breakage of single chain
Sacchuromyces cerevisiae Insects DNA in individual cells
Sch izosacchuromyces pombe using gradient or elution
Neurospora crassu Drosophilu melanogaster techniques
Aspergillus niduluns Bombyx mori Single strand DNA breaks
and alkali labile damage
Insect test Mammalian cells in vitro individual cells using
Drosophila Chinese hamster ovary cells single cell gel electro-
Human fibroblasts phoresis (COMET assay)
Plants Human lymphocytes
Tradescantiu huir
Mammalian cells in vivo
Mammalian cells in vitro The rat bone marro metaphase
HPRT ase locus assay
Chinese hamster V79 The micronucleus test
Chinese hamster ovary Dominant lethal assay
Human lymphocytes The heritable translocation
Human fibroblasts test
TK ase locus Direct spermatocyte test
L5 178Y Sex chromosome losses
P388F The fertilised oocyte test
Mammalian cells in vivo

The spot test
The specific locus test
The dominant skeletal test
The cataract test
Others
Biochemically based tests
involving protein change,
histocompatability loci,
Pigment loci
Inversion techniques to
measure recessive lethals
Sperm morphology changes
Transgenic mice
Diana Anderson 247
Backward (reverse) mutation is generally studied in mutant cells for

which the base-pair substitutions or frameshift mutations are known.
Normal functional activity is re-established following a new substitution,
or a second deletion and insertion near the first one. Backward mutation
is highly specific in the type of DNA interaction required, and a chemical
compound that does not increase the frequency of this type of mutation
may nevertheless still cause other genetic effects.
Forward mutations may arise from effects which range from the
substitution, insertion, or deletion of the nucleotide bases of a gene, to
the deletion of the entire gene and neighbouring genes. They are
detected when there is a loss of an enzyme function (auxotrophy) or from
acquired resistance to various toxic chemicals, substrate analogues and
agents such as heat. In some instances where only one specific locus may
be affected, the forward mutation can be just as specific as the backward
mutation.
The specificity of some of these mutation systems, together with
secondary factors such as the effect of the particular nucleotide sequence
near the original mutation, makes a quantitative comparison or a
between-species extrapolation of the mutagenicity of different chemicals
very difficult. For a list of other factors that can influence effects in vitro,
see Ashby and Styles (1978) and Brusick (1987).
Bacterial Tests
Point mutations can be detected in various bacterial species. Those most
commonly used are Salmonella typhimurium and Escherichia coli.
In addition to its use in the reverse mutation assay of Ames, detecting
independence to histidine as a medium nutrient (Ames et al., 1973; 1975;
McCann et al., 1975a and b; Ashby and Tennant, 1988; McCann and
Ames, 1976; Maron and Ames, 1983; Tennant et al., 1987; Zeiger, 1987)
S. typhimurium can be used to detect forward mutations to 8-azaguanine
resistance (Skopek et al., 1978) and arabinose resistance (Peuyo, 1978).
The Ames test has been reviewed in the EPA Gene-Tox Programme (Kier
et al., 1986) and by Gatehouse et al., 1990. Strains most commonly used
are TA1535, TA1537, TA1538, TA97, TA98, TA100, and TA102.
E. coli has also been reviewed in the EPA Gene-Tox Programme
(Brusick et al., 1980) and can detect (a) the induction of mutations from
arginine dependence to independence (base-pair substitution reverse
mutation) (Mohn et al., 1974) or from tryptophan dependence to
independence (base-pair and frameshift reverse mutation) (Green and
Muriel, 1976; Venitt and Crofton-Sleigh, 1981; Venitt et al., 1983); (b)
the induction of forward mutations from inability to ability to ferment
galactose (Saedler et al., 1968) and from sensitivity to resistance to
5-methyltryptophan (Mohn, 1973) or streptomycin (Wild, 1973). These
forward systems allow the detection of base-pair substitutions, frameshift
mutations, or small deletions; (c) the induction of prophage (inductest),
which induces lysis of bacterial cells harbouring a latent bacteriophage,
through changes that lead to the destruction or deactivation of the phage

repressor. These effects can be observed in a single strain which carries
the necessary genetic markers and the bacteriophage in the prophage
form (Ho and Ho, 1981; Moreau et al., 1976; Speck et al., 1978); (d)
reverse fluctuation test (Green et al., 1976).
Tests using DNA-repair deficient bacterial strains have been reviewed
by Leifer et al. (1981) (Gene-Tox).
Fungal Tests
The yeasts Saccharomyces cereuisiae and Schizosaccharomyces pombe
and the Ascomycetes Neurospora crassa and Aspergillus nidulans are
among the fungi most used for the detection of point mutations.
In the yeasts Saccharomyces cereuisiae and Schizosaccharomyces
pombe, mutations at each of the two genetic loci that control adenine
biosynthesis cause a red pigmentation of the colonies. In Saccharomyces
cerevisiae, the effect has been used to develop forward mutation tests
either in haploid or diploid strains (Mortimer and Manney, 1971). In this
system it is possible to distinguish (by established morphological criteria)
the effect of the mutation from that of mitotic recombination (Zimmer-
man, 1973; 1975; Zimmerman et al., 1984, Gene-Tox).
Schizosaccharomyces pombe has been used to detect forward mutation
in wild or adenine-dependent haploid strains where mutations have been
introduced that increase sensitivity to chemical mutagens (Mortimer and
Manney, 1971). The use of these two yeasts for mutagenicity assays has
been extensively reviewed (Loprieno et al., 1974; Loprieno et al., 1983,
Gene-Tox; Zimmerman, 1973; 1975). A system of forward mutation from
canavanine sensitivity to resistance (Brusick, 1972) has also been
developed in Saccharomyces cereuisiae and the organism can also be used
in a fluctuation test, for example, the D-7 strain is used for detecting
isoleucine dependence (Parry, 1977a and b; Parry et al., 1984).
Strains of Neuruspora crassa can also detect the induction of forward
mutation in each of two genetic loci which control adenine biosynthesis
(Ong and de Serres, 1972). In addition, strains have been developed that
allow identification of recessive point mutations at each of the two
genetic loci mentioned, dominant lethal mutations in the genetic region
in which the two loci are situated and recessive mutations in the whole
genome (Brockman et al., 1984, Gene-Tox; de Serres and Malling, 1971;
Ong and de Serres, 1972). Strains of aspergillus can be used for the
induction of forward mutation to 8-azaguanine and p -fluorophenylalanine
and back mutation to methionine independence (Roper, 1971). Asper-
gillus haploid and diploid strains, respectively, have been reviewed in the
gene-tox programme (Kafer et al., 1982; Scott et al., 1982).
Insect Tests
The most widely used point mutation test in Drosophila melanogaster is
that which detects the induction of sex-linked recessive lethals. These
Diana Anderson 249
become evident in the second generation of treated individuals and can

be observed at much lower doses than those needed to induce chromo-
some loss or dominant lethal mutations in this system. The presence of
the X chromosome (which is about 20% of the genome) makes it possible
to classify most of the sex-linked recessive lethals as gene mutations or
multi-gene deletions. Autosomal recessive lethal mutations can be
observed in flies of second generation. The induction of visible mutations
is measured by crossing wild males (treated or control) with females that
are homozygous for the visible mutations. The function of Drosophila in
genetic toxicology testing has been described by Abrahamson and Lewis
(1971), Vogel and Sobels (1976) and others (Auerbach, 1977; Baker et
al., 1976; Bootman and Kilbey, 1983; Lee et al., 1983, Gene-Tox; Millet
and Weilenmann, 1976; Mollet and Wurgler, 1974; Valencia et al., 1984,
Gene-Tox; Vogel, 1977; Wurgler et al., 1977).
Mammalian Cell Tests In Vitro

There are various genetic systems for the detection of point mutations in
cultures of mammalian cells (Cole et al., 1990). They are based on the use of
phenotypic markers generally resulting from resistance to antimetabolites
or temperature sensitivity. In addition to gene mutation, variants of somatic
cells in culture can be generated by other effects such as gene suppres-
sion. Hamster, mouse, and human diploid fibroblasts are among the cell
lines most often used to detect genetic changes. For some markers many cell
lines may be used. For example, the locus controlling the hypoxanthine
guanine phosphoribosyl transferase (HPRT) enzyme occurs in
Chinese hamster embryo lung cells [V-79 cells] (Arlett, 1977; Bradley et
al., 1982, Gene-Tox; Chu et al., 1976; Cole et al., 1983; Fox et al., 1976;
Huberman and Sachs, 1974; Krahn and Heidelberg, 1977), Chinese
hamster ovary cells [CHO cells] (Hsie et al., 1981, Gene-Tox; Neil1 et al.,
1977) and primary cultures of lung cells (Dean and Senner, 1977), human
fibroblast (Jacob and Mars, 1977) and lymphoblast cells (Thilly et al.,
1976). The locus controlling adenine phosphoribosyl transferase (APRT)
is present in human fibroblasts (de Mars, 1974). The thymidine kinase
(TK) locus occurs in L5178Y (Clive, 1973;Clive et al., 1972;Clive etal., 1983,
Gene-Tox; Clive and Spector, 1975; Cole et al., 1983; Knaap and Simons,
1975;Nakamura et al., 1977) and P388F mouse lymphoma cells (Anderson,
1975a;Anderson and Cross, 1982;Anderson and Cross, 1985;Anderson and
Fox, 1974; Fox and Anderson, 1976).
A fluctuation test (Cole et al., 1976) and a host-mediated assay
(Fischer et al., 1974) have been developed using L5178Y cells, and other
cell lines from the mouse have been employed for mutagenesis assays
(Anderson, 1975b).
Tests for measuring cell transformation are also available (Heidelber-
ger et al., 1983, Gene-Tox).
Mammalian Cell Tests In Vivo

The ‘spot’ test (Fahrig, 1978; Russel et al., 1981b, Gene-Tox). This test
detects somatic mutations. Mice that are heterozygous for certain coat
colour genes are treated in utero by administration of the test compound
to the mother. If a mutation occurs during the development of the hair
follicles, this may be observed in the mouse when the coat is fully
developed. Thus, the female is treated during pregnancy and the
offspring grown until the coat colour has fully developed. Each mouse is
then examined for the coat markings indicative of a clone arising from a
single mutant cell. This system has the advantage of being one of the few
assays for gene mutations in mammals. The main disadvantages are that
the somatic mutations for coat colour may occur as a consequence, not of
gene mutations, but of chromosomal mutations, and the coloured spots
can be difficult to distinguish from the natural pigmentation zones of the
coat.
The specific locus test (Russel, 1951; Russel et al., 1981a, Gene-Tox;
Searle, 1975). This test in mice is at present the only one available that
can detect heritable mutations in mammalian germ cells. The test uses
mouse strains, homozygous for certain dominant wild-type alleles and
strains that are homozygous for recessive alleles at these loci. The
dominant, wild-type strains are treated and then mated with the recessive
strains. Changes induced at any of the loci cause visible changes in the
offspring, in respect of eye colour, ear size, and coat colour. Although it
is recognised that these events are gene mutations, their exact nature
cannot be identified in most instances and they are assumed to be small
multi-locus deletions. The main disadvantage of the specific locus test is
that the spontaneous rate for these loci is low. Hence, for a negative
result to be recorded, large numbers of offspring (up to 30,000) need to
be scored, making the test extremely costly and laborious.
Biochemically -based tests. There are a number of biochemical tests for
detecting gene mutations in somatic and germ cells in mammals. Enzyme
activity can be used as a genetic marker to detect induced microlesions in
mammals. They depend on variations in specific proteins caused by
mutations in the genes controlling the synthesis of the enzymes or
proteins (Mays et al., 1978; Neel, 1979; Neel et al., 1979; Valcovic and
Malling , 1973).
Other approaches to detection of mutants in mice include changes in
mandible shape (Festing and Wolfe, 1979), skeletal mutations (Ehling,
1970; Selby and Selby, 1977), histocompatibility loci (Kohn, 1973),
recessive lethals measured by the inversion technique (Evans and
Phillips, 1975; Roderick, 1979) and pigment loci (Searle, 1979).
Transgenic animals tests. Transgenic animals provide efficient systems for

detecting gene mutations in vivo. Such systems rely on the use of a
bacteriophage lambda shuttle vector which has been integrated into the
Diana Anderson 25 1
genome of an inbred mouse via microinjection so that as division occurs

every somatic and germ cell in the animal contains this gene which is
quickly recovered using standard laboratory techniques. Two systems
are commercially available-Mutamouse and the Big Blue mouse.
Mutamouse was created by incorporating DNA from the bacteriophage
lambda gtlO lacZ into the genome of the CD2 hybrid mouse. Each
Mutamouse cell contains approximately 40 copies of the transgene
arranged head-to-tail at a single insertion site. Mutamouse has become
homozygous for the transgene and contains approximately 80 lac Z sites
within every normal diploid cell and 40 per haploid gamete.
In the case of the Big Blue mouse the target organ is the Laclq gene in
C57B 16/6 mouse strain.
These Lac genes are susceptible to mutation like any other endogenous
genes in the cell. After the transgenic mouse is exposed to a test compound,
genomic DNA is isolated from the desired tissues such as bone marrow,
germ tissue, and liver. The integrated lambda vectors are then easily and
efficiently recovered by subjecting the genomic DNA to an in vitro
packaging extract (proprietorial) which excises the vector DNA and
packages it into a lambda phage head. Each packaged phage is then used
to infect E. coli cells. The packing extract is used to give high target gene
rescue efficiencies.
When E. coli infected with the lambda vectors are plated and incubated
(1 8 hr) on indicator agar dishes containing the chromagenic substrate
X-gal, mutations with the Laclq target gene cause the formation of blue
plaques, while unmutated or intact Laclq targets result in clear plaques. In
the case of the gtlO LacZ gene (Mutamouse) mutations are the opposite
from those in the Big Blue mouse, i.e. they form clear plaques while the
unmutated ones are blue.
Details of the transgenic systems have been described in several
publications (Kohler et al., 1990; Kohler et al., 1991a and b).
It has been found, for example, that for male germ cells the spontaneous
mutation frequency is lower in these cells than in somatic tissues (Kohler et
al., 1991a).
Transgenic rats containing the same lambda/Lacl shuttle vector have
been developed for inter-species comparison of mutagenesis testing
results, which may offer a better understanding of the specific mechanisms
involved in mutagenesis at the molecular level in vivo (Provost et al., 1993).
The Use of Metabolic Activation Systems

Microbial and mammalian cell systems in vitro generally do not possess
the metabolic capability of cells of the intact animal. Several supplemen-
tary systems have been developed for use with microbial and mammalian
test systems.
The most widely used systems for metabolic activation is the rat liver
S-9 fraction supplemented with the co-factors NADP and glucose-6-
phosphate (Ames et al., 1973; Ames et al. ? 1975; Maron and Ames, 1983;
Kier et al., 1986, Gene-Tox). The S-9 fraction is the supernatant fraction
of the liver homogenate obtained by centrifugation at 9000 g for 10 min.
It contains a preponderance of microsomes and the attendant oxidases.
Before the fraction is prepared, the animals are commonly treated with
inducing agents such as polychlorinated biphenyls (PCB) , phenobarbital
(PB) , or 3-methylcholanthrene. PCB can induce microsomal enzymes
capable of metabolising many different types of environmental mutagens
and carcinogens, but a combination of P-naphthoflavone and PB is to be
preferred (Matsushima et al. 1976).
?
Rat liver is the most widely used source but the livers from other
species, including man, have also been used. The use of human liver
might be important in evaluating the risk of environmental mutagens and
carcinogens to man (Bartsch et al., 1975; Tang and Friedman, 1977).
Several experimental procedures exist for microbial systems: (a) the
liquid method. The bacteria are incubated with the S-9 mix and the test
substance, washed, and the revertants and surviving bacteria counted; (b)
In the plate method (Ames et al., 1975; Maron and Ames, 1973), the
mixture of bacteria, S-9 mix and test substance is added directly to
molten-soft agar and poured onto hard agar. This method is relatively
simple and quick; (c) In a variant of this procedure (Nagao et al., 1978;
Sugimura et al., 1976) the mixture of bacteria, test substance and S-9 mix
is incubated for 20min at 37°C before adding molten-soft agar. With this
pre-incubation method, dimethylnitrosamine, a definite mutagen and
carcinogen, gives positive results (Sugimura et al., 1976) which are not
readily detected by other methods.
An alternative approach, where mammalian cells in culture constitute
the test system, utilises ‘feeder’ layers of mammalian cells (Huberman,
1975). These are usually of embryonic origin, with their enzyme system
fully intact but with cell replication inactivated by radiation or a high
dose of an antimitotic agent such as mitomycin C.
A method involving in vivo metabolic activation of a drug has been
reported (Legator et al., 1982, Gene-Tox; Legator and Malling, 1971;
Legator et al., 1977). The test substance is administered to animals by
gastric tube or subcutaneous injection and the test bacteria (indicator
organisms) are injected intraperitoneally, withdrawn after several hours,
and examined for mutation using an in vitro assay. This ‘host-mediated
assay’ has been used to investigate the mutagenicity of undefined or
unsterilised material thought to contain mutagens and carcinogens. The
test organism may also be given intravenously or into the testes. This
assay is relatively insensitive by comparison with the plate incorporation
assay. High doses of test compound have to be used to produce a
response but it is conceivable that this method gives a realistic measure of
the mutagenic impact in vivo. Clearly the antibacterial activity of the host
may be a confounding factor. Salmonella typhimurium (Legator and
Malling, 1971), Escherichia coli (Mohn and Ellenberger, 1973),
Diana Anderson 253
Saccharomyces cerevisiae (Fahrig , 1974; 1977), Schizosaccharomyces

pombe (Legator and Malling, 1971), Neurospora crassa (Legator and
Malling, 1971), and mouse lymphoma cells (Fischer et al., 1974) have all
been used as indicator organisms.
Chemical substances which have undergone metabolism are excreted in
the urine as degradation or conjugated products and as such could be
identified by suitable microbial indicators, after deconjugation (Com-
moner et al., 1974; Durston and Ames, 1974). Although of little value in
the detection of new mutagens, analysis of the mutagenic activity of the
urine or body fluids (Legator et al., 1982, Gene-Tox; Legator et al., 1976;
Siebert, 1973) or faeces (Bruce et al., 1977; Combes et al., 1984) might be
of value to screen workers with known exposures.
Metabolic activation systems have been described with tests for gene
mutation, but can also be applied with tests for chromosome damage and
DNA repair.
B Chromosome Damage Tests

Chromosome damage is defined as a modification of the number or
structure of chromosomes. It can be detected both by cytological and
genetic methods.
Variations in the number of chromosomes (aneuploidy and polyploidy)
may result from endoreduplication (continued chromosome division),
metaphase arrest, anaphase retardation, and non-disjunction in mitosis
and meiosis.
Structural changes (Figure 1) are mainly the result of breaks in the
chromatid arms. Depending on the number of breaks and the way in
which they may be rejoined a series of unstable structural modifications
may arise that are not transmitted to successive cellular generations.
These include ‘gaps’ (achromatic interruptions, but see Anderson and
Richardson, 1981), breaks of one or both the chromatids, chromatid
interchanges, acentric fragments, ring and dicentric chromosomes. Stable
structural modifications that are transmissible may also arise, such as
inversions, translocations, and deletions (Evans, 1976; Buckton and
Evans, 1973; Preston et al., 1981, Gene Tox; Scott et al., 1983; Scott et al.,
1990).
Microbial Tests
It is possible to measure non-disjunction in micro-organisms. Tests for
detecting mitotic non-disjunction have been developed using the fungus
Aspergillus nidulans (Bignami et al., 1974; Kafer et al., 1976; Kafer et al.,
1982, Gene-Tox; Scott et al., 1982, Gene-Tox) and the yeast Saccharomy-
ces cerevisiae (Parry et al., 1977b; Parry et al., 1984; Parry and
Zimmerman, 1976; Sora and Magni, 1988; Zimmerman, 1975); meiotic
non-disjunction can be measured in the fungus Sordaria brevicollis (Bond,
1976).
Chromatid gap Chromatid break Chromosome gap
Fragment Dicentric Ring Chromatid interchange
Translocation Inversion
Figure 1 Diagram of diferent categories of chromosome damage
Mitotic chromosome loss has also been examined in Saccharornyces

cerevisiae (D61M) (Zimmerman et al., 1985). The test relies upon the
recovery and expression of multiple recessive markers ieflecting the
presumptive loss of the chromosome VII homologue carrying the
corresponding wild type alleles. Deviations from the conventional aneup-
loid or diploid state often causes serious perturbations in normal growth
and development. Whittaker et al. (1989; 1990) have carried out an inter-
laboratory assessment of this assay. Albertini ( I 989) has examined
chromosomal mal-segregation and aneuploidy and Albertini et al. (1 993)
have examined various chemicals in the EEC aneuploidy programme.
Plant Tests
The radical apices (root tips) of Vicia faba, broad bean, (Kihlman, 1971;
Te-Hsiu Ma, 1982a, Gene-Tox) Allium cepa, common onion (Gran,
1982, Gene-Tox; Kihlman, 1971) and some species of the genus
Tradescantia (Marimuthu, 1970; Te-Hsiu Ma, 1982b, Gene-Tox; Van’t
Hof and Schairer, 1982, Gene-Tox) can be used for the detection of
chromosome aberrations, but such systems may not be suitable for
extrapolation to mammalian systems because different degrees of chrom-
osome damage have been recorded in onion root tips when compared
with Chinese hamster cells, due possibly to variation in the efficiency of
the repair system (Kihlman, 1971). Other plants, such as barley, offer
systems for the analysis of chromosomal aberrations (Constantin and
Diana Anderson 255
Nilan, 1982a and b, Gene-Tox; Ehrenberg, 1971; Nilan and Vig, 1976;
Plewa, 1982, Gene-Tox; Vig, 1982, Gene-Tox). The role of plants for
genetic and cytogenic screening has been reviewed by Constantin and
Owens (1982, Gene-Tox).
Insect Tests
Effects at the chromosome level such as non-disjunction, loss of the X
chromosome, deletions, translocations, dominant lethal mutations, and
mitotic and meiotic recombination have been observed in Drosophila
melanogaster (see earlier references, also Valencia et al., 1984, Gene-
Tox; Wurgler et al., 1977). Changes are generally shown by observing the
phenotypes of the progeny from the appropriate crosses. For example,
reciprocal translocations may be detected in the second generation of
treated individuals carrying defined recessive markers in the autosomes.
The absence of any of the expected phenotypes in the progeny indicates
that the translocations occurred in the parent reproductive cells. The loss
or acquisition of a chromosome, or the loss of part of a chromosome, can
be detected by observing phenotypes in the progeny of crosses in which
only one of the parents have been treated.
Mammalian Cell Tests In Vitro

Chromosome dumuge can easily be observed by cytological methods in
mammalian cells in culture (Preston et al., 1981, Gene-Tox; Scott et al.,
1983). Human fibroblasts, lymphocytes, and rodent cells (hamster
fibroblasts and mouse lymphocytes) are used. Translocations, inversions,
and other stable rearrangements indicate genetic damage that can be
inherited. Gaps probably indicate early toxic effects unless accompanied
by other types of damage (Anderson and Richardson, 1981) but have in
fact persisted for three months when other categories of damage have
been eliminated after vinyl chloride treatment.
Sister Chromatid exchange. Sister chromatid exchange in somatic mam-
malian cells is a measure of the reciprocal exchange in segments of
homologous loci between sister chromatids, i.e. the number of crossovers
during replication that occur between paired chromatids following
treatment with the substance (Latt et al., 1981, Gene-Tox; Perry and
Evans, 1975; Perry et al., 1984; Stetka and Wolff, 1976). Cells are
exposed to 5-bromodeoxyuridine for one round of replication. A second
round without it follows, so that one arm of the chromosome is a
substituted chromatid (these are recognised by differential staining
techniques) and the frequency of such chromatid exchanges is increased
by mutagens. It is considered to be a sensitive method for detecting
chromosome damage but the significance of the test is not yet fully
understood. It is considered as a test that indirectly measures DNA
damage which involves repair.

Methods exist for the assay of chromosomal aberrations in somatic cells
after the administration of a test chemical to the intact animal (Albanese
et al., 1984; Preston et al., 1981, Gene-Tox; Richold et al., 1990; Topham
et al., 1983). Peripheral lymphocytes (Lilly et al., 1975) and bone marrow
cells (Cohen and Hirschorn, 1971; Legator et al., 1973; Nichols et al., 1973;
Richold et al., 1990; Schmid et al., 1971; Tijo and Whang, 1962) are
commonly used. Bone marrow cells are a naturally proliferating cell
population but peripheral lymphocytes are a synchronised population of
Gocells that rarely proliferate and have to be stimulated by mitogens for
analysis. Such systems have large numbers of cells available and show
good potential for analysis of human cells, provided solid baseline data are
available (Anderson et al., 1988).
The micronucleus test. In anaphase chromosomes, abnormal chromosome
or chromatid fragments lag behind the other migrating chromosomes,
and deformed anaphase bridges occur. At telophase, these lagging
elements form micronuclei (Heddle, 1973; Heddle et al., 1983, Gene-
Tox; Maier and Schmid, 1976; Salamone et al., 1980; Schmid, 1976)
because they do not become incorporated into the nuclei of the daughter
cells. Micronuclei are most often observed in polychromatic erythrocytes
but they can be observed in embryonic and other cells in culture by
suitable staining techniques. The test is also of value in the detection of
mitotic spindle poisons. There has been debate about the number of cells
to be analysed, up to 2000 recommended for suitable sensitivity (Ashby
and Mirkova, 1987).
Dominant lethal test. Basically, this assay depends on an increase in the
foetal abortion rate due to defective sperm or ova, measured as a
decrease in the numbers of uterine implants or an increase in early foetal
death, i. e. pre- and post-implantation loss, respectively (Anderson et al.,
1983; Bateman, 1977; Bateman and Epstein, 1971; Ehling et al., 1978;
Epstein and Rohrborn, 1970; Green et al., 1985, Gene-Tox). An effect
on fertility may be measured in the same study. Pre-implantation losses
can occur due to other than genetic reasons (including lack of fertility)
but the early implantation death is thought to be due to chromosomal
abnormalities produced in germ cells. In order to sample all stages of
sensitivity of the mating germ cell, an 8-week mating period is required in
the mouse and a 10-week period in the rat (Epstein and Rohrborn, 1970),
these being the times taken to produce mature sperm from the germinal
cells.
The assay may be extended by examining surviving offspring for
congenital malformations (Anderson et al., 1993; Frances et al., 1990;
Jenkinson et al., 1987; Jenkinson et al., 1990; Knudsen et al., 1977).
The dominant lethal assay can also be carried out using Drosophila
(Wurgler et al., 1977) (see earlier references).
Diana Anderson 257
Sperm -head morphology. Some mutagens give rise to abnormally shaped

spermatocytes (Topham, 1980b; Wyrobek and Bruce, 1975) and this can
result in inherited sperm abnormalities after parental treatment in
animals (Topham, 1980b; Wyrobek et al., 1983a, Gene-Tox). An
evaluation has been made of human sperm as indicators of chemically-
induced alterations of spermatogenic function (Wyrobek et al., 1983b,
Gene-Tox; Wyrobek et al., 1984).
Heritable translocation test. The presence of a heritable translocation (i.e.
a balanced chromosome translocation) is thought to confer sterility or
partial sterility on the F1 offspring. Its presence can be confirmed either
cytogenetically or by mating on several occasions suspect males who will
continue to produce litters of reduced numbers (Cachiero et al., 1974;
Generoso et al., 1977, Gene-Tox; Generoso et al., 1980).
Direct spermatocyte test. Spermatogonia are examined at or near meta-
phase l for the abnormal mitotic figures that result from reciprocal
translocation (Leonard, 1975; 1977).
Sex chromosome losses. The test also measures the heritability of
chromosomal damage by scoring the progeny of exposed animals for
infertility (Russel, 1979).
The fertilised oocyte test. The first cleavage embryos from treated male
and female germ cells are scored for structural chromosome damage. As
the male and female pronuclei condense at different times, both genomes
can be examined. Chromosome aberrations in the male pronuclei have
been shown to correlate with dominant lethality (Brewen et al. 1975;
Albanese, 1987).
The various germ cell methods have been reviewed by Albenese
(1987).
C DNA Damage and Repair Tests

When DNA is damaged, repair normally follows (Cleaver, 1975; 1977;
Larsen et al., 1982, Gene-Tox). The initial lesions in DNA may be lethal,
may remain without being repaired, may be repaired correctly to restore
a normal genome, or incorrectly to produce errors and an abnormal
genome. There are several types of repair possibilities.
Excision repair. Cut and patch or pre-replication repair occurs when a
lesion, recognised by an endonuclease, is excised by an exonuclease and
the missing part re-synthesised by a polymerase to reconstitute the
original strand. The new part is joined to existing DNA by a ligase.
Excision repair is normally error-free and is known to occur in human
cells. Hydroxyurea has been used to inhibit normal DNA replication,
which then allows the detection of excision repair, but it is now suspected
that this compound itself may have some effect on excision repair so that
the results from experiments using hydroxyurea should be considered

with caution (Pearson and Styles, 1984).
Post-replication (by-pass) repair. The lesion is by-passed in newly
synthesised daughter DNA, thus leaving a gap that is sealed by insertion
of a DNA segment by recombination into the new daughter DNA. This
process is error-prone. The post-replication gap may be filled by DNA
synthesised de nouo and thus correct errors copied by replication.
Recombination repair is a post-replication repair process in bacteria
which involves the recombination of daughter strands of DNA to
reconstruct the correct genome. This process is error-free, but error can
occur if the repair requires de nouo synthesis. This has been established
in bacteria but has not been demonstrated conclusively in mammalian
cells.
Damage to the DNA molecule may be considered as a primary lesion
that could be involved in the process of mutation and the extent of the
repair is an indicator of the amount of damage that has occurred to
DNA.
Several methods exist for detecting DNA repair phenomena. These
include differential zones of inhibition or killing in bacterial strains with
and without repair processes (Ichinotsubo et al., 1977; Leifer et al., 1981,
Gene-Tox; Tanooka, 1977) gene conversion in strains of yeast (Zimmer-
man, 1973; Zimmerman et al., 1984, Gene-Tox), sister chromatid ex-
change in mammalian cells (Latt et al., 1981, Gene-Tox; Perry and
Evans, 1975; Perry et al., 1984; Steka and Wolff, 1976) and the direct
measurement of DNA damage and repair (Cleaver, 1975; 1977; Larsen et
al., 1982, Gene-Tox).
Bacterial Tests
‘Spot’ tests measuring diflerential zones of inhibition (Tweats et al., 1984),
A Petri dish is seeded or streaked with the test organism (Salmonella
typhimurium) (Ames et al., 1975; Leifer et al., 1981, Gene-Tox) , Bacillus
subtilis (Tanooka, 1977), Escherichia coli (Leifer et al., 1981, Gene-Tox;
Sugimura et al., 1977), Saccharomyces cereuisiae (Zimmerman, 1975;
Zimmerman et al., 1984 Gene-Tox) or Aspergillus nidulans (Kafer et al.,
1976; Kafer et al., 1982, Gene-Tox; Roper, 1971). The test compound is
placed in the dish and the inhibition zone or lethal effect produced by the
compound is evaluated in two different strains of the test organism, one
being the wild-type strain and one being deficient in a DNA repair system
(e.g. pol A - in B. subtilis and rec A - and uvr- in E. coli). When a
greater zone of inhibition is produced in the repair-deficient strain than in
the wild-type strain, the compound is considered to be capable of
affecting DNA. The assay can be carried out with and without metabolic
activation (S-9 mix) incorporated in the agar. If minimal medium is used,
both mutation and inhibition zones can be detected.
Diana Anderson 259
Fungal Tests Measuring Mitotic Recombination or Gene

Conversion
In eukaryotes, it is possible to measure an increase in the frequency of
mitotic recombination or gene conversion when a recessive phenotype is
expressed in the transition from a heterozygote to a homozygote
situation. These tests are thought to be related to the exchange following
breakage of the chromatids of two homologous chromosomes. Such
changes may allow for the expression of recessive mutation, as it is
known that meiotic recombination allows the expression of recessive gene
mutations in man.
Such tests are carried out in the yeast Saccharomyces cereuisiae and
Schizosaccharomyces pombe. A fluctuation test can also be used to detect
mitotic-gene conversion (Zimmerman et al., 1984). Other fungi such as
Aspergillus nidulans have also been used (Kafer et al., 1982, Gene-Tox).

Tests to measure sister chromatid exchange in mammalian cells have
already been described.
Tests to measure D N A repair by synthesis. DNA damage can be detected
by determining unscheduled DNA synthesis which occurs as a result of
DNA excision repair. One of the techniques reveals repair synthesis by
determining radioactivity (tritiated thymidine) incorporated into DNA
during the repair process (Mitchell et al., 1983, Gene-Tox; Waters et al.,
1984). Tests to measure unscheduled synthesis in liver cells in vivo are now
suggested if a negative result has been obtained in the bone marrow assay
[see DOH 1989 Guidelines (Figure 2)].
The radioactivity can be measured either by autoradiography or by
direct counting of the incorporated thymidine by liquid scintillation.
Metabolic activation can also be included with such systems.
Another technique measures the breakage of the single D N A chain in
alkali, using gradient or elution techniques. The gradient technique is the
most sensitive but alkaline elution is simple and faster, measuring the
rate of elution of DNA through a filter, a function of the relative
molecular mass of the DNA (Cleaver, 1975; 1977).
The COMET or single cell gel electrophoresis assay has more recently
become established as a useful technique for detecting DNA damage (e.g.
Green and Lowe, 1992). It detects single strand breaks and alkali labile
damage in individual cells. Cells are electrophoresed under alkaline
conditions. Stained nuclei with increased DNA damage display increased
migration of single stranded DNA towards the anode. The length of
migration of the tail can be measured with a graticule. Density profiles can
also be measured using image analysis. This method has been reviewed by
McKelvey-Martin et al. (1993).
STAGE 1 In Vitro Tests

Initial Screening (a) Bacterial assay for gene mutation
Two tests required (a + b) except
where human exposure would be (b) Test for clastogenicity in mammalian
expected to be extensive and/or cells, for example metaphase analysis
sustained, and, difficult to avoid,
when all three tests are necessary (c) Test for gene mutation in mammalian
cells (for example the L5178Y
T K + / - assay)
STAGE 2 In Vivo Tests
Tests for: (a) Bone marrow assay for chromosome
Compounds positive in one or more damage (metaphase analysis or
tests in Stage 1, micronucleus test)
and
All compounds where high, or Plus, if above negative, and any in vitro
moderate, prolonged levels of human test positive
exposure are anticipated
(b) Test(s) to examine whether
mutagenicity or evidence of DNA
damage can be demonstrated in other
organs (e.g. liver, gut, etc.)
STAGE 3 In Vivo Tests for Germ Cell Eflects
If risk assessent for germ cell effects is (a) Tests to show interaction with DNA
justified (on basis of properties
including pharmacokinetics, use, and Dominant lethal assay (most useful)
anticipated exposure). Cytogenetics in spermatogonia
One cell embryo test
(b) Tests to show potential for inherited
effects
Dominant lethal assay gives indication
of likelihood of inherited effects
Cytogenetics in spermatocytes for
reciprocal translocations
Non-disjunction in the mouse (10-day
embryo)
(c) Test for quantitative assessment of
heritable effects
Ouantitative studies need strong Mouse heritable translocation test
jistification in view of their complexity, Mouse specific locus test
long duration, costs, and use of large
numbers of animals
* General guidance only is given in this flow diagram. Decisions regarding, for example,
whether a specific compound is expected to produce high or moderate, but prolonged,
exposure would normally be taken by Regulatory Authorities having regard to other
relevant data, and on a case-by-case basis.
There may be instances where alternative tests to those specified might be more appropriate,
and it is important that a flexible approach is adopted. Each compound should be
considered on a case-by-case basis with regard to the selection of tests as well as the
interpretation of results.
Figure 2 Flow diagram f o r testing strategy for investigating mutagenic properties
of substance
(After Department of Health Mutagenicity Guidelines, HMSO, 1989)
Diana Anderson 261
Conclusion
Not all of these systems described have been equally well-studied and
some are used more for specialised study than screening chemicals. The
most widely used screening system today is the Ames test. Its use is
widespread because of its extensive validation and the assertion of its
potential to detect genotoxic carcinogens.
3 Usefulness of Some of the Systems for

Screening Purposes
Microbial assays. The major advantage of the microbial methods is that
they are rapid, inexpensive and relatively simple to carry out for the
experienced scientist, though not for the novice.
A Bacterial Assays
The plate incorporation test of Salmonella typhimurium (the Ames test)
is the test used routinely for screening purposes. It was shown in early
blind trials to have correlations with known carcinogens and non-
carcinogens as high as 90% and also to detect carcinogen and non-
carcinogen pairs equally well (Purchase et al., 1978). It is a test which can
be carried out by many laboratories (de Serres and Ashby, 1981; Ashby
et al., 1985). It has a stable phenotype which is demonstrated by its lack
of genetic drift (Anderson et al., 1984).
Correlations with known carcinogens have been much lower in recent
years (Zeiger, 1987; Tennant et al., 1987) but this is primarily because
non-genotoxic carcinogens have been included in the testing programme.
The assay still has high predictivities for genotoxic carcinogens (Ashby
and Tennant, 1988). This will be discussed in more detail later-see
Prediction of Human Carcinogens.
The liquid incubation or the pre-incubation method has been used for
detecting those mutagens difficult to determine in the plate incorporation
assay. The fluctuation assay has been suggested as being more sensitive
than the plate incorporation assay, in that it can detect compounds at
lower dose levels of the test compound. A forward mutation assay is also
available for Salmonella but these deviations from the standard Ames test
method have not been validated. The repair-deficient/proficient micro-
bial tests and spot tests, where a ‘spot’ of the test compound is placed in
the centre of the bacterial plate, are really only suitable as pre-screens.
The Ames test can be completed in about three days.
B Yeast Assays
Both Saccharomyces cereuisiae and Schizosaccharomyces pombe are
suitable for use in routine screening assays. Strains can be cultivated in
both the diploid and haploid phases, which allows for the detection of a
wide range of mutation events. Yeasts can be used to detect both point
mutations and DNA repair events in terms of mitotic recombination as
evaluated by crossing-over and gene conversion. A disadvantage is that
the chromosomes are too small for direct cytological observation but
chromosome damage can be measured by tests for non-disjunction. Yeast
systems take a few days longer than bacterial systems for colony growth
but the overall time scale involved is not greatly different. They tend to be
used as supplementary assays.
C Plant Assays
Plant systems can detect most types of damage. They have short
generation times and the cost, handling, and space requirements are
relatively small; genetics of seeds can be investigated under a wide range
of environmental conditions such as pH, water content, and temperature,
and chromosomal organisation is similar to the human system (Nilan and
Vig, 1976). Difficulty is experienced, however, in extrapolating the
results to mammalian systems, including man. Several agents, such as
cytosine arabinoside, daunomycin, and adriamycin, are known to be
ineffective on the plant genomes and yet cause severe genetic damage to
mammalian cells. This may be because the cell wall inhibits absorption
and because of greater ability to repair DNA lesions. Such systems are
probably not satisfactory for predicting potential human mutagens. They
are used as supplementary assays but may be useful for testing chemicals
which are sprayed on plants such as pesticides.
D Insect Assays
Drosophila has a short generation time of 10-12 days and is cheap and
easy to breed in large numbers with relatively simple facilities. Extensive
studies on the metabolism of insecticides performed over 15 years
(Wurgler, 1977) have revealed that insect microsomes are capable of
similar enzymic activity to those of the mammalian liver, but insects do not
have any specific organ in which the enzymes are predominantly located.
In Drosophila, mutagenic activity can be tested at different germ cell
stages which is important where mutagens have specificity of action.
Drosophila permits the scoring for the whole spectrum of genetic effects.
The observation that the lowest effective concentration (LEC) values
(and therefore the highest mutagenic effectiveness) have been recorded
for recessive lethals indicates the superior discriminating power of this
test. (The X chromosome represents a fifth of the whole genome.) By
comparison, the test for dominant lethality is of limited value. High doses
are required, and dominant lethals sometimes fail to arise when agents
cause the induction of recessive lethals. Changes in hatchability some-
times produce false-positives. However, Drosophila is a good ‘catch-all’
Diana Anderson 263
system owing to the variety of genetic and end-points that can be

detected.
Vogel (1987) points out that, based on Gene-Tox Report data and two
international collaborative trials, the sex-linked recessive lethal test does
not have a high ability to detect carcinogens when genotoxins other than
direct acting and simple pro-mutagens are included. However, it has a
high predictability for non-carcinogens. The tests detecting somatic
mutation/mitotic recombination (SMART) have higher predictibilities
than the sex-linked recessive lethal assays. Drosophila assays, although
they were included in some regulatory guidelines, e.g. DHSS, 1981,are not
currently included in those of DOH, 1989 (Figure 2).
E Mammalian Cell Assays In Vitro

Mammalian cell systems are generally considered more valid than
non-mammalian systems in terms of extrapolation to man, because
mammalian DNA is more similar to that of man. However, the use of
mammalian cell mutation assays is currently in debate. Few chemicals are
known which are detected exclusively in mammalian cell mutation assays,
and are not detected by the similar (in terms of cost) mammalian cell in
vitro cytogenetic assays. For this reason the use of mammalian cell
mutation assays is questioned. In addition, the V79 and the CHO cell
mutation systems, whilst suitable for detecting a positive response are
basically inadequate in terms of cell numbers for detecting a negative
response. By increasing cell numbers, whilst the mutant fraction remains
constant , there would be sufficient cell numbers available for statistical
purposes. Suitable cell numbers are only available at present in the cell
suspension assay of mouse lymphoma-the L5178Y system. The L5178Y
cell system of Clive (Clive et al., 1983, Gene-Tox) detects both large and
small colonies. The large are thought to arise from gene mutation and the
small from chromosome damage. It has been suggested that this system
may be too sensitive and lack discriminating power for the detection of
carcinogens and non-carcinogens. The thymidine kinase (TK) and
hypoxanthine guanine phosphoriboxyl transferase (HPRT) loci are
commonly used in mammalian cell assays but the latter, which is used in
V79 and CHO cells, is less sensitive than the TK locus (McGregor,
personal communication). If mutation cell systems are to be used,
however, cells grown in suspension are easier to handle. They do not
require trypsinisation, are easily sub-cultured and are not subject to
metabolic co-operation, so do not suffer from the reduced sensitivity that
occurs through metabolic co-operation when mutated cells are in close
contact with non-mutated cells.
[Current DOH guidelines (1989) suggest that chemicals giving negative
responses in three in vitro systems (bacterial mutation, chromosomal and
cellular point mutation assays) with and without metabolic activation d o
not require further testing in animal systems.]
F Mammalian Assays In Vivo

The advantage to in vivo studies is that the test compound is metabolised
in the animal. Both the bone marrow metaphase and micronucleus assays
thus have this advantage. Positive results in such assays indicate that the
test chemical is a mutagen to somatic mammalian cells in vivo. Results
correlate well with carcinogenicity studies, particularly human car-
cinogens (Shelby, 1988).
The dominant lethal assay in rodents has been claimed to be insensitive
to detecting chemicals, but the lack of sensitivity may reflect the real
situation because of the so-called ‘testes barrier’ formed by the Sertoli cells
surrounding the germ cells. In addition to the pharmacokinetic hurdles
and organ specificity, etc., and the short half-life of the chemical, the
blood-testes barrier may be important. Even if the compound reaches the
testes, it may not be metabolised. With dimethylnitrosamine, for example,
there is less alkylation of the DNA in the testes than in any other organ
(Swann and Magee, 1968) and the compound gives a negative result in the
dominant lethal assay (Propping et al., 1972).
These considerations, of course, also affect the other assays concerned
with the germ cells, such as the specific locus and heritable translocation
assays. The drawbacks to the latter two assays are that they require vast
numbers of animals in order to detect a response and thus are very costly
in terms of resources and time. However, the mammalian in vivo assays
are required for determining genetic hazard and risk estimates. The
dominant lethal assay although not useful for predicting carcinogens
(Green et al., 1985, Gene-Tox) could be useful for predicting heritable
hazard and the specific locus and heritable translocation assays are useful
for quantitative risk assessment.
Kirkland (1987) discusses the implications of germ cell cytogenetic tests
in the regulatory process. One of the responses to a questionnaire sent to
regulators was that germ cell tests are rarely requested as a matter of
course. Where germ cell tests are indicated, a dominant lethal test (most
often) or a heritable translocation test (sometimes) would be seen as
helpful in elucidating germ cell effects. The mouse specific locus test is
rarely requested due to the large number of animals involved and the
small number of laboratories with the relevant experience.
The current DOH guidelines (1989) (Figure 2) recommend germ cell
assays only for risk assessment as the last tier. In previous guidelines
(DHSS, 1981) germ cell assays were included in a battery of assays.
G DNA Repair Assays

Measurements of excision repair are determined as mean values for a
population of cells, whereas mutation is a rare event in individual cells.
The amount of excision repair after exposure to an agent will depend on
several factors, such as the extent of reaction with the DNA (the total
Diana Anderson 265
number of damaged sites), the number of sites that can be excision

repaired, the size of the repaired regions, the kinetics of excision repair
as functions of time and dose, and the extent to which chemical
interactions modify other sites, and possibly inhibit excision repair
(Cleaver, 1975; 1977). The amount of excision repair will therefore be
greatest for mutagens that induce the greatest proportion of extensively
damaged sites requiring repair by the large substitution. The number of
mutations depends on the severity of pre- and post-replication damage.
Studies of the relationship between DNA damage, excision repair,
post-replication repair, and mutagenesis must take account of the
numbers and varieties of lesions involved in mutagenesis and the modes
of repair. Exclusion reliance on any one measurement is useless. It is best
to consider DNA repair, for example, only in conjunction with some
other parameters before assessing the possible mutagenicity of an agent.
However, the COMET assay is proving useful for the rapid measurement
of genetic damage (McKelvey-Martin, 1993).
4 Cell Transformation Assays

These assays are thought to ‘bridge the gap’ between mutagenicity and
carcinogenicity. Cell transformation has been defined as the induction in
cultured cells of certain phenotypic alterations that are related to
neoplasia (Barrett et al., 1986). There are several types of endpoints for cell
transformation, including loss of anchorage dependence and alterations in
morphology, viral dependence, and altered growth in agar (McGregor
and Ashby, 1985). Morphological transformation and altered growth
have been used most extensively. These systems are the Syrian hamster
(SHE) assay (Barrett and Lamb, 1985; Berewald and Sachs, 1963; DiPaolo
et al., 1969a and b; 1971; 1972; Huberman and Sachs, 1966) and the
mouse C3H/IOT1/2 and mouse BALB/c3T3 assays (Heidelberger et al.,
1983; Kakunaga, 1973; Reznikoff et al., 1973). Of these, the SHE cell
transformation assay is unique in that it uses normal diploid cells (Berwald
and Sachs, 1963; 1965). The other two systems are based on established
cell lines which have undergone some adaptive changes in culture,
resulting in an aneuploid karyotype and a potentially preneoplastic
phenotype. A recommended protocol for all three assays based on a
survey of current practice has been suggested (Dunkel et al., 1991). SHE
cells have a limited lifespan in culture and rarely become tumorigenic
unless treated with carcinogens (Barrett et al., 1977).Therefore the cellular
events underlying the morphological transformation of SHE cells might
be indicative of earlier neoplastic changes compared with those underlying
cell transformation observed in the other cell lines. The acquisition of a
fully neoplastic phenotype in these cells in a multistep process analogous
to that in vivo and this system is therefore particularly useful for studying
the cellular and molecular events involved in neoplastic development
(Barrett et al., 1986; Koi and Barrett, 1986). Fitzgerald and Yarnasaki
(1990) have addressed the issue of tumour promotion describing models

and assay systems.
In the SHE transformation assay morphologically transformed colo-
nies are identified by a disorientated pattern of piled up cells (Berwald and
Sachs, 1965). The cells have a considerable range of metabolic activities
but chemicals requiring further metabolic activation can be studied with
incorporation of appropriate sub-cellular fractions or a second cell type.
5 Test significance and Interpretation

The available sub-mammalian test systems, used without mammalian in
vivo studies, would not be acceptable by any governmental authority for
estimating risk to man. They do provide a useful tool for a preliminary
screening of possible human mutagens. A positive result in a well
constructed and validated system is generally regarded as a warning sign.
However, when considering the many and diverse chemicals which are of
unknown mutagenic potential which require metabolic activation and
which may react differently with different cells and organs before
producing genotoxic effects, it is not surprising that many give equivocal
data that cannot be resolved from experience gained from classical
studies.
By comparison, the handling of positive data is much more clear-cut,
but it is desirable that dose-response curves should be established in
routine testing and results should be reproducible. Difficulties may arise
when mutagens have a strong killing effect so that a genetic effect is
obviously not as readily detected. This can be exemplified by mammalian
cell mutation assay systems where the activity of a chemical can produce
an absolute increase in the number of mutants per treated cell or an
absolute decrease if there is a strong killing effect. In the former case the
rate of increase of mutants (over the spontaneous) is then greater than
the rate of inactivation, per unit dose; in the latter, this is not so. Weak
mutagens are more difficult to evaluate when there is an increase only in
mutants per surviving cell (that is, after correction for survival) and not
per treated cell. The apparently weak positive effect could be due to the
induction of new mutants but may be due to a greater resistance of
spontaneous mutants to inactivation by the agent used. Reconstruction
experiments (where known numbers of mutant and wild-type cells are
mixed together) can solve the former problem, but to the latter there is
no good solution.
Dose-response curves may have linear or diphasic shape, or may be
diphasic at high dosage with a linear function at low doses. Thresholds
may exist at some chemical concentrations when the chemical is without
effect below a certain concentration. At higher doses, dose-response
curves tend to flatten or plateau or decline when the killing effect
overrides the mutagenic effect. When a dose-response relationship is
Diana Anderson 267
established it is easier to reach a conclusion regarding the mutagenicity of

a chemical.
In in uitro mammalian cell systems, particular emphasis should be
placed on using doses of chemicals which are not too high and so do not
alter the pH or affect the osmolality of the test system. The problems
arising in such instances, where false positives can be generated, have
been highlighted in a special issue of Mutation Research (Brusick, 1987).
6 Strategies for the Protection of Man

Auerbach (1975) stated that the procedures for the estimation of possible
hazards from genotoxins are full of uncertainties. This is still the case. It
is, nevertheless, important to attempt to develop approaches that will
allow risk estimation in men exposed to potentially genotoxic agents. The
first approach is to use a group of short term tests to detect possible
mutagenic activity while recognising that such an approach is subject to
limitations in terms of test variability, species sensitivity, and interspecies
extrapolation (but this is true for toxicological tests in general).
The second is to measure induced genetic damage directly in man.
The third is to use results from an evaluation of effects in the gonads
which can be combined with the pattern of expected human exposure
from which a judgement as to the amount of risk can be estimated. This
can be attempted by combining human exposure data with (a) the known
dose-response observed in animal studies, or (b) measured target/germ
cell concentrations combined with mutagenic responses defined by the
best understood in vitro assay systems. Concepts such as radiation-
equivalents and doubling doses may be of value in this approach.
A Approach 1-Use of Short Term Test Battery

Prediction of Human Mutagens
Developments in short term test procedures for genotoxicity have
generally focused on their ability or inability to identify potential
carcinogens. However, in regulatory practice in the UK and some other
European countries, such tests are conducted to determine mutagenicity
per se with a view to identifying potential human mutagens.
There are no examples of induced mutations in man with proven
causality. Cigarette smoke, vinyl chloride, lead, and anaesthetic gases are
the agents which are best documented but no effects in man are yet
unequivocally determined. Thus a validated mutational assay for detect-
ing potential human mutagens is not attainable, although attempts are
being made to address this issue.
Before this can be achieved, there is a need to examine the data on the
performance of the test systems and to establish how consistent these data
are when derived from different sources.
Within test variation. An assessment of the variation that is obtained

within a test is most easily achieved by comparing the performance of tests
which independently assay the same chemical. Over a decade ago, three
such studies were completed (de Serres and Ashby, 1981; Dunkel, 1979;
Poirier and de Serres, 1979). All showed errors of about 10% in detecting
positive effects and about the same in detecting negative effects. It is worth
considering the consequences of a 10% discrepancy when 6 test systems
are used. Thus if a single test is used, 90% of the mutagens will be correctly
identified. If it is assumed that each one of the six tests provides an
independent assessment of the mutagenicity of the chemical, and that the
error rate of each test for both positive and negative results is lo%, six tests
will identify 99.999% of the mutagens. At the same time, however, only
56% of the non-mutagens will be negative in all six test systems.
Between test variation. When considering test systems which, although
they may assess the same genetic end-point, use different organisms, the
range of variability is wide. One study (Poirier and de Serres, 1979) using
three assay systems found agreement for fifty-five chemicals (twenty-five
all negative and thirty all positive) and disagreement for forty-four.
Similar findings occurred in an international study (de Serres and Ashby,
1981, as shown in Table 2).
The use of such schemes for regulatory purposes was much discussed
(Brusick, 1982; Purchase, 1980) but as yet has still not been satisfactorily
resolved. The reproducibility of the Salmonella typhimurium and
Escherichia coli mutagenicity assays has been examined by Dunkel et al.
(1985). It is generally recognised that a single positive is not sufficient to
define mutagenicity, but it was proposed that a single response from
some test systems might have a greater weighting than from others
(Brusick, 1982). However, as more work has been done with these tests it is
realised that some of these weightings may not apply. Assessment panels
have addressed the evaluation of mutagenicity assays for genetic risk
assessment (Brusick et al., 1992; Russel et al., 1984), whilst Ray et al.
(1987) have examined the various assays for identifying classes of
chemicals.
Evaluation of the performance of short term tests in identifying germ cell
mutagens. Waters et al. (in press), evaluated the performance of various
STTs in identifying germ cell mutagens. Using a combined data set derived
from the US EPA/IARC Genetic Activity Profile (GAP) database and the
US EPA Gene-Tox database, a total of 56 germ cell mutagens were
identified. These chemicals had given positive results in one or more of the
following assays: the mouse specific locus test; in vivo tests for chromo-
somal aberrations in the germ cells; the dominant lethal test in mice or rats;
and the mouse heritable translocation test. The same two databases were
used to provide information (where available) on the activity of these
chemicals in bacterial mutagenicity assays, in two in vitro mammalian cell
assays (one for chromosome aberrations, another for gene mutation), and
Diana Anderson 269
in two in vivo tests on the bone marrow (chromosome aberrations or

micronucleus formation). The performance of the various STTs is
summarised in Figure 3. Although the sample size was only small, the data
indicated that the two in vitro assays with mammalian cells were able to
identify 86% (gene mutation) and 93% (chromosome aberration) of the
germ cell mutagens, whilst bacterial mutagenicity assays were slightly less
sensitive (75%). [Unfortunately, in the absence of any data on chemicals
that are not germ cell mutagens, it is not possible to assess the specificity of
the assays (i.e. their ability to correctly identify such chemicals as
non-mutagenic to the germ cells)]. The sensitivity of the individual assays
was increased when the results from two or three of the assays were
combined. Of the 36 germ cell mutagens that had been tested in a bone
marrow assay (for chromosomal aberrations or micronucleus formation),
33 gave positive results, and the evidence for germ cell mutagenicity of two
of the three not identified was called into question. The problem of strain
variability among rodents was raised as a possible source of discrepancy.
This study indicates that STTs can provide valuable information on the
potential of a chemical to induce germ cell mutations.
Ashby (1986) proposed a testing strategy to detect genotoxic agents in
vivo, where after a positive response in Salmonella or in vitro cytogenetics,
a chemical should be tested in a mouse micronucleus or a rat liver
unscheduled DNA synthesis assay in vivo. This strategy was debated by
Garner and Kirkland (1986) and Gatehouse and Tweats (1986). However,
the guidelines for mutagenicity testing (DOH, 1989) use a similar scheme,
but in viuo tests for germ cell effects may be justified on the basis of
properties including pharmacokinetics, use, and anticipated exposure
(Figure 2). Guidelines in most countries and internationally are constantly
under scrutiny (e.g. OECD Guidelines Harmonisation Meeting at the 6th
International Environmental Mutagen Meeting in Melbourne 1993).As a
result of such moves, it is hoped that better models for evaluating human
mutagenic risk may emerge.
In Vitro In Vivo
Bone Marrow Cytogenetics
19/25 = 76% 29/34 = 85%

Positive Positive
Figure 3 Test performances are given,for the germ cell mutagensfrom the combined EPAjIARC G A P and
GENE-TOXdatabases. Performance is indicated by thefraction of agents with positive test results divided
by the number ojagenis tested and is expressed also as the percentage positive
[After Waters et al., (in press)]
Prediction of Human Carcinogens

The use of genetic toxicology assays in predicting the carcinogenicity of
chemicals to rodents or humans has come under a great deal of scrutiny. In
a 1975 publication in which 300 chemicals were tested in the Ames test,
McCann et al. found that 90% of the 174 carcinogens were mutagenic.
Similar figures were provided by other workers (see Introduction), and it
seemed only a matter of time before complementary short-term tests
(STTs) would be developed to detect the remaining 10% of carcinogens
that were ‘missed’ by the Ames test.
Through the late 1970s and early 1980s there was a period of activity as
the matrix of STTs and chemicals tested increased, and various groups of
workers endeavoured to show that their favoured assays could identify
known mammalian carcinogens. A series of internationally coordinated
(through WHO/IPCS, UNEP, and ILO) validation studies (funded partly
by the US National Institute of Environmental Health Sciences and the
UK Medical Research Council) were conducted (Ashby et al., 1985; 1988).
The various STTs were assessed on their sensitivities (the percentage of
known carcinogens correctly detected), their specificities (the percentage
of non-carcinogens correctly identified), and their concordances (the
overall accuracy in their identifications). These studies revealed that the
predictive values of the tests were no longer as high as they had been in
earlier investigations. As more and more rodent carcinogens were
identified, mainly under the National Toxicology Program (NTP), it was
found that the predictive power of the Ames test, in particular, declined.
While the activity of most of the long-standing and well-established rodent
carcinogens could be rationalised in terms of their electrophilic properties
(because this was the primary stimulus for their initial selection for
carcinogenicity testing), an increasing proportion of the newly-identified
carcinogens (not pre-selected in this way) were both without a supporting
chemical rationale for their activity and were non-mutagenic to
Salmonella. The fact that these chemicals were appearing positive in the
animal carcinogenicity studies but were consistently negative in STTs was
of great concern, and had regulatory implications.
This can now be explained in terms of the two types of carcinogen that
are thought to exist, those acting by a genotoxic mechanism (that would
have produced positive results in the STTs), and those acting by a
non-genotoxic mechanism (generally negative in the STTs).
Analysis of’concordance between STT results and carcinogenicity data. The
inconsistency between STT results and carcinogenicity findings prompted
an investigation by the Cellular and Genetic Toxicology Branch of the
NTP, the aim of which was to assess the ability of prokaryotic and
eukaryotic STTs to detect rodent carcinogens. The resultant publication
(Tennant et al., 1987) revealed that four of the most used STTs (the Ames
test, the mouse lymphoma L5178Y mutagenicity assay, and tests for
chromosomal aberrations and sister chromatid exchange in Chinese
hamster ovary cells) were poor predictors of carcinogenic activity. The
assessment of 73 chemicals that had recently been tested for carcinogenic-
ity by the NTP revealed that the concordances between the carcinogenicity
results and the genotoxicity findings were only about 60% for each STT.
Diana Anderson 271
The individual tests exhibited different data profiles, the mouse lymphoma
and SCE assays giving more positive results than the Ames or chromo-
somal aberration assays. Although some chemicals always gave consistent
results, there were eleven chemicals which showed only a single positive
STT result. No combination of the tests improved the accuracy of the
cancer prediction: for instance, defining a chemical as positive if it gave a
positive result in any of the four tests increased the overall sensitivity but
decreased the specificity. The most difficult problem from the genotoxicity
viewpoint was the failure of six carcinogens to show any positive results at
all, despite the fact that three of these (dioxin, reserpine, and a
polybrominated biphenyl mixture) were the most potent carcinogens in
the whole group, at least on the basis of the dose producing statistically
significant increases in tumour incidences. In an update of this study,
Zeiger et al. (1990) tested a further 41 NTP chemicals and found a wider
variation in the level of concordance for the four STTs (ranging from 54%
for the SCE assay to 73% for the Ames test), but this was not considered to
represent any substantial improvement in the predictive power of the
Ames test. Again, no combinations of STT improved upon the concord-
ance and predictivity of the Ames test alone.
The 1987 publication by Tennant and colleagues provoked immediate
responses and counter-responses in the genetic toxicology journals (e.g.
Ashby, 1988a; Ashby and Purchase, 1988; Auletta and Ashby, 1988;
Brockman and DeMarini, 1988; Haseman et al., 1988; Kier, 1988;Trosko,
1988; Young, 1988). The 73 compounds tested were said to be a distorted
sample, as they represented compounds that had only recently been tested
by the NTP and had been selected for testing on the basis of production
volumes, degrees of human exposure and suspicion of carcinogenic
potency. They were certainly more representative of the more subtle 1980s
type of carcinogen than the classic potent carcinogens of the 1950s and
1960s. Brockman and DeMarini (1988) criticised the approach of blind
testing, because no account was taken of what might already be known
about a chemical’s properties or those of structurally related compounds.
Individually ‘customized’ protocols, it was argued, might have given
results that more closely matched the carcinogenicity findings. They also
considered that the results of animal carcinogenicity bioassays (particular-
ly negative ones) did not deserve the exalted position that the scientific
community had generally assigned to them because they had low
statistical power, were rarely replicated, were seldom done under different
sets of experimental conditions, and had many limitations that were
unlikely to be overcome in the near future. Other investigators have also
questioned the scientific validity of lifetime animal feeding studies in which
the chemical is administered at the maximum tolerated dose (see later for
further discussion) (Ames and Gold, 1990a and b; Ashby and Morrod,
1991). The problem with the numerical approach of Tennant et al. (1987)
is that the mathematical sophistication does not make up for the inherent
limitations of the carcinogenicity and genotoxicity assays. The expression
of carcinogenicity bioassay findings as a simple positive or negative result

is, in many cases, an over-simplification or misrepresentation of a
chemical’s true carcinogenic activity. Similarly, the results of a number of
STTs for a particular chemical often include conflicting or equivocal
findings which do not allow a simple positive or negative categorisation.
Mammalian cell assays are said to present a particular problem in
producing so-called ‘false positive’ results (Adler et al., 1989; Scott et al.,
1991). Such limitations can seriously affect the interpretation of the data
obtained, although various attempts may subsequently be made to
rationalise the discordant results in the different STTs, as occurred for the
1987 study by Tennant et al. (e.g. Ashby, 1988b; Myhr and Caspary, 1991;
Prival and Dunkel, 1988).
Despite the limitations, the mathematical approach has been used by a
number of other investigators (e.g. Auletta and Ashby, 1988; Benigni and
Giuliani, 1988; Ennever and Rosenkranz, 1989; Klopman and Rosen-
kranz, 1991; Kuroki and Matsushima, 1987; Loprieno et al., 1991; Parodi
et al., 1991). Kuroki and Matsushima (1987) evaluated the performance of
a range of STTs in detecting 71 established, probable or possible human
carcinogens (classified in IARC Groups 1, 2A and 2B respectively) and
concluded that the chromosome aberration test in mammalian cells in
vitro provided results that were complementary to those obtained in the
Ames test. No figures could be derived to assess the specificity of the
chromosome aberration test because of the lack of clearly identified
human non-carcinogens. Sorsa et al. (1992) evaluated the available
cytogenetic data on 27 proven, 10 probable, and 15 possible human
carcinogens and found that 19/27, 6/10, and 5/15 induced chromosomal
aberrations, sister chromatid exchanges, and/or micronuclei in humans. A
large prospective cohort study suggested that chromosomal aberrations,
but not sister chromatid exchanges, are significant for prospective cancer
risk (Sorsa et al., 1992).
Klopman and Rosenkranz (1991) selected 253 compounds that had
been tested by the NTP for carcinogenicity to animals, and used
probability calculations to assess the predictivity of the Ames test and of
assays for sister chromatid exchange (SCE) and chromosomal aberrations
(CA). The unscheduled DNA synthesis (UDS) assay was similarly
evaluated for 130 chemicals, the corresponding cancer bioassay data being
taken from Williams et al. (1989). The analysis revealed concordances of
62% (Ames), 61.3% (SCE), 56.5% (CA), and 56.2% (UDS). The
probabilities of a carcinogen being positive in individual tests were 56%
(Ames), 88% (SCE), 75% (CA), and 53% (UDS), while the probabilities
of a non-carcinogen being positive in individual tests were 27% (Ames),
89% (SCE), 78% (CA), and 33% (UDS). Although the former might
suggest that the SCE and CA assays were better than the Ames and UDS
assays at detecting carcinogens, the latter indicate that they may in fact be
too sensitive to be of use in predicting carcinogenicity, as a high
proportion of non-carcinogens also gave positive results in these assays.
Diana Anderson 273
The concordance figures reflect the fact that none of the four assays is
particularly good at predicting carcinogenicity.
As noted previously, Tennant et al. (1987) and Zeiger et al. (1990) found
that various combinations of in vitro STTs (Ames, CA, SCE, and the
mouse lymphoma assay) were no better at predicting carcinogenicity than
was the Ames test alone. However, Jenssen and Ramel (1980) and Shelby
(1988) proposed that a combination of two genotoxicity assays, the Ames
test and the (in vivo) bone marrow micronucleus test, could be used for the
detection of genotoxic chemicals that might be predicted to be carcino-
gens. In an analysis of 23 chemicals designated by the International
Agency for Research on Cancer (IARC) as Group 1 compounds
(carcinogenic to humans), Shelby (1988) found that 20 of the 23
carcinogens (87%) were active in one or both STTs. Seventeen of the 22
that were tested in the Ames assay gave positive results, and the untested
chemical (treosulphan) was considered likely to be active on structural
grounds (subsequently confirmed by Zeiger and Pagano, 1989); 12 of these
17 were also active in the in vivo bone marrow test, and four of the
remaining five were considered as likely positives due to structural
similarities to known bone marrow clastogens. Three other chemicals
(benzene, diethylstilboestrol, and arsenic) were inactive in the Ames test
but gave positive results in the bone marrow test, while of the remaining
three chemicals not tested in the bone marrow test, treosulphan was again
considered a likely positive (as was subsequently demonstrated by Gulati
et al., 1990 and Shelby et al., 1989), and asbestos and conjugated
oestrogens (both negative in the Ames test) were not expected to affect the
bone marrow. Thus, the latter two carcinogens would not be detected by
this combination of assays, nor would their carcinogenicity be anticipated
on structural grounds (Shelby, 1988).
Since that time, numerous studies have been conducted to assess the
various possible protocols for the in vivo bone marrow test, including
single or multiple exposures and different sampling times (e.g. Adler and
Kliesch, 1990;cihak and Vontorkova, 1990; George et al., 1990; Gulati et
al., 1990; Mavournin et al., 1990; Mirkova, 1990; Tice et al., 1990). The
various investigators reached different conclusions on which protocol was
most effective in detecting carcinogens, but as there was some evidence
that a three-exposure protocol might be more effective than a single-
exposure protocol (e.g. Gulati et al., 1990; Tice et al., 1990), Shelby et al.
(1993) went on to test 49 NTP chemicals (25 rodent carcinogens and 24
non-carcinogens) using a three-exposure intraperitoneal protocol with a
single sampling time. Only five of the 25 rodent carcinogens gave positive
results, although a further two were found to be positive in a single-
exposure protocol, Two of the seven (benzene and monuron) would not
have been suspected from Ames test data or from their chemical structure.
Four of the 24 chemicals that had shown no evidence of carcinogenicity in
rodent bioassays (ascorbic acid, phenol, titanium dioxide, and 2,6-
toluenediamine) were found to be active in the micronucleus test; all but
2’6-toluenediamine were non-mutagenic in the Ames test. This study

indicates that the three-exposure (ip) protocol with single sampling time is
not satisfactory for distinguishing between carcinogens and non-carcino-
gens; further work is needed before the micronucleus test can provide
additional meaningful information on a chemical’s genotoxic activity.
In an evaluation of a mammalian gene mutation assay, the Chinese
hamster ovary/hypoxanthine guanine phosphoribosyl transferase (CHO/
HGPRT) assay, Li et al. (1988) found that 40 out of 43 reported animal
carcinogens (93%) gave positive results, while the only definitive non-
carcinogen, caprolactam, was negative. Since then, a further nine
chemicals that gave negative results in NTP bioassays have been tested in
this assay; seven gave negative results, one (2-chloroethanol) was positive,
and one (benzoin) was equivocal (Li et al., 1991). The investigators
concluded that the CHOlHGPRT assay seemed to have high specificity as
well as high sensitivity, and that it could be used to complement the Ames
test and cytogenetic assays. Oshiro et al. (1991) tested ten compounds
deemed non-carcinogenic in the literature, and concluded that the
CHO/HGPRT assay produced more relevant results than other
genotoxicity tests in mammalian cells: only two of the chemicals
[dichlorvos and 2-(chloromethyl)pyridine], were positive in this assay, and
the non-carcinogenic status of both was considered questionable.
In an analysis by Loprieno et al. (1991), a test battery of two in vitro
assays (Ames and CA tests) and one in vivo assay (the rodent bone marrow
micronucleus test) was reported to have greater predictivity than the Ames
test alone. In an analysis of 716 chemicals that had been tested in cancer
bioassays, the accuracy of the three STTs was 68.6% for the Ames test
(number of test chemicals, n = 544)’ 64.3% for the CA test (n = 449, and
70.6% for the in vivo micronucleus test (n = 163). This was increased to
+
71.6% for Ames CA (n = 310)’ 85.0% for Ames + micronucleus
+
(n = 113), 87.9% for CA micronucleus (n = 107), and 92.5% for a
combination of all three STTs (n = 93). Thus the three STTs together
correctly identified 43 out of 45 carcinogens and 43 out of 48 non-
carcinogens. The 7 16 chemicals that had been tested for carcinogenicity
were part of a much larger data set of 3389 chemicals, 2898 of which had
been tested in the Ames test (85.5%), 1399 (41.3%) in the CA test, and 319
(9.4%) in the in vivo micronucleus test. For the 270 chemicals that had
been tested in all three STTs, 107 were positive in both in vitro assays, of
which just over half (56) were also positive in vivo; a smaller proportion
(25% or less) of the 16 or 71 chemicals that were positive only in the Ames
test or only in the CA test (respectively) also gave positive results in vivo;
and eleven of the 76 chemicals (14.5%) that gave an indication of
genotoxicity in vitro were found to be active in the micronucleus test. Five
of the eleven (chlorobenzene, ortho- and para-dichlorobenzene, toluene
and trichloroethylene) have also demonstrated carcinogenic activity in
rodents, while the other six (isoxaben, 1,3,5-, 1,2,4-, and 1’2’3-trich-
lorobenzene, trimethoprim, and vincristine) have not been adequately
Diana Anderson 275
tested. The eleven chemicals were thought by the investigators to represent

a class of compound for which the mechanism of genotoxicity could not be
fully applied, and more information was being collected on them
(Loprieno et al., 1991).
Various supplementary assays have been developed, including tests for
recombination, gene conversion and aneuploidy (chromosome loss) in
different strains of the yeast, Saccharornyces cerevisiae. These have not
been well validated, and inconsistencies have been reported, for example,
in an aneuploidy test system using strain D61.M (Albertini, 1991).
Nevertheless, there is some indication that tests for recombination
(Schiestl, 1989; Schiestl et al., 1989) or for gene mutation (Morita et al.,
1989) in yeast cells might be able to detect carcinogens that are ‘missed’ by
the Ames test.
Ashby and Leigibil (1992) authored a commentary on the use of
transgenic mouse mutation assays in the context of genotoxic and
non-genotoxic mutagens. Recommendations for dose levels and treat-
ment protocols for these mice were presented and the authors asked for
further discussion of dosing regimens. Mirsalis (1993) and Ashby and
Leigibil(l993) debated the proposed dosing approach for the detection of
genotoxic and non-genotoxic carcinogens. Gunz et al. (1993) further
examined whether non-genotoxic carcinogens can be detected with the
lac19 transgenic mouse assays.
Use of the Arnes test andlor structural alerts in predicting carcino-

genicity. A non-computational chemistry based method for identifying
carcinogens has been proposed (Ashby, 1985). [This has been progressed
in parallel with developments in quantitative structure activity relation-
ships (QSAR)-see review by Phillips and Anderson (1993)l. This
chemistry based approach utilises the electrophilic theory of carcino-
genesis propounded by the Millers in the 1970s and relies on personal skill
to identify any electrophilic centre(s) within the molecule under consider-
ation. Such a feature constitutes a ‘structural alert’ for reaction with a
nucleophilic site in DNA.
Following the 1987 study by Tennant et al., Ashby and Tennant (1988)
and Ashby et al. (1989) examined the Ames data, carcinogenicity verdicts
and chemical structures of a set of 264 compounds tested in the NTP’s
rodent cancer bioassays. Chemicals were classified by the breadth of their
cancer activity. For instance, 66 were assigned to Group A because they
caused cancer in both rats and mice at one or more sites. Group B
chemicals (19 in all) produced tumours at multiple sites in one species.
Group C (27 compounds) and Group D (26 compounds) resulted in
cancer at a single site either in both sexes or a single sex (respectively) in a
single species. The two remaining classes consisted of 94 chemicals where
there was no indication of carcinogenicity in either species (Group F) and
32 where there was an equivocal or uncertain result (Group E). Aspects of
the chemical structure that were suggestive of electrophilic activity either
of the compound or its metabolites were then identified. The investigators

assessed how well the results of the Ames test and the presence or absence
of structural alerts predicted the NTP carcinogenicity results. The
concordance between the four carcinogenic and non-carcinogenic classes
(A, B, C, D, and F) and the Ames test was 64%, as low as Tennant’s earlier
results. The concordance with the structural alerts was also 64%. The
number of mismatches between the structural alerts and the Ames test
results was 31 (12%). Thus, although there was a high correlation between
the presence or absence of structural alerts in the 264 chemicals and their
mutagenic activity in the Ames test, neither of these was particularly
effective in predicting carcinogenicity. The observed sensitivity of the
Ames test (in correctly identifying carcinogens) was only 58%, but this
increased to 72% when the 66 two-species carcinogens (Group A) were
considered on their own (Ashby et al., 1989). This compared with only
about 50% (Groups B and D) or 32% (Group C) for the 72 single-species
carcinogens. The evaluation of a further 39 chemicals had little effect on
these figures for sensitivity (Tennant and Ashby, 1991).
These differences between the four classes of carcinogens provided the
impetus for a further analysis of all the chemicals that had been included in
the three earlier surveys (i.e. Ashby et al., 1989; Ashby and Tennant, 1988;
Tennant and Ashby, 1991). The combined data set consisted of 301
chemicals, which were split into broad classes based, this time, on their
known or projected chemical reactivity (Ashby and Tennant, 1991).
Roughly half of the chemicals (154) were structurally-alerting chemicals,
and these were further subdivided into aromatic amino/nitro-type com-
pounds (84 chemicals), natural electrophiles including reactive halogens
(46), and miscellaneous structurally-alerting groups (24). The 147 non-
alerting chemicals were broadly categorised as compounds devoid of
actual or potential electrophilic centres (6 l), compounds containing a
non-reactive halogen (50), or compounds previously classed as non-
altering in structure but with minor concerns about a possible structural
alert (36). When the six classes were each analysed separately, Ashby and
Tennant reported a clear distinction between the three groups of
structurally-alerting chemicals and the three groups of non-alerting
chemicals. Thus 65-70% of the structurally-alerting chemicals were
carcinogenic compared with 25-52% of the non-alerting chemicals, and
6 5 4 6 % were mutagenic in the Ames test compared with only 0.02-9% of
the non-alerting chemicals. The sensitivity of the Ames test in predicting
carcinogenicity was 93% for the nitro/amino compounds, 83% for the
electrophiles, and 56% for the miscellaneous structurally-alerting com-
pounds; however, for all three groups there was a high false-positive rate,
as 71%, 55% (6/11), and 75% (3/4) of the non-carcinogens in the
respective groups gave positive results in the Ames test. For the
non-alerting chemicals, the sensitivity was low (13%) or non-existent,
while the high specificity (with only one non-carcinogen in each group
showing mutagenic activity) was of little value.
Diana Anderson 277
Genotoxic and non-genotoxic carcinogens. Ashby and Tennant (199 1)

distinguished between structurally-alerting carcinogens, i.e. chemicals
with one or more electrophilic groups that could potentially interact with
and damage the cell’s DNA in some fairly immediate way (‘genotoxic
carcinogens’), and non-alerting carcinogens that had no obvious elec-
trophilic centre and were thought to induce tumours by some mechanism
that did not involve early attack of the chemical or a direct metabolite on
the cell’s DNA (putative ‘non-genotoxic carcinogens’). Some examples of
non-genotoxic (or epigenetic) routes to tumour formation include en-
hanced cell proliferation resulting from cytotoxicity or mitogenesis
(induced cell division), hormonal changes, and peroxisome proliferation.
At a 1990 meeting on early indicators of non-genotoxic carcinogenesis,
several possible mechanisms were described (Anon., 1991).
Since the various STTs are specifically designed to detect genotoxic
activity, any chemical that induces tumours by a non-genotoxic mechan-
ism will not (or at least in theory ought not to) be active in these STTs.
Jackson et al. (1993) indicated that the situation is not that simple.
Many chemicals that are described as non-genotoxic carcinogens in the
literature have in fact shown evidence of genotoxic activity in appropriate
STTs. In an analysis of 39 putative non-genotoxic carcinogens which had
been tested in five or more STTs for gene mutation, chromosomal
aberrations and/or aneuploidy, 14 showed evidence of activity (seven in
vivo and seven in vitro), and a further ten showed limited evidence (in vivo
and/or in vitro). Only two were considered to have been sufficiently tested
to warrant the description non-mutagenic (i.e. non-genotoxic). The
remaining 13 chemicals had not been adequately tested in vitro and in vivo
for each of the three broad categories of DNA alterations (i.e. gene
mutation, chromosome aberration, and aneuploidy). Thus, although all
but one of the 39 chemicals gave predominantly negative results in the
Ames test, and most were also devoid of structural alerts, over half showed
genotoxic potential in other STTs, and another third might also do so if
they were to be adequately tested. Whether their genotoxic activity plays
an important role in their carcinogenicity is questionable, but the
investigators noted that four out of six compounds that induce peroxi-
some proliferation, two out of five cytotoxic carcinogens and one out of
three mitogens all demonstrated some mutagenic activity.
Distinguishing between genotoxic and non-genotoxic carcinogens on the

basis of their turnour profile. Ashby et al. (1989) and Ashby and Tennant
(1988, 1991) distinguished between two types of carcinogen: multi-species
(and usually multi-site) carcinogens which could generally be predicted by
either a positive result in the Ames test or (equally well) by a structural
alert (the so-called genotoxic carcinogens); and the single-species (often
single-site, and/or single-sex) carcinogens which were less likely to be
mutagenic or structurally-alerting (the so-called non-genotoxic carcino-
gens). Analysis of the patterns of carcinogenic response provided by these
two classes of carcinogen had revealed that for many tissues (notably the
mouse liver), equal sensitivity to genotoxic and non-genotoxic carcino-
gens was observed, but that certain tissues (notably the Zymbal’s gland in
the rat) appeared only to be sensitive to genotoxic carcinogens (Ashby et
al., 1989; Ashby and Tennant, 1988). When Ashby and Tennant (1991)
evaluated the target sites for 59 carcinogenic nitro/amino compounds
(93% of which were Ames-positive) and 30 carcinogenic electrophilic
agents (83% Ames-positive), they were all said to have been previously
associated with genotoxic carcinogens. In contrast, the target sites for the
57 carcinogenic non-alerting chemicals (only two of which were mutagenic
in the Ames test) were said to be almost exclusively confined to tissues
connected with non-genotoxic carcinogens. This in fact may be an
over-simplification. Whilst the 1988 study had identified 16 tissues that
were affected only by genotoxic carcinogens, the 1991 study revealed a
marked reduction in this number, as some non-genotoxic carcinogens also
induced tumours in these tissues. Only the Zymbal’s gland (rat) and the
lungs (both species) appeared to be exclusively targeted by genotoxic
carcinogens. The authors’ post-analysis rationalisation of certain putative
non-genotoxic carcinogens which were active in tissues previously only
associated with genotoxic carcinogens did not explain away all of the
discrepancies.
In a further analysis of the pattern of the carcinogenic response among
tissues and test groups, Ashby and Paton (1993) aimed to test the tentative
conclusions of the 1991 paper on a larger data set, restricting their analysis
of genotoxicity to a simple consideration of chemical structure. The data
for the analysis were taken from the carcinogen database compiled by
Gold et al. (1991), which detailed the site of carcinogenesis for 522
chemical carcinogens. Of the 5 1 1 rodent carcinogens, 25 1 had been tested
in both rats and mice, 168 in rats only and 92 in mice only. One chemical,
benzene, was eliminated from the analysis because of its ‘exceptional’
carcinogenic effects (it induces an extensive range of tumours in both sexes
of rats and mice often at unusual sites, and yet no useful structural alert
can be derived without condemning all benzenoid chemicals). Around
70% of the remaining 510 rodent carcinogens were structurally-alerting,
i.e. 300 of the 418 chemicals tested in rats and 236 of the 342 tested in mice.
When these were analysed in terms of the number of tissues affected,
27.7% (rat) and 18.6% (mouse) of the structurally-alerting chemicals were
active in more than two tissues, compared with only 5.9% (rat) and 0%
(mice) of the non-alerting chemicals. In contrast, a much higher percen-
tage of the non-alerting chemicals affected only a single site (rat, 58.8%;
mouse 65.9%), compared with the structurally-alerting chemicals (rat,
34.2%; mouse, 34.6%). Of the 250 chemicals tested in both species,
roughly half of the 43 rat-specific carcinogens and of the 64 mouse-specific
carcinogens were structurally-alerting chemicals, compared with 77% of
the 143 two-species carcinogens. The investigators considered this a
confirmation of their earlier hypothesis, that structurally-alerting (or
Diana Anderson 279
genotoxic) carcinogens tend to be active in more than one species and at

multiple sites, whereas the non-alerting (putatively non-genotoxic) car-
cinogens are more likely to affect only a single species and a single tissue.
In an earlier paper, Ashby and Purchase (1992) had selected five ‘geno-
toxic carcinogens’ (positive in the Ames test and with structural alerts)
and five ‘non-genotoxic carcinogens’ (negative Ames and no structural
alerts), and had shown that all five genotoxins were carcinogenic to both
species and both sexes, and that four of the five were multi-site
carcinogens; in contrast, the five non-genotoxins were active in only one
species, three of the five affected only a single sex, and four were single-site
carcinogens. These ten compounds well supported their earlier hypothesis.
However, when Ashby and Paton (1993) tried to demonstrate the same
effect amongst over 500 carcinogens, the findings were not so clear cut.
Thus, around 34% of the structurally-alerting carcinogens were active in
only a single tissue and 35-40% of the non-alerting chemicals were active
in more than one tissue. Furthermore, although 110 of the 143 two-species
carcinogens were structurally-alerting, this still left 33 which had no
structural alerts. For the 107 carcinogens that were tested in both species
and found to be active in only one, 53 were structurally-alerting and 54
were not. In a separate study by Gold et al. (1993), involving an analysis of
35 1 mutagenic (in the Ames test), or non-mutagenic rodent carcinogens
from the Carcinogenic Potency Database, it was found that 42% of the
single-site, single-species carcinogens were mutagenic while 3 1% of the
two-species carcinogens were not. This does not greatly support the
distinction between genotoxic carcinogens (multi-species, multi-site) and
non-genotoxic carcinogens (single-species, single-site).
Gold et al. (1993) also compared the target organs for the mutagenic
and non-mutagenic carcinogens, and concluded that both groups induced
tumours in a wide variety of sites, that most organs were target sites for
both, and that the same sites tended to be the most common targets for
both. When the more unusual tumour sites had been excluded (because the
number of mutagenic and/or non-mutagenic chemicals that affected them
was too small to be meaningful), Gold et al. found that only the Zymbal’s
gland was targeted exclusively by mutagenic carcinogens. The one
discordant chemical was benzene, which Ashby and Paton (1993) had
eliminated from their own analysis because it was clastogenic but non-
alerting and non-mutagenic. The latter investigators also reported the
Zymbal’s gland to be the sole tissue that was affected only by structurally-
alerting chemicals (after benzene had been excluded). Ashby and Tennant
(1988) attributed this ‘property’ to 15 specific sites in 1988, but ten of the
sites were subsequently shown also to be susceptible to non-alerting
chemicals (Ashby and Paton, 1993); it remains to be seen whether the
other sites (including the Zymbal’s gland, subcutaneous tissues, clitoral
gland, spleen, and ovary) will also be shown, eventually, to be susceptible
to non-genotoxic carcinogens. Gold et al. (1993) noted that out of 351
rodent carcinogens, 20 induced Zymbal’s gland tumours in rats and two
induced them in mice (these two being included in the 20 that were active in
the rat). The 20 chemicals were also multi-site carcinogens. Of the 230
carcinogens in the overall Carcinogenic Potency Database that were tested
in both rats and mice, only 42 (1 8%) affected multiple sites in both species;
but of the 14 that were tested in both species and that induced Zymbal’s
gland tumours, ten (71%) affected multiple sites in both species. This
indicates that chemicals that induce Zymbal’s gland tumours are generally
multi-site, multi-species carcinogens.
It can be seen that successive attempts to distinguish between genotoxic
and non-genotoxic carcinogens on the basis of their tumour profile have
become less convincing as the number of carcinogens analysed has
increased.
Screening for non-genotoxic carcinogens. Whilst there are a number of
STTs that can be used to detect genotoxic chemicals, the problem remains
of how to screen for non-genotoxic carcinogens when the actual mechan-
isms involved are not known in any detail. Carcinogenesis is a complex,
multi-stage process, and there are many ways in which its onset and
progress may be affected (Green, 1991). The various stages of cancer
development have been defined operationally as initiation, promotion,
progression, and metastasis. Initiation appears to represent damage to key
genes involved in the regulation of cell growth, and genotoxic chemicals
are believed to contribute to tumour initiation as a result of their
damaging effect on DNA. Other factors such as viruses, UV light, ionising
radiation, and error-prone DNA replication may also lead to initiation.
The clonal expansion of these initiated cells may result from the action of,
for example, growth factors, hormones, or many non-genotoxic carcino-
gens.
1 Cell proliferation
The induction of cell proliferation appears to be a key factor in
non-genotoxic carcinogenicity, and may involve either a direct mitogenic
effect on the target without apparent cytolethality or a cytotoxic effect
which produces cell death in the target tissue followed by regenerative cell
proliferation (Butterworth, 1990). Cell proliferation is a key factor in
genotoxic as well as non-genotoxic carcinogenicity, as mutagenic activity
may occur as a secondary event in the carcinogenic process (Butterworth
et al., 1992). As noted previously, use of the maximum tolerated dose
(MTD) in animal carcinogenicity bioassays has been the subject of much
debate; such a high dose may cause cell death and subsequent cell
proliferation, the development of tumours being secondary to this
excessive organ-specific toxicity. Thus, tumours that are induced only at
the MTD are of questionable relevance to humans who are generally
exposed at much lower levels (Ames and Gold, 1990a and b; Ashby and
Morrod, 1991). Information on a chemical’s capacity to induce cell
proliferation may therefore be very useful in setting bioassay doses or in
Diana Anderson 281
evaluating bioassay results. Two measures that have been used to assess
the extent of cell proliferation are the mitotic index and the labelling index.
These indicate, respectively, the fraction of cells that are in the process of
mitosis and the percentage that have taken up a radiolabelled DNA
precursor (Butterworth et al., 1992). More complex studies may be
conducted to identify the specific receptors that mediate the mitogenic
action of some non-genotoxic carcinogens, or to understand how the
various proto-oncogenes and tumour suppressor genes that regulate cell
growth are influenced by non-genotoxic carcinogens (Green, 1991). These
types of study are based on mechanistic considerations and have not been
applied to any great extent in testing and validation programmes.
2 Cell transformation assays

The most well-established assays for detecting non-genotoxic (and also
genotoxic) carcinogens are the various cell transformation assays, which
are based on carcinogen-induced loss of contact inhibition of cultured cells
resulting in a piling up of transformed cells in a criss-cross fashion. These
assays have been critically evaluated by various investigators and expert
groups, who considered them highly relevant for the process of carcino-
genesis in vivo because they involve the same endpoint (i.e. the transformed
cells are tumorigenic in appropriate hosts) and because they may well
share many of the same cellular and molecular mechanisms (Dunkel et al.,
1981; Heidelberger et al., 1983; IARC/NCI/EPA Working Group, 1985).
Studies evaluating the performance of cell transformation tests in
predicting animal or human carcinogens have generally given promising
results (Barrett and Lamb, 1985; DiPaolo et al., 1972; Fitzgerald et al.,
1989; Jones et al., 1988; Pienta et al., 1977; Swierenga and Yamasaki,
1992), although the lack of a dose response, the low transformation
frequency and the difficulties in obtaining consistently reproducible results
in repeat assays have impeded the development of this system for routine
use (Jones et al., 1988; Tu et al., 1986).
In an assessment of the Syrian hamster embryo (SHE) cell transform-
ation assay, Jones et al. (1988) tested 18 coded chemicals in three different
laboratories using the same basic protocol. Rodent carcinogenicity data
were available for 16 of the 18 chemicals. When the four chemicals that
gave discordant transformation responses in two laboratories were
counted as positive, the investigators found the results of the two systems
(for carcinogenicity and cell transformation) to be in agreement for 14 of
the 16 chemicals. Four rodent carcinogens that were inactive in the Ames
test gave positive results in the SHE assay. However, two of the eight
purported rodent non-carcinogens, caprolactam and geranyl acetate, were
found to be active in the SHE assay in all three laboratories. A
non-carcinogen might turn out to be carcinogenic, however, if tested in the
appropriate species or strain and under appropriate conditions, but it is
interesting to note that caprolactam is the only chemical classified by the
International Agency for Research on Cancer (IARC, 1987) as a Group 4
compound (probably not carcinogenic to humans).
Swierenga and Yamasaki (1992) evaluated the performance of trans-

formation tests in identifying IARC Group 1 compounds (carcinogenic to
humans) and Group 2A compounds (probably carcinogenic to humans).
The data from cell transformation tests were considered collectively, so
that a chemical was positive if it gave a positive response in any such assay.
Out of 28 Group 1 carcinogens for which data were available, 25 (89%)
gave a positive response in the cell transformation assay, while for the
three chemicals that were negative, only a single test result was identified.
For comparison, the Ames test and the chromosome aberration test
identified 25 out of 41 (61 YO)and 25 out of 30 (83%) Group 1 carcinogens,
respectively. For Group 2A compounds, the cell transformation assay
again showed a high level of sensitivity, with 21 of the 25 (84%) giving
positive results; the corresponding figures for the Ames test and the
chromosome aberration test were 34 out of 41 (83%) and 30 out of 31
(96%), respectively. The specificity of the cell transformation assays (in
identifying human non-carcinogens) is more difficult to assess because of
the lack of clearly identified non-carcinogens, but Ennever et al. (1987)
reported that only four out of twelve probable human non-carcinogens
(33%) gave negative results in at least one cell transformation assay.
According to Swierenga and Yamasaki (1992), some of these probable
human non-carcinogens have since shown evidence of carcinogenic
activity in animals, but the exclusion of these from the analysis apparently
does not increase the specificity of the cell transformation test. These
investigators nevertheless concluded that the importance of cell trans-
formation tests may have been underestimated in previous IARC
evaluations while that of the Ames test may have been overestimated.
They considered the various assays for all transformation to have
generally shown good agreement, and suggested that further refinements
were needed to include the use of human cells, human xenobiotic
metabolism and appropriate tissue-specific target cells (Swierenga and
Yamasaki, 1992).
3 Gap-junction intercellular tests

It has been suggested that, since normal cells surrounding those that
contain transforming oncogenes are able to suppress the transformed
phenotype, the disruption of intercellular communication via gap junc-
tions may play a role in carcinogenesis (Green, 1991; Yamasaki, 1990). In
an evaluation of the performance of gap junction intercellular tests,
Swierenga and Yamasaki (1992) reported that three out of four Group 1
compounds (established human carcinogens) and four out of five Group
2A compounds (probable human carcinogens) gave positive results in this
test. For the two compounds that were not detected, one, crystalline silica,
was thought to exert its carcinogenic effects without disturbing cell-to-cell
communication via the gap junction, and the other, diethylstilboestrol was
thought to be cell-type specific, as two steroidal oestrogens gave positive
results in this assay. Four out of five organochlorine pesticides also gave
positive results, and three of the four were classified by IARC as Group 2B
Diana Anderson 283
compounds (possible human carcinogens). Whilst these findings are

encouraging, the data set is obviously far too small for any definitive
conclusions on this type of test. Even so, Swierenga and Yamasaki (1992)
considered that the results of such tests should be accorded greater weight
than they currently receive.
4 Use of toxicity data

An alternative approach might be to use existing toxicological data on a
chemical as an indication of its carcinogenic potential. A relationship
between carcinogenicity and systemic toxicity might be expected, given
that the highest dose used in carcinogenicity tests is selected on the basis of
toxicity. The possible tautologous nature of this relationship has been
discussed by Bernstein et al. (1985) and Crouch et al. (1987). In fact, only
limited correlations have been shown between LD,, values and carcino-
genic potencies of known animal carcinogens (McGregor, 1992; Metzger
et al., 1989; Zeise et al., 1984). McGregor (1992) suggested that better
correlations might be obtained if the maximum tolerated dose (MTD)
were used as a measure of toxicity rather than the LD,, value, but that
even this assumed that the mechanisms of death and of carcinogenicity
were the same in animals exposed to a particular chemical, Haseman and
Seilkop (1992) looked for such an association between the MTD and
carcinogenicity in an analysis of 326 NTP studies conducted in rodents,
and they found the overall concordance between toxicity and carcino-
genicity to be only 56%. When 130 NTP carcinogenicity studies, involving
around 1500 sex-species-exposuregroups, were analysed to see if there was
a direct causal relationship between organ toxicity and carcinogenicity, it
was concluded that the available data did not support a correlation
between chemically-induced toxicity and carcinogenicity (Hoel et al.,
1988; Huff, 1992; Tennant et al., 1991). Some chemicals caused organ
toxicity without cancer, others induced site-specific cancer with no
associated toxicity, and a third group caused toxicity and cancer in the
same organ; only seven of the 53 positive carcinogenicity studies were said
to have exhibited the types of target organ toxicity that could have been
the cause of all observed carcinogenic effects (Hoel et al., 1988). Huff
(1992) concluded that it would be scientifically premature to make any
inference about the influence of toxicity or of cell replication on chemical
carcinogenesi s.
A wider approach suggested by Travis et al. (1990a and b, 1991)
involves making use of all available information on a compound’s
biological activity; they found that a combination of acute and reproduc-
tive toxicity data, mutagenicity data and sub-chronic and chronic
tumorigenicity data provided a far better prediction of carcinogenic
potency than did mutagenicity data alone. Unfortunately the method
cannot be used to distinguish between carcinogens and non-carcinogens,
as it can only predict the carcinogenic potency of known mouse
carcinogens.
It can be seen that, for non-genotoxic carcinogens, mechanistic consider-

ations must play a major role in any assessment of risk. Until the various
mechanisms of action are more fully understood, it is difficult to see how a
suitable battery of short-term tests can be developed to screen against the
carcinogenic effects of such chemicals. This area merits intensive study if it
is to be systematised to the extent necessary to be of use in carcinogen
regulation (Clayson and Arnold, 1991).
Despite all the problems identified, the short term tests do have a role in
predicting carcinogens. IARC is currently using data from short term
tests, alongside animal and epidemiological evidence to reclassify carcino-
gens. Until 1987, there were 23 human chemical carcinogens and 7 from
industrial processes. Now as a result of reclassification, there are 55
recognised human carcinogens.
In addition to their use in predicting carcinogens, short term tests when
used with due consideration, have a role to play in mechanistic studies, for
examining complex mixtures and air pollutants and for the early
identification of the genetic toxicity of new chemical products.
B Approach %The Direct Measurement of Genetic

Damage in Man
An approach to measuring genetic damage in man has been to investigate
the incidence of chromosome damage in peripheral lymphocytes of
subjects exposed to a given chemical (Anderson, 1988; Kilian et al., 1975;
Purchase et al., 1978). It has been reported by several authors, for
example, that workers occupationally exposed to vinyl chloride have an
increase in the incidence of chromosome damage (Ducatman et al., 1975;
Fleig and Theiss, 1977; Kilian et al., 1975; Natarajan et al., 1978; Szentesi
et al., 1976). In the study of Purchase et al. (1978b), 81 workers were
investigated (57 VC exposed and 24 controls) and effects found in the
exposed group. Eighteen months later, the incidence of chromosome
damage was still present, though it had decreased in those workers who
had changed jobs. At 42 months the incidence of damage had returned to
control values (Anderson et al., 1980). Many other chemicals had been
investigated in a similar way, e.g. ethylene oxide, epichlorohydrin,
styrene, butadiene, acrylonitrile, asbestos, benzene, chloroprene, cyclo-
phosphamide, chromium, lead, etc.
Such studies have to be well-controlled with age- and sex-matched
individuals and confounding factors, such as drinking and smoking, taken
into account (Anderson et al., 1990; Brinkworth et al., 1992; Dewdney et
al., 1986). As yet, the studies yield results which can only be interpreted on
a group basis and not used for individual counselling. In these circumstan-
ces, chromosome analysis is useful for determining whether exposure
levels of a chemical do or do not induce chromosome damage. At a
meeting in Luxembourg in 1987 (convened by the EEC, IPCS, WHO,
IARC, and Institute of Occupational Health, Helsinki) on Human
Diana Anderson 285
Monitoring, a concensus opinion was that increased levels of chromosome

damage were indicative of an increased risk of cancer even though no
causal relationship had yet been established between chromosome damage
and cancer. However, most of the agents causing chromosome damage
have been shown to be carcinogenic to animals. Sorsa et al. (1990) in a
preliminary study, however, did suggest a causal relationship between
chromosome damage and cancer, but not between SCE and cancer.
Other techniques for direct application to man are available, such as
the use of urine or blood plasma from exposed workers in combination
with a microbial assay (Legator et al., 1982, Gene-Tox), electrophoretic
monitoring of enzymatic markers in man (Neel, 1979b) detection of
variants in haemoglobin molecules (Nute et al., 1976; Popp et al., 1979),
determining increases in the formation of haemoglobin adducts (Farmer,
1982; Farmer et al., 1986; Tornquist et al., 1988) and DNA adducts
(Pfeifer et al., 1993; Phillips et al., 1988; Weston, 1993), investigations of
sperm morphology (Wyrobek et al., 1983b, Gene-Tox), increases in YY
bodies in sperm (Kapp et al., 1979), thioguanine mutant frequency in
human lymphocytes (Albertini et al., 1988; Cole et al., 1988), and
oncoproteins in plasma (Brinkworth et al., 1992).
A special issue of Mutation Research reviews Human Monitoring
methods (Anderson, 1988) and there is another issue in preparation
(Anderson, 1994).
C Approach 3-Measurement of Gonadal Effects in

Mammalian Systems and Extrapolation to Man Based
on Principles Determined from Radiation Genetics
The virtual universality of DNA as the genetic material furnishes a
rationale for using various sub-mammalian and non-human test systems
for such predicting mutagenic potential. Nevertheless, some organisms
may be considered more similar to man than others and thus more
suitable for the purpose of evaluating human genetic risk. In man,
however, there are many pharmacokinetic factors which affect the
mutagenic efficiency of a compound.
The assessment of heritable mutagenic risk to humans involves
determining gonadal effects and acknowleding the influence of many
variables. Results can be assessed on the basis of expected human
exposure and known mutagenic potential from such such studies as
dominant lethal, heritable translocation, or specific locus assays. Simi-
larly, the concentration of the mutagen in germ cells can be assessed
biochemically, extrapolated to human exposure levels, and combined
with mutagenicity data from the most-used and well-understood in vitro
test systems.
It has been suggested that the population effects of chemical
mutagens should take radiation as an equivalent and equate the
population dose of those mutagens to the radiation dose admissible for
that population (Bridges, 1974; Crow, 1973); or should not exceed a dose
which doubles the spontaneous mutation rate of that population. Both
concepts, however, have been criticised (Auerbach, 1975; Schalet and
Sankaranarayanan, 1976; Sobels, 1977). An approach currently regaining
favour is the parallelogram approach of Sobels (1977) where with data
from rodent germ cells and human and rodent somatic cells, human germ
cell data can be predicted/estimated. An EEC/US EPA collaboration has
been instigated (1993) to examine this approach.
7 General Conclusions
The present methods available for the testing for mutagenicity are not
equally reliable or reproducible. Even the systems most frequently used
(e.g. Ames test) give different results in different circumstances.
Since no one test system satisfactorily detects all genetic end-points, a
combination of tests is required. Such a ‘battery’ should preferentially
consist of a microbial test and a test detecting chromosome damage. This
combination would be the minimum required. Currently in some coun-
tries an in vitro assay for gene mutation is also carried out (Arlett and Cole,
1988). If a larger ‘battery’ is to be used to identify those chemicals which
are potentially hazardous to man, a mammalian in viuo system should be
considered. By making the test battery too large a greater number of false
results may be generated (Purchase et al., 1971; 1978a; 1980; Tennant et
al., 1987). It is possible that the ‘gap’ that exists between mutagenicity and
carcinogenicity may be partially bridged at an early stage of testing with a
cell transformation assay (Heidelberger et al., 1983; Meyer et al., 1984;
Dunkel, 1991; DOH COC Guidelines, HMSO, 1991; Fitzgerald and
Yamasaki, 1990).
The need for safety evaluation in general toxicological testing is well
recognised and this is certainly true in the field of genetic toxicology.
However, it is not certain that positive or negative results in laboratory
model test systems are relevant to man because of man’s unique
metabolism and because of the absence of any convincing ‘no-effect’ level
data for animals or man. Epidemiological evidence for germ cell
mutation after chemical exposure (or, in fact, any agent) is lacking. In
the industrial situation, it is often difficult to identify the exact chemical
or agent that may be causing a problem. Unbiased abortion rates are
difficult to determine by comparison with control or unexposed popula-
tions. Not all abortions are recorded.
The limitations of the simpler short term tests, which are more
concerned with the concept of somatic mutation than of heritable genetic
damage, are now better understood (Purchase, 1980; Tennant et al.,
1987; Ashby and Tennant, 1988; Ashby and Paton, 1993; Waters et al.,
1993) but to extrapolate to man in terms of heritable damage, in vivo
animal studies are required. The logistics of tests for this purpose (the
specific locus test and heritable translocation test), however, are difficult to
satisfy.
Dianu Anderson 287
There is a school of opinion which holds that just as mutagenicity tests

might detect carcinogens, so might carcinogenicity tests detect mutagens.
This opinion is based on the concept that cancer may be an easily
observable phenotype of DNA mutation. If this is so, then the car-
cinogenicity and mutagenicity data for a chemical should be considered
together. This is probably true for genotoxic carcinogens and IARC is
using this approach.
It is hoped that as research progresses, our understanding and
techniques will improve so that the results generated in our model
systems will become unequivocal in terms of hazard to man. To achieve
this aim, attention will have to be given to studies aimed at assessing the
significance to man of positive mutagenic responses produced by a test
system for a given chemical, in addition to the search for better assay
procedures and the understanding and interpretation of effects from
chemicals producing mutation by indirect means.
References
Abrahamson, S. and Lewis S. B. (1971). The detection ofmutations in Drosophila:
In: ‘Chemical Mutagens, Principles and Methods for their Detection’, Ed. A.
Hollaender, Plenum Press, New York, London, Vol. 2, pp. 461-489.
Adler, I. D., Ashby, J., and Wurgler, F. E. (1989). Screening for possible human
carcinogens and mutagens: a symposium report. Mutat. Res., 213, 27-39.
Adler, I.-D. and Kliesch, U. (1990). Comparison of single and multiple treatment
regimens in the mouse bone marrow micronucleus assay for hydroquinone
(HQ) and cyclophosphamide (CP). Mutat. Res., 234, 115-123.
Albanese, R. (1987). Mammalian male germ cell cytogenetics. Mutagenesis, 2,
79-87.
Albanese, R., Topham, J. C., Evans, E., Clare, M. G., and Tease, C. (1984).
Mammalian germ-cell cytogenesis. The report of the UKEMS Sub-Com-
mittee on Guidelines for Mutagenicity Testing. Part 2, pp. 145-172.
Albertini, S. (1989). Influence of different factors on the induction of chromosome
mal-segregation in Saccharomyces cerevisiae D6 1 .M by baviston and assess-
ment of its genotoxic property in the Ames test and in Saccharornyces
cerevisiae D7. Mutat. Res., 216, 327-340.
Albertini, S. (1991). Re-evaluation of the 9 compounds reported conclusive
positive in yeast Saccharomyces cerevisiae aneuploidy test systems by the
Gene-Tox Program using strain D61 .M of Saccharomyces cerevisiae. Mutat.
R ~ s .260,
, 165-180.
Albertini, S., Brunner, M., and Wiirgler, F. E. (1993). Analysis of six additional
chemicals for in vitro assays of the European Economic Communities (EEC)
aneuploidy programme using Saccharomyces cerevisiae D61 .M and the in
vitro Porcine Brain Tubulin Assembly assay. Environ. Mol. Mutagen., 21,
180-1 92.
Albertini, R. J., Sullivan, L. M., Berman, J. K . , Greene, C. J., Stewart, J. A.,
Silveira, J. M., and O’Neill, J. P. (1 988). Mutagenicity monitoring in humans
by autoradiographic assay for mutant T lymphocytes. Mutat. Res., 204,
48 1-492.
Ames, B. N., Durston, W. E., Yamasaki, E., and Lee, F. D. (1973a). Carcinogens
are mutagens: a simple test system combining liver homogenates for
activation and bacteria for detection. Proc. Natl. Acad. Sci. ( U S A ) , 70,
2281-2285.
Ames, B. N. and Gold, L. S. (1990a). Chemical carcinogenesis: too many rodent
carcinogens. Proc. Natl. Acad. Sci ( U S A ) , 87, 7772-7776.
Ames, B. N. and Gold L. S. (1990b). Too many rodent carcinogens: mitogenesis
increases mutagenesis. Science, 249, 970-97 1.
Ames, B. N., Lee, F. D., and Durston, W. E. (1973b). An improved bacterial test
system for the detection and classification of mutagens and carcinogens. Proc.
Natl. Acad. Sci. ( U S A ) , 70, 782-786.
Ames, B. N., McCann, J., and Yamasaki, E. (1975). Methods for detecting
carcinogens and mutagens with the Salmonella/mammalian microsome
mutagenicity test. Mutat. Res., 31, 347-364.
Anderson, D. (1 975a). The selection and induction of 5-iodo-2-deoxyuridine and
thymidine variants of P388 mouse lymphoma cells with agents with are used
for selection. Mutat. Res., 33, 399-406.
Anderson, D. (1975b). Attempts to produce systems for isolating spontaneous
and induced variants in various mouse lymphoma cells using a variety of
selective agents. Mutat. Res., 33, 407-416.
Anderson, D. (1988). Human Monitoring-Special Issue. Mutat. Res., 204,
No. 3.
Anderson, D. and Cross, M. F. (1982). Studies with 4-chloromethylbiphenyl in
P388 cells resistant to 5-iodo-2-deoxyuridine. Mutat. Res., 100, 257-261.
Anderson, D. and Cross, M. F. (1985). Suitability of the P388 mouse lymphoma
system for detecting potential carcinogens and mutagens. Food Chem.
Toxicol., 23, 115-1 18.
Anderson, D., Dewdney, R. S., Jenkinson, P. C, Lovell, D. P., Butterworth, K. R.,
and Conning, D. M. (1985). Sister chromatid exchange in 106 control
individuals. In : ‘Monitoring of Occupational Genotoxicity’. Proceedings of
the Fourth ICEM Satellite Meeting, Helsinki, June 30-July 2. Eds M. Sorsa
and H. Norppa, pp. 38-58. Alan R. Liss, and abstract in Hum. Toxicol., 4,
79-1 16.
Anderson, D. and Fox, M. (1 974). The induction of thymidine and IUdR resistant
variants in P388 mouse lymphoma cells by X-rays, UV, and mono- and
bifunctional alkylating agents. Mutat. Res., 25, 107-122.
Anderson, D., Francis, A. J., Godbert, P., Jenkinson, P. C., and Butterworth, K.
R. (1990). Chromosome aberrations (CA), sister chromatid exchanges (SCE)
and mitogen induced blastogenesis in cultured peripheral lymphocytes from
48 control individuals sampled 8 times over 2 years. Mutat. Res., 250,
467 -476.
Anderson, D., Green, M. H. L., Mattern, 1. E., and Godley, M. J. (1984). An
international collaborative study of ‘Genetic drift’ in Salmonella typhimurium
strains used in the Ames test. Mutat. Res., 130, 1-10.
Anderson, D., Jenkinson, P. C., Dewdney, R. S., Francis, A. J., Godbert, P., and
Butterworth, K. R. (1988). Chromosome aberrations, mitogen-induced
blastogenesis and proliferative rate index in peripheral lymphocytes from 106
control individuals of the UK population. Mutat. Res., 204, 407-420.
Anderson, D., Jenkinson, P. C . , Edwards, A. J., Evans, J. G., and Lovell, D. P.
(1993). Male-mediated F, abnormalities. In : ‘Reproductive Toxicology’, Ed.
M. Richardson. VCH Verlagsgesellschaft, Weinheim, pp. 101-1 16.
Diana Anderson 289
Anderson, D., McGregor, B. D., and Bateman, A. J. (1983). Dominant lethal

mutation assays. Report of the UKEMS sub-committee on guidelines for
mutagenicity testing, pp. 43-1 64.
Anderson, D. and Purchase, I. F. H. (1983). The mutagenicity of food. In: ‘Toxic
Hazards in Food’, Eds D. M. Conning and A. B. G. Lansdown. Croom Helm
Ltd., London, pp. 145-182.
Anderson, D. and Richardson, C. R. (1981a). Issues relevant to the assessment of
chromosome damage in viuo. Mutat. Res., 90,261-272.
Anderson, D., Richardson, C . R., and Davies, P. J. (1981b). The genotoxic
potential of bases and nucleosides. Mutat. Res., 91, 265-272.
Anderson, D., Richardson, C. R., Purchase, I. F. H., Weight, T. M., and Adams,
W. G. F. (1980). Chromosomal analyses in vinyl chloride exposed workers.
Mutat. Res., 70, 151-162.
Anon. (1991). Early indicators of non-genotoxic carcinogenesis. Mutat. Res., 248,
21 1-376 (Special Issue).
Arlett, C. F. (1977). Mutagenicity testing with V79 Chinese hamster cells.
‘Handbook of Mutagenicity Test Procedures’, Eds B. J. Kilbey, M. Legator,
W. Nichols, and C. Ramel. Elsevier Biomedical Press, North-Holland,
Amsterdam, pp. 175-1 9 1.
Arlett, C. F. and Cole, J. (1988). The role of mammalian cell mutation assays in
mutagenicity and carcinogenicity testing. Mutagenesis, 3, 455 -458.
Ashby, J. (1985). Fundamental structural alerts to potential carcinogenicity or
non-carcinogenicity. Environ. Mutagen., 7, 9 19-92 1.
Ashby, J. (1986a). The prospects of simplified and internationally harmonized
approach to the detection of possible human carcinogens and mutagens.
Mutagenesis, 1, 3-16.
Ashby, J. (1986b). Letter to the Editor, Mutagenesis, 1, 309-317.
Ashby, J. (1988a). The separate identities of genotoxic and non-genotoxic
carcinogens. Mutagenesis, 3, 365-366.
Ashby, J. (1988b). An opinion on the significance of the 19 non-clastogenic
gene-mutagens reported by Tennant et al. (1987). Mutagenesis, 3,463-465.
Ashby, J., de Serres, F. J., Draper, M., Ishidate, M., Jr., Margolin, B. H., Matter,
B. E., and Shelby, M. D. (1985). Evaluation of short term tests for
carcinogens. Report of the International Programme on chemical safety’s
collaborative study on in vitro assays, Vol. 5. Elsevier-Science Publishers,
Amsterdam, Oxford, New York.
Ashby, J., de Serres, F. J., Shelby, M. D., Margolin, B. H., Ishidate, M., and
Becking, G. (1988). Evaluation of short term tests for carcinogens. Report of
the International Programme on Chemical Safety’s Collaborative Study on in
vitro assay, Vol. 1; and in vivo assays, Vol. 1 1. Cambridge University Press on
behalf of the World Health Organisation.
Ashby, J. and Liegibel, U. (1992). Transgenic mouse mutation assays. Potential
for confusion of genotoxic and non-genotoxic carcinogenesis. A proposed
solution. Environ. Mol. Mutagen., 20, 145-147.
Ashby, J. and Liegibel, U. (1993). Dosing regimens for transgenic animals
mutagenesis assays. Environ. Mol. Mutagen., 21, 2, 120-121.
Ashby, J. and Mirkova, E. (1987). The activity of MNNG in the mouse bone
marrow micronucleus assay. Mutagenesis, 2, 199-205.
Ashby, J. and Morrod, R. S. (1991). Detection of human carcinogens. Nature
(London), 352, 185-186.
Ashby, J. and Paton, D. (1993). The influence of chemical structure on the extent
and sites of cacinogenesis for 522 rodent carcinogens and 55 different human
carcinogen exposures. Mutut. Res., 286, 3-74.
Ashby, J. and Purchase, 1. F. H. (1988). Reflections on the declining ability of the
Sulmonellu assay to detect rodent carcinogens as positive. Mutat. Res., 205,
51-58.
Ashby, J. and Purchase, 1. F. H. (1992). Non-genotoxic carcinogens: an extension
of the perspective provided by Perera. Environ. Hlth. Perspect., 98,223-226.
Ashby, J. and Styles, J. A. (1978). Factors influencing mutagenic potency in vitro.
Nuture (London), 274, 20-22.
Ashby, J. and Tennant, R. W. (1988). Chemical structure, Salmonellu mutagenic-
ity and extent of carcinogenicity as indicators of genotoxic carcinogenesis
amoung 222 chemicals tested in rodents by the US NCI/NTP. Mutut. Res.,
204, 17-1 15.
Ashby, J. and Tennant, R. W. (1991). Definitive relationships among chemical
structure, carcinogenicity and mutagenicity for 30 1 chemicals tested by the
US NTP. Mutut. Res., 257, 229-306.
Ashby, J., Tennant, R. W., Zeiger, E., and Stasiewicz, S. (1989). Classification
according to chemical structure, mutagenicity to Salmonella and level of
carcinogenicity of a further 42 chemicals tested for carcinogenicity by the US
National Toxicology Program. Mutat. Res., 223, 73-103.
Auerbach, C. (1 975). The effects of six years of mutagen testing on our attitude to
the problems posed by it. Mutut. Res., 33, 3-10.
Auerbach, C. (1977). The role of Drosophilu in mutagen testing. I n : ‘Topics in
Toxicology, Mutagenesis in Sub-mammalian Systems. Status and Signifi-
cance’, Ed. G. Paget. MTP Press Ltd., Lancaster, pp. 13-20.
Auerbach, C., Robson, J. M., and Carr, J. G. (1947). The chemical production of
mutations. Science, 105, 243.
Auletta, A. and Ashby, J. (1988). Workshop on the relationship between
short-term test information and carcinogenicity; Williamsburg, Virginia,
January 20-23, 1987. Environ. Mol. Mutagen., 11, 135-145.
Baker, B. S., Boyd, J. B., Carpenter, A. T. C., Green, M. M., Nguyen, T. D.,
Tipoll, P., and Smith, P. D. (1976). Genetic controls of meiotic recombination
and somatic DNA metabolism in Drosophilu melunogaster. Proc. Nutl. Acud.
Sci. ( U S A ) , 73, 4140-4144.
Barrett, J. C. (1985). Cell culture models of multistep carcinogens. In : ‘Age-related
factors in carcinogenesis’, Eds A. Likhacker, V. Arisimov, and R. Mon-
tesano. IARC, Lyon, pp. 181-202.
Barrctt, J. C., Crawford, B. D., Grady, D. L., Hester, L. D., Jones, P. A., Benedict,
W. F., and Ts’o, P. 0. P. (1977). Temporal acquisition of enhanced
fibrinolytic activity by Syrian hamster embryo cells following treatment with
benzo(u)pyrene. Cancer Res., 37, 3815-3823.
Barrett, J. C., Kakunaga, T., Kuroki, T., Neubert, D., Troske, J. E., Vasilieu, J.
M., Williams, G. M., and Yamasaki, H. (1986). Mammalian Cell Transform-
ation in Culture. In: ‘Long and short term test assays for carcinogens. A
critical appraisal’, Eds R. Montesano, H. Bartsch, H. Vainio, J. Wilbourn,
and H. Yamasaki. IARC Sci Publ. No. 83, Lyon, pp. 267-286.
Barret, J. C. and Lamb, P. W. (1985). Tests with the Syrian hamster embryo cell
transformation assay. In : ‘Progress in Mutation Research. Volume 5.
Evaluation of Short-Term Tests for Carcinogens. Report of the International
Diana Anderson 29 1
Programme on Chemical Safety’s Collaborative Study on In Vitro Assays’,

Eds J. Ashby, F. J. de Serres, M. Draper, M. Ishidate Jr., B. H. Margolin, B.
E. Matter, and M. D. Shelby. Elsevier, Amsterdam, pp. 623-628.
Bateman, A. J. (1977). The dominant lethal assay in the mouse. In: ‘Handbook of
Mutagenicity Test Procedures’, Eds B. J. Kilbey, M. Legator, W. Nichols,
and C. Ramel. Elsevier Biochemical Press, North-Holland, Amsterdam,
pp. 225-255.
Bateman, A. J. and Epstein, S. S. (1971). Dominant lethal mutations in mammals.
In : ‘Chemical Mutagens, Principles and Methods for their Detection’.
Plenum Press, New York, Vol. 2, pp. 541-568.
Bartsch, H., Malaveille, C., and Montesano, R. (1975). Human rat and mouse
liver-mediated mutagenicity of vinyl chloride in S . typhimurium strains. Int. J .
Cancer, 15, 429-437.
Benigni, R. and Giuliani, A. (1988). Statistical exploration of four major
genotoxicity data bases: an overview. Environ. Mol. Mutagen., 12, 75-83.
Bernstein, L., Gold, L. S., Ames, B. N., Pike, M. C., and Hoel, D. G . (1985). Some
tautologous aspects of the comparison of carcinogenic potency in rats and
mice. Fund. Appl. Toxicol., 5, 79-86.
Berwald, Y. and Sachs, L. (1963). In vitro transformation with chemical
carcinogens. Nature (London), 200, 1 182-1 184.
Berwald, Y. and Sachs, L. (1965). In vitro transformation of normal cells to
tumour cells by carcinogenic hydrocarbons. J . Natl. Cancer Inst., 35,
64 1-66 1.
Bignami, M., Morpugo, C., Pagliani, R., Carere, A., Conte, G., and Di Guiseppe,
G. (1974). Non-disjunction and crossing-over induced by pharmaceutical
drugs in Aspergillus nidulans. Mutat. Res., 26, 159-1 70.
Bond, D. J. (1976). A system for the study of meiotic non-disjunction using
Sodaria breuicollis. Mutat. Res., 39, 21 3-220.
Bootman, J. and Kilbey, B. J. (1983). Recessive lethal mutations in Drosophilu.
The report of the UKEMS sub-committe on guidelines for mutagenicity
testing. Part 1.
Bradley, M. O., Bhuyan, B., Francis, M. C., Langenbach, R., Peterson, A., and
Huberman, E. (1981). Mutagenesis by chemical agents in V79 Chinese
hamster cells: A review and analysis of literature. A report of the US
Environmental Protection Agency Gene-Tox program. Mutat. Res., 87,
81-142.
Brewen, J. C., Payne, H . S., Jones, K. P., and Preston, R. J. (1975). Studies on
chemically-induced dominant lethality. 1. The cytogenetic basis of MMS-
induced dominant lethality in post-meiotic male germ cells. Mutat. Res., 33,
239-250.
Bridges, B. A. (1974). The three-tier approach to mutagenicity screening and the
concept of the radiation equivalent dose. Mutat. Res., 26, 335-340.
Brinkworth, M. H., Yardley-Jones, A., Edwards, A. J., Hughes, J. A., and
Anderson, D. (1992). A comparison of smokers and non-smokers with
respect to oncogene products and cytogenetic parameters. J. Occup. Med., 34,
1181-1 188.
Brockman, H. E., de Serres, F. J., Ong, T.-M., DeMarini, D. M., Katz, A. J.,
Griffiths, A. J. F., and Stafford, R. S. (1984). Mutation tests in Neurosporu
crussu. A report of the US Environmental Protection Agency Gene-Tox
Program. Mutat. Res., 133, 87-134.
Brockman, H. E. and DeMarini, D. M. (1988). Utility of short-term tests for

genetic toxicity in the aftermath of the NTP’s analysis of 73 chemicals.
Environ. Mol. Mutagen., 11, 421-435.
Bruce, W. R., Varghese, A. J., Furrer, R., and Land, P. C. (1977). A mutagen in
the faeces of normal human. In: ‘Origins of Human Cancer’, Eds H. H. Hiatt,
J. D. Watson, and J. A. Winsten. Cold Spring Harbor Laboratory, New
York, Vol. C, pp. 1641-1646.
Brusick, D. J. (1972). Induction of cyclohexamide resistant mutants in Sacchar-
omyces cerevisiae with N-methyl-N-nitronitrosoguanidineand ICR-70. J.
Bacteriol., 109, 1 134.
Brusick, D. J. (1982). Genetic toxicology. In: ‘Principles and Methods of
Toxicology’, Ed. A. W. Hayes. Raven Press, New York, pp. 223-272.
Brusick, D. J. (1987). Genotoxicity produced in mammalian cell assays by
treatment conditions. Special issue. Mutat. Res., 189.
Brusick, D. J. and Auletta, A. (1 985). Development status of bioassays in genetic
toxicology. A report of Phase I1 of the US Environmental Protection Agency
Gene-Tox Program. Mutat. Res., 153, 1-10.
Brusick, D. J., Gopalan, H. N. B., Heseltine, E., Huismans, J. W., and Lohman, P.
H. M. (1992). Assessing the risk of genetic damage. UNEPIZCPEMC,
Hodder and Stoughton, pp. 1-52.
Brusick, D. J., Simmons, V. F., Rosenkranz, H. S., Ray, V. A,, and Stafford, R. S.
(1980). An evaluation of the Escherichia coli WP2 and WP2 uvrA reverse
mutation assay. Mutat. Res., 76, 169-190.
Buckton, K. E. and Evans, H. J. (1973). Methods for the analysis of human
chromosome aberrations. WHO Publication, Geneva.
Butterworth, B. E. (1990). Consideration of both genotoxic and nongenotoxic
mechanisms in predicting carcinogenic potential. Mutat. Res., 239, 1 17-132.
Butterworth, B. E., Popp, J. A., Conolly, R. B., and Goldsworthy, T. L. (1992).
Chemically induced cell proliferation in carcinogenesis. In : ‘Mechanisms of
Carcinogenesis in Risk Identification’, Eds H. Vainio, P. N. Magee, D. B.
McGregor, and A. J. McMichael. IARC, Lyon, pp. 279-305.
Cachiero, N. L. A,, Russel, L. B., and Swarthout, M. S. (1974). Translocation, the
predominant cause of total sterility in sons of mice treated with mutagens.
Genetics, 76, 73-9 1.
Carter, C. 0. ( 1 977). The relative contribution of mutant genes and chromosome
aberrations in genetic ill health. In : ‘Progress in Genetic Toxicology’, Eds D.
Scott, B. A. Bridges, and F. H. Sobels. Elsevier Biomedical Press, North-
Holland, Amsterdam, pp. 1-14.
Chu, E. G. Y., Brimer, P., Jacobson, K. B., and Merriam, E. V. (1976).
Mammalian cell genetics. 1. Selection and characterisation of mutations
auxotrophic for 1-glutamhe or resistance to 8-azaguanine in Chinese hamster
cells in vitro. Genetics, 62, 359-377.
cihak, R and Vontorkova, M. (1990). Activity of acrylamide in single-, double-,
and triple-dose mouse bone marrow micronucleus assays. Mutat. Res., 234,
125- 127.
Clayson, D B. and Arnold, D. L. (1991). ICPEMC Publication No. 19. The
classification of carcinogens identified in the rodent bioassay as potential risk
to humans: What type of substance should be tested next? Mutat. Res., 257,
9 1-1 06.
Cleaver, J. E. (1 975). Methods for studying repair of DNA damaged by physical
Diana Anderson 293
and chemical carcinogens. In: ‘Methods of Cancer Research’, Ed. H. Busch.

Academic Press, New York, Vol. 9, pp. 123-165.
Cleaver, J. E. (1977). Methods for studying excision repair of DNA damaged by
physical and chemical mutagens. In : ‘Handbook of Mutagenicity Test
Procedures’, Eds B. J. Kilbey, M. S. Legator, W. Nichols, and C. Ramel.
Elsevier Biomedical Press, North-Holland, Amsterdam, pp. 19-47.
Clive, D. W. (1973). Recent developments with the L5178TK heterozygote
mutagen assay system. Environ. Hlth. Perspect., 6 , 119-125.
Clive, D. W., Flamm, W. G., Machesko, M. R., and Bernheim, N. H. (1972).
Mutational assay system using the thymidine kinase locus in mouse
lymphoma cells. Mutat. Res., 16, 77-87.
Clive, D. W., McCuen, R., Spector, J. F. S., Piper, C., and Mavournin, K. H.
(1983). Specific gene mutations in L5 178Y cells in culture. A report of the US
Environmental Protection Agency Gene-Tox Program. Mutat. Res., 155,
225-25 1.
Clive, D. W. and Spector, J. F. S. (1975). Laboratory procedure for assessing
specific locus mutations at the TK locus in cultured L5178Y mouse
Cohen, M. M. and Hirschhorn, K. (1971). Cytogenetic studies in animals. In:
‘Chemical Mutagens, Principles and Methods for their Detection’, Ed. A.
Hollaender. Plenum Press, New York, Vol. 2, pp. 515-534.
Clode, S. A. and Anderson, D. (1988a). High performance liquid chromato-
graphic analysis of bases and nucleosides in mouse testes following in vivo
thymidine administration. Mutat. Res., 200, 63-66.
Clode, S. A. and Anderson, D. (1988b). Germ and somatic cell abnormalities
following in vivo administration of thymidine and adenine. Mutat. Res., 200,
249-254.
Clode, S. A., Anderson, D., and Gangolli, S. D. (1986). Studies with unbalanced
precursor pools in vivo. In : ‘Genetic Toxicology of Environmental Chemicals,
Part A, Basic principles and mechanisms of action’, Eds C. Ramel, B.
Lambert, and J. Magnusson, Liss, New York, pp. 551-561.
Cole, J. and Arlett, C. F. (1976). Ethyl methanesulphonate mutagenesis with
L5 178Y mouse lymphoma cells. A comparison of ouabain, thioguanine, and
excess thymidine resistance. Mutat. Res., 24, 507-531.
Cole, J., Arlett, C. F., and Green, M. H. L. (1976). The fluctuation test as a more
sensitive system for determining induced mutation in L5 178Y mouse
Cole, J., Fox, M., Garner, R. C., McGregor, D. B., and Thacker, J. (1983). Gene
mutation assay in cultured mammalian cells. The report of the UKEMS
sub-committee on guidelines for mutagenicity testing. Part 1, pp. 65-102.
Cole, J., Green, M. H. L., James, S. E., Henderson, L., and Cole, H. (1988). A
further assessment of factors influencing measurements of thioguanine-
resistant mutant frequency in circulating T-lymphocytes. Mutat. Res., 204,
493 -507.
Cole, J., McGregor, D. B., Fox, M., Thacker, J., Garner, R. C. (1990). Gene
mutation assays in cultured mammalian cells. In : ‘Basic Mutagenicity Tests.
UKEMS Recommended Procedures’, Ed. D. J. Kirkland. Cambridge
University Press, Cambridge, pp. 87-1 14.
Combes, R. D., Anderson, D., Brooks, T. M., Neale, S., and Venitt, S. (1984).
Mutagens in urine, faeces and body fluids. The report of the UKEMS
sub-committee on guidelines for mutagenicity testing. Part 2, pp. 203-243.
Commoner, B., Vithayathal, A. J., and Henry, J. 1. (1974). Detection ofmetabolic

carcinogen intermediates in urine of carcinogen fed rats by means of bacterial
mutagenesis. Nature (London), 249, 850.
Constantin, M. J. and Nilan, R. A. (1982a). Chromosome aberration assays in
barley (Hordeurn vulgare). A report of the US Environmental Protection
Agency Gene-Tox Program. Mutat. Res., 99, 13-36.
Constantin, M. J. and Nilan, R. A. (1982b). The chlorophyll-deficient mutant
assay in barley (Hordeurn vulgure). A report of the US Environmental
Protection Agency Gene-Tox Program. Mutat. Res., 99, 37-49.
Constantin, M. J. and Owens, E. T. (1982). Introduction and perspectives of plant
genetic and cytogenetic assays. A report of the US Environmental Protection
Crouch, E., Wilson, R., and Zeise, L. (1987). Tautology or not tautology? J .
Toxicol. Environ. Hlth., 20, 1-10.
Crow, J. (1973). Impact of various types of genetic damage and risk assessment.
Environ. Hlth. Perspect., 6, 1-5.
Dean, B. J. and Senner, K. R. (1977). Detection of chemically induced mutation in
Chinese hamsters. Mutat. Res., 46, 403-405.
DeMars, R. (1974). Resistance of cultured human fibroblasts and other cells to
purine and pyrimidine analogues in relation to mutagenesis detection. Mutat.
R ~ s .24,
, 335-364.
de Serres, F. J. and Ashby, J. (Eds) (1981). Evaluation of short term tests for
carcinogenesis. In : ‘Chemical Mutagens. Principles and Methods for their
Detection’. Elsevier Biomedical Press, North-Holland, Amsterdam.
de Serres, F. J. and Malling, H. V. (1971). Measurement of recessive lethal damage
over the entire genome and at two specific loci in the Ad-3 region of a two
component heterokaryon of Neurospora crussa. In : ‘Chemical Mutagens,
Principles and Methods for their Detection’, Ed. A. Hollaender. Plenum
Press, New York, London, Vol. 2, pp. 31 1-341.
Dewdney, R. S., Lovell, D. P., Jenkinson, P. C., and Anderson, D. (1986).
Variation in sister chromatid exchange using 106 members of the general UK
population. Mutat. Res., 171, 43-51.
DHSS (1981), No. 24. Guidelines for the testing of chemicals for mutagenicity,
Report on Health and Social Subjects. Committee on mutagenicity of
chemicals in food, consumer products and the environment. Department of
Health and Social Security (8 July, 1981), London, HMSO.
DOH (1989), No. 35. Guidelines for the testing of chemicals for mutagenicity.
Committee on mutagenicity of chemicals in food, consumer products and the
environment. Department of Health and Social Security (April, 1989),
London, HMSO.
DOH (1991), No. 42. Guidelines for the evaluation of chemicals for carcinogenic-
ity. Committee on carcinogenicity of chemicals in food, consumer products
and the environment. Department of Health and Social Security (July, 199l),
London, HMSO.
DiPaolo, J. A., Donovan, P. J., and Nelson, R. L. (1969a). Qualitative studies of in
vitro transformation by chemical carcinogens. J . Natl. Cancer Inst., 42,
867-876.
DiPaolo, J. A., Nelson, R. L., and Donovan, P. J. (1969b). Sarcoma producing cell
clones derived from clones transformed in vitro by benzo(u)pyrene. Science,
167, 917-918.
DiPaolo, J. A., Nelson, R. L., and Donovan, P. J. (1971). Morphological,
Diana Anderson 295
oncogenic and karyological characteristics of Syrian hamster embryo cells

transformed in vitro by carcinogenic polycyclic hydrocarbons. Cuncer Res.,
31, 11 18-1 127.
DiPaolo, J. A., Nelson, R. L., and Donovan, P. J. (1972). In vitro transformation
of Syrian hamster embryo cells by diverse chemical carcinogens. Nature
(London), 235, 278-280.
Ducatman, A., Hirshhorn, K., and Selikoff, I. J. (1975). Vinyl chloride exposure
and human chromosome aberrations. Mutat. Res., 31, 163-1 68.
Dunkel, V. (1979). Collaborative studies on the Salmonella microsome mutagenic-
ity assay. J Assoc. Anal. Chem., 62, 874-882.
Dunkel, V. C., Pienta, R. J., Sivak, A., and Traul, K. A. (1981). Comparative
neoplastic transformation responses of Balb/3T3 cells, Syrian hamster
embryo cells, and Rauscher murine leukemia virus-infected Fischer 344 rat
embryo cells to chemical carcinogens. J . Nutl. Cancer Inst., 67, 1303-1315.
Dunkel, V. C., Rogers, C., Swierenga, S. H. H., Brillinger, R. L., Gilman, J. P. W.,
and Nestmann, E. R. (1991). Recommended protocols based on a survey of
current practice in genotoxicity testing laboratories. 111. Cell transformation
in C3H/lOT; mouse embryo cell, BALB/c3T3 mouse fibroblast and Syrian
hamster embryo cell cultures. Mutat. Res., 246, 285-300.
Dunkel, V., Zeiger, E., Brunswick, D., McCoy, E., McGregor, D., Mortelmans,
K., Rosenkranz, H. S., and Simmons, V. F. (1985). Reproducibility of
microbial mutagenicity assays. 11. Testing of carcinogens and non-carcino-
gens in Salmonella typhimurium and Escherichia coli. Environ. Mutugenesis,
77, (Suppl. 5), 1-248.
Durston, W. and Ames, B. N. (1974). A simple method for detection of mutagens
in urine, studies with the carcinogen 2-acetylaminofluorene. Proc. Nutl. Acud.
Sci. ( U S A ) , 71, 737-741.
Ehling, U. H. (1970). Evaluation of presumed dominant skeletal mutations. In:
‘Chemical Mutagenesis in Mammals and Man’, Eds F. Vogel and F.
Rohrborn. Springer-Verlag, Berlin, Heidelberg, New York, pp. 162-1 66.
Ehling, U. H., Machemer, L., Buselmaier, W., Dycka, J., Frohbert, H.,
Kratochvilova, J., Lang, R., Lorke, D., Muller, K. D., Pehn, J., Rohrborn,
G., Rollm, R., Schulze-Schencking, M., and Wiseman, H. (1978). Standard
protocol for the dominant lethal test on male mice set up by the work group
‘Dominant Lethal Mutations of the ad hoe committee chemogenetics’. Arch.
Toxicol., 39, 173-185.
Ehrenberg, L. (1971). Higher plants. In: ‘Chemical Mutagens, Principles and
Methods for their Detection’, Ed. A. Hollaender. Plenum Press, New, York,
London, Vol. 2, pp. 365-386.
Ennever, F. K., Noonan, T. J., and Rosenkranz, H. S. (1987). The predictivity of
animal bioassays and short term genotoxicity tests for carcinogenicity and
non-carcinogenicity to humans. Mutagenesis, 2, 73-78.
Ennever, F. K. and Rosenkranz, H. S. (1989). Application of the carcinogenicity
prediction and battery selection method to recent National Toxicology
Program short-term test data. Environ. Mol. Mutagen., 13, 332-338.
Ennever, F. K. and Rosenkranz, H. S. (1987). Prediction of carcinogenic potency
by short term genotoxicity tests. Mutugenesis, 2, 39-44.
Epstein, S. S. and Rohrborn, G. (1970). Recommended procedures for testing
genetic hazard from chemicals based on the induction of dominant lethal
mutation in mammals. Nature (London), 230, 264.
Evans, E. P. and Phillips, R. J. S. (1975). Inversion heterozygosity and the origin

of XO daughters of Bpa/ + female mice. Nature (London), 256, 40-41.
Evans, H. J. (1 976). Cytological methods for detecting chemical mutagens. In :
Hollaender. Plenum Press, New York, London, Vol. 4, p. 1.
Fahrig, R. (1974). Development of host-mediated mutagenicity tests. 1. Differen-
tial response of yeast cells injected into testes of rats and peritoneum of mice
and rats to mutagens. Mutat. Res., 26, 29-36.
Fahrig, R. (1977). Host-mediated mutagenicity tests-yeast systems. Recovery of
yeast cells out of testes, liver, lung and peritoneum of rats. In : ‘Handbook of
Mutagenicity Test Procedures’, Eds B. J. Kilbey, M. Legator, W. Nichols,
and C . Ramel. Elsevier Biomedical Press, North-Holland, Amsterdam, pp.
135-147.
Fahrig, R. (1978). The mammalian spot test: A sensitive in vivo method for the
detection of genetic alterations in somatic cells of mice. In: ‘Chemical
Mutagens, Principles and Methods for their Detection’, Eds A. Hollaender
and F. de Serres. Plenum Press, New York, Vol. 5, pp. 152-176.
Farmer, P. B. (1982). Monitoring of human exposure to carcinogens. Chem. Br.,
18, 790-794.
Farmer, P. B., Bailey, S. M., Gorf, M., Tornqvist, S., Osterman-Golkar, A.,
Kautianen, and Lewis-Enright, D. P. (1986). Monitoring human exposure to
ethylene oxide by the determination of haemoglobin adducts using gas
chromatography-mass spectrometry. Carcinogenesis, 7, 637-640.
Festing, M. F. W. and Wolff, G. L. (1976). Quantitative characteristics of
potential value in studying mutagenesis. Genetics, 92, No. 1, Part 1, S. 173.
Felton, S. J. and Knize, M. G. (1991). Occurrence, identification, and bacterial
mutagenicity of heterocyclic amines in cooked food. Mutut. Res., 259,
205-2 17.
Fielding, R. (1987). Letter to the Editor. Mutagenicity Testing. Mutagenesis, 2,
31 5-316.
Fischer, G. A., Lee, S. Y., and Calabresi, P. (1974). Detection of chemical
mutagens using a host-mediated assay L5 178Y mutagenesis system. Mutat.
Res., 16, 501-511.
Fitzgerald, D. J., Piccoli, C., and Yamasaki, H. (1989). Detection of non-
genotoxic carcinogens in the BALB/c3T3 cell transformation/mutation assay
system. Mutagenesis, 4(4), 286-291.
Fitzgerald, D. J. and Yamasaki, H. (1990). Tumour promotion: models and assay
systems. Terutogen. Carcinogen. Mutagen., 10, 89.
Fleig, I. and Theiss, A. M. (1977). External chromosome studies undertaken on
persons and animals with VC Illness (Abstract 2nd Int. Conf. Environ.,
p. 219). Mutat. Res., 52, 187.
Fox, M. and Anderson, D. (1974). Induced thymidine and 5-iodo-2-deoxyuridine
resistant clones of mouse lymphoma cells. Mutat. Res., 25, 89-105.
Fox, M., Boyle, J. M., and Fox, B. W. (1976). Biological and biochemical
characterisation of purine analogue resistant clones of V79 Chinese hamster
cells. Mutat. Res., 35, 289-310.
Francis, A. J., Anderson, D., Evans, J. G., Jenkinson, P. C., and Godbert, P.
(1990). Tumours and malformations in the adult offspring of cyclophos-
phamide-treated and control male rats. Preliminary Communication, 229,
239-246.
Diana Anderson 297
Funes-Cravioto, F. B,, Lambert, B., Linsten, L., Ehrenberg, L., Natarajan, A. T.,
and Osterman-Golkar, S. (1975). Chromosome aberrations in workers
exposed to vinyl chloride. Lancet, i, 459.
Garner, R. C. and Kirkland, D. J. (1986). Reply-Letter to the Editor.
Gatehouse, D. G., Rowland, I. R., Wilcox, P., Callander, R. D., and Forster, R.
(1990). Bacterial mutation assays in basic mutagenicity tests, UKEMS
Recommended Procedures. Ed. D. K. Kirkland, Cambridge University Press,
Cambridge, pp. 13-61.
Gatehouse, D. G. and Tweats, D. J. (1986). Letter to the Editor. Mutagenesis, 1,
307-309.
Generoso, W. M., Bishop, J. B., Gosslee, D. G., Newell, G. W., Sheu, C.-J., and
von Halle, E. (1980). Heritable translocation test in mice. A report of the US
191-21 5.
Generoso, W. M., Cain, K. T., Huff, S. W., and Gosslee, D. G. (1977). Heritable
translocation in mice. In : ‘Chemical Mutagens, Principles and Methods for
their Detection’, Ed. A. Hollaender, Plenum Press, New York, Vol. 5, p. 55.
George, E., Wootton, A. K., and Gatehouse, D. G. (1990). Micronucleus
induction by azobenzene and 1,2-dibromo-3-chloropropanein the rat:
evaluation of a triple-dose protocol. Mutat. Res., 234, 129-1 34.
Gold, L. S., Slone, T. H., Manley, N. B., and Bernstein, L. (1993). Comparison of
target organs of carcinogenicity for mutagenic and non-mutagenic chemicals.
Mutat. Res., 286, 75-100.
Gold, L. S., Slone, T. H., Stern, B. R., and Bernstein, L. (1991). Target organs in
chronic bioassays of 533 chemical carcinogens. Environ. Hlth. Perspect., 93,
233-246.
Grant, W. F. (1982). Chromosome aberration assays in Allium. A report of the US
273-291.
Green, M. H. L. and Lowe, J. E. (1992). Dietary vitamin C modulates the response
of human lymphocytes to ionising radiation: Studies with the COMET Assay.
‘6th International Conference of Environmental Mutagens’, Melbourne.
Abstract. 27-1.
Green, M. H. L. and Muriel, W. J. (1976). Mutagen testing using tryp + reversion
in Escherichia coli. Mutat. Res., 38, 3-32.
Green, M. H. L., Muriel, W. J., and Bridges, B. A. (1976). Use of a simplified
fluctuation test to detect low levels of mutagens. Mutat. Res., 38, 33-42.
Green, S. (199 1). The search for molecular mechanisms of non-genotoxic
carcinogens. Mutat. Res., 248, 371-374.
Green, S., Auletta, A., Fabricant, J., Kapp, R., Manandhar, M., Sheu, C.-J.,
Springer, J., and Whitfield, B. (1985). Current status of bioassays in genetic
toxicology-the dominant lethal assay. A report of the US Environmental
Gulati, D. K., Wojciechowski, J. P., and Kaur, P. (1990). Comparison of single-,
double-, or triple-exposure protocols for the rodent bone marrow/peripheral
blood micronucleus assay using 4-aminobiphenyl and treosulphan. Mutat.
Res., 234, 135-1 39.
Gunz, D., Shephard, S. E., and Lutz, W. K. (1993). Can non-genotoxic
carcinogens be detected with the lac 1 transgenic mutation assay? Environ.
Mol. Mutagen., 21.
Hansteen, I., Hillestad, L., Thiis-Evensen, E., and Heldaas, S. S. (1978). Effect of
vinyl chloride in man. A cytogenic follow-up study. Mutat. Res., 51,172-178.
Haseman, J. K., Margolin, B. H., Shelby, M. D., Zeiger, E., and Tennant, R. W.
(1988). Do short-term tests predict rodent carcinogenicity? Response.
Science, 241, 1233.
Haseman, J. K. and Seilkop, S. K. (1992). An examination of the association
betwen maximum-tolerated dose and carcinogenicity in 326 long-term studies
in rats and mice. Fund. Appl. Toxicol., 19, 207-213.
Heddle, J. A. (1973). A rapid in vivo test for chromosomal damage. Mutut. Res.,
18, 187-190.
Heddle, J. A., Hite, M., Kirkhart, B., Mavournin, K., MacGregor, J. T., Newell,
G. W., and Salamone, M. F. (1983). The induction of micronuclei as a
measure of genotoxicity. A report of the US Environmental Protection
Agency Gene-Tox Program. Mutat. Res., 123, 61-1 18.
Heidelberger, C., Freeman, A. E., Pienta, R. J., Sivak, A., Bertram, J. S.,Casto, B.
C., Dunkel, V. C., Francis, M. W., Kakunaga, T., Little, J. B., and
Schechtman, L. M. (1983). Cell transformation by chemical agents-a review
and analysis of the literature. A report of the US Environmental Protection
Ho, Y. L. and Ho, S. K. (1981). Screening of carcinogens with the prophage clts
857 induction test. Cancer Res., 41, 532-536.
Hoel, D. G., Haseman, J. K., Hogan, M. D., Huff, J., and McConnell, E. E.
(1988). The impact of toxicity on carcinogenicity studies: implications for risk
assessment. Carcinogenesis, 9, 2045-2052.
Hsie, A. W., Casciano, D. A., Couch, D. B., Krahn, D. F., O’Neill, J. P., and
Whitfield, B. L. (1991). The use of Chinese hamster ovary cells to quantify
specific locus mutation and to determine mutagenicity of chemicals. A report
of the US Environmental Protection Agency Gene-Tox Program. Mutat.
Res., 86, 193-214.
Huberman, E. (1975). Mammalian cell transformation and cell-mediated
mutagenesis by carcinogenic polycyclic hydrocarbons. Mutat. Res., 29,
285-291.
Huberman, E. and Sachs, L. (1966). Cell susceptibility to transformation and
cytotoxicity by the carcinogenic hydrocarbon benzo(a)pyrene. Proc. Natl.
Acad. Sci. USA, 56, 1123-1 129.
Huberman, E. and Sachs, L. (1974). Cell-mediated mutagenesis of mammalian
cells with chemical carcinogens. Int. J . Cancer, 13, 326-333.
Huff, J. E. (1992). Chemical toxicity and chemical carcinogenesis. Is there a causal
connection? A comparative morphological evaluation of 1500 experiments.
In: ‘Mechanisms of Carcinogenesis in Risk Identification’, Eds H. Vainio, P.
N. Magee, D. B. McGregor, and A. J. McMichael. IARC, Lyon, pp.
437-475.
IARC (1987). IARC Monographs on the Evaluation of Carcinogenic Risks to
Humans. Overall Evaluations of Carcinogenicity: An updating of IARC
Monographs Volumes 1 to 42. Supplement 7. p. 55. International Agency for
Research on Cancer, Lyon.
IARC/NCI/EPA Working Group (1985). Cellular and molecular mechanisms of
cell transformation and standardization of transformation assays of esta b-
lished cell lines for the prediction of carcinogenic chemicals: overview and
recommended protocols. Cancer Res., 45, 2395-2399.
Dim a A nde rson 299
IPCSjWHO (1 992). Final Report of the Steering Groups on Naturally Occurring

toxins in plant antigen.
Ishinotsubo, D., Mower, H. F., Setliff, J., and Mandel, M. (1977). The use of REC
bacteria for testing carcinogenic substances. Mutat. Res., 46, 53-62.
Jackson, M. A., Stack, H. F., and Waters, M . D. (1993). The genetic toxicology of
putative non-genotoxic carcinogens. Mutat. Res., 296, 241-277.
Jacob, L. and de Mars, R. (1977). Chemical mutagenesis with diploid human
fibroblasts. In: ‘Handbook of Mutagenicity Test Procedures’, Eds. B. J.
Kilbey, M. Legator, W. Nichols, and C. Ramel. Elsevier Biomedical Press,
North-Holland, Amsterdam, p. 193-220.
Jenkinson, P. and Anderson, D. (1 990). Malformed foetuses and karyotypic
abnormalities in the offspring of cyclophosphamide and ally1 alcohol-treated
male rats. Mutat. Res., 229, 173-184.
Jenkinson, P., Anderson, D., and Gangolli, S. D. (1987). Increased incidence of
abnormal foetuses in the offspring of cyclophosphamide treated mice. Mutat.
Res., 188, 57-62.
Jenssen, D. and Ramel, C. (1980). The micronucleus test as part of a short-term
mutagenicity test program for the prediction of carcinogenicity evaluated by
143 agents tested. Mutat. Res., 75, 191-202.
Jones, C. A., Huberman, E., Callahan, M. F., Tu, A., Halloween, W., Pallota, S.,
Sivak, A., Lubet, R. A., Avcry, M. D., Kouri, R. E., Spalding, J., and
Tennant, R. W. (1988). An interlaboratory evaluation of the Syrian hamster
embryo cell transformation assay using eighteen coded chemicals. Toxic. In
Vitro, 2, 103-116.
Kafer, E., Marshall, P., and Cohen, G. (1976). Well marked strains of Aspergillus
for tests of environmental mutagens. Identification of induced recombination
and mutation. Mutat. Res., 38, 141-148.
Kafer, E., Scott, B. R., Dorn, G. L., and Stafford, R. (1982). Aspergillus nidulans:
Systems and results of tests for chemical induction of mitotic segregation and
mutation. 1 . Diploid and duplication assay systems. A report of the US
1-48.
Kakunaga, T. (1 973). A quantitative system for assay of a malignant transform-
ation by chemical carcinogens using a clone derived from Balb/3T3. Int. J.
Cancer, 12, 463-473.
Kapp, R. W., Picciano, D. J., and Jacobson, C. B. (1979). Y-chromosomal
non-disjunction in dibromochloropropane exposed workmen. Mutat. Res.,
64,47-51.
Kier, L. D. (1988). Comments and perspective on the EPA Workshop on ‘The
relationship between short-term test information and carcinogenicity’. En-
viron. Mol. Mutagen., 11, 147-1 57.
Kier, L. D., Brusick, D. J., Auletta, A. E., von Halle, E. S., Simmons, V. F.,
Brown, M. M., Dunkel, V. C., McCann, J., Mortelmans, K., Prival, M . J.,
Rao, T. K., and Ray, V. A. (1986). The Salmonella typhimurium/mammalian
microsome mutagenicity assay. A report of the US Environmental Protection
Kihlman, B. A. (1971). Root tips for studying the effects of chemicals on
chromosomes. In : ‘Chemical Mutagens, Principles and Methods for their
Detection’, Ed. A. Hollaender. Plenum Press, New York, Vol. 2, pp. 365-514.
Kilian, D. J., Picciano, D. J., and Jacobson, C. B. (1975). Industrial monitoring: a
cytogenetic approach. Ann. N . Y . Acad. Sci., 249, 41 1.

Kirkland, D. (1987). Regulatory aspects of genotoxicity testing. Implications of
germ cell cytogenetic tests in the regulatory process. Mutagenesis, 2, 61-67.
Klopman, G. and Rosenkranz, H. S. (1991). Quantification of the predictivity of
some short-term assays for carcinogenicity in rodents. Mutat. Res., 253,
237 -240.
Knaap, A. G. A. C. and Simmons, J. W. I. M. (1975). A mutational assay system
for L5 178Y mouse lymphoma cells using hypoxanthine guanine phos-
phoribosyl transferase deficiency as a marker. The occurrence of a long
expression time for mutations induced by X-rays and EMS. Mutat. Res., 30,
97-109.
Knudsen, I., Hansen, E. V., Meyer, 0. A., and Poulsen, E. (1977). A proposed
method for the simultaneous detection of germ cell mutations leading to
foetal death/dominant lethality and malformations (Male Teratogenicity) in
mammals. Mutat. Res., 48, 267-270.
Kohler, S. W., Provost, G. S., Fieck, A., Kretz, P. L., Bullock, W. O., Sorge, J. A.,
Putnam, D. L., and Short, J. M. (1991a). Spectra of spontaneous and
mutagen induced mutations in the lac1 gene in transgenic mice. Proc. Natl.
Acad. Sci. U S A , 17, 4707-4716.
Kohler, S. W., Provost, G. S., Fieck, A., Kretz, P. L., Bullock, W. O., Putnam, D.
L., Sorge, J. A., and Short, J. M . (1991b). Analysis of spontaneous and
induced mutations in transgenic mice using a Lambda ZAP/lacl shuttle
vector. Environ. Mol. Mutagen., 18, 316-321.
Kohler, S. W., Provost, G. S., Kretz, P. L., Dycaico, M. J., Sorge, J. A., and Short,
J. M . (1990). Development of a short term in vivo mutagenesis assay. The
effects of methylation on recovery of a Lambda phage shuttle vector from
transgenic mice. Nucleic Acids Res., 18, 3007-301 3.
Kohn, H. I. (1973). H-gene (Histocompatibility) mutations induced by triethy-
lene-melamine in the mouse. Mutat. Res., 20, 235-247.
Koi, M. and Barrett, J. C. (1986). Loss of tumour-suppressive function during
chemically induced neoplastic progression of Syrian hamster embryo cells.
Proc. Natl. Acad. Sci. U S A , 83, 5992-5996.
Krahn, D. F. and Heidelberger, C. (1977). Liver homogenate-mediated mutagen-
esis in Chinese hamster V79 cells by polycyclic 'Aromatic' hydrocarbons and
aflatoxins. Mutat. Res., 46, 27-44.
Kuroki, T. and Matsushima, T. (1987). Performance of short term tests for
detection of human carcinogens. Mutagenesis, 2, 33-37.
Larsen, K. H., Mavournin, K. H., Brash, D., Cleaver, J. E., Hart, R. W., Maher,
V. M., Painter, R. B., and Sega, G. A. (1982). DNA repair assays as tests for
environmental mutagens. A report of the US Environmental Protection
Latt, S. A., Allen, J., Bloom, S. E., Carrano, A., Falke, E., Kram, D., Schneider,
E., Schreck, R., Tice, R., Whitfield, B., and Solff, S. (1981). Sister-chromatid
exchanges: A report of the US Environmental Protection Agency Gene-Tox
Lee, W. R., Abrahamson, S., Valencia, R., von Halle, E. S., Wurgler, F. E., and
Zimmering, S. (1983). The sex-linked recessive lethal test for mutagenesis in
Drosophila melanogaster. A report of the US Environmental Protection
Legator, M S., Bueding, E., Batzinger, R., Connor, T. H., Eisenstadt, E., Farrow,
Dianu Anderson 30 1
M . G., Fiscor, G., Hsie, A., Seed, J., and Stafford, R. S. (1982). An evaluation
of the host-mediated assay and body fluid analysis. A report of the US
3 19-374.
Legator, M. S. and Malling, H. V. (1971). The host-mediated assay, a practical
procedure for evaluating potential mutagenic agents in mammals. In :
Legator, M. S., Palmer, K. A., and Adler, I. A. (1973). A collaborative study of in
vivo cytogenetic analysis. 1. Interpretation of slide preparations. Toxicol.
Appl. Pharmacol., 24, 337.
Legator, M. S., Pullin, T. G., and Connor, T. H. (1977). The isolation and
detection of mutagenic substances in body fluid and tissues of animals and
body fluid of human subjects. In: ‘Handbook of Mutagenicity Test Pro-
cedures’, Eds B. J. Kilbey, M. Legator, w. Nichols, and C. Ramel. Elsevier
Biomedical Press, North-Holland, Amsterdam, pp. 149-1 59.
Legator, M. S., Zimmering, S., and Connor, T. H. (1976). The use of indicator
systems to detect mutagenic activity in human subjects and experimental
animals. In : ‘Chemical Mutagens, Principles and Methods for their Detec-
tion’. Ed. A. Hollaender. Plenum Press, New York, Vol. 4, pp. 171-192.
Leifer, Z., Kada, T., Mandel, M., Zeiger, E., Stafford, R., and Rosenkranz, H . S.
(198 1). An evaluation of tests using DNA repair-deficient bacteria for
predicting genotoxicity and carcinogenicity. A report of the US Environ-
mental Protection Agency Gene-Tox Program. Mutat. Res., 87, 21 1-297.
Leonard, A. (1975). Tests for heritable translocation in male mammals. Mutat.
Res., 31, 291-298.
Leonard, A. (1977). Observations on meiotic chromosomes of the male mouse as a
test of the potential mutagenicity of chemicals. In: ‘Chemical Mutagens,
Principles and Methods for their Detection’, Ed. A. Hollaender, Plenum
Press, New York, Vol. 3, pp, 21-56.
Li, A. P., Aaron, C. S., Auletta, A. E., Dearfield, K. L., Riddle, J . C., Slesinski, R.
S., and Stankowski, Jr., L. F. (1991). An evaluation of the roles of
mammalian cell mutation assays in the testing of genotoxicity. Regulut.
Toxicol. Pharmacol., 14, 24-40.
Li, A. P., Gupta, R. S., Heflich, R. H., and Wassom, J. S. (1988). A review and
analysis of the Chinese hamster ovary/hypoxanthineguanine phosphori bosyl
transferase assay to determine the mutagenicity of chemical agents. A report
of Phase 111 of the US Environmental Protection Agency Gene-Tox Program.
Mutat. Res., 196, 17-36.
Lilly, L. J . , Bahner, B., and Magee, P. N. (1975). Chromosome aberrations
induced in rat lymphocytes by N-nitroso compounds as a possible basis for
carcinogen screen. Nature (London), 258, 61 1-612.
Loprieno, N., Barale, R., Bauer, C., Baroncelli, S., Bronzetti, G., Cammellini, A.,
Cini, A., Corsi, G., Leporini, C., Nieri, R., Mozzolini, M., and Serra, C.
(1974). The use of different systems with yeasts for the evaluation of
chemically-induced gene conversion and gene mutations. Mutat. Res., 25,
197-21 7.
Loprieno, N., Barale, R., von Halle, E. S., and von Borstel, R. C. (1983). Testing
of chemicals for mutagenic activity with Schizosaccharomyces pombe. A
report of the US Environmental Protection Agency Gene-Tox Program.
Mutat. Res., 115, 215-223.

Loprieno, N., Boncristiani, G., Loprieno, G., and Tesoro, M. (1991). Data
selection and treatment of chemicals tested for genotoxicity and carcinogenic-
ity. Environ. Hlth. Perspect., 96, 121-1 26.
Maier, P. and Schmid, W. (1976). Ten model mutagens evaluated. Mutat. Res., 50,
325-338.
Marimuthu, K. M., Sparrow, A H., and Schairer, L. A. (1970). The cytological
effects of space flight factors, vibration, clinostat and radiation on root tip
cells of Tradescantia. Radiat. Res., 42, 105.
Maron, D. and Ames, B. N . (1983). Revised methods for the Salmonella
Matter, B. E. (1976). Problems of testing drugs for potential mutagenicity. Mutat.
Res., 38, 243-258.
Matsushima, T., Sawamura, M., Hara, K., and Sugimura, T. (1976). A safe
substitute for polychlorinated biphenyls as an inducer of metabolic activation
systems. I n : ‘In vitro Metabolic Activation of Mutagenesis Testing’, Eds F. J.
de Serres, J. R. Fout, J. R. Bend, and R. M. Philpot. Elsevier Biochemical
Press, North-Holland, Amsterdam, pp. 85-88.
Mavournin, K. H., Blakey, D. H., Cimino, M. C., Salamone, M. F., and Heddle,
J. A. (1990). The in vivo micronucleus assay in mammalian bone marrow and
peripheral blood. A report of the US Environmental Protection Agency
Mays, J., McAninch, J., Feurs, R. J., Burkhart, J., Mohrenweiser, H., and
Casciano, D. A. (1978). Enzyme activity as a genetic marker to detect induced
microlesions in mammals. Mutat. Res., 58, No. 1, Abstract 58.
McCann, J. and Ames, B. N . (1976). Detection of carcinogens as mutagens in the
Sal~zonella/microsometest: assay of 300 chemicals: Dicussion. Proc. Natl.
Acad. Sci. (USA), 73, 950-954.
McCann, J., Choi, E., Yamasaki, E., and Ames, B. N. (1975a). Detection of
carcinogens as mutagens in the Salmanella/microsome test: assay of 300
chemicals. Proc. Natl. Acad. Sci. (USA), 72, 5 135- 5 139.
McCann, J., Spingarn, N. E., Kobori, J., and Ames, B. N. (1975b). Detection of
carcinogens as mutagens: bacterial tester strains with R factor plasmids. Proc.
Natl. Acad. Sci. (USA), 72, 979-983.
McGregor, D. B. (1992). Chemicals classified by IARC: their potency in tests for
carcinogenicity in rodents and their genotoxicity and acute toxicity. In:
‘Mechanisms of Carcinogenesis in Risk Identification’, Eds H. Vainio, P. N.
Magee, D. B. McGregor, and A. J. McMichael, IARC, Lyon, 323-352.
McGregor, D. and Ashby, J. (1985). Summary report on the performance of the
cell transformation assays. In : ‘Evaluation of short-term tests for carcino-
gens’, Eds J. Ashby, F. de Serres, M. Draper, M. Ishidate, B. H. Margolin, B.
E. Matter, and M. D. Shelby. Elsevier, New York, pp. 103-115.
McKelvey-Martin, V. J., Green, M. H. L., Schmezer, P., Pool-Zobel, B. L., De
Meo, M. P., and Collins, A. (1993). The single cell gel electrophoresis assay
(COMET assay). Mutat. Res., 288, 47-63.
McKusick, V. (1975). ‘Mendelian inheritance in Man’, 4th Edn, John Hopkins
Press, Baltimore.
Metzger, B., Crouch, E., and Wilson, R. (1989). On the relationship between
carcinogenicity and acute toxicity. Risk Analysis, 9, 169-177 (cited in Travis
et al., 1990a).
Diunu Anderson 303
Meuth, M. (1984). The genetic consequences of nucleotide precursor pool

imbalance in mammalian cells. Mutat. Res., 126, 107-1 12.
Meyer, A. L., McGregor, D. B., and Styles, J. A. (1984). In vitro cell
transformation assays. The report of the UKEMS Sub-committee on
Guidelines for Mutagenicity Testing. Part 2, pp. 123-145.
Mirkova, E. (1990). Activity of the human carcinogens benzidine and 2-
naphthylamine in triple- and single-dose mouse bone marrow micronucleus
assays: results for a combined test protocol. Mutat. Res., 234, 161-163.
Mirsalis, J. (1993). Dosing regimes for transgenic animal mutagenesis assays.
Environ. Mol. Mutagen., 21, 118-1 19.
Mitchell, A. D., Casciano, D. A., Meltz, M. L., Robinson, D. E., San, R. H. C.,
Williams, G. M., and von Halle, E. S. (1983). Unscheduled DNA synthesis
tests. A report of the US Environmental Protection Agency Gene-Tox
Mohn, G. (1973). 5-Methyl-tryptophan resistance mutations in Escherichia coli
K- 12. Mutagenic activity of monofunctional alkylating agents, including
organophosphorus insecticides. Mutat. Res., 20, 7-1 5.
Mohn, G. and Ellenberger, J. (1973). Mammalian blood-mediated mutagenicity
tests using a multipurpose strain of Escherichiu coli K-12. Mutat. Res., 19,
187-196.
Mohn, G., Kerklaan, R., and Ellenberger, J. (1984). Methodologies for the direct
and animal mediated determination of various genetic effects in derivatives of
strain 343/113 of E. coli K-12. In: ‘Handbook of Mutagenicity Test
Procedures’, 2nd Edition, Eds B. J. Kilbey, M. Legator, W. Nichols, and
C. Ramel. Elsevier, Amsterdam, pp. 189-214.
Mollet, P. and Wurgler, F. E. (1974). Detection of somatic recombination and
mutations in Drosophila. A method for testing genetic activity of chemical
compounds. Mutat. Res., 25, 421-424.
Moreau, P., Bailone, A., and Devoret, R. (1976). Prophage induction in
Escherichiu coli K-12 enVA uvrB: A highly sensitive test for potential
carcinogens. Proc. Natl. Acad. Sci. ( U S A ) , 73, 3700-3704.
Morita, T., Iwamoto, Y., Shimizu, T., Masuzawa, T., and Yanagihara, Y. (1989).
Mutagenicity tests with a permeable mutant of yeast on carcinogens showing
false-negative in Salmonella assay. Chem. Pharm. Bull., 37, 407-409.
Mortimer, R. K. and Manney, T. R. (1971). Mutation induction in yeast. In:
Miiller, H. J. (1927). Artificial transmutation of the gene. Science, 66, 84-87,
Myhr, B. C . and Caspary, W. J. (1991). Chemical mutagcnesis at the thymidine
kinase locus in L5178Y mouse lymphoma cells: results for 31 coded
compounds in the National Toxicology Program. Environ. Mol. Mutagen.,
18, 51-83.
Nagao, M., Yahagi, T., and Sugimura, T. (1978). Differences in effects of
Norharman with various classes of chemical mutagens and amounts of S-9.
Biochem. Biophys. Res. Commun., 83, 373-378.
Nakamura, J., Suzuki, N., and Okada, S. (1977). Mutagenicity of furylfuramide, a
good preservative tested by using alanine-requiring mouse L5 178Y cells in
vitro and in vivo. Mutat. Res., 46, 355-364.
Natarzjan, A. T., van Bud, P. W., and Raposa, T. (1 978). An evaluation of the use
of peripheral blood lymphocyte systems for assessing cytological effects
induced in vivo by chemical mutagens. In : ‘Mutagen Induced Chromosome

Damage in Man’, Eds H. C. Evans and D. C. Lloyd. Edinburgh University
Press, Edinburgh, pp. 268-276.
Neel, J. V. (1978). Some trends in the study of spontaneous and induced mutation
in man. Genetics, 92, No. 1, Part 1, S25.
Neel, J. V., Kohrenweiser, H., Satoh, C., and Hamilton, B. H. (1979). A
consideration of two biochemical approaches to monitoring human popula-
tions for a change in germ cell mutation rate. In: ‘Genetic Damage in Man
Caused by Environmental Agents’, Ed. K. Berg. Academic Press, New York,
London, pp. 29-47.
Neill, J. P., Brimer, P. A., Machanoff, R., Hirsch, G. P., and Hsie, A. W. (1977). A
quantitative assay of mutation induction at HGPRT locus in Chinese
hamster ovary cells: development and definition of the system. Mutat. Res.,
45, 91-101.
Nesnow, S., Argus, M., Bergman, H., Ohu, K., Firth, C., Helmes, T.,
McGaughey, R., Ray, V., Staga, T. J., Tennant, R., and Weisburger, E.
(1983). Chemical carcinogens: A review and analysis of the literature of
selected chemicals and the establishment of the Gene-Tox carcinogen
data-base. A report of the US Environmental Protection Agency Gene-Tox
Nichols, W. W., Moorhead, P., and Brewen, G. (1973). Chromosome methodolo-
gies in mutation testing. Toxicol. Appl. Pharmacol., 24, 337.
Nikaido, 0. and Fox, M. (1976). The relative effectiveness of 6-thioguanine and
8-azaguanine in selecting resistant mutants from two V79 Chinese hamster
clones in vitro. Mutat. Res., 35, 279-288.
Nilan, R. A. and Vig, B. K. (1976). Plant test systems for detection of chemical
mutagens. In: ‘Chemical Mutagens, Principles and Methods for their
Detection’, Ed. A. Hollaender. Plenum Press, New York, Vol. 4, pp. 143-1 70.
Nute, P. E., Wood, N. E., Stammatoyannopoulos, G., Olivery, C., and Failkaw,
P. J. (1976). The Kenya form of hereditary persistence of foetal haemoglobin:
Structural studies and evidence for homogeneous distribution of haemo-
globin F using fluorescent F antibodies. Br. J. Huernatol., 32, 55.
OECD (1983). Organisation for Economic and Cultural Development. Guide-
lines for testing of chemicals. Genetic Toxicology Guidelines.
Oficial Journal of the European Communities. Council Directive 18 September
1979 Amending for the Sixth Time Directive 67/548/EEC on the approxi-
mation of the Laws, Regulations, and Administrative Provisions relating to
the classification, packaging, and labelling of dangerous substances.
Ong, T. M. and de Serres, F. J. (1972). Mutagenicity of chemical carcinogens in
Neurospora crassa. Cancer Res., 32, 1890-1 893.
Oshiro, Y., Piper, C. E., Balwierz, P. S., and Soelter, S. G. (199 1). Chinese hamster
ovary cell assays for mutation and chromosome damage: data from
non-carcinogens. J. Appl. Toxicol., 11, 167-1 77.
Parodi, S., Malacarne, D., and Taningher, M. (1991). Examples of uses of
databases for quantitative and qualitative correlation studies between
genotoxicity and carcinogenicity. Environ. Hlth. Perspect., 96, 61-66.
Parry, J. M. (1977a). The use of yeast cultures for the detection of environmental
mutagens using a fluctuation test. Mutat. Res., 46, 165-176.
Parry, J. M. (1977b). The detection of chromosome non-disjunction in the yeast
Saccharomyces cerevisiae. In : ‘Progress in Genetic Toxicology’, Eds D. Scott,
Diunu Anderson 305
B. A. Bridges, and F. H. Sobels. Elsevier Biomedical Press, North-Holland,

Amsterdam, pp. 223-29.
Parry, J. M., Brooks, T. M., Mitchell, I. de G., and Wilson, P. (1984). Gentoxicity
studies using yeast cultures. Report of the UKEMS Sub-committee on
Guidelines of Mutagenicity Testing. Part 2, pp. 27-62.
Parry, J. M. and Zimmerman, F. K. (1976). The detection of monosomic colonies
produced by mitotic chromosome non-disjunction in the yeast Sacchuromy-
ces cerevisiae. Mutat. Res., 36, 49-66.
Pearson, C. M. and Styles, J. A. (1984). Effects of hydroxyurea consistent with the
inhibition of repair synthesis in UV irradiated HeLa cells. Cancer Lett., 21,
247-252.
Perry, P. and Evans, H. J. (1975). Cytological detection of mutagen-carcinogen
exposure by sister chromatid exchange. Nature (London), 258, 121-1 25.
Perry, P. E., Henderson, L., and Kirkland, D. J. (1984). Sister chromatid exchange
in cultured cells. In: ‘Report of the UKEMS Sub-committee on Guidelines of
Mutagenicity Testing’. Part 2, pp. 89-123.
Pfeifer, G. P., Drouin, R., and Holmqvist, G. P. (1993). Detection of DNA
adducts at the DNA sequence level of ligation-mediated PCR. Mutat. Res.,
288, 39-46.
Phillips, D. H., Hemminki, K., Alhonen, A,, Hewer, A., and Grover, P. L. (1988).
Monitoring occupational exposure to carcinogens. Detection by 32P-post-
labelling of aromatic DNA adducts in white blood cells from iron foundry
workers. Mutat. Res., 204(3), pp. 531-541.
Phillips, J. C. and Anderson, D. (1993). Predictive toxicology Structure Activity
relationships and carcinogens. Occup. Hlth. Rev., 41, 27-30.
Pienta, R. J., Poiley, J. A., and Lebherz, W. B., 111 (1977). Morphological
transformation of early passage golden Syrian hamster embryo cells derived
from cryopreserved primary cultures as a reliable in vitro bioassay for
identifying diverse carcinogens. Int. J. Cancer, 19, 642-655.
Plewa, M. J. (1982). Specific-locus mutation assays in Zea mays. A report of the
US Environmental Protection Agency Gene-Tox Program. Mutat. Res., 99,
317-337.
Poirier, L. A. and de Serres, F. J . (1979). Initial National Cancer Institute studies
on mutagenesis as a pre-screen for chemical carcinogens: an appraisal. J.
Natl. Cancer Inst., 62, 919-926.
Popp. R. A., Hrisch, G. P., and Bradshaw, B. S. (1979). Amino acid substitution:
Its use in detection and analysis of genetic variants. Genetics, 92, No. 1 , Part 1,
s.39.
Preston, R. J., Au, W., Bender, M. A., Brewen, J. G., Carrano, A. V., Heddle, J.
A., McFee, A. F., Wolff, S., and Wassom, J. S. (1981). Mammalian in vivo and
in vitro cytogenetic assays. A report of the US Environmental Protection
Prival, M. J. and Dunkel, V. C. (1989). Reevaluation of the mutagenicity and
carcinogenicity of chemicals previously identified as ‘false positives’ in the
Salmonella typhimurium mutagenicity assay. Environ. Mol. Mutagen., 13,
1-24.
Propping, P., Rohrborn, G., and Buselmaier, W. (1972). Comparative investiga-
tions on the chemical induction of point mutations and dominant lethal
mutations in mice. Mol. Gen. Genet., 117, 197.
Provost, G., Kretz, P. L., Hammer, R. T., Matthews, C. D., Rogers, B. J.,
Lundberg, K. S., Dycaico, M. J., and Short, J. M. (1993). Transgenic systems

for in viva mutation analysis. Mutat. Res., 288, 123-131.
Pueyo, C. A. (1978). A forward mutation assay using 1-Arabinose sensitive strain.
Abstracts (Ab- American Environmental Mutagen Society.
Purchase, I. F. H. (1980). Appraisal of the merits and shortcomings of tests of
mutagenic potental. In : ‘Mechanisms of Toxicity and Hazard Evaluation’,
Eds B. Holmstedt, R. Lauwerys, M. Mercier, and M. Roberfroid. Elsevier
Biochemical Press, North-Holland, Amsterdam, pp. 105-1 19.
Purchase, I. F. H., Longstaff, E., Ashby, J., Styles, J. A., Anderson, D., Lefevre, P.
A., and Westwood, F. R. (1978a). Evaluation of six short-term tests for
detecting organic chemical carcinogens and recommendations for their use.
Nature (London), 263, 624-627; Br. J . Cancer, 37, 873-903.
Purchase, I. F. H., Richardson, C. F., Anderson, D., Paddle, G. M., and Adams,
W. G . F. (1978b). Chromosomal analysis in vinyl-chloride exposed workers.
Mutat. Res., 57, 325-334.
Ray, V. A., Kier, L. D., Kannan, K. L., Haas, R. T., Auletta, A. E., Wassom, J. S.,
Nesnow, S., and Waters, M. D. (1987). An approach to identifying specialised
batteries of bioassays for specific causes of chemicals and analysis using
mutagenicity and carcinogenicity relationships and phylogenetic concord-
ance and discordance patterns. 1. Comparison and analysis of the overall
data base. A report of the US Environmental Protection Agency Gene-Tox
Reznikoff, C.A., Bertram, J. S., Brankow, D. W., and Heidelberger, C. (1973).
Quantitative and qualitative studies of chemical transformation of cloned
C34 mouse embryo cells sensitive to post-confluence inhibition of cell
division. Cancer Res., 33, 3239-3249.
Richold, M.,Chandley, A., Ashby, J., Gatehouse, D. G., Bootman, J., and
Henderson, L. (1990). In vivo cytogenetics assays, UKEMS Recommended
Procedures, Ed. D. J. Kirkland. Cambridge University Press, Cambridge,
pp. 115-141.
Roderick, J. H. (1979). Chromosomal-inversions-Studies of mammalian
mutagenesis. Genetics, 92, No. 1, Part 1, S. 121.
Roper, J. A. (1971). Aspergillus. In: ‘Chemical Mutagens, Principles and Methods
for their Detection’, Ed. A. Hollaender. Plenum Press, New York, Vol. 2, pp.
343-365.
Rowland, I. R., Rubery, E. D., and Walker R. (1984). Bacterial assays for
mutagens in food. In: ‘Report of the UKEMS Sub-committee on Guidelines
of Mutagenicity Testing’, Part 2, pp. 173-202.
Russell, L. B. (1979). In vivo somatic mutations in the mouse. Genetics, 92,No. 1,
Part 1, S.153.
Russell, L. B., Aaron, C. S., de Serres, F. J., Generoso, W. M., Kannan, K. L.,
Shelby, M. P., Spring, J., and Voytek, P. (1984). Evaluation of mutagenicity
assays for purposes of genetic risk assessment. A report of the US
143-1 57.
Russell, L. B., Selby, P. B., von Hallk, E., Sheridan, W., and Valcovic, L. (1981a).
The mouse specific-locus test with agents other than radiation: interpretation
of data and recommendations for future work. Mutat. Res., 86, 329-354.
Russell, L. B., Selby, P. B., von Hallk, E., Sheridan, W., and Valcovic, L. (1981b).
Use of the mouse spot test in chemical mutagenesis: interpretation of past
Diana Anderson 307
data and recommendations for future work. Mutat. Res., 86, 355-359.
Russell, W. L. (1951). Specific locus mutations in mice. Cold Spring Harbor
Laboratory. Quant. B i d . , 16, 237.
Saedler, H., Gullon, A., Fiethen, L., and Starlinger, P. (1968). Negative control of
galactose operon in Escherichia coli. Mol. Gen. Genet., 102, 79.
Salamone, M., Heddle, J., Stuart, E., and Katz, M. (1980). Towards an improved
micronucleus test. Studies on 3 agents, Mitomycin C, cyclophosphamide and
dimethylbenzanthracene. Mutat. Res., 35, 347-370.
Schantel, A. P. and Sankaranarayanan, K. (1976). Evaluation and re-evaluation
of genetic radiation hazards in man. Mutat. Res., 35, 341-370.
Schiestl, R. H. (1989). Non-mutagenic carcinogens induce intrachromosomal
recombination in yeast. Nature (London), 337, 285-288.
Schiestl, R. H., Gietz, R. D., Mehta, R. D., and Hastings, P. J. (1989).
Carcinogens induce intrachromosomal recombination in yeast. Carcinogen-
esis, 10, 1445-1455.
Schmid, W. (1973). Chemical mutagen testing on in vivo somatic cells. Agents
Actions, 3, 77-85.
Schmid, W. (1976). The micronucleus test for cytogenetic analysis. In: ‘Chemical
Mutagens, Principles and Methods for their Detection’, Ed. A. Hollaender.
Plenum Press, New York, Vol. 4, pp. 31-53.
Schmid, W., Arakaki, D. T., Breslau, N. A., and Culbertson, J. C. (1971).
Chemical mutagenesis. The Chinese hamster bone marrow as an in vivo test
system. Humangenetik, 11, 103-1 18.
Scott, D., Danford, N. D., Dean, B. J., Kirkland, D. J., and Richardson, C. R.
(1983). In vitro chromosomal aberratidn assays. In: ‘Report of the UKEMS
sub-committee on Guidelines of Mutagenicity Testing,’ Part 1, pp. 41-64.
Scott, D., Dean, B. J., Danford, N. D., and Kirkland, D. J. (1990). Metaphase
chromosome aberration assays in vitro. In : ‘Basic Mutagenicity Tests
UKEMS Recommend Procedures’, Cambridge University Press, Cambridge,
pp. 62-86.
Scott, B. R., Dorn, G. L., Kafer, E., and Stafford, R. (1982). Aspergillus nidulans:
Systems and results of tests for induction of mitotic segregation and
mutation. 11. Haploid assay systems and overall response of all systems. A
Mutat. Res., 98, 49-94.
Scott, D., Galloway, S. M., Marshall, R. R., Ishidate, M., Brusick, D., Ashby, J.,
and Myhr, B. C. (1991). Genotoxicity under extreme culture conditions.
Mutat. Res., 257, 147-206.
Searle, W. G. (1975). The specific locus test in the mouse. Mutat. Res., 31,
277-290.
Searle, W. G. (1979). The use of pigment loci for detecting reverse mutations in
somatic cells of mice. Arch. ToxicoZ., 38, 105-108.
Selby, P. B. and Selby, P. R. (1977). Gamma-ray induced dominant mutations
that cause skeletal abnormalities in mice. Mutat. Res., 43, 357-376.
Shelby, M. D. (1988). The genetic toxicity of human carcinogens and its
implications. Mutat. Res., 204, 3-1 5 .
Shelby, M, D., Erexson, G. L., Hook, G. J., and Tice, R. R. (1993). Evaluation of a
three-exposure mouse bone marrow micronucleus protocol: results with 49
chemicals. Environ. Mol. Mutagen., 21, 160-179.
Shelby, M. D., Gulati, D. K., Tice, R. R., and Wojciechowski, J. P. (1989). Results
of tests for micronuclei and chromosomal aberrations in mouse bone marrow

cells with the human carcinogens 4-aminobiphenyl, treosulphan, and mel-
phalan. Environ. Mol. Mutagen., 13, 339-342.
Siebert, D. (1973). Induction of mitotic conversions of Saccharomyces cerevisiae
by lymph fluid and urine of cyclophosphamide-treated rats and human
patients. Mutat. Res., 21, 202.
Skopek, T. R., Liber, J. L., Krowleski, J. J., and Thilly, W. G. (1978). Quantitative
forward mutation assays in Salmonella typhimurium using 8-azaguanine
resistance as a genetic marker. Proc. Natl. Acad. Sci. ( U S A ) , 75, 410-41 1.
Sobels, F. H. (1977). Some thoughts on the evaluation of environmental
mutagens. From the Address presented to the 7th Annual EMS Meeting in
Atlanta, Georgia.
Sora, S. and Magni, G. E. (1988). Induction of meiotic chromosomal malsegrega-
tion in yeast. Mutat. R e x , 201, 375-384.
Sorsa, M., Ojajarui, A,, and Salomaa, S. (1990). Cytogenetic surveillance of
workers exposed to genotoxic chemicals: preliminary experience from a
prospective cancer study in a cytogenetic cohort. Teratogen. Carcinogen.
Mutagen., 10, 21 5-221.
Sorsa, M., Wilbourn, J., and Vainio, H. (1992). Human cytogenetic damage as a
predictor of cancer risk. In: ‘Mechanisms of Carcinogenesis in Risk
Identification’, Eds H. Vainio, P. N. Magee, D. B. McGregor, and A. J.
McMichael. IARC, Lyon, pp. 543-554.
Speck, W. T., Santella, R. M., and Rosenkranz, H. S. (1978). An evaluation of the
prophage induction (inductest) for the detection of potential carcinogens.
Mutat. Res., 54, 101-104.
Stetka, E. G. and Wolff, G. (1976). Sister chromatid exchange as an assay for
genetic damage by mutagen-carcinogens: In vivo test for compounds
requiring metabolic activation. Mutat. Res., 41, 333-342.
Stich, H. F., San, R. H. C., Lam, P., Koropatnick, J., and Lo, L. (1977).
Unscheduled DNA synthesis of human cells as a short term assay for
chemical carcinogens. In: ‘Origins of Human Cancer’, Eds H. H. Hiatt, J. D.
Watson, and J. A. Winsten. Cold Spring Harbor Laboratory, New York, Vol.
C, pp. 1499-1 5 12.
Sugimura, T. (1978). Let’s be scientific about the problems of mutagens in cooked
foods. Mutat. Res., 55, 149-152.
Sugimura, T., Kawachi, T., Matsushima, T., Nagao, M., Sato, S., and Yahai, T.
(1977). A critical review of sub-mammalian systems for mutagen detection.
In: ‘Progress in Genetic Toxicology’, Eds D. Scott, B. A. Bridges, and F. H.
Sobels. Elsevier Biochemical Press, North-Holland, Amsterdam, pp. 25-140.
Sugimura, T., Yahagi, T., Nagao, M., Takuichi, M., Kawacki, T., Hara, K.,
Yamasaki, E., Matsushima, T., Hashimoto, Y., and Okada, M. (1976).
Validity of mutagenicity testing using microbes as a rapid screening method
for environmental carcinogens. In : ‘Screening Tests in Chemical Carcino-
gens’, Eds R. Montesano, H. Bartsch, and T. Tomatis. IARC, Lyon, 12,
81-101.
Swann, P. F. and Magee, P. N. (1968). Nitrosamine-induced carcinogenesis. The
alkylation of nucleic acids of the rat by N-methyl-N-nitrosourea, dimethyl-
nitrosamine dimethyl sulphate, and methyl methanesulphonate. Biochem.
J., 110, 39-47.
Swierenga, S. H. H. and Yamasaki, H. (1992). Performance of tests for cell
Diana Anderson 309
transformation and gap-junction intercellular communication for detecting

nongenotoxic carcinogenic activity. In : ‘Mechanisms of Carcinogenesis ir,
Risk Identification’, Eds H. Vainio, P. N. Magee, D. B. McGregor, and A. J.
McMichael. IARC, Lyon, pp. 165-193.
Szentesi, I., Hornyak, E., Unvary, G., Czeizel, A., Bognar, Z., and Timor, M.
(1976). High rate of chromosomal aberrations in PVC workers. Mutat. Res.,
37, 313-316.
Tang, T. and Friedman, M. A. (1977). Carcinogen activation of human liver
enzymes in the Ames mutagenicity test. Mutat. Res., 46, 387-394.
Tanooka, J. (1977). Development and applications of Bacillus subtilis test system
for mutagens involving DNA repair deficiency and suppressible auxotrophic
mutations. Mutat. Res., 48, 367-370.
Te-Hsui, Ma. (1982a). Vicia cytogenetic test for environmental mutagens. A
Mutat. Res., 99,293-302.
Te-Hsui, Ma. (1982b). Tradescantia cytogenetic tests (root-top mitosis, pollen
mitosis, pollen mother-cell mitosis). A Report of the US Environmental
Tennant, R. W. and Ashby, J. (1991). Classification according to chemical
structure, mutagenicity to Salmonella and level of carcinogenicity of a further
39 chemicals tested for carcinogenicity by the US National Toxicology
Tennant, R. W., Elwell, M. R., Spalding, J. W., and Griesemer, R. A. (1991).
Evidence that toxic injury is not always associated with induction of chemical
carcinogenesis. Mol. Carcinogen., 4, 420-440.
Tennant, R. W., Margolin, B. H., Shelby, M. D., Zeiger, E., Haseman, J. K.,
Spalding, J., Caspary, W., Resnick, M., Stasiewicz, S., Anderson B., and
Minor R. (1987). Prediction of chemical carcinogenicity in rodents from in
vitro genetic toxicity assays. Science, 236, 935-941.
Thilly, W. G., de Luca, J. G., Hoppe, I. V. H., and Penmann, B. W. (1976).
Mutation of human lymphoblasts by methylnitrosourea. Chem. Biol. Inter-
act., 99, 293-302.
Tice, R. R., Erexson, G. L., Hilliard, C. J., Huston, J. L., Boehm, R. M., Gulati,
D., and Shelby, M. D. (1990). Effect of treatment protocol and sample time
on the frequencies of micronucleated polychromatic erythrocytes in mouse
bone marrow and peripheral blood. Mutagenesis, 5, 3 13-32 1.
Tijo, J. W. and Whang, J. (1 962). Chromosome preparation of bone marrow cells
without prior in vitro culture of in vivo colchicin administrations. Stain
Technol., 37, 17-20.
Topham, J. C. (1 980). Chemically-induced transmissible abnormalities in sperm
head shape. Mutat. Res., 70, 109-114.
Topham, J. C., Albanese, R., Bootman, J., Scott, D., and Tweats, D. (1983). In:
‘Report of the UKEMS Sub-committee on Guidelines of Mutagenicity
Testing’, Part 1, pp. 119-142.
Tornqvist, M., Osterman-Golkar, S., Kautianen, A., Naslund, M., Calleman, C .
J., and Ehrenberg, L. (1988). Methylations in human haemoglobin. Mutat.
Res., 204, 521-529.
Travis, C. C . ,Richter Pack, S. A., Saulsbury, A. W., andYambert, M. W. (1990a).
Prediction of carcinogenic potency from toxicological data. Mutat. Res., 241,
21-36.
3 10 Genetic Toxicology
Travis, C. C., Saulsbury, A. W., and Richter Pack, S. A. (1990b). Prediction of

cancer potency using a battery of mutation and toxicity data. Mutagenesis, 5,
21 3-219.
Travis, C. C., Wang, L. A., and Waehner, M. J. (1991). Quantitative correlation of
carcinogenic potency with four different classes of short-term test data.
Trosko, J. E. (1988). A failed paradigm: carcinogenesis is more than mutagenesis.
Tu, A., Hallowell, W., Pallotta, S., Sivak, A., Lubet, R. A., Curren, R. D., Avery,
M. D., Jones, C., Sedita, B. A., Huberman, E., Tennant, R., Spalding, J., and
Kouri, R. E. (1986). An interlaboratory comparison of transformation in
Syrian hamster embryo cells with model and coded chemicals. Environ.
Mutagen., 8, 77-98.
Tweats, D. J., Bootman, J., Combes, R. D., Green, M. H. L., and Watkins, P.
(1984). Assays for DNA repair in bacteria. In: ‘Report of the UKEMS
sub-committee on Guidelines of Mutagenicity Testing’, Part 11, pp. 5-27.
Valcovic, L. R. and Malling, H. V. (1973). An approach to measuring germinal
mutations in the mouse. Environ. Hlth. Perspect., 6, 201-206.
Valencia, R., Abrahamson, S., Lee, W. R., von Hall&,E. S., Woodruff, R. C . ,
Wurgler, F. E., and Zimmering, S. (1984). Chromosome mutation tests for
mutagenesis in Drosophila rnelanogaster. A report of the US Environmental
Van’t Hof. J. and Schairer, L. A. (1982). Tradescantia assay system for gaseous
mutagens. A report of the US Environmental Protection Agency Gene-Tox
Venitt, S. and Crofton-Sleigh (1981). Mutagenicity of 42 coded compounds in a
bacterial assay using Escherichia coli and Salmonella typhimurium. In :
‘Evaluation of Short Term Test for Carcinogens, Progress in Mutation
Research’, Vol. 11, Eds F. J. de Serres and J. Ashby. Elsevier Biochemical
Press, North-Holland, Amsterdam, pp. 351-360.
Venitt, S., Forster, R. C., and Longstaff, E. (1983). Bacterial mutation assays. In:
‘Report of the UKEMS Sub-committee on Guidelines of Mutagenicity
Testing’, Part I, pp. 5-40.
Vig, B. K. (1982). Soybean [Glycinemax (L.) merrill] as a short term assay for the
study of environmental mutagens. A report of the US Environmental
Vogel, E. (1976). The function of Drosophila in genetic toxicology testing. In:
Vogel, E. (1977). Identification of carcinogens by mutagen testing in Drosophila.
The relative reliability for the kinds of genetic damage measured. In: ‘Origins
of Human Cancer’, Eds H. H. Hiatt, J. D. Watson, and J. A. Winsten. Cold
Spring Harbor Laboratory, New York, Vol. C, pp. 1483-1497.
Vogel, E. (1987). Evaluation of potential mammalian genotoxins using
Drosophila: the need for a change in a test strategy. Mutagenesis, 2,161-171.
Wakabayashi, K., Ushiyama, H., Takahashi, M., Nukaya, H., Kim, S. B., Horse,
M., Ochiai, M., Sugimura, T., and Nagao, M. (1993). Exposure to
heterocyclic amines. Environ. Hlth. Perspect., 99, 129-1 33.
Waters, M. D., Stack, H. F., Jackson, M. A., and Bridges, B. A. (in press). Hazard
identification-efficiencyof short-term tests in identifying germ cell mutagens
Diana Anderson 31 1
and putative nongenotoxic carcinogens. Environ. Hlth. Perspect.

Waters, R., Ashby, J., Barrett, R. H., Coombes, R. D., Green, M. H. L., and
Watkins, P. (1984). Unscheduled DNA synthesis. In: ‘Report of the UKEMS
Sub-committee on Guidelines of Mutagenicity Testing’, Part 2, p. 63-89.
Weston, A. (1993). Physical methods for the detection of carcinogen-DNA
adducts in humans. Mutat. Res., 288, 19-29.
Whittaker, S. C., Zimmermann, F. K., Dicus, B., Piegorsch, W. W., Resnick, M.
A., and Fogel, S . (1990). Detection of induced mitotic chromosome loss in
Saccharomyces cerevisiae-interlaboratory assessment of 12 chemicals.
Mutat. Res., 241, 225-242.
Wild, D. (1973). Chemical induction of streptomycin-resistant mutations in
Escherichia coli. Dose and mutagenic effect of dichlorvos and methyl
methanesulphate. Mutat. Res., 19, 33-41.
Williams, G. M., Mori, H., and McQueen, C. A. (1989). Structure-activity
relationships in the rat hepatocyte DNA-repair test for 30 chemicals. Mutat.
Res., 221, 263-286.
Wurgler, F. E., Sobels, F. H., and Vogel, E. (1977). Drosophila as an assay system
for detecting genetic changes. In: ‘Handbook of Mutagenicity Test Pro-
cedures’, Eds B. J. Kilbey, M. Legator, w. Nichols, and C. Ramel. Elsevier
Biochemical Press, North-Holland, Amsterdam, pp. 335-373.
Wyrobek, A. J. and Bruce, W. R. (1975). Chemical induction of sperm
abnormalities in mice. Proc. Natl. Acad. Sci. ( U S A ) , 72, 4425-4429.
Wyrobek, A. J., Gordon, L. A., Burkhart, J. G., Francis, M. W., Kapp, R. W. Jr.,
Letz, G., Malling, H. V., Topham, J. C., and Whorton, M . D. (1983a). An
evaluation of the mouse sperm morphology test and other sperm tests in
non-human mammals. A report of the US Environmental Protection Agency
Wyrobek, A. J., Gordon, L. A., Burkhart, J. G., Francis, M. W., Kapp, R. W. Jr.,
Letz, G., Malling, H. V., Topham, J. C., and Whorton, M. D. (1983b). An
evaluation of human sperm as indicators of chemically induced alterations
of spermatogenic function. A report of the US Environmental Protection
Wyrobek, A. J., Watchmaker, G., and Gordon, L. (1984). An evaluation of sperm
tests and indicators of germ cell damage in men exposed to chemical or
physical agents. Teratogen, Carcinogen. Mutagen., 4, 83-107.
Young, S. S. (1988). Do short-term tests predict rodent carcinogenicity? Science,
241, 1232-1233.
Zeiger, E. (1987). Carcinogenicity of mutagens: Predictive capability of the
Salmonella mutagenesis assay for rodent carcinogenicity. Cancer Res., 47,
1287-1 296.
Zeiger, E., Haseman, J. K., Shelby, M. D., Margolin, B. H., and Tennant, R. W.
(1990). Evaluation of four in vitro genetic toxicity tests for predicting rodent
carcinogenicity: confirmation of earlier results with 41 additional chemicals.
Environ. Mol. Mutagen, 16, (Suppl. 18), 1-4.
Zeiger, E. and Pagano, D. A. (1989). Mutagenicity of the human carcinogen
treosulphan in Salmonella. Environ. Mol. Mutagen., 13, 343-346.
Zeise, L., Wilson, R., and Crouch, E. (1984). Use of acute toxicity to estimate
carcinogenic risk. Risk Analysis, 4, 187-199 (cited in Travis et al., 1990a).
Zimmermann, F. K. (1975). Procedures used in the induction of mitotic
recombination and mutation in the yeast Saccharomyces cerevisiae. Mutat.

Res., 31, 71-86.
Zimmermann, F. K., Mayer, V. W., Scheel, I., and Resnick, M. A. (1985).
Acetone, methyl ethyl ketone, ethyl acetate, acetonitrile and other polar
aprotic solvents are strong inducers of aneuploidy in Saccharomyces
cerevisiae. Mutat. Res., 149, 339-351.
Zimmermann, F. K., von Borstel, R. C., von Hall&,E. S., Parry, J. M., Siebert, D.,
Zetterberg, G., Barale, R., and Loprieno, N. (1984). Testing of chemicals for
genetic activity with Saccharomyces cerevisiae: a report of the US Environ-
mental Protection Agency Gene-Tox Program. Mutat. Res., 133, 199-214.
CHAPTER 14
Molecular Toxicology
PAUL RUMSBY
1 Introduction
The techniques of gene cloning, nucleic acid manipulation, and detection
known as molecular biology have infiltrated most branches of the
biological and medical sciences. Toxicology has in general been slow in
utilising this technology but the last 5 years have increasingly seen the
application of basic knowledge of gene structure and regulation to answer
questions in this field. Much of this technologyis now common knowledge
among undergraduates in the biological sciences so rather than explain it
in depth, the basic ‘tools’will be outlined and examples given of their use in
toxicology. For those readers wishing for greater depth of knowledge on
specific techniques various good text books and reviews have been
suggested.
Due to tremendous advances in defining the genetic events in cancer,
many of the best examples for toxicologists involve the assessment of
chemically-induced carcinogenesis.
2 Techniques of Molecular Biology

A Gene Cloning
Gene cloning is the insertion of a DNA sequence into a vector, which can
then be introduced into a host such as a bacterium for propagation (Figure
1). The host-vector system used depends on the size of DNA to be cloned
and whether there is a need for expression of the gene. A list of the many
and varied vector systems now available is beyond the scope of this
chapter but a perusal of the catalogues of biotechnology companies such
as Clontech, Stratagene, and Promega will give a flavour of the technology
involved.
313
314
- Molecular Toxicology
Fragment of DNA
Vector
l Ligation
Recombinant DNA molecule
I Transformation of host
C-g)
O%O
00
I
el@)
00
Multiilicati of host and plasmid
p j )
00
00
I Extraction of plasmid
Figure 1 Basic principles of gene cloning. Fragment of foreign D N A and plasmid

vector are cut with an appropriate restriction enzyme and then ligated. The
recombinant D N A is transported into the host cell, in this case a bacterium, where
there is multiplication of the recombinant molecule and division of the host cells
Paul Rumsby 315
The vectors are usually based on plasmids for smaller fragments (up to
5-10 kb), bacteriophage lambda, and a combination of the two, called a
cosmid (a plasmid with cohesive ends for efficient bacteriophage-like
infection) for larger fragments (50 kb). A newer system, based on the Yeast
Artificial Chromosome (YAC) for the isolation of DNA of up to 1 Mb,
has proved vital to the Human Genome Mapping Project. The vector with
its DNA insert will replicate autonomously within the host to produce
multiple copies of the original DNA sequence.
The donor DNA for cloning can be isolated from the genome of any
source or from complementary DNA (cDNA) synthesised from RNA by
the action of the enzyme, reverse transcriptase.
A collection of clones representative of a population of cDNA or
genomic DNA is called a library from which specific sequences can be
isolated.
B Restriction Enzymes
The cloning of DNA was made possible by the isolation of a number of
DNA modifying enzymes mainly from bacteria which enabled the cutting
and joining of DNA. Restriction endonucleases, more than 100 of which
are commercially available, cut DNA molecules at specific nucleotide
sequences (typically 4 or 6 base-pairs). For example, the restriction
enzyme, EcoRl will only cut DNA at the sequence GAATTC which
occurs in the mammalian genome, on average, once every 4096 base-pairs.
SGAATTC3’ G AATTC
3’ C T TAAG 5‘ CTTAA -k G
This enzyme gives ‘overhanging’ ends which are the same in all
EcoR1-treated DNA and can be easily joined in the presence of the
enzyme, DNA ligase to construct recombinant DNA molecules. Most
vectors now incorporate a number of common restriction enzyme
recognition sites. Thus the vector and donor DNA are cut by a specific
restriction enzyme, ligated by DNA ligase, and this mix is used to
transform a recipient bacterium. The vector carries an antibiotic resistance
gene so growth of the bacteria in the antibiotic confirms the presence of
plasmid (Figure 1).
This is a simple outline of gene cloning. Further guides to its intricacies
can be found elsewhere (e.g. Berger and Kimmel, 1987) including the
molecular biologist’s ‘bible’, ‘Molecular Cloning-A Laboratory Man-
ual’ (Sambrook et al., 1989). The expansion of this manual from its single
volume first edition in 1982 (Maniatis et al., 1982) to its present 3-volume
form is, if anything, an understatement of the growth of techniques in the
field.
Restriction enzymes are also useful for the detection of changes in DNA
sequence (polymorphism) leading to gain or loss of restriction recognition
sites (restriction fragment length polymorphism, RFLP). Examples of the
use of RFLPs will be outlined later in the chapter.
316 Molecular Toxicology
C Detection of DNA and RNA

Over the last 20 years the chief means of analysing nucleic acids have been
the Southern blot (Southern et al., 1975) for DNA and the Northern blot
for RNA (Alwine et al., 1977). The basis for these two techniques is very
similar and is outlined in Figure 2.
The probes used for the detection of DNA or RNA can be cloned DNA,
RNA, or short synthesised sequences specific for the gene of interest.
Typically these are radiolabelled using 32P(a p-emitter of short half-life,
14 days). The labelling of the probe can be by a number of methods
including random priming (Feinberg and Vogelstein, 1983) or endlabel-
ling. Recently, there has been much work on non-radioactive labelling
techniques using chemiluminescence or colour changes in an attempt to
circumvent the short half-life and potential radiation hazard of 32P.
Separate nucleic acids by agarose gel electrophoresis

a-
-0
-0
-0
+
-0
-0
1 Transfer to nylon membrane
/-\
btottingpaper - , -weight
- nylon membrane
d
wick
\
1
Hybridise with labelled probe
Wash and autoradiograph

Figure 2 Basic principles of Southern and Northern blotting. D N A can be cut with
restriction enzymes before separation while R N A is separated by electrophoresis
through a denaturing gel. DNAIRNA is then transferred to a nylon membrane which
is incubated with a labelled probe and washed under controlled conditions.
Autoradiography indicates where the labelled probe has bound to a complementary
sequence on the blot.
Paul Rumsby 317
The conditions under which DNA sequences will anneal to a comple-

mentary strand is a science in itself (Young and Anderson, 1985). In
practice hybridisation of two strands of DNA takes place at one set of
conditions and separate conditions govern the separation of this duplex so
that only the ‘best fit’ remains annealed.
D The Polymerase Chain Reaction (PCR)

The mid-1980s saw the development of PCR as a method for the
amplification of a specific region of DNA (Saiki et al., 1985). This basic
technique has proved to be the most important contribution to molecular
biology since gene cloning was first described. Briefly, PCR consists of a
cycle of events (Figure 3). Firstly, the double-strands of DNA are
separated by heating. Next, primers, usually oligonucleotides of 20-30
base-pairs, designed to complement sequences either side of the region of
Target sequence
3‘ 5‘
I I
II
5
” 3’
DefBatUat5al
Extension
I I,
.
I
Figure 3 Thepolymerase chain reaction as described in the test. The second cycle of
primer annealing defines the 3’ end so the PCR product is the length of the target
sequence
interest are annealed to the single-stranded DNA at an appropriate

temperature. Finally, the action of Taq polymerase in the presence of
excess deoxynucleotides leads to the formation of a new complementary
strand of DNA commencing at double-stranded DNA where the primers
have annealed to the template. The DNA is then heated again to separate
the new and old strands and a further cycle of amplification commences.
25 -40 cycles of strand separation, primer annealing, and elongation lead
to an amplification of up to a million-fold of the region of interest.
These techniques can be used in very many ways to facilitate various
aspects of gene cloning and the study of gene regulation. Of potential
importance to toxicologists is the detection of very low levels of gene
expression. This technique, usually called reverse transcriptase-PCR
(RT-PCR), involves the production of complementary DNA (cDNA)
from cellular RNA samples and the subsequent amplification of cDNA by
PCR, thus increasing the sensitivity of the assay. A further method for the
detection of DNA mutations of potential importance in toxicology is the
use of small samples of tissues such as scrapes from archival paraffin-fixed
sections (Shibata et al., 1988). Examples of the use of PCR will be given
later in the chapter. A number of books describing the method and its
applications have been produced (Ehrlich, 1989; Innis et al., 1990)
although new uses are continually being described.
3 Applications of Molecular Biology to Toxicology

Although the techniques of molecular biology have now been utilised in all
aspects of toxicology, it is in the field of chemical carcinogenesis that its
weight has been most felt.
A Metabolism of Chemicals
Many chemicals require enzymatic activation to become carcinogenic.
The microsomal P450 oxygenases often metabolise chemicals to DNA
damaging (genotoxic) electrophilic intermediates. These enzymes are part
of a seemingly ever-expanding gene superfamily, divided into 13 families
with 65 different protein sequences (for a review see Gonzalez, 1989). The
genes encoding many of these members have been cloned. The gene for
aryl hydrocarbon hydroxylase (AHH now called CYPlA1) is a good
example of how molecular biology has elucidated gene structure and
regulation.
Figure 4 shows the structure of the CYPlAl gene in humans. In
prokaryotes such as bacteria, an individual gene consists of a single coding
sequence whereas in eukaryotes such as mammals, the DNA of each gene
usually consists of exon (coding) sequences interspersed with intron
(non-coding) sequences. These non-coding sequences are spliced from the
primary RNA transcript leaving an mRNA which is ultimately translated
to form the functional protein. The CYPl A1 gene consists of seven exons
Paul R urnsby 319
Exon 1 2 3 45 6 7
Figure 4 Map of the human CYPIAI gene showing exons as black boxes with
introns in between. The binding sites f o r regulatory products include BTE-basic
transcriptase element and XRE-xenobiotic response element. ATG and TAG mark
the start andfinish of the coding region; GT and AG, the boundaries of introns, and
A A T A A , the signal for the poly A tail. Msp 1 indicates the polymorphic site
described in the text
and spans 63 11 base-pairs (Jaiswal et al., 1985) while the mRNA is 2596
base-pairs long and encodes a protein of 5 12 amino acids (MW58 151). The
gene sequence contains a series of signals, for instance, to mark the start
(ATG) and finish of the gene (TAG) and a sequence to mark the addition
of a poly (A) tail to the mRNA. All six of the introns begin with GT and
end with AG.
Sequencing of the 5’ end of the gene (known as upstream: mRNA is
synthesised 5’ + 3’ so the 5’ end retains a phosphate) reveals a series of
sequences which are conserved in a number of species. Some of these are
common to the majority of genes such as the TATA box promoter lying
20-30 base-pairs from the site of transcriptional initiation. However, a
number of other sequences are binding sites for proteins which control the
transcription of the gene. How the presence of toxic chemicals affects the
proteins which regulate genes such as CYPlAl is currently the subject of
much research and one to which I will return later in the Chapter.
About 20 years ago it was suggested that high inducibility of the
CYPlAl enzyme was an important risk factor in lung cancer. Induction
was more frequently observed in lung cancer patients than those with
benign disease implying that these patients could more readily activate the
carcinogens in cigarette smoke (McLemore et al., 1990).
The human CYPlAl gene shows a sequence change (RFLP) revealed
by the restriction enzyme, Msp 1 (Figure 5). Japanese patients
homozygous for the rarer Msp 1 form of the gene (i.e. both copies of the
gene have this uncommon sequence) have been shown to be susceptible to
the type of lung cancer associated with smoking (Kawajiri et al., 1990;
Nakachi et al., 1991). This homozygous form appears to be rare in
Caucasian populations (Tefre et al., 1991).
Another member of the P450 family, debrisoquine hydroxylase
(CYP2D subfamily) also has mutant forms, detectable by RFLP analysis,
which are associated with altered ability of humans to metabolise certain
Figure 5 Agarose gel electrophoresis of products from PCR of the 3'jlanking

region of the human CYPIAI gene. The band of 340 base-pairs (bp)represents the
common sequence of the gene while the bands at 200 bp and 140 bp represent an
uncommon sequence recognised by the presence of an M s p 1 restriction site (RFLP).
Thegel shows PCR analysis of DNAfrom the white blood cells of 10 volunteers. Four
have one copy (allele) of the common sequence and one copy of the uncommon
sequence. There are no volunteers homozygous for the uncommon sequence. These
would have no band at 340 bp.
The ladder of bands at each site of the gel are molecular weight markers
drugs and chemicals. There is now some evidence that certain forms of the
gene product which lead to poor metabolism of the model substrate,
debrisoquine, may have some protective effects against cigarette smoking-
induced lung cancer (Gough et al., 1990).
It is now clear that from the study of the structure and regulation of
genes such as the P450 family, we can develop fairly simple procedures
(PCR of the relevant regions of the gene and restriction enzyme digestion)
to look at the genetic susceptibility of individuals to environmental
chemicals such as cigarette smoke. Such differencesin individual suscepti-
bility to hazardous chemicals may be a major variable in the assessment of
risk (Idle, 1991).
B Risk and Hazard Assessment

The detection of potential chemical hazards and the assessment of their
risk to an exposed workforce and the general population, particularly in
the field of carcinogenesis, has been a preoccupation of toxicologists for
many years. The use of molecular biology to uncover fundamental genetic
Paul Rumsby 32 1
events during the development of human cancer will lead to improved

model systems and tests for toxicologists. At present, carcinogenic risk is
assessed by in vitro cell mutation assays and rodent bioassay. Both of these
components will be improved by recent discoveries using molecular
biology. A continuing problem lies in the interpretation of tumours found
in rodents after treatment with chemicals which are found to be
non-mutagenic (non-genotoxic) in bacterial and/or mammalian cell
mutation assays. Discovery of the mechanisms involved in the develop-
ment of tumours induced by these non-genotoxic carcinogens should
enable a better interpretation of their importance in risk assessment.
C Genetic Events in Human Cancer

Oncogenes
It was first shown by Rous in 191 1 that sarcomas of chicken could be
transmitted by the inoculation of tumour extracts (Rous, 191 1). The first
real breakthrough in defining genetic events in cancer came in the late
1970s from the study of transforming retroviruses (RNA tumour viruses).
A large number of retroviruses have been identified and consist of a few
genes which encoded the protein coat of the virus, a polymerase for
synthesising a DNA copy of the virus for incorporation into the host
genome and a transforming gene. The first of these genes to be described
was called src (Hanafusa and Hanafusa, 1966) but subsequently many
others were discovered and given 3 letter names derived from the parent
retrovirus. It was then found that all these transforming genes had cellular
counterparts, cellular or proto-oncogenes (Stehelin et al., 1976).
Proto-oncogene products are involved in the whole range of mechan-
isms controlling growth and proliferation of the cell: (a) Growth factors
and receptors such as c-sis (platelet-derived growth factor) and erb B
(EGF receptor); (b) transduction of signals from the cell surface to the
nucleus including the ras family and tyrosine kinases such as src; and (c)
nuclear proteins such as transcriptional activators, myc and fos.
These genes are highly conserved, for example, the ras family is found in
all species from yeast to man. The connection between the proto-
oncogenes and human cancer was not made however until the develop-
ment of DNA transfection assays (Shih et al., 1979) in which DNA
isolated from human tumour tissue is applied to a cell line. When treated in
one of a number of ways (originally as a calcium phosphate precipitate)
DNA is taken into the cells usually NIH3T3 mouse fibroblasts. Cells then
showing a transformed phenotype, e.g. ‘piling up’ of cells to form foci or
tumourigenicity in nude mice, are analysed for the presence of foreign
DNA. DNA sequences consistently present in these transformed cells
were considered to code for transforming factors. Most of these factors
from tumours proved to be members of the ras family, Ha-ras, Ki-ras, and
one not found in a retrovirus, N-ras. These ras genes appear to be
activated in many cases by point mutation at one of 3 ‘hotspots’, codons

12’13,or 61 of the coding sequence. Mutations such as these are present in
10-20% of all human cancers (Weinberg, 1989). The gene product of ras,
the p21 protein is located on the inside of the cell membrane and its
function appears to be the transmission of signals from receptors at the cell
surface via ‘cascade’ systems to the nucleus. In its active form, the p21
protein is bound to GTP and mutations in the gene yield a constantly
activated protein.
Oncogenes may also be activated by inappropriate changes in the level
of gene expression, either by gene amplification (increasing the number of
gene copies) or by transcriptional activation. A number of oncogenes
including ras and myc are amplified in human tumours (Weinberg, 1989),
although the significance to tumour development is unknown in most
cases. However, in childhood neuroblastoma, Stage I of the disease is
localised, treatable, and N-myc amplification never seen while Stages I1
and 111 are widespread, usually fatal with N-myc amplification present in
50% of the cases. Stage 111s is a variant with widespread disease but
reversible with the tumour redifferentiating to nerve cells. N-myc amplifi-
cation is never seen in this variant.
Transcriptional activation has been reported in a number of cases due to
the translocation of oncogenes. The best known example is the transloca-
tion of c-myc in Burkitt’s Lymphoma from its normal position on
chromosome 8 to a region of chromosome 14 where an immunoglobulin is
situated and under strong regulation (Rabbitts and Rabbitts, 1989). Other
rarer variants of the disease involve chromosomes 2 and 22 where further
immunoglobulin genes are located thus suggesting that this mechanism is
important in Burkitt’s Lymphoma.
A further case of chromosomal translocation is seen in the Philadelphia
chromosome characteristic of chronic myelogenous leukaemia. In this
case, the c-abl proto-oncogene is translocated from chromosome 9 to the
breakpoint cluster region (bcr) of chromosome 22. A fusion protein,
bcr-abl, is formed of 210 kd (as compared to the normal 145kd) which has
abnormal tyrosine kinase activity (De Klein, 1987).
From these examples it can be seen that oncogenes are present in normal
cells as proto-oncogenes and when expressed aberrantly either as the result
of mutation or transcriptional activation, are involved in the malignant
transformation of the cell.
Tumour Suppressor Genes

Theodore Boveri first postulated in 1914 (Boveri, 1929) that chromosomal
abnormalities might be a cause of cancer. Evidence for this theory
accumulated over several decades and includes that described above for
Burkitt’s Lymphoma and chronic myelogenous leukaemia. Further
evidence has recently come from research into two cancers, retinoblas-
toma and Wilms’ Tumour, characterised by chromosomal deletions, 13q
Paul Rumsby 323
and 11p respectively, which can be inherited and predisposes the carrier to
cancer. These deletions gave rise to the hypothesis of tumour suppressor
genes which when lost give rise to cancer. For instance, cancer would occur
after the loss of the second copy of the gene somatically in patients with a
germline deletion. A rarer sporadic form of the disease consists of a
somatic mutation/loss of both copies of the gene. The genes involved in
retinoblastoma, Rb-1 and Wilms’ Tumour, WT-1 have both been isolated
by detailed mapping of the deleted chromosomal region, a process known
as reverse genetics. A further tumour suppressor gene, p53 was already
isolated but its true nature not revealed until its location on chromosome
17 was found to coincide with a region often lost in human neoplasm,
particularly colorectal cancer. Alterations in the p53 gene have proved to
be the most common genetic event yet found in human cancer, occurring
in up to 70% of tumours (Hollstein et al., 1991). Many of the mutations
are very specific to certain cancer types suggesting organ specificity
perhaps due to DNA repair sensitivity or the effects of environmental
mutagens. A specificmutation (GC to TA) in the p53 gene found in human
hepatocellular carcinoma in Southern Africa and China may reflect the
exposure of the population to aflatoxins (Hsu et al., 1991; Bressac et al.,
1991).
The best example of a sequence of genetic events in the development of
human cancer is shown in Figure 6. Seven histological stages and six
genetic or epigenetic events have been described in colon cancer (Fearon
and Vogelstein, 1990) involving both the oncogene, Ki-ras and a number
of tumour suppressor genes on chromosome 5 (mutated in colon cancer,
MCC, and gdenomatous polyposis coli, APC), 17 (p53), 18 (deleted in
colon cancer, DDC), andrecently changes to a gene on chromosome 2
-
(familial colon cancer, FCC; Marx, 1993). How these multistages can be
interpreted in terms of models for human risk assessment remains a
challenge to toxicologists. It would seem that simple models of initiation,
promotion, and progression may be no longer helpful in analysis of
carcinogenesis; for example, does each of the genetic events in colorectal
cancer require an initiation and promotion stage?
D Animal Studies
Do the animal models used for the assessment of chemically-induced
cancer reflect the genetic events seen in the human disease? Many of these
model systems have been analysed for oncogene mutation particularly the
ras gene. The formation of mouse skin papillomas and carcinomas is the
best documented example of multistage carcinogenesis in animals. This
system can be divided into 3 distinct stages; initiation, promotion, and
progression. Treatment with one of a number of genotoxic carcinogens
such as dimethyl-benzanthracene (DMBA) or methyl-N-nitroso guan-
idine (MNNG) leads to formation of genetically altered (‘initiated’) cells.
Treatment with a further agent, frequently a phorbol ester such as
Chromosome Alteration
Normal
I
Hyperproliferative
4 5q
2P
Mutationfloss
(?)
APC/MCC
FCC (?)
epithelium
hypomethylation
1 4 12p Mutation Ki-rae
Adenoma Intermediate
,
1
Late
- 16q Mutation/Loss DCC
+ 17p Mutation/Loss p53
Carcinoma
0- other
alterations
Metastasis
Figure 6 The genetic and histological events in the development of human colon
cancer
12-0-tetradecanoyl-phorbol- 13-acetate (TPA) leads to the promotion of

these initiated cells to form a precancerous lesion, a papilloma. Promoters
applied alone or before the initiator do not lead to papillomas. Prolonged
application of the promoter may lead to the spontaneous development of
malignant carcinomas. This event can be accelerated by treatment with a
further mutagen. Therefore, 2 genetic events, one early and one late may
be the minimum required plus promotion of the original initiated cells.
Almost all of the tumours induced by DMBA possess a specific
mutation (AT to TA) in codon 61 of the Ha-ras oncogene, whereas those
induced by MNNG have a different mutation (GC to AT, Balmain et al.,
1984; Quintanilla et al., 1986). These mutations are consistent with the
specific action of the chemical in forming adducts with DNA bases and so
activation of the oncogene appears to be the initiating event in this cancer
(Quintanilla et al., 1986).
In rat mammary tumours, induced by methylnitrosourea (MNU), there
is a 100% incidence of GC to AT transition in codon 12 of Ha-ras. Using
DNA analysis this mutation can be found in mammary cells with no overt
signs of pathological abnormality within 2 weeks of treatment (Sukumar
et al., 1983; Kumar et al., 1991).
One of the main targets for chemically-induced tumours in the rodent
bioassay is the mouse liver. A number of non-genotoxic carcinogens only
give tumours in this organ. For this reason the mechanism of tumour
Paul Rumsby 325
development induced by genotoxins and non-genotoxins in the mouse

liver has been the subject of much study. A number of groups have found
mutations at codon 61 of the Ha-ras oncogene in both spontaneous and
genotoxic carcinogen-induced tumours (Reynolds et al., 1986, 1987;
Wiseman et al., 1986) although the frequency appears low in mouse
strains resistant to chemically-induced hepatocarcinogenesis (Buchmann
et al., 1991; Dragani et al., 1991; Rumsby et al., 1992). With a number of
chemicals the spectrum of mutations appears specific to the compound
suggesting that the Ha-ras mutation is an initiating event (Wiseman et al.,
1986; Rumsby et al., 1991). However, non-genotoxic agents, such as
chloroform, ciprofibrate and phenobarbitone, all have a frequency of
mutations lower than that found in spontaneous tumours (Fox et al.,
1990; Rumsby et al., 1991). It has been suggested that non-genotoxic
carcinogens act by promoting spontaneous lesions. If this is true in the
mouse liver, these lesions must be Ha-ras mutation free or there is a
separate mechanism of tumour development. It does appear that muta-
tions in the ras oncogene are initiating events in a number of animal
systems and can be matched to the action of the chemical carcinogen. So
the measurement of this genetic event, e.g. H a m s codon 61 mutations in
the mouse liver, may give some indication of the mechanisms involved
which is of value in assessing carcinogenic risk to man. Some of the
molecular methods used for the detection of oncogene mutations are
described in Figure 7.
It appears that oncogene activation is often not the initiating event in
human cancer. For example, late adenomas in human colorectal cancer
have frequent Ki-ras mutations (58%) whereas it is present in only 9% of
the early adenomas.
The tumour suppressor genes have not as yet been so extensively studied
in animal models. Mutations in p53 have been shown to be involved in the
progression of papilloma to carcinoma in the mouse skin (Bremner and
Balmain, 1990)but appear not to be involved in the development of mouse
liver tumours (Kress et al., 1992; Goodrow et al., 1992). However, p53
mutations are found in rat liver tumours induced chemically where ras
mutations are rare. Toxicologists must try to understand such specific
differencesand use them to analyse the value of animal models in assessing
human toxicity.
Although mutations in the p53 tumour suppressor genes do appear to
cluster at ‘hot spots’ in exons 5, 7, and 8, these regions are too large to
enable the use of allele-specific hybridisation to screen for mutation. A
recent technical advance called single-strandconformation polymorphism
(SSCP, Orita et al., 1989) together with PCR has proved useful for such
screening. This method takes advantage of the observations that single-
stranded DNA’s (usually PCR products) with a single base change have an
altered conformation configuration which can be detected as a change in
electrophoretic mobility in a polyacrylamidegel. Samples showing altered
mobility in the SSCP screen can then be sequenced to define the mutation.
Figure 7 Detection of Ha-ras mutations. This figure indicates the methods for the
analysis of the Ha-ras codon 61 mutations. PCR is used to amplify the region of exon
2 around codon 61.
1. PCR products are blotted onto a nylon membrane which is hybridised to a panel
of radiolabelled oligonucleotides (20 bases long) complementary to single base
mutations at codon 61. Stringent washes mean that only a perfect match
anneals the radiolabelledprobe to the dot while a 1 base mismatch washes 08.
In this example, each dot represents DNA from a tumour having at least one
copy of the gene with the normal sequence, CAA at codon 61. Hybridisation with
Paul Rumsby 327
E In Vitro Studies
A number of in vitro cell culture assays have been developed for the
measurement of mutagenicity. However, both the Ames Salmonella assay
and the mammalian short-term tests have been criticised for poor
correlation with the result of long-term carcinogenicity bioassays due
mainly to non-genotoxic carcinogens. These assays detect mutation in
target genes leading to phenotypic changes detected by survival in a
selective medium.
Molecular analysis of the mutation in mammalian cell assay such as the
Chinese hamster hprt (Fuscoe et al., 1983; Thacker et al., 1985; Vrieling et
al., 1985; Davies et al., 1993) and mouse lymphoma tk (Applegate et al.,
1990; Clive et al., 1990; Yandell et al., 1990; Davies et al., 1993a) has
i-dicated that certain chemicals act mainly as point mutagens while others
cause large-scale chromosomal damage. However, the assays differ in their
ability to detect a range of damage; for example, large-scale damage to
regions around the hprt gene essential for growth which are present as a
sole copy on the X chromosome may lead to cell death and no detection of
the damage. The tk gene and its environs are present in two copies (one tk
allele is disabled) and so large scale damage to one copy of an essential
region is not invariably fatal and can be detected as a mutant. Such a
system may also detect changes involving two alleles such as recombina-
tion. Further molecular analysis of these assays would enable the
assessment of their ability to detect events we know to be important in the
development of human cancer and to devise more accurate in vitro assays.
F Transgenic Animals
The techniques of molecular biology have led to the widespread use of
transgenic animals to study expression of individual genes. Genes can be
introduced into the germline of rodents and thus into every cell of the
offspring. The use of promoters leads to the expression of the gene in
specific sites, and it is now also possible to selectively delete genes. This
technology has also been used to develop in vivo mutation assays
(Stratagene, ‘Big Blue’; Hazleton, ‘Mutamouse’).A mouse with a ‘target’
other oligonucleotides shows some tumours with the sequences A A A , CGA or

C T A in the other copy of the gene.
The above screen can be ver$ed by direct sequencing of the PCR product. Each
track represents a base and by reading each track in turn from the bottom, a
normal C A A sequence can be detected (as in the left-hand picture) or the
presence of two bases in the same position indicates one copy of the gene with
C A A and the other copy with either A A A (as in the middle picture) or C T A (right
hand picture).
The detection of RFLPs can prove a sensitive method if a new restriction site is
produced by the mutation. In this example, the normal C A A sequence gives a
107 bp PCRproduct. Mutation at codon 61 to C T A in one copy of the gene gives
the X b a 1 recognition site, TCTAGA andso bands of 72pb and 35 bp are seen in
lanes 1-4 and 9-1 1.
bacterial transgene incorporated is treated with a test chemical. The

‘target’ gene can then be recovered from any organ, transformed into a
bacterium, and mutations detected by a colour reaction. These assays have
the advantage of measuring organ-specific mutations in a system using in
vivo metabolism of the chemical.
G Non-genotoxic Carcinogenesis
There has recently been marked progress in describing events which may
be important in non-genotoxic carcinogenesis. Much work has centred on
measurement of cellular proliferation mediated by non-genotoxic chemi-
cals. Increased proliferation may increase the likelihood of spontaneous
mutational events, the importance of which has been previously outlined.
Receptors have been described (and their genes isolated) which mediate
the action of groups of non-genotoxic chemicals; the aromatic hydro-
carbons such as 2,3,7,8-tetra-chlorodibenzo-p-dioxin (TCDD) via the A,
receptor (Burbach et al., 1992), and peroxisome proliferators via PPAR
(peroxisome proliferator activated receptor, Isseman and Green et al.,
1990). The latter, a member of the steroid superfamily (oestrogen also has
mitogenic effects), apears to bind to a ligand, the peroxisome proliferator,
and also to specific sites on DNA (Isseman and Green, 1990). The A,,
receptor action is more complicated as there seems to be two proteins, the
A, receptor which binds ligand and specific DNA sites (Burbach et al.,
1992) and ARNT (A,-receptor nuclear translocator) which may form a
dimer with the A, receptor and aid DNA binding (Hoffman et al.,1991).
There is now evidence that such receptors bind to the regulatory sites of
the genes for proteins such as the P450s leading to increased transcription.
The role of such receptors in the normal regulation of cellular control (for
instance, the identification of an endogenous ligand) and how this is
disrupted in chemically-induced cancer is the next step for molecular
biology and toxicology.
4 Future Developments
Although the use of molecular biology in toxicology is most advanced in
the field of chemically-induced cancer, the techniques are now being used
in all branches of toxicology.
In immunology, molecular techniques are being used to dissect the
network of cytokines which regulate the cells of the immune system.
Toxicologists are using these sensitive techniques to examine changes in
this regulation mediated by toxic chemicals in, for example, allergic
reactions and sensitisation. This new field has been reviewed by Meredith
(1992).
Developmental biology has undergone a revolution in understanding
using the techniques of molecular biology to study development of the
embryo and foetus. Regions of spatial and temporal development have
Paul Rumsby 329
been defined by the expression of highly conserved genes called homeobox

genes. Expression of some of these genes in transgenic animals has led to
foetal abnormalities similar to those seen with teratogenic chemicals such
as the retinoids (Balling et al., 1989). The effect of toxic compounds on the
expression of homeobox genes may open the way for mechanistic studies
in teratology and possibly in vitro screens using altered gene expression.
A number of different processes, such as apoptosis and stress response,
are currently of intense research interest and may help to explain some
fundamental problems in toxicology. Apoptosis is the way in which cell
populations control their normal development by programmed cell death.
A number of genes have been implicated in this process including the
oncogenes myc, bcl-2, and the p53 tumour suppressor gene, thus
suggesting a link with carcinogenesis. It has been suggested that p53 may
be a ‘molecular policeman’ (Lane, 1992), extending the GI stage of the cell
cycle in damaged cells to allow for DNA repair and in some way leading
cells to apoptosis, if the damage is too severe. Mutations leading to
malfunction and no p53 response would mean no extended repair time or
apoptosis in time of DNA damage with mutations present and the
inherent risk of the clonal expansion of initiated cells.
The aim of toxicologists must be the development of a series of universal
principles to explain the response of a cell to toxic insult whether chemical,
including metals, or physical agents such as radiation and heat. A number
of stress response proteins have been identified which may fulfil these
criteria and in the future be the ‘target’ for sensitive analysis as a screen for
toxic insult (Welch, 1993). The heat shock protein (hsp) family appears to
act as ‘molecular chaperones’ in binding protein which becomes denatured
as a result of toxins. The transcription factors, myc and particularlyfos are
also very early responses to modulatory stimuli. The analysis of these
stress response genes is at an early stage and their manipulation in
toxicology is for the future, as one of the many challenges which have been
set to toxicologists by the advances brought about by molecular biology.
References
Alwine, J. C., Kemp, D. J., and Stark, G. R. (1977). Method for detection of
specific RNAs in agarose gels by transfer to diazobenzyloxy-methyl paper
and hybridization with DNA probes. Proc. Natl. Acad. Sci. (USA), 74,
5350-5354.
Applegate, M. L., Moore, M. M., Broder, C. B., Burrell, A., Juhn, G., Kasweck,
K. L., Lin, P.-F., Wadhams, A., and Hozier, J. C. (1990). Molecular
dissection of mutations at the heterozygous thymidine kinase locus in mouse
lymphoma cells. Proc. Natl. Acad. Sci. ( U S A ) , 87, 51-55.
Balling, R., Mutter, G., Gruss, P., and Kessel, M. (1989). Craniofacial abnorm-
alities induced by ectopic expression of the homeobox gene Hox-1.1 in
transgenic mice. Cell, 58, 337-347.
Balmain, A., Ramsden, M., Bowden, G. T., and Smith, J. (1984). Activation of the
mouse cellular Harvey-ras gene in chemically-induced benign skin papil-

lomas. Nature (London), 307, 658-660.
Berger, S. L. and Kimmel, A. R. (Eds), (1987). Guide to molecular cloning
techniques. Methods Enzymol., 152. Academic Press, San Diego, CA.
Boveri, T. (1929). ‘The origin of malignant tumors’. Williams and Watkins,
Boston.
Bremner, R. and Balmain, A. (1990). Genetic changes in skin tumor progression:
correlation between presence of a mutant ras gene and loss of heterozygosity
on mouse chromosome 7. Cell, 61, 407-417.
Bressac, B., Key, M., Wands, J., and Ozturk, M. (1991). Selective G to T
mutations of p53 in hepatocellular carcinoma from Southern Africa. Nature
(London), 350, 429-43 1.
Buchman, A., Bauer-Hofmann, R., Mahr, J., Drinkwater, N. R., Luz, A., and
Schwarz, M. (1991). Mutational activation of the c-Ha-ras gene in liver
tumors of different rodent strains: correlation with susceptibility to hepa-
tocarcinogenesis. Proc. Natl. Acad. Sci. (USA), 88, 91 1-915.
Burbach, K. M., Poland, A., and Bradfield, C. A. (1992). Cloning of the
A,,-receptor cDNA reveals a distinctive ligand-activated transcription factor.
Proc. Natl. Acad. Sci. (USA), 89, 8185-8189.
Clive, D., Glover, P., Applegate, M., and Hozier, J. (1990). Molecular aspects of
chemical mutagenesis in L5 178Y/TK+’- mouse lymphoma cells.
Davies, M. J., Phillips, B. J., Anderson, D., and Rumsby, P. C. (1993). Molecular
analysis of mutation at the hprt locus of Chinese hamster V79 cells induced by
ethyl methanesulphonate and mitomycin C.Mutat. Res., 291, 1 17-1 24.
Davies, M. J., Phillips, B. J., and Rumsby, P. C. (1993a). Molecular analysis of
mutations at the tk locus of L5 178Y mouse lymphoma cells induced by ethyl
methanesulphonate and mitomycin C. Mutat. Res. In press.
De Klein, A. (1987). Oncogene activation by chromosomal rearrangement in
chronic myelocytic leukaemia. Mutat. Res., 186, 161-1 72.
Dragani, T. A., Manenti, G., Colombo, B. M., Falvella, F. S., Garibondi, M.,
Pierotti, M. A., and Della Porta, G. (1991). Incidence of mutations at codon
61 of the Ha-ras gene in liver tumors of mice genetically susceptable and
resistant to hepatocarcinogenesis. Oncogene, 6, 333-338.
Ehrlich, H. A. (Ed.),(1989). ‘PCR Technology’, Stockton Press, New York.
Fearon, E. R. and Vogelstein, B. (1990). A genetic model for colorectal
tumorigenesis. Cell, 61, 759-767.
Feinberg, A. P. and Vogelstein, B. (1983). A technique for radiolabeling DNA
restriction endonuclease fragments to high specific activity. Anal. Biochern.,
132, 6-13.
Fox, T. R., Schumann, A. M., Watanabe, P. G., Yano, B. L., Maher, V. M., and
McCormick, J. J. (1990). Mutational analysis of the Ha-ras oncogene in
spontaneous C57BL/6 x C3H/He mouse liver tumors and tumors induced
with genotoxic and non-genotoxic hepatocarcinogens. Cancer Res., 50,
4014-4019.
Fuscoe, J. C., Fenwick, R. G., Ledbetter, D. H., and Caskey, C. T. (1983).
Deletion and amplification of the hgprt locus in Chinese hamster cells. Mol.
Cell Biol., 3, 1086-1096.
Gonzalez, F. J. (1989). The molecular biology of cytochrome P450s. Pharmacol.
Rev., 40, 243-287.
Paul Rumsby 33 1
Goodrow, T. L., Storer, R. D., Leander, K. R., Prahalada, S. R., van Zwieten, M.
J., and Bradley, M. 0.(1992). Murine p53 intron sequences 5-8 and their use
in polymerase chain reaction/direct sequencing analysis of p53 mutations in
CD-1 mouse liver and lung tumors. Mol. Carcinog., 5, 9-15.
Gough, A. C., Miles, J. S., Spurr, N. K., Moss, J. E., Gaedigk, A., Eichelbaum,
M., and Wolf, C. R. (1990). Identification of the primary gene detect at the
cytochrome P450 CYP2D locus. Nature (London), 347, 773-776.
Hanafusa, H. and Hanafusa, T. (1966). Determining factor in the capacity of
Rous sarcoma virus to induce tumors in animals. Proc. Natl. Acad. Sci.
( U S A ) , 55, 531-535.
Hoffman, E. C., Reyes, H., Chu, F.-F., Sander, F., Conley, L. H., Brooks, B. A.,
and Hankinson, 0. (1991). Cloning of a factor required for activity of the A,
(Dioxin) receptor. Science, 252, 954-958.
Hollstein, M., Sidransky, D., Vogelstein, B., and Harris, C. C. (1991). p53
mutations in human cancers. Science, 253,49-53.
Hsu, I. C., Metcalf, R. A., Sun, T., Welsh, J. A., Wang, N. J., and Harris, C. C.
(1991). Mutational hotspot in the p53 gene in human hepatocellular
carcinomas. Nature (London), 350, 427-428.
Idle, J. R. (199 1). Is environmental carcinogenesis modulated by host poly-
morphism? Mutat. Res., 247, 259-266.
Innis, M. A., Gelfand, D. H., Sninsky, J. J., and White, T. J. (Eds), (1990). ‘PCR
Protocols: a guide to methods and applications’, Academic Press, San Diego
Issemann, I. and Green, S. (1990). Activation of a member of the steroid hormone
receptor superfamily by peroxisome proliferators. Nature (London), 347,
645-650.
Jaiswal, A. K., Gonzalez, F. J., and Nebert, D. W. (1985). Human P,-450 gene
sequence and correlation of mRNA with genetic differences in benzo[a]-
pyrene metabolism. Nucl. Acids Res., 13, 4503-4520.
Kawajiri, K., Nakachi, K., Imai, K., Yoshi, A., Shimoda, N., and Watanabe, J.
(1990). Identification of genetically high risk individuals to lung cancer by
DNA polymorphism of the cytochrome P450 IAI gene. FEBS Lett., 263,
131-133.
Kress, S . , Konig, J., Schweizer, J., Lohrke, H., Bauer-Hofmann, R., and
Schwarz, M. (1992). p53 mutations are absent from carcinogen-induced
mouse liver tumors but occur in cell lines established from these tumors. Mol.
Carcinogen., 6, 148-158.
Kumar, R., Sukumar, S., and Barbacid, M. (1990). Activation of ras oncogenes
preceding the onset of neoplasia. Science, 248, 1101-1 104.
Lane, D. P. (1992). p53, guardian of the genome. Nature (London),358, 15-16.
Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982). ‘Molecular cloning-a
laboratory manual’, Cold Spring Harbor Laboratory, Cold Spring Harbor,
New York.
Marx, J. (1993). New colon cancer gene discovered. Science, 260,7512-752 and
others in the same edition of Science.
McLemore, T. L., Adelberg, S., Liu, M. C., McMahon, N. A., Yu, S. J.,
Hubbard, W. C., Czerwiniski, M., Wood, T. G., Storeng, R., Lubet, R. A.,
Egglestone, J. C., Boyd, M. R., and Hines, R. N. (1990). Expression of
CYPlAl gene in patients with lung cancer: evidence for cigarette smoke-
induced gene expression in normal lung tissue and for altered gene regulation
in primary pulmonary carcinomas. J. Natl. Cancer Inst., 82, 1333-1 339.
Meredith, C. (1992). Molecular Immunotoxicology. In: ‘Principles and practice of

immunotoxicology’. Eds. K. Miller, J. L. Turk, and S. Nicklin. Blackwell
Scientific Publications, Oxford, pp. 344-356.
Nakachi, K., Imai, K., Hayashi, S.-I., Watanabe, J., and Kawajiri, K. (1991).
Genetic susceptibility to squamous cell carcinoma of the lung in relation to
cigarette dose. Cancer Res., 51, 5177-5180.
Orita, M., Suzuki, Y., Sekiya, T., and Hayashi, K. (1989). Rapid and sensitive
detection of point-mutation and DNA polymorphisms using the polymerase
chain reaction. Genomics, 5, 874-879.
Quintanilla, M., Brown, K., Ramsden, M., and Balmain, A. (1986). Carcinogen-
specific mutations in mouse skin tumours: ras gene involvement in both
initiation and progression. Nature (London), 322, 78-80.
Rabbitts, T. H. and Rabbitts, P. H. (1989). Molecular pathology of chromosomal
abnormalities and cancer genes in human tumors. In: ‘Oncogenes’. Eds.
D. M. Glover and B. D. Hames. IRL Press, Oxford, pp. 67-111.
Reynolds, S. H., Stowers, S. J., Maronpot, R. R., Anderson, M . W., and
Aaronson, S. A. (1986). Detection and identification of activated oncogenes
in sponyneously occurring benign and malignant hepatocellular tumours of
B6C3F1 mouse. Proc. Natl. Acad. Sci. ( U S A ) , 83, 33-37.
Reynolds, S.H., Stowers, S. J., Patterson, R., Maronpot, R. R., Aaronson, S. A.,
and Anderson, M. W. (1987). Activated oncogenes in B6C3F1 mouse liver
tumors: implications for risk assessment. Science, 237, 1309-1 3 16.
Rous, P. (19 11). Transmission of a malignant new growth by means of a cell-free
filtrate. J . Am. Med. Assoc., 56, 198.
Rumsby, P. C., Barrass, N. C., Phillimore, H. E., and Evans, J. G. (1991).
Analysis of the Ha-ras oncogene in C3H/He mouse liver tumours derived
spontaneously or induced with diethylnitrosamine or phenobarbitone.
Carcinogenesis, 12, 233 1-2336.
Rumsby, P. C., Evans, J. G., Phillimore, H. E., Carthew, P., and Smith, A. G.
(1992). Search for Ha-ras codon 61 mutations in liver tumours caused by
hexachlorobenzene and Aroclor 1254 in C57BL/10 ScSn mice with iron
overload. Carcinogenesis, 13, 1917-1920.
Saiki, R. K., Scharf, S., Faloona, F., Mullis, K. B., Horn, G. T., Erlich, H. A., and
Arnheim, N. (1985). Enzymatic analysis of P-globin sequences and restriction
site analysis of sickle-cell anemia. Science, 230, 1350-1 352.
Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989). ‘Molecular cloning-a
laboratory manual’, Vols. 1, 2, and 3, 2nd Edn. Cold Spring Harbor
Laboratory, Cold Spring Harbor, New York.
Shibata, D., Martin, W. J., and Arnheim, N. (1988). Analysis of DNA sequences
in forty-year-old thin-tissue sections: a bridge between molecular biology and
classical histology. Cancer Res., 48, 4564-4566.
Shih, C., Shilo, B.-Z., Goldfarb, M. P., Dannenberg, A., and Weinberg, R. A.
(1979). Passage of phenotypes of chemically transformed cells via transfection
of DNA and chromatin. Proc. Natl. Acad. Sci. ( U S A ) , 76, 5714-5718.
Southern, E. M. (1975). Detection of specific sequences among DNA fragments
separated by gel electrophoresis. J . Mol. Biol., 98, 503-507.
Stehelin, D., Varmus, H. E., Bishop, J. M., and Vogt, P. K. (1976). DNA related to
the transforming gene(s) of avian sarcoma viruses is present in normal avian
DNA. Nature (London), 260, 170-1 72,
Paul Rumsby 333
Sukumar, S., Notario, V., Martin-Zanca, D., and Barbacid, M. (1983). Induction
of mammary carcinomas in rats by nitroso-methylurea involves malignant
activation of Ha-ras-1 locus by single point mutation. Nature (London), 306,
658-661.
Tefre, T., Ryberg, D., Hangen, A., Nebert, D. W., Skaug, V., Brogger, A., and
Borressen, A. L. (1991). Human CYPIAI (cytochrome P450) gene: lack of
association between the Msp 1 restriction fragment length polymorphism and
incidence of lung cancer in a Norwegian population. Pharmacogenetics, 1,
20-25.
Thacker, J. (1985). The molecular nature of mutations in cultured mammalian
cells, a review. Mutat. Res., 150, 431-442.
Vrieling, H., Simons, J. W., Arwent, F., Natarajan, A. T., and van Zeeland, A. A.
(1985). Mutations induced by X-rays at the hprt locus in Chinese hamster cells
are mostly large deletions. Mutat. Res., 144, 281-286.
Weinberg, R. A. (Ed.), (1989). ‘Oncogenes and the molecular origins of cancer’,
Cold Spring Harbor Laboratory, Cold Spring Harbor, New York.
Welch, W. J. (1993). How cells respond to stress. Sci. Am., 268, 34-41.
Wiseman, R. W., Stowers, S. J., Miller, E. C., Anderson, M. W., and Miller, J. W.
(1986). Activating mutations of the c-Ha-ras proto-oncogene in chemically-
induced hepatomas of the male B6C3F1 mouse. Proc. Natl. Acad. Sci. (USA),
83, 5825-5829.
Yandell, D. W., Dryja, T. P., and Little, J. B. (1990). Molecular genetic analysis of
recessive mutations at a heterozygous autosomal locus in human cells. Mutat.
Res., 229, 89-102.
Young, B. D. and Anderson, M. L. (1985). Quantitative analysis of solution
hybridisation. In : ‘Nucleic acid hybridisation-a practical approach’, Eds
B. D. Haines and S. J. Higgins, IRL Press, Oxford, pp. 73-1 11.
CHAPTER 15
Testing for Carcinogenicity

P. GRASS0
1 Introduction
Over the years, experience has shown that animals, like man, are afflicted
by the development of tumours as they approach the end of their
lifespan. The natural history of the animal tumours has many features in
common with human tumours; some grow rapidly, metastasise widely,
are readily invasive and rapidly fatal, whereas others grow slowly, do
not invade or metastasise, and do not threaten life. There is also some
resemblance in the histological structure between the animal and
human tumours that arise from the same type of tissue; but there may be
striking differences-structures that in man are strongly suggestive of a
malignant potential may not have the same meaning in animals. The
reverse may also be true.
Certain chemicals have caused cancer to develop in both animals and
man. The susceptibility of animals to cancer development by those
chemicals which are known to be human carcinogens has been the main
justification for using laboratory animals to investigate the carcinogenic
potential of chemicals. At the moment there are some forty chemicals and
processes which are known to cause cancer in man and all of these, except
benzene and arsenic, have caused cancer in animals (see Table 1 for
examples).
2 Methods Employed in Carcinogenicity Studies

The routes employed in carcinogenicity studies are summarised in Table
2. There is, at the moment, some general agreement among scientists
that the route most suitable for testing the carcinogenicity of a substance
is that by which the chemical in question is likely to reach man. Although
this principle is sound, it could give misleading results. For example, the
inhalation of cigarette smoke appears to be carcinogenic to man, yet
studies in animals have failed to produce lung tumours (other than
pulmonary adenomata in some strains of mice) by the inhalation of
cigarette smoke. On the other hand, cigarette smoke condensate has
334
P
Table 1 Chemical carcinogens of established activity in man and laboratory rodents
Man Lab. rodent

Route of Route of
Chemical Target organ exposure Target organ Lab. rodent exposure
Aflatoxin liver oral liver rats, mice oral
(benign oral
tumours
only)
4-Aminobiphenyl bladder inhalation bladder
skin
oral
Asbestos lung inhalation lung mice, rats inhalation
pleura pleura hamsters and intra-
pleural
injection
Auramine bladder inhalation liver mice oral
(manufacture of) skin
oral
Benzidine bladder inhalation liver rats, hamsters oral
skin
oral
NJV-bis(2-chloroethyl) bladder oral lung mice intraperitoneal
-2-naphthylamine
Bis(chloromethy1)ether lung inhalation lung mice, rats inhalation w
mice topical appl. w
skin Lh
Chromium compounds lung inhalation lung rats intratracheal
(unspecified in man, installation
chromium, strontium
and zinc chromate in
animals)
Cyclophosphamide bladder oral or lung mice, rats intraperi t oneal
parenteral lymphomas intravenous
injection (rats only)
Diethylstilbestrol vagina oral to mammary mice, rats oral
(of mother mother glands hamsters
treated in uterus
first 3
months)
Melphalan bone-marrow parenteral lymphosar- mice, rats intraperitoneal
(leukemia) injection comas injection
Mustard gas lung inhalation lung mice inhalation
or intraperitoneal
injection
2-Napht hylamine bladder inhalation bladder hamsters oral
liver mice
Nickel (unspecified lung inhalation lung rats inhalation
compound in man,
nickel subsulfide and
nickel carbonyl in
rodents)
Phen acetin renal pelvis oral (in urinary rats oral
analgesic tract
abuse)
Soots, tars and skin topical skin mice topical application
mineral oils (? other (inhalation)
sites)
Vinyl chloride liver inhalation liver rats, mice inhalation
(other sites) hamsters
P. Grasso 337
Table 2 Route employed in carcinogenicity testing

(1) Oral High dose administration by gavage or mixed in diet for
‘lifespan’ at 2 or more levels.
(2) Parenteral (i) Subcutaneous or (ii) Intraperitoneal-oil or water are
used as solvents for test substance. Administration at the
same site once or twice weekly for a limited period or
‘lifespan’. (iii) Intravenous-used for a limited period only.
(iv) Intrapleural-used mainly for dusts.
(3) Inhalation Route employed for testing dusts, aerosols, and vapour.
Maximum exposure 5 hr day-l, 5-6 days a week for
‘lifespan’. Special expertise required to establish, maintain,
and monitor dose of test substance inhaled.
(4) Topical Test substance usually dissolved in an aqueous or organic
application solvent and applied topically to skin of mice. Mineral oils
are usually applied neat. Application usually once or twice
weekly for a limited period or for ‘lifespan’.
(5) Intravaginal Test substance applied onto epithelium of vagina or cervix
application in female mice once or twice weekly or for ‘lifespan’.
been shown to be powerfully carcinogenic when painted on the skin of

mice, a test system which is radically different from inhalation (Roe,
1981). The type of information sought should be clear before choosing
the route of exposure. In exploring the carcinogenic potential of a
chemical, the route of application should be the one likely to yield the
optimum response for that particular type of chemical. On the other
hand, in assessing the risk to man, the carcinogenic potential of the
chemical by a route of exposure similar to that of man has been found of
some value.
A number of points need to be considered when planning car-
cinogenicity studies, and many have been the subject of much debate in
recent years. Although no final answer has been found, there would
appear to be a general consensus among scientists on the best way of
resolving the uncertainties.
The important points that need to be paid particular attention are:
(a) The test chemical
(1) Chemical class
(2) Physical properties
(3) Methods for analysis in animal feed
(4) Suitability for a particular route
(5) Choice of vehicle
(b) The test animal
(1) Species and strain
(2) Background incidence of tumours
(3) Longevity and resistance to disease
(4) Diet and caging
(5) Number of animals per group
(c) The test procedure
(1) Route of administration
338 Testing for Carcinogenicity
(2) Dose-levels and frequency of administration. Volume (by ga-

vage or injection)
(3) The meaning of the maximum tolerated dose
(4) Criteria for terminating the study
(5) Clinical observations-e.g. bodyweight, general behaviour, etc.
(6) Termination of study
(7) In utero exposure
(8) Post-mortem procedure-record keeping, fixation for histology,
which organs to be sampled if no tumour is present
(9) Pathological expertise and method of reporting
(10) Statistical considerations: randomisation at beginning of the
study, statistical analysis of data at termination.
A Choice of Species and Strain

It is often recommended that a species should be chosen which closely
resembles man in its physiology and in the metabolism of the compound
to be tested. While this recommendation is ideal, it is rarely achieved.
For purely practical reasons studies are restricted to three species, the
mouse, hamster, and rat. Other species present serious difficulties,
principally because of their size and long lifespan. The larger animals
require much more floor space for accommodation than the laboratory
rodents and their lifespan is usually so long that the test may embrace the
period of employment of two or more toxicologists. These two factors
immeasurably add to the cost of the test and to the complexity of
administration since ‘hand overs’ are seldom completely effective. The
restriction of choice to the three rodent species is not necessarily a
drawback in a screening test for carcinogenicity, since experience has
shown that they readily develop tumours with a large number of
chemicals including those which have produced tumours in man (IARC,
1979).
There has been much debate on the use of outbred or inbred strains for
testing carcinogenic activity. In the case of hamsters and rats, inbred
strains are not as easy to obtain as are such strains in the mouse, so that
the choice is limited to one of a number of outbred strains. In the case of
the mouse, the number of inbred strains available is very large indeed
and the choice is usually wide.
In theory, inbred strains are preferable because the tumour incidence is
generally known with some degree of precision, whereas in outbred mice
the incidence usually varies markedly from one generation to the next. In
practice, outbred mice are more frequently used than the inbred variety
because it has been the traditional practice for several years, and some
influential governmental agencies have advocated their use (FDA, 1971;
Ministry of Health and Welfare Canada, 1973), and it may happen that
the inbred mouse may metabolise the substance in a way which avoids
the production of the ultimate carcinogenic species (Page, 1977). It is
P. Grasso 339
possible to circumvent this pitfall by employing several inbred strains, on

the lines advocated by Festing , to test pharmacological or biochemical
responses to a chemical in mice (Festing, 1975). The advice is quite sound
since using several strains reduces the likelihood of missing some
important response. It does imply, however, that a particular test has to
be done in several strains and although this may be possible in the sort of
short-term tests that are usually employed in pharmacology or biochem-
istry, it is rarely practicable in a long term carcinogenicity test where
funds usually put strict limits on the design, scope and size of the test.
B Diet
It is usual to allow animals free access (ad libitum) to food and water
throughout the carcinogenicity studies. The food currently made avail-
able to the test animals usually contains an adequate amount of vitamins
and minerals, but is usually rich in fat and protein. It is a common
experience that rodents, particularly those in the control group, gain
rapidly in weight in their first weeks on treatment and become obese as
the experiment progresses. The obese mice and rats lose their sleek clean
appearance as well as their agility and general spontaneous activity. The
rapid onset of obesity indicates that the animals ‘overeat and oversleep’.
There is now clear and fully documented evidence that some benign
tumours in both these species may be increased in incidence by excessive
food intake (chiefly calorie intake) (Roe, 1979). It is clearly undesirable
to have a high background incidence of tumours since it may affect the
validity of the results in carcinogenicity studies. It is not possible however
at this stage to advocate a reduction of calorie intake in long term tests
since there is no knowledge of whether it might alter the sensitivity of the
animals to carcinogenic agents. Obviously further research is urgently
needed in this important field.
C Caging
Conventionally, rodents are housed 5 to a cage: the animals in a single
cage belong to the same sex. This type of communal caging has resulted
in problems, the principal ones being loss of animals from fighting
(particularly in certain mice strains) (I. F. Gaunt, personal communica-
tion, 1982) and cannibalism of dead animals (IARC, 1980). Some
authorities advise caging singly to overcome difficulties of this sort but
this may cause problems of a different nature. For example, the incidence
of liver tumours has been known to increase in single-caged (C,H) mice
compared with conventionally caged ones (Peraino, Fry, and Staffeldt,
1973).
D The Test Substance

Carcinogenicity studies in rodents were launched by the demonstration
that a complex mixture of substances (coal tar) was carcinogenic when
painted repeatedly on the skin of rabbits (Yamigawa and Ichikawa,

1918). It soon became obvious that such complex mixtures contributed
very little to the identification of specific chemicals possessing car-
cinogenic properties so that the tendency in scientific circles has been to
test only single chemicals, with particular attention to purity. Although
this practice has been found of considerable value in the pharmaceutical,
food additive, and chemical industries, other sections such as the oil,
tobacco, and cosmetics industries, are compelled to test complex
mixtures .
In selecting chemicals or products for carcinogenicity testing it is
essential to have some basic information on:
(a) Chemical structure, physical properties, and impurities
(b) Stability, particularly in the medium in which it is intended for
administration
(c) Probable daily exposure level for man and the route by which man is
exposed
(d) If man is exposed to a mixture of substances ( e . g mineral oil or
tobacco smoke) then a representative specimen should be selected.
This advice may be difficult to follow since it is not always possible to
reach agreement on what would constitute a representative specimen
of, for example, a mineral oil.
The degree of purity of the chemical to be tested very often raises
considerable difficulties in the selection of the source of the chemical for
carcinogenicity testing. While pharmaceutical chemicals are, in general,
of a high degree of purity, most chemicals in industrial use contain a
variable amount of impurities. Such chemicals are known as ‘technical
grade’ chemicals to distinguish them from the much more purified
‘laboratory grade’ of chemical. Man is exposed to the ‘technical grade’ of
a chemical, and there has been some debate whether this grade or the
‘pure’ form should be tested. Ideally, of course, both grades should
be tested simultaneously (IARC, 1980) but in a limited programme of
testing, it would appear preferable to test the ‘technical grade’ in the first
instance. In testing ‘technical grade’ materials there is always the chance
of missing the carcinogenic activity of some component which is present
in amounts too small to exert a clear effect in the sort of rodent study that
is normally carried out. If such a situation is suspected, then the
impurities may have to be tested separately.
E Route of Administration
There is some agreement among oncologists that the route of administra-
tion should be closely similar to the one by which human exposure occurs
(IARC, 1980). This rule, however, need not be inflexible if, for any
reason, there is a strong suspicion that by following such a rule there is
the risk of obtaining false results. In such circumstances, some other
more promising route should be attempted. For example, it is not
advisable to test for carcinogenicity the aerosol from a new type of
P. Grasso 34 1
cigarette or from an oil obtained from a new refining process by

inhalation in rodents when topical application of the condensate may
yield much more acceptable results.
Oral Exposure
This route is appropriate for most substances ingested by man. The
substance may be administered mixed in the diet, dissolved in the
drinking water, or by gavage, dissolved or suspended in a suitable
medium-most substances given by gavage are suspended or dissolved in
water or some edible oil. Administration by gavage has the advantage of
a fairly accurate control of dosage. On the other hand, it results in a high
concentration of the test chemical in the stomach and intestines with the
possibility of causing serious mucosal damage. Such high concentrations
in the gut enhance absorption and result in high blood levels, although
these are of limited duration. Incorporation of the test substance in food
or water avoids the possibility of these undesirable local pathological
effects and of the large fluctuations of blood concentrations. It allows a
much more uniform level of exposure and, in general, a larger daily dose
can be administered by dietary incorporation of the test substance than
by gavage.
Inhalation Exposure
This route is usually advocated to test the carcinogenicity of dusts, mists,
aerosols, and vapours. The technology required to generate an atmos-
phere containing the required concentration of any of these test
substances is complicated and requires considerable expertise and special
equipment (WHO, 1978b). The design of the experiment, the type of
equipment necessary, and the precautions to ensure an acceptable result
must be carried out in collaboration with an inhalation toxicologist and a
respiratory physiologist (IARC, 1980). In conducting such a test it is
important to bear in mind:
(1) The respiratory rate is, on average, 18 per minute for man, 120 for
rat and 150 for the mouse (WHO, 1987a)
(2) The respiratory exchange of the rat is approximately 150-
200ml/min, that of the mouse is 25ml/min and that of man is
8.51 l/min (Mauderly and Kritchevsky, 1979; Sanockij , 1978)
(3) Rodents respire only through the nose, while man may respire
through the mouth as well
(4) Airway resistance to flow may develop in experimental animals if
the inhaled material is too irritant (Mead, 1960)
( 5 ) ‘Head only’ exposure is recommended to avoid oral intake of the
test substance from licking of one another’s fur or to prevent the
animals from burying the nose in their neighbour’s fur thus filtering
the inhaled air. Both of these undesirable complications could
occur in total body exposure of animals grouped conventionally

(for example, 5 to a cage) (IARC, 1980)
(6) ‘Head only’ exposures usually involve a considerable degree of
restraint and are not normally conducted for periods longer than 6
hours a day, five days per week. Longer periods of exposure may
lead to serious damage to the animal’s health from ‘stress’ (Powell
and Hosey, 1965).
Topical Application
Skin painting is usually conducted in mice to assess the carcinogenicity of
substances that normally would come into contact with human skin, for
example mineral oils (Grimmer, Dettbarn, Brune, Deutsch-Wenzel, and
Misfeld, 1982). It is also a useful method for detecting the carcinogenic
activity of polycyclic hydrocarbons (Clayson, 1962), alkylating agents,
and other proximate carcinogens (Grasso and Crampton, 1972). Other
rodent species are less sensitive.
The test substance is usually applied in a vehicle which favours its
passage across the keratin layer of the skin so that it comes into contact
with the epidermal layer of cells. It is not usually recommended to apply
the material daily; 2 or 3 times weekly would normally suffice. A mild
degree of irritation produced locally by the applied material is permis-
sible but ulceration of the skin must be avoided as this may destroy the
tissue destined to develop into tumours.
A preliminary experiment to ‘gauge’ the epidermal reaction to the test
substance is usually recommended.
Topical application is suitable for experimental designs intended to
detect ‘initiating’ or ‘promoting’ potency of a particular chemical (Beren-
blum, 1955).
Parenteral Injection
This involves subcutaneous, intraperitoneal, or intravenous injection,
Subcutaneous injection was very popular at one time because of the
advantages it offers: accurate control of dosage, prevention of aerial
oxidation, and the small amounts that may be used-an important
consideration for a newly synthesised and very expensive chemical. The
test substance is usually administered for 2 to 5 days a week. It is usually
dissolved in a vehicle which may be water or a vegetable oil. Other
vehicles are to be avoided as they are generally very irritant. Substances
in liquid form have on occasion been administered undiluted but this
practice is rarely commendable unless the test material is non-irritant
(Grasso and Golberg, 1966).
Solids have often been implanted into the subcutaneous tissues of the
rat for carcinogenicity testing. This practice is now no longer pursued
(Bischoff and Bryson, 1964).
P. Grasso 343
Repeated intraperitoneal injection has on occasion been used in place

of repeated subcutaneous injection. There is always a serious risk of
inducing a ‘chemical peritonitis’ if the substance is too irritant or of
puncturing the gut with the consequence of inducing an infective type of
peritonitis. The use of this route is to be discouraged even though it
offers the advantages of rapid diffusion of the substances throughout the
tissues.
Occasionally carcinogenicity testing has been attempted by the in-
travenous route. As in the intraperitoneal route , substances administered
intravenously diffuse rapidly throughout the body. Unfortunately, only a
limited number of intravenous injections can be given to rodents because
the accessible veins soon become thrombosed. The total dose of a
chemical which can be given by this route is necessarily much smaller than
that which can be given by other routes, so that intravenous administra-
tion has a very limited scope in testing substances for carcinogenicity.
Other Routes of Exposure

Chemicals have been tested by other routes of exposure for some specific
reason. For example , some anti-fertility compounds intended for local
application have been applied intravaginally to mice (Boyland, Roe, and
Mitchley, 1966), dusts have been applied intrapleurally (Wagner, 1962)
and many compounds have been applied intratracheally in order to
maximise the reaction of the lung to the test substance or to bypass the
nasal passages of the experimental animal (Grimmer, Dettbarn, Brune,
Deutsch-Wenzel, and Misfeld, 1982). Although the results from these
routes are insufficient to determine the carcinogenic activity of a
chemical , they often yield valuable information and are particularly
suited to study some specific aspect of the carcinogenic process by a
particular substance.
F Selection of Dose
The dose selected for lifetime administration during carcinogenicity
testing has been the subject of considerable scientific debate. It has been
stated that ‘high test dose selection is the most controversial and perhaps
the most important element’ in designing the protocol for a car-
cinogenicity test (Food Safety Council, 1978).
It has been the practice for the last two decades to select the highest
dose compatible with long term survival, (Maximum Tolerated Dose).
Doses of this sort are essential in order to ensure that the tissues of the
test animal are exposed to the highest possible dose of the chemical. Such
a precaution is essential in these experiments because the numbers of
animals employed are, for logistic reasons, strictly limited. The doses
applied must be large enough, therefore, to ensure that optimum
conditions exist for tumour development to take place in statistically
significant numbers. The alternative is to use very large numbers of
animals (milli- or mega-mouse/rat) experiments. In theory, such large

numbers would ensure that a sufficient number of sensitive animals,
which would develop tumours with relatively low doses of chemicals,
have been included in the test groups. For many reasons, large
experiments of this sort are not practicable. For instance, in an
experiment carried out on a total of 24,000 mice (the so-called EDol
experiment), there was no clear evidence that tumours of the bladder or
liver were induced by the low doses (35 and 30 p.p.m.) of AAF adminis-
tered, despite the fact that at each of these dose levels over a thousand
mice had survived for 24 months.
The definition of a ‘high dose’ for any particular substance is, however,
a difficult task. This topic has been discussed in detail in several
publications and there seems to be some agreement on what is to be
achieved and what is to be avoided (Food Safety Council, 1978; Munro,
1977).
In general, the toxic potential of the chemical should be established by
conventional toxicity tests and some knowledge of the pharmacokinetics
and metabolism should be available. These should form the basis for
establishing the dose levels to be administered. Effects to be avoided
include: depression of growth rate greater than 10% of untreated
controls, recognisable injury of the target tissue, excessive stimulation of
glandular activity through normal or pathological mechanisms, and
abnormal pharmacological effects of excessive dosage (Grasso, 1979).
For substances which are non-toxic, e.g. some food colours, it is usual
to limit the administration to 5% of the diet (Grasso, 1979).
There is at the moment no established procedure for estimating the
dose at which novel foods (e.g. from single cell protein) could be
administered to rodents for carcinogenicity testing. It may, perhaps, be
more accurate to talk of ‘levels’ since they usually replace a substantial
proportion (up to 50%) of the test animal’s diet (de Groot, Til, and
Feron, 1970). Some of the major factors that may influence the choice of
the level at which such substances are to be administered are: the relative
contents of phospholipid and protein, the nature of the protein and the
amount of nucleic acid residues present (de Groot, Til, and Feron, 1970).
G Duration of Experiment
Exposure to the test chemical usually commences when the animals are a
few weeks old (4-6 weeks) and have been weaned for about 2 weeks. It
is then continued for the lifespan of the animal (IARC, 1980). Strictly
speaking, ‘lifespan’ means until the animal dies either from cancer
induced by the chemical or from some naturally occurring tumour or
other disease. Experience has shown that this type of ‘lifespan’ study is
impracticable. In every experiment, there are usually some animals which
substantially exceed the average lifespan and a ‘lifetime’ experiment in
rats could easily go into the fourth year, adding considerably to the cost.
P. Grasso 345
There is a more cogent reason against this sort of practice. Normally by

27 months, most strains of rats do not exhibit an incidence of tumours
greater than 25% (approximately the same as that of man). After this
date, the incidence rises exponentially and in one ‘lifetime’ study
conducted at the British Industrial Biological Research Association
(BIBRA) on nitrosamines, the tumour rate in untreated controls reached
a level of 80% in animals that survived beyond 27 months and up to 34
years (P. Grasso, personal observation). This would mean that the longer
the animals are allowed to live, the greater will be the ‘background noise’
and the less ‘sensitive’ in statistical terms will the experiment become. It
is likely that similar problems occur in experiments with mice and
hamsters.
Aware of this difficulty, some investigators advocate the termination of
the experiment at 2 years in rats and at 18 months in mice and hamsters.
Such a regime is, however, too inflexible and might mean that the
animals are killed at the time of development of tumours induced by the
test substance.
The International Agency for Research on Cancer (IARC), in their
Supplement 2 (1980) (IARC, 1980) draw attention to the difficulties
involved in deciding when to terminate a long term test and suggest a
sensible compromise which is worth citing: ‘the survivors in all groups are
killed and the whole experiment is terminated if mortality in the control
or low dose group ever reaches 75%. However,. . . . the study is not
continued beyond week 130 of age of rats, 120 for mice and 100 weeks
for hamsters, irrespective of mortality’. They also add this important
rider: ‘An experiment is not really satisfactory . . . . if the mortality in
the control or low dose group is higher than 50% before the end of week
104 of age for rats, week 96 for mice and week 80 for hamsters’. This may
perhaps be a counsel of perfection but it is worth bearing in mind.
H The Numbers per Group

It is essential to have sufficient animals at the end of a carcinogenicity
study to carry out a valid statistical analysis. Experience has shown that,
in general, this can normally be achieved by having group sizes of 50
males and 50 females per dose level at the beginning of the experiment,
provided that no excessive deaths from natural causes occur. Smaller
numbers may make it impossible to evaluate differences in tumour
incidences between test and control even if natural wastage is average,
while increasing group sizes beyond the numbers given will not improve
matters to an extent commensurate with the increase in effort (and cost).
In spite of these considerations, it is not always prudent to adhere
rigidly to these numbers, particularly in the case of mice and hamsters,
where, on occasion, an unexplained higher than expected mortality may
occur. Ideally, groups sizes of about 80 to 100 of each sex per dose group
in both rats, mice, and hamsters at the start of the experiment would be
recommended.
An important issue in the design of carcinogenicity testing is the

number of animals to be allocated to control groups and the type of
controls that should be ‘carried’ in an experiment. Obviously, untreated
control groups of the same number per sex as the test groups are essential
to act as a reference point (i.e. to give some idea about the natural
incidence of tumours) but it is unrealistic to have a single control group if
3 or more dose levels are contemplated, since this would increase the
chance of obtaining a false ‘negative’ or ‘positive’ result. In fact, 2 control
groups (i.e. double the test group size) should be included if the number
of dose levels in the experiment is 3 or more.
The term ‘untreated’ does not mean that the animals are left entirely
alone. To do this would mean that ‘non-specific stress’ factors which
usually accompany certain test procedures, such as the restraint essential
in ‘head only’ exposure in inhalation tests, repeated injection in
subcutaneous tests, or the handling necessary in gavage experiment, will
affect only the test animals. Since such procedures could affect the
incidence of tumours it is essential that the controls be subject to
procedures of this sort as well. In some instances it may be necessary to
have also one control group which is essentially left untreated and
undisturbed.
I Positive Controls
It has sometimes been the practice to include another group or groups of
animals treated with a known carcinogen. This group then served as a
‘positive control’ designed to demonstrate that the strain of animals used
can respond to carcinogens. This practice is less favoured today and
some authorities go as far as to recommend against it (Committee on
Carcinogenicity of Chemicals in Food, Consumer Products and the
Environment, 1978; Page, 1977). Nevertheless, if the chemicals under
test are close in chemical structure to that of known carcinogens, it may
be advisable to include a positive control of an additional small group of
animals (30 of each sex).
J Historical Controls
It is often recommended that as much information as possible is obtained
on the ‘background’ incidence of tumours in the particular strain of
animals employed in a specific test (IARC, 1980). This background
incidence can be obtained from the breeder or from other long term tests
using the same strain. It provides useful information on tumour incidence
and can be of value in the interpretation of a borderline or doubtful
‘negative’ or ‘positive’ result , particularly if there is an unusual incidence
of tumours (high or low) in concurrent controls. The information
obtained from sources of this sort are not as valuable as those obtained
from the concurrent control because the diet of animals or the ‘non-
specific stress’ factors may have been different. Both of these, but
P . Grasso 347
particularly the diet, may influence profoundly the incidence of certain

tumours.
K Analysis of Diet
The basic composition of the diet needs to be known with the aim of
ensuring that an adequate amount of essential minerals and vitamins is
present. Analysis for the main components can often reveal the presence
of excess of fat or proteins (IARC, 1980). Knowledge of the composition
of the feed is of particular importance when the test material itself is a
nutrient (e.g. industrially treated protein or starch, single-cell protein, or
irradiated food). This is essential to ensure that an appropriately
balanced diet is provided, since these materials may be incorporated at
levels as high as 20-60% (de Groot, Til, and Feron, 1970).
Analysis of diet should include a search for common dietary con-
stituents which may influence carcinogenesis. Some of the constituents to
be looked for are antioxidants, chlorinated hydrocarbons, substances
with oestrogen-like activity, nitrites and nitrates, nitrosamines, heavy
metals, polycyclic aromatic hydrocarbons, and mycotoxins.
Analysis of diet is also important to ensure that the concentration of
the test substance is as close to the desired level as possible. It may reveal
that some degradation of the test compound has occurred with the
production of substances that may possibly interfere with the conduct of
the test.
L Two-generation Studies
These have been advocated in the belief that a carcinogen stands a better
chance of being detected if exposure is commenced in utero (IARC,
1980). With this in view, it is usual to mate animals dosed with the
required dose of the test substance and then to commence treatment of
the offspring soon after weaning. The offspring are then treated with the
carcinogen in the conventional way.
Despite the theoretical attraction of this model, based on evidence that
foetal tissues are more sensitive to carcinogens than those of adults of the
same strain (Toth, 1968), it may not be possible to obtain meaningful
results in a specific test. First because the transplacental passage is known
to vary considerably for the same compound within the same litter and it
is likely to differ even more so between different litters (Ruddick,
Ashanullah, Craig, and Stavric, 1978). Secondly, the rate and site of
metabolism of the compound may influence considerably the availability
of the compound itself or of its reactive metabolites for transplacental
passage. Finally, there is the possibility that the compound may be
inactivated by the maternal tissues, by the placenta, or by the mammary
glands (before weaning) (IARC, 1980).
The imponderables in such studies outweigh the theoretical sensitivity
of the foetal tissues to carcinogens. Considering the substantial increase

in the duration of the experiment and in cost, this type of approach is not
recommended.
M Observations During the Test

A careful check needs to be kept of the food and water consumption, of
the body weights of the animals, and of their state of health and general
behaviour. These are standard procedures in any well kept animal house
so that no further details need to be given here. Data from urinalysis, if
required, could be obtained without undue discomfort to the animals.
However, the infliction of some injury is unavoidable in obtaining blood
for haematological data and may cause some degree of stress.
If other types of investigations are required, for example the use of
radiolabelled tracers, it is advisable to have separate or ‘satellite’ groups.
The same advice applies if it is necessary to carry out interim kills. These
extra groups should be kept apart from the main experiment and
reported on separately.
N Conducting the Necropsy

The necropsy is probably one of the most important events in the course
of carcinogenicity testing. Unless it is properly conducted, valuable
information may be lost and this may not only reduce the value of the
test but may make it uninterpretable. Every effort should be made to
describe carefully, preferably with diagrams or pictures (the Polaroid-
type camera is a useful gadget in this respect) of every ‘lump and bump’,
and a tentative diagnosis made. This sort of information can be of great
help later on when examining the tissues microscopically. In fact some
authors (Roe, 1981) think (and for many a good reason) that thorough-
ness at the necropsy table is just as important as proficiency at
histological examination, and probably more so. In order to minimise
errors at necropsy the following points are worth considering:
(a) A careful watch should be kept on the state of health of animals so
that those which show signs of being ‘in extremis’ are killed before
they die naturally. This will diminish the risk of autolysis.
(b) The necropsy team for a particular study should consist as far as
possible of the same individuals (even at weekends) in order to
ensure uniformity.
(c) A standard number of specimens should be taken from each organ
and as far as possible from the same region (e.g. thryoid gland to
include parathyroids, adrenals to include the medualla, etc. ).
(d) If, for any reason, more than the standard number of samples are
taken from any organ, additional samples should be taken from
controls.
(e) It is desirable to give an opinion whether the tumour found was the
cause of death or not.
P. Grasso 349
0 The Histological Examination

It is usual to take tissues from a pre-determined set of organs and to have
a standard number of sections from each tissue sample. Where tumours
are large it is prudent to take a number of sections from various sites to
ensure that the sections read are truly representative. The histologist
must make every effort to be consistent in the diagnostic nomenclature.
This may prove more difficult with unfamiliar tumours at the ‘first time
round’ and may necessitate re-reading of the slides. The histologist must
also grade the tumours into benign or malignant and, if possible, give the
grade of malignancy. Any evidence of invasion or metastasis is of crucial
importance and should be reported.
The pathologist must confirm or revise the opinion given at necropsy
about the cause of death. If it is uncertain whether the tumour is the
cause of death, some form of words may be used to express this
uncertainty. A proposed form of words would be:
(a) cause of death
(b) probable cause of death
(c) probably not the cause of death
(d) incidental finding - not cause of death.
The principles involved have been described in detail by Glaister (1986).
3 Evaluation of Results
The objective of a carcinogenicity study is to determine the ability of a
substance to enhance tumours in animals. An increase in tumour
incidence in the test animals, as compared with controls, throws some
degree of suspicion on the compound’s ability to cause cancer but is not
sufficient to attribute causally the increased tumour incidence to the test
compound.
A number of factors need to be considered before such a causal link
can be established. An important first step is to establish whether any
difference in tumour incidence between test and controls could have
occurred by chance. There are a number of statistical methods that could
be used to achieve this end. [This subject is discussed in detail in IARC
Supplement 2 (IARC, 1980)l. But even if the difference is statistically
significant, it may not be sufficient to establish a cause-and-effect
relationship. An important consideration is the dose-response
relationship.
A Dose-Response Relationship
Over the last 3 or 4 decades, several experiments have shown that the
larger the dose of a carcinogen, the greater the number of tumours it will
induce. In this respect, carcinogens resemble most other toxic and
pharmacologic substances which elicit a stronger response the higher the
dose. The basis of the dose-response relationship of carcinogens was
3 50 Testing for Carcinogenicity
established by Druckrey in the late 1950s and has been confirmed amply
in subsequent experiments (Druckrey and Schmahl, 1962). This link
between dose and response is one of the most important pieces of
evidence causally linking a test compound with an increased incidence of
tumours. At least three dose levels are essential in the experimental
design to make it possible to establish such a link, although the greater
the number of dose levels (within reason) the better can such a link be
established.
Data from dose-response relationships are sometimes employed to
estimate the expected tumour incidence at dose levels very much lower
than those which could possibly be employed in conventional experi-
ments. A number of mathematical models have been employed in such
estimates (FDA, 1971; IARC, 1980). They include (1) the tolerance
distribution models-the probit and logit method are the best known
(Cornfield, Carlborg, and Van Ryzin, 1978; Hill, 1963); (2) simple linear
extrapolation (Gross, Fitzhugh, and Mantel, 1970); (3) various hit
models-for example the ‘one hit’ (Hoel, Gaylor , Kirschtein, Saffiotii,
and Schneiderman, 1975), the ‘multi-hit’ model (Cornfield, Carlborg, and
Van Ryzin, 1978); (4) the Armitage-Doll multi-stage model (Armitage
and Doll, 1961); (5) the Weibull model (Peto and Lee, 1973); (6) a
simplified statisticopharmacokinetic model (Cornfield, 1977); (7) time to
tumour (Hogan and Hoel, 1982).
The value of these models in yielding reasonably reliable estimates of
risk of cancer from low levels of carcinogens were critically assessed by
the Scientific Committee of the Food Safety Council (Food Cosmet.
Toxicol., 1980, 18, 711). The Committee expressed the view that the
choice of the mathematical model for low-dose extrapolations has no firm
biological basis and must, to some extent, be arbitrary. However, they
recommend the use of the more flexible models such as the Armitage-
Doll, Weibull, or multi-hit. In particular, such models are recommended
where a linear response at low doses would be difficult to imagine, for
example in those instances where a carcinogenic response is thought to
result from prolonged tissue damage. On the other hand, the Committee
is of the opinion that in the case of classical electrophilic carcinogens (e.g.
mutagenic agents), then the models which assume a linear response at
very low doses may well have an important role to play. Nevertheless,
data from rodent carcinogenicity could itself be misleading due to the
high levels given to animals when compared to the low levels of human
exposure (Clayson, 1987).
4 Nature and Type of Turnour Induced

A Malignant and Benign Tumours
In order to establish the carcinogenic property of a test compound, some
of the tumours induced must be malignant. Although there is no general
agreement with this view it would seem illogical to label a compound as a
P. Grasso 351
‘carcinogen’ unless it has been shown to induce ‘cancer’ in experimental

animals and ‘cancer’ is widely held to be the occurrence of a fatal
malignant tumour.
This statement does not mean that benign tumours are to be ignored.
In the experience of most workers in this field, chemical carcinogens
induce both benign and malignant tumours. Many hold that benign
tumours may progress into malignant tumours (Foulds, 1975), while
others hold the view that benign tumours represent an ‘end-point’ and do
not progress further. At the moment it is doubtful whether this
controversy could be resolved, since the evidence for either view is
inconclusive, but the fact that both types of tumours of the same
histogenetic origin occur in animals treated with carcinogens indicates
that both must be taken into account when assaying chemical car-
cinogenesis. Benign tumours do not have the same significance as
malignant tumours, however, and could be viewed with less concern. The
scheme recently devised by Squires supports this view and attempts to
give some numerical rating to distinguish the gravity between these two
types of tumours (Squires, 1981).
Hormones, compounds with hormone-like actions, and a gross im-
balance in normal hormonal homeostasis, however it is brought about,
can enhance both benign and malignant tumours (Foulds, 1975). The
proportion of each type of tumour induced varies with the organ
involved. For example, tumours of the pituitary and of the islet cells of
the pancreas tend to be principally benign, but mammary tumours tend
to have a fairly high proportion of malignancies, at least in some strains
of rats and mice (Foulds, 1975).
5 Factors Affecting Tumour Incidence

Certain types of tumour are less reliable as an index of carcinogenicity
than other types, irrespective of their benign or malignant nature. Such
tumours often have a high and variable natural incidence. Their incidence
often can be changed by clearly definable factors unconnected with the
test substance, or else it can be determined by factors which are clearly
connected with local pathological changes in the target organ.
Some of t h e more important factors which can influence the incidence
of tumours in a carcinogenicity experiment will be outlined and some
indication will be given of the way in which they could act as
‘confounding’ factors in the evaluation of results:
A Genetic
The development of three commonly occurring tumours in mice (pulmo-
nary adenoma, mammary adenocarcinoma, and lymphoma) appears to
be under some form of genetic control. The pulmonary adenoma appears
to be the best studied in this respect. This tumour originates from the
Type I1 pneumocyte and is, generally speaking, benign in nature. The
high natural incidence of this tumour (e.g. 7% in C57BL and 71% in
Strain A) has led several investigators to study the genetic factors
involved in its development. According to Bentvelzen and Szalay (1966)
a single gene is reported to have a major effect in determining the high
incidence of this tumour in some strains, while others seem to have found
evidence that more than one pair of genes are involved (Shimkin and
Stoner, 1975). Earlier experiments had shown that hybrids of a high-
tumour and low-tumour strain resulted in F1 generations that resembled
the high-tumour strain. The back-cross generations resembled the strain
to which the F1 mice were mated. This was interpreted as indicating that
susceptibility to the development of pulmonary tumours was inherited in
a dominant manner (Falconer and Bloom, 1964; Heston and Ulahakis,
1961).
Genetic factors play a less dominant, but none the less important, role
in murine mammary neoplasia. This type of tumour was found as long
ago as 1911 (Murray, 1911), to occur in a high incidence in the female
progeny of mice whose mothers also displayed a high incidence of this
tumour. Investigations by several authors have now established that a
genetic basis for mammary tumour development does exist but that other
factors also play an important role (Nandi and McGrath, 1973).
The gene operations involved in the development of lymphoreticular
tumours in mice were for a long time overshadowed by the active role of
tumour viruses in the development of this tumour. It would now appear
that at least three genes are involved, namely Fv-1 which governs the
likelihood that leukaemia virus is successful in producing the disease, the
Fv-2 which determines the capacity of cell transformation, and H-2 which
determines the host response to the tumour antigens (Lilly and Pincus,
1973).
Although genetic factors in relation to tumour development have not
been as clearly defined in the rat as they are for some of the mouse
tumours, there is some reason for suspecting that a genetic basis exists for
the development of some of the rat tumours, particularly the leukaemias
in F344 rats and the mammary gland tumours of Sprague-Dawley rat
because of their high incidence (Gala and Loginsky, 1973; Moloney,
Boschetti, and King, 1970).
B Hormonal
Under naturally occurring conditions, hormones appear to play a part in
determining the incidence of mammary tumours in mice. It has long been
known that mammary gland turnours occur with much greater frequency
in female than in male mice of any strain (Bittner, 1957). Furthermore,
there is considerable experimental evidence to show that oestrogens
(both synthetic and naturally occurring) and prolactin can enhance the
P. Grasso 353
incidence of this type of tumour (Gardner, Pfeifter, and Trentin, 1959).

It is now acknowledged, however, that hormones are only one of the
factors that determine the incidence of this tumour, and that genetic and
viral factors operate as well (Nandi and McGrath, 1973).
Natural or synthetic oestrogens are known to enhance the development
of mammary tumours in certain strains of rat, e.g. Sprague-Dawley and
Charles River, under experimental conditions (Gibson, Newberne,
Kuhn, and Elsea, 1967). It would seem reasonable to assume that they
may also have some role in the development of these tumours under
natural conditions.
C Viruses
Although several different types of viruses are known to be capable of
inducing tumours in a variety of organs in the mouse, when experimen-
tally introduced in the host, it would seem that under natural conditions
only the lymphoma virus causes tumours in any substantial numbers
(Grasso, Crampton, and Hooson, 1977). Lymphoma viruses in mice have
been known to exist since their discovery by Gross in 1951 (Gross, 1951).
It is now known that such viruses are widespread in mice, although they
produce the disease in only a relatively small proportion of infected
animals (Grasso, Crampton, and Hooson, 1977). The ‘stress’ imposed by
the experimental procedure (e.g. restraint during an inhalation experi-
ment) may cause a latent infection to become manifest (Lemonde, 1959).
A test chemical may well have the same effect through some mechanism
other than ‘stress’ (such as immune suppression) which would tip the
balance in favour of virus activity.
It is not known whether a similar situation occurs in rats. When
oncogenic viruses are introduced experimentally in rats, tumours develop
in the tissue sensitive to the particular virus (Gross, Schidlovsky,
Feldman, Dreyfus, and Moore, 1975) but there is no substantial evidence
as yet to indicate that an analogous situation exists under natural
conditions. There is some suspicion that the high incidence of mono-
nuclear cell leukemia in F344 rats may have some viral etiology but no
virus has been isolated as yet (Moloney, Boschetti, and King, 1970). The
high incidence of mammary tumours in Sprague-Dawley rats could also
conceivably have a viral origin since C-type virus readily induces
mammary tumours when injected into rats but there is no experimental
evidence to support this suggestion (Gross, Schidlovsky , Feldman,
Dreyfus, and Moore, 1975).
D Diet
The incidence of tumours of the pituitary, the liver, and mammary glands
in mice can be influenced by both the calorie and protein content of the
diet (Roe, 1979). It has been shown by a number of authors that
3 54 Testing for Carcinogenicity
reduction of food intake results in a marked reduction of these tumours

and in a better survival of the animals (Conybeare, 1980). The beneficial
effect apparently results from a restriction of the calorie content rather
than from the reduction of any particular type of food (Tannenbaum,
1959). Quite modest reductions in food intake (6%) are sufficient to
achieve a substantial reduction of tumours and to increase longevity
(Tucker, 1979).
Increase in protein and fat intake, however, has the reverse effect,
particularly in the case of liver tumours: substantial increases in tumour
incidence have been shown to result from an increase in these two dietary
components (Gellatly, 1975). A high fat intake can also result in a high
incidence of mammary tumours (Tannenbaum, 1959).
Similar results occur in rats in the case of pituitary and mammary
tumours. Other tumours, including the liver, do not appear to be affected
in this species (Conybeare, 1980).
E Other Factors
Apart from diet, hormones, viruses, and genetic constitution of the
animals, there are a number of other factors which may make the
induction of certain tumours uninterpretable in terms of risk to man.
Such factors are usually physico-chemical in nature and do not involve
damage to the DNA, but they cause injury to the tissue in which cancer
will ultimately develop. Factors of this type operate in:
(a) The induction of subcutaneous sarcoma in rodents by the surgical
implantation of solids of diverse chemical composition in the
subcutaneous tissue. It has been shown by many investigators that
induction of these tumours occurs when large pieces of solid,
biologically inert material (e.g. gold, glass) (approximately 2 cm’)
are implanted, but the same material implanted in shredded or
powder form does not produce any local tumours (Bischoff and
Bryson, 1964).
(b) The induction of subcutaneous sarcoma by repeated injection of
solutions or suspensions of diverse materials at the same site for
several weeks. Glucose and table salt in hypertonic solution,
hydrochloric acid at pH 1, and surface active agents have been
shown to induce local sarcomas under these conditions (Grasso,
1976).
(c) Transitional cell carcinoma of the urothelium from stones formed
endogenously in the bladder or from foreign bodies implanted
surgically. Glass beads and cholesterol crystals produce tumours
readily when implanted in this viscus (Ball, Field, Roe, and
Walters, 1964).
(d) Carcinoma of thyroid by compounds which interfere with the
synthesis of secretion of thyroxine (Innes, Ulland, Valerio, et al.,
1969).
P. Grasso 355
(e) Lympho-reticular tumours from immunosuppression by anti-

lymphocyte serum (Gleichmann and Gleichmann, 1973).
(f) In cancer of the liver when chronic liver injury is caused by
non-genotoxic agents. Injury of this sort may involve repeated
cycles of necrosis and proliferation such as may occur when large
doses of hepatotoxic agents (e.g. CC14) are administered or it may
involve some biochemical change in the liver which, in some way
which is not quite understood, leads to nodular hyperplasia and
then to cancer (e.g. hypolipidemic agents and peroxisome prolifer-
ation) (Cohen and Grasso, 1981).
It is not clear whether physico-chemical factors operate as well in other
tissues, e.g. skin, but the evidence available from subcutaneous tissue,
urothelium, and the other tissues mentioned in the preceding paragraph
indicate that some caution needs to be exercised in the interpretation to
be given to tumour induction in rodents in conventional long term tests.
In particular, one should seek to exclude the operation of the
‘confounding’ factors outlined in this section before concluding that a
compound is carcinogenic. In order to do this, one has to apply common
sense and knowledge gained from experience. Only then will the
biological significance of any excess tumour incidence in test over control
become apparent.
6 Conclusions
Carcinogenicity studies require careful planning and execution and an
even greater care in interpretation. A false positive result may cause
unnecessary worry among those who are exposed to a widely used
chemical, which is quite harmless but which has given rise to tumours
because of the factors discussed above. It may also lead to the loss of a
valuable chemical. A false negative result, on the other hand, may give
rise to a false ‘sense of security’ and may result in the tragic development
of tumours among those exposed.
References
Armitage, P. and Doll, R. (1961). Stochastic models for carcinogenesis. In:
‘Proceedings of the 4th Berkeley Symposium on Mathematical Statistics and
Probability’, No. 4. Eds Lecam and Neyman.
Ball, J. K . , Field, W. E. H., Roe, F. J. C . , and Walters, M. (1964). The
carcinogenic and co-carcinogenic effects of paraffin wax and glass beads in
the mouse bladder. Br. J . Urol., 36, 225.
Bentvelzen, P. A. J. and Szalay, G. (1966). Some genetic aspects of difference in
susceptibility to the development of lung tumours between inbred strains of
mice. In: ‘Lung Tumours in Animals’. Proceedings of the 3rd Quadrennial
Conference on Cancer, University of Perugia, 1965. Ed. L. Severi. p. 835.
Berenblum, I. (1955). The significance of the sequence of irritating and
promoting actions on the process of skin carcinogenesis in the mouse. Br. J .
Cancer, 9, 268.
Bischoff, F. and Bryson, G. (1964). Carcinogenesis through solid state surfaces.

In: ‘Progress in Experimental Tumour Research’. Vol. 5. Ed. F. Homberg.
p. 85. S. Karger, Basle.
Bittner, I. 1. (1957). Recent studies on the mouse mammary tumour agent CMTA.
Ann. N . Y . Acad. Sci., 68, 636.
Boyland, E., Roe, F. J. C., and Mitchley, B. C. V. (1966). Tests of certain
constituents of spermicide for carcinogenicity in genital tract of female mice.
Br. J . Cancer., 20, 18.
Cairns, T. (1979). The EDo, Study: Introduction, objectives and experimental
design. J . Exp. Pathol. Toxicol., 3, 1.
Clayson, D. B. (1962). In: ‘Chemical Carcinogenesis’. p. 135. J. and A. Churchill
Ltd., London.
Clayson, D. B. (1987). ICPEMC Publ. No. 13. The need for biological risk
assessment in reaching decisions about carcinogens. Mutat. Res., 185,
243-269.
Cohen, A. J. and Grasso, P. (1981). Review of the hepatic response to
hypolipidaemic drugs in rodents and assessment of its toxicological sig-
nificance to man. Food Cosmet. Toxicol., 19, 585-605.
Committee on Carcinogenicity of Chemicals in Food, Consumer Products and the
Environment (1978). Guidelines on carcinogenicity testing. London.
Conybeare, G. (1980). Effect of quality and quantity of diet on survival and
tumour incidence in outbred Swiss mice. Food Cosmet. Toxicol., 18, 65.
Cornfield, J. (1977). Carcinogenic risk assessment. Science, 198, 693.
Cornfield, J., Carlborg, F. W . , and Van Ryzin, J. (1978). Setting tolerance on the
basis of mathematical treatment of dose-response data extrapolated to low
doses. Proc. First Int. Toxicology Congress.
de Groot, A. P., Til, H. P., and Feron, V. J. (1970). Safety evaluation of yeast
grown on hydrocarbons. I. One year feeding study in rats with yeast grown
on gas-oil. Food Cosmet. Toxicol., 8, 267-276.
Druckrey , H. and Schmahl, D. (1962). Quantitative analyse der experimentellen-
tellen Knebserzeugung. Naturwissenschaften, 49, 19.
Falconer, D. S. and Bloom, J. L. (1964). Changes in susceptibility to urethane
induced lung tumours produced by selective breeding in mice. Br. J . Cancer,
18, 322.
Festing, F. W. (1975). A case for using inbred strains of laboratory animals in
evaluating the safety of drugs. Food Cosmet. Toxicol. 13, 369.
Food and Drug Administration Advisory Committee on Protocols for Safety
Evaluation (1971). Panel on Carcinogenesis Report on Cancer Testing in the
Safety of Food Additives and Pesticides.
Food Safety Council (1978). Chronic toxicity testing. Proposed system for food
safety assessment. Food Cosmet. Toxicol., 16, Suppl. 2, 97-108.
Foulds, L. (1975). ‘Neoplastic Development’. Vol. 2. p. 6. Academic Press,
London.
Foulds, L. (1975). ‘Neoplastic Development’. Vol. 2. p. 345. Academic Press,
London.
Gala, R. R. and Loginsky, S. J. (1974). Correlation between serum prolactin
levels and incidence of mammary tumours induced by 7,12-dimethylbenz[a]-
anthracene in the rat. J. Natl. Cancer Inst., 51, 593-597.
Gardner, W, U., Pfeifter, C. A., and Trentin, I. I. (1959). Hormone factors in
experimental carcinogenesis. In : ‘The Physiopathology of Cancer’. 2nd Edn.
p. 152. Ed. F. Homburger. Hoeber Harper, New York.
P. Grasso 357
Gellatly, I. B. M. (1975). The natural history of hepatic parenchymal nodule

formation in a colony of C57BL mice with reference to the effect of diet. In:
‘Mouse Hepatic Neoplasia’. Ed. W. H. Butler and P. M. Newberne. Elsevier
Scientific Publishing Company, Amsterdam.
Gibson, J. P., Newberne, J. W., Kuhn, W. L., and Elsea, J. R. (1967).
Comparative chronic toxicity of three oral oestrogens in rats. Toxic. Appl.
Pharmacol., 11, 489.
Glaister, J. (1986). ‘Principles of Toxicological Pathology’. Taylor and Francis,
London and Philadelphia. pp. 1-223.
Gleichmann, H. and Gleichmann, E. (1973). Immunosuppression and neoplasia.
I. A critical review of experimental carcinogenesis and the immuno-
surveillance theory. Klin. Wochensch., 51, 255.
Grimmer, G., Dettbarn, G., Brune, H., Deutsch-Wenzel, R., and Misfeld, J.
(1982). Quantification of the carcinogenic effect of polycyclic aromatic
hydrocarbons in used engine oil by topical application onto the skin of mice.
Int. Arch. Occup. Environ. Health, 50, 95.
Grasso, P. (1976). Review of tests for carcinogenicity and their significance to
man. Clin. Toxicol., 9, 745-760.
Grasso, P. (1979). Carcinogenic risk from food-real or imaginary? Chem. Ind.
(London), 3rd February.
Grasso, P. and Crampton, R. F. (1972). The value of the mouse in car-
cinogenicity testing. Food Cosmet. Toxicol., 10, 418.
Grasso, P., Crampton, R. F., and Hooson, J. (1977). The mouse and car-
cinogenicity testing. The British Industrial Biological Research Association,
Woodmansterne Road, Carshalton, Surrey, UK.
Grasso, P. and Golberg, L. (1966). Subcutaneous sarcoma as an index of
carcinogenic potency. Food Cosmet. Toxicol., 4, 297.
Gross, L. (1951). Spontaneous leukaemia developing in C3H mice following
inoculation, in infancy, with AK-leukaemic extracts, or AK-embryos. Proc.
Soc. Exp. Biol. Med., 76, 27.
Gross, L., Schidlovsky, G., Feldman, D. M . , Dreyfus, T., and Moore, L. A.
(1975). Proc. Natl. Acad. Sci. (USA), 72, 3240-3244.
Gross, M. A . , Fitzhugh, 0. G., and Mantel, N. (1970). Evaluation of safety for
food additives: an illustration involving the influence of methyl salicylate on
rat reproduction. Biometrics, 26, 101.
Heston, W. E. and Ulahakis, G. (1961). Elimination of the effect of the AYgene
on pulmonary tumours in mice by alteration of its effects on normal growth.
J . Natl. Cancer Inst., 27, 1189.
Hill, B. M. (1963). The three parameter lognormal distribution and Bayesian
analysis of a point-source epidemic. J . A m . Stat. Assoc., 58, 72.
Hoel, D. G., Gaylor, D. W., Kirschstein, R. L., Saffiotii, V., and Schneiderman,
M. A. (1975). Estimation of risks of irreversible delayed toxicity. J . Toxicol.
Environ. Health, 1, 133.
Hogan, M. D. & Hoel, D. G. (1982). Extrapolation to man. In: ‘Principles and
Methods of Toxicology’. Ed. A. W. Hayes. Raven Press, New York.
IARC Working Group (1979). ‘Monographs on the Evaluation of the Car-
cinogenic Risk of Chemicals to Humans’. Suppl. l. International Agency for
IARC Working Group (1980). ‘Monographs on the Evaluation of the Car-
cinogenic Risk of Chemicals to Humans’. Suppl. 2 . Long term and short term
screening Assays for carcinogens: a critical appraisal. International Agency

for Research on Cancer, Lyon.
Innes, J. R. M., Ulland, B. M., Valerio, M. G . , Petrucelli, L., Fishbein, L.,
Hart, E. R., Pallotta, A. J., Bates, R. R., Falk, H. L., Gart, J. J., Klein,
M., Mitchell, I. and Peters, J. (1969). Bioassay of pesticides and industrial
chemicals for tumorigenicity in mice. A preliminary note. J . Nutl. Cancer
Inst., 42, 1101.
Kipling, M. D. (1976). Scots, tars and oils are causes of occupational cancer. In:
‘Chemical Carcinogens’. Ed. C. E. Searle. ACS Monograph 173. American
Chemical Society, Washington, DC.
Lemonde, P. (1959). Influence of fighting on leukaemia in mice. Proc. SOC. Exp.
Biol. Med., 102, 292.
Lilly, F. and Pincus, T. (1973). Genetic control of murine viral leukaemogenesis.
Adu. Cancer Res., 17, 231.
Mauderly, J. L. and Kritchevsky, J. (1979). Respiration of unsedated F-344 rats
and the effect of confinement in exposure tubes. In ‘ITRI’ Annual Report
1978-1979. LF-69 pp. 475-478. Available from the National Technical
Information Service, Springfield, Va.
Mead, J. (1960). Control of respiratory frequency. J . Appl. Physiol., 15, 325.
Ministry of Health and Welfare Canada (1973). The testing of chemicals for
carcinogenicity, mutagenicity and teratogenicity.
Moloney, W. C., Boschetti, A. E. and King, V. P. (1970). Spontaneous
leukaemia in Fischer rats. Cancer Res., 30, 41-43.
Munro, I. C. (1977). Considerations in chronic toxicity testing: the chemical, the
dose, the design. J . Envir. Path. Toxicol., 1, 183.
Murray, I. A. (1911). Cancerous ancestry and the incidence of cancer in mice.
Sci. Rep. Invest. Imp. Cancer Res. Fund, 4, 114.
Nandi, S. and McGrath, C. M. (1973). Mammary neoplasia in mice. Adv. Cancer
Res., 17, 353.
Page, N. P. (1977). Current concepts in a bioassay programme in environmental
carcinogenesis. In: ‘Environmental Cancer’. p. 114. Eds H. F. Kraybill and
M. A. Mehlman. John Wiley and Sons, New York, London.
Page, N. P. (1977). Concepts of a bioassay programme in environmental
carcinogenesis. In : ‘Environmental Carcinogenesis’. Eds H. Kraybill and M.
Mehlman. p. 7-171. Hemisphere Publishers, Washington.
Peraino, C., Fry, R. D. M., and Staffeldt, E. (1973). Enhancement of
spontaneous hepatic tumourigenesis in C3H mice by dietary phenobarbital.
J . Natl. Cancer Inst., 51, 1349.
Peto, R. and Lee, P. N. (1973). Weibull distributions for continuous-
carcinogenesis experiments. Biometrics, 29, 457-470.
Powell, C. H. and Hosey, A. D. (1965). The industrial environment, its
evaluation and control (US DHS Publ. 614).
Roe, F. J. C. (1979). Food and Cancer. J . Hum. Nutr., 33,405-415.
Roe, F. J. C. (1981). Testing in viuo for general chronic toxicity and
carcinogenicity. In: ‘Testing for Toxicity’. Ed. J. W. Gorrod. Taylor and
Francis Ltd., London.
Ruddick, J. A., Ashanullah, M., Craig, J., and Stavric, B . (1978). Uptake and
distribution of 14C-saccharin in the rat foetus. In: Proceedings, Canadian
Federation of Biological Societies, 21st Annual Meeting’, London, Ontario.
Abstract 635, p. 159.
P . Grasso 359
Sanockij, I. V. ( E d . ) (1978). Methods for determining toxicity and hazards of

chemicals. Moscow. Medicina. pp. 62-63 (in Russian); cited in WHO
Environmental Health Criteria..
Shimkin, M. B. and Stoner, G. D. (1975). Lung tumours in mice: application to
carcinogenesis bioassay. Adv. Cancer Res., 21, 1.
Squires, R. A. (1981). Ranking Animal Carcinogens. A proper regulatory
approach. Science, 214, 877-880.
Tannenbaum, A. (1959). Nutrition and cancer. In: ‘Physiopathology of Cancer’.
2nd Edn. Ed. F. Homberger. pp. 517-62. Hoeber-Harper, New York.
Toth, B. (1968). A critical review of experiments in chemical carcinogenesis using
new born animals. Cancer Res., 28, 727-738.
Tucker, M. I. (1979). The effects of long-term food restriction on tumours in
rodents. Znt. J . Cancer, 23, 803-807.
Wagner, J. C. 1962). Experimental production of mesothelial tumours of the
pleura by implantation of dusts in laboratory animals. Nature, London, 196,
180.
WHO (1978). Environmental Health Criteria 6 . ‘Principles and Methods for
Evaluating the Toxicity of Chemicals’. Part I. World Health Organisation,
Geneva.
WHO (1978a). Factors influencing the design of toxicity studies. In: ‘Principles
and Methods for Evaluating the Toxicity of Chemicals’. Part 1. World Health
Organisation, Geneva. pp. 62-94.
WHO (1978b). Inhalation exposure. In : ‘Principles and Methods for Evaluating
the Toxicity of Chemicals’. Part 1. World Health Organisation, Geneva. pp.
199-235.
Yamigawa, K . and Ichikawa, K. (1918). Experimental study of the pathogenesis
of carcinoma. J . Cancer Res., 3, 1.
CHAPTER 16
In Vitro Methods for

Teratology Testing
DIANA ANDERSON
1 Introduction
Concern with congenital disease and malformations is ancient (Mark 9,
17-19) but prior to 1960 the only recommended protocol for testing
chemicals during the reproductive cycle of animals was the 6 weeks’
toxicity test in male and female rodents, which was evaluated over two
pregnancies with a subsequent assessment of foetal survival. Following
the thalidomide episode more stringent guidelines were adopted for
reproductive studies for safety evaluation. Testing models for teratologic
effects of foreign chemicals continue to rely chiefly upon mice, rats, and
rabbits but they are, of necessity, time consuming and expensive to
perform. Extrapolating findings from mammalian teratology testing
models to those agents with the potential to cause human birth defects
can never be absolute, since negative results can give no guarantee that a
chemical will lack teratogenic effects in man, due to species variation in
teratogenic response (Wilson, 1977). Effects at high doses will not
necessarily be the same as at low doses. This, of course, is true for most
branches of toxicology.
Because of inherent problems in models for mammalian teratology , the
search continues for identifying predictive tests and various in uitro
models have been established. For simplicity, in uitro here refers to the
use of test subjects other than the pregnant mammal. The purpose of in
vitro testing would be to identify compounds which require further whole
animal testing rather than to define the pre-natal toxicity of the
compounds. Smith et al. (1983) have suggested a selection of candidate
compounds which might be used for in vitro test validation although
Jensen et al. (1989) have criticised the list of compounds.
Selection of a representative tissue is important in the development of
360
Diana Anderson 36 1
Table 1 In Vitro Teratogenicity Test Systems

Sub-mammalian systems Mammalian systems
Invertebrate Embryos ModiJied Whole Mammal System

Nematodes The Chernoff Test
Coelenterates
Isolated Whole Embryos
Echinoderms
Pre-implantation
Insects
Post-implantation
Vertebrate Embryos
Isolated Embryonic Organs
Pisces
Limb buds
Amp hibia
Palate
Aves
Isolated Cell Systems
Rat embryo limb and mid-brain cells
Neural crest cells
Chinese hamster ovary V79 cells and human
embryonic and palatal mesenchyme
Ascitic mouse ovarian tumour cells
Other cell lines
an in vitro assay. The test systems vary as to the type of target employed
and can be categorised according to whether DNA, single cells, organs,
or whole embryos are used to assess pre-natal toxicity. The main problem
is that little is known of the initial insults which lead to abnormal
development. The insult may result in a change of one or more of the
normal cellular functions of embryogenesis, such as proliferation, deter-
mination, aggregation, organisation, migration, and death. This may
result from intracellular damage, perhaps involving genetic damage or
inhibited enzymes, or extracellular damage such as altered cell mem-
branes. In vitro systems are essentially too simple in biological terms to
provide a complete model for a teratological event. Their value lies in
isolating and investigating the constituent elements of teratological
activity. However, some test systems have correctly predicted in vivo
teratogenicity (ECETOC, 1989).
A test system should optimally incorporate all of the processes of
growth and differentiation and detect retardation, malformation, and
death. Two types of systems mainly satisfy these criteria. Those using
developing embryos or embryo portions and those using regenerating
tissue. Examples are sub-mammalian systems using invertebrate embryos
[nematodes, coelenterates (Hydra), echinoderms (sea urchin), and in-
sects (Drosophila)] and vertebrate embryos [Pisces (zebra fish), am-
phibia (frog, salamander) and aves (chick)]; and mammalian systems
utilising isolated whole embryos (pre- and post-implantation) and isolated
embryonic organs (palate and limb-buds) and isolated cells in culture
(Table 1).
362 In Vitro Methods for Teratology Testing
2 Sub-mammalian Systems
A Invertebrates
Large numbers of adult and embryonic and developmental stages can be
maintained in the laboratory. The cell lineage of the nematode
Caenorhabditis elegans from the fertilised egg to 900 cell adult has been
elucidated (Sulston and Horvitz, 1977; Deppe et al., 1978) and may prove
a useful model. Planarians can also be used. Planarians can regenerate
damaged fragments; absent, abnormal, or retarded regeneration, or loss
of normal morphological appearances of intact animals are used as
indicators of toxicity. Planarians appear to respond well to mammalian
teratogens (Best and Monita, 1982). Similarly, coelenterates have proved
useful. The adult Hydra (Attenuata) can be disrupted into individual
cells, which on centrifugation form a pellet which may reform into a new
organism. These artificial ‘embryos’ (Johnson, 1980) together with the
adult have been used in an investigation of known teratogens and the
data correlate well with mammalian data (Johnson et al., 1982). Certain
echinoderms, particularly the sea urchin (Gross et al., 1972) can be
cultured in large numbers and have been used in experimental embryol-
ogy, and antimetabolite effects have been determined for embryogenesis
(Lallier, 1965). The fruit fly Drosophila has been used extensively for
short term tests for mutagenicity (Wurgler et al., 1977; de Serres and
Ashby, 1981) but there are a great deal of genetic and developmental
data for this organism and it may be suitable for teratogenicity studies
(Schuler et al., 1982). The mated female deposits eggs in a nutritive
medium containing the test substance. It takes 6 days, after passing
through larval and pupal stages, for the adults to emerge. These are
anaesthetised and examined for abnormal development with a binocular
microscope. In a study of 100 chemicals with single cell suspensions of
Drosophilia eggs, 43 out of 47 teratogens and 2 non-teratogens were
positive (Bournias-Vardiabasis et al., 1982; 1983).
Despite the large phylogenetic difference between these non-mam-
malian species and man, some of these in vitro tests have performed well in
predicting in vivo teratogenic activity (Collins, 1987; Daston and
D’Amato, 1989; ECETOC, 1989).
B Vertebrates
Fish , Amphibia
An integral part of toxicology studies on fish is the evaluation of the
effects of chemicals on the development of eggs, embryos, and larvae
(e.g. Birge et al., 1979) and early experimental teratology studies often
used fish as models (see Wilson, 1978). Fish such as zebra fish or
minnows are easy to maintain in the laboratory, as are amphibia such as
newts, salamander, and frogs. Dumart and co-workers (1982) have
developed a teratogenicity screen known as FETAX (Frog Embryo
Diana Anderson 363
Teratogenesis Assay) using Xenopus embryos. Mid- to late-blastula stage

embryos are exposed to test agents in water for up to 4 days and are
assessed for mortality, pigmentation, growth, extent of development, and
abnormalities. Preliminary validation of FETAX with various com-
pounds including known teratogens and non-teratogens suggests it may be
comparable to other screens (Dumont and Epler, 1984; Dawson et al.,
1989; Fort et al., 1988; 1989).
Birds
The avian embryo, particularly that of the chick, has been used to
investigate the direct effect of substances on developing morphogenetic
systems since maternal influences are excluded. The chick embryo
possesses its own intact metabolic system and comparative studies have
claimed a predictive value of the chick embryo system comparable to
other in uitro systems, as well as to whole animal systems (Jelinek, 1982).
WHO stated in 1967 that the use of the chick embryo for screening for
teratogenicity is not to be recommended due to the absence of maternal-
foetal relations, pharmacokinetic dissimilarities in the closed character of
the avian egg, and high non-specific sensitivity. Jelinek (1982) argues that
these three objections have arisen mainly from neglecting teratological
principles, poor standardisation of test subjects, and inadequate ad-
ministration techniques. Of late, the chick embryo has received more
favourable reviews (Wilson, 1978) and this model would seem to require
definitive validation.
The method is relatively simple. Eggs (generally White Leghorn) are
incubated at 30°C. The test agent is administered through a hole bored in
the shell, which may be resealed later with parafilm or wax. The chemical
may be placed in the yolk sac, sub-germinal cavity, allantois, amnion, or
air chamber (Wilson, 1978). Opinions differ as to appropriate treatment
time, varying from 0 to 96 hours, but a detailed summary of these
schedules and sacrifice times is available. The chick may be examined for
abnormalities during incubation, hatching, or at full maturity to evaluate
functional normality.
A more rapid ‘chick embryotoxicity screening test’ (CHEST) was
proposed (Jelinek and Rychter, 1980) based upon the caudal mor-
phogenetic system. Here, agents are administered directly below the
caudal region of the embryo. According to Jelinek (1982) the chick
embryo carries a complete set of morphogenetic control systems and has
an advantage over other in vitro systems that employ isolated embryonic
tissues which may have limited survival or lack a developing vascular bed,
a frequent target for teratogens. Jelinek et al. (1984; 1985) tested 130
compounds, of these 119 exerted dose-related embryotoxic effects and
the method permitted discrimination between substances toxic in low and
high doses, suggesting a discrimination for teratogenic potency. KuEera
and Burnand (1987) described an assay using Warren hen eggs.
3 Mammalian Systems
The various alternative test methods have been discussed by ECETOC
(1989) and Daston and D’Amato (1989).
A A Modification of a Whole Mammal System-The

Chernoff Test
This is a simplified in uiuo teratogenicity assay where pregnant mice are
treated at a single dose level (LDIO) with the test compound on days
8-12 of gestation. The new born are examined on days 1 and 3
post-partum for weight, litter size, and general appearance, and maternal
weight during treatment is also recorded (Chernoff and Kavlock, 1980).
Initially, 10 known teratogens affected one of the parameters measured
whilst 11 out of 12 non-teratogens did not, and an independent validation
with 55 compounds claimed a very good correspondence between
teratogens and non-teratogens (Brown et al., 1983). This test was
independently validated under contract to NIOSH and has been reviewed
(Hardin, 1987). Nevertheless, many mammals are still required to
perform the assay even though they are half the number required for the
usual animal study. It is at a different level than the other assays
described here and may be useful to confirm data from a pre-screen.
B Whole Embryo Cultures

The two major test systems involving whole embryo culture use pre- or
post-implantation embryos. The methodology of the two systems are well
reviewed by Flynn (1987).
Pre -implantation
Culture techniques for the early mammalian conceptus from the one cell
to blastocyst stages are well established (Daniel, 1971). The conceptus can
be removed from the dam, maintained in uitro for several days, and
re-implanted into a surrogate dam with development in term (Staples,
1971). Some authors have reported that the pre-implantation embryos
are refractory to treatment with teratogenic agents (Staples, 1975) and
high doses can result in embryo death, but instances of congenital
malformation have resulted from insult to the blastocyst (Wilson, 1978).
Hsu (1973; 1980) has perfected techniques to allow mouse blastocysts to
be grown in uitro to the early somite stage, but this test is as yet too
complicated for routine use and suffers from poor residual viability.
Various biochemical and cytogenetic parameters can be assessed fairly
readily (ECETOC, 1989). The DNA repair and metabolising capacity of
the system after treatment with the test compound should be assessed
(Spielmann and Vogel, 1989).
Diana Anderson 365
Post -implantation
Most studies of this kind have used the rat. Day 1 is defined as the time
when a post-vaginal smear or a copulatory plug is detected. According to
Brown and Fabro (1982), implantation is completed by 7.5-8 days of age.
The embryo is then at the egg cylinder stage. Subsequent development
proceeds through the following stages: age 8.5 days, primitive streak; 9
days, pre-somite neurula; 9.5 days, first somites and brainfolds; 10.5
days, 10-14 somites, yolk sac circulation; 11.5 days, tail bud embryo,
26-30 somites, forelimb bud; 12.5 days, complete embryo, 40-42
somites; 13.5 days, early foetus 48-50 somites. Techniques are available
which support embryonic growth for 1-4 days. The dam is killed, the
uterus removed and implantation sites dissected from the uterine wall
with the aid of a dissecting microscope and microforceps. The maternal
decidual tissue is removed, leaving the conceptus enclosed with the
trophoblastic membrane. This is then removed and the conceptus
transferred to culture medium using static or circulating methods. The
latter enables the medium to be circulated around the immobilised
conceptus and allows greater dilution of homologous rat or human
serum. The growth of embryos explanted at 9.5 days and cultured over 48
hours is indistinguishable from that in utero over the equivalent period.
Growth of older embryos is retarded due to a lack of placental flow of
nutrients (the yolk sac placentation is sufficient for earlier embryos). The
survival of embryos explanted earlier than 9.5 days is reduced compared
with those explanted at 9.5 days, but they require lower oxygen
concentrations (5% as opposed to 95% after 12.5 days). Tesh (1988)
showed that the rat embryo has only limited metabolic competency at this
stage of development.
A metabolising system may be incorporated into the system by using
an hepatic microsomal fraction [the S-9 mix system of Ames et al. (1975)
and Maron and Ames (1983)l. Thus Greenaway et al. (1982) found that
10 day embryos in uitro were unaffected by cyclophosphamide concentra-
tions as high as 250 pg ml-' alone, but 6.25 pg ml-' of the drug produced
abnormal embryos if S-9 mix was present. Another method uses serum
from exposed animals or humans which is assumed to contain metabo-
lites. For example, Chatot et al. (1980) showed that the serum of patients
undergoing therapy with chemotherapeutic agents exerted significant
embryotoxic effects on cultured rat conceptuses. Clearly such experi-
ments require careful control and are subject to great problems of species
specificity, dose determination, and uncertainty concerning the presence
of other active factors that may be present in the serum. Anderson et al.
(1986) and Jenkinson et al. (1986) have shown that not only chemicals,
direct-acting or requiring metabolic activation, but also oxygen radical
species can effect the development of rat embryos in culture.
Embryonic growth is relatively easy to measure and various parameters
are suitable such as crown-rump length, head length, and embryonic
366 In Vitro Methods f o r Teratology Testing
protein or DNA content (Brown et al., 1979). Yolk sac effects must be
taken into account and data are best presented on those conceptuses
which have a functional yolk sac circulation at the end of the culture
period. When presenting incidences of malformation , some indication
must be given of the proportion coming from abnormal yolk sacs. The
exact relationship of yolk sac damage in vitro to actions of the test
compound in vivo has yet to be clarified.
Cicurel and Schmid (1988a and b) concluded that this system had high
predictive value for teratogens using 25 chemicals from 46 reference
compounds (Smith et al., 1983; Jensen et ul., 1989), however, criticised the
use of these reference compounds. Van Maele-Fabry and Picard (1 987)
investigated the post-implantation embryo culture system in the mouse.
C Organ Culture
Organs from Embryos and Foetuses
Various organs have been maintained in culture (see Brown and Fabro.
1982; Lyng, 1989; 1990; Whitby, 1987) such as palatal shelves (mouse),
palate (whole-embryo mouse), bone (mouse, rat), miillerian ducts (rat),
testis, ovaries (rabbit, rat), and teeth (mice). The vast majority of these
systems would not be appropriate as testing techniques since organs are
explanted at an advanced stage of differentiation. They may be useful in
studies of biochemical and histological development.
Limb buds. Limb buds are the most amenable organ to in uitro culture
(Aydelotte and Kochlar, 1972; Kochlar et al., 1982). Mouse is the main
source of material but rat and rabbit limb buds have also been used. The
best development is achieved with fore-limb buds explanted from
embryos at the 40-43 somite stage. The limb buds, dissected from
explanted embryos, are placed on membrane filters supported by metal
grids in petri dishes. Sufficient culture medium is added to wet the
underside of the filter and the culture is incubated in a humidified 5%
C02-air atmosphere at 30°C. Maximum limb bud growth is obtained on
day six of culture when cartilaginous rudiments of phalanges, ulna,
radius, humerus, and scapula are formed. Neubert and Barrach (1977)
maintained limb buds explanted in a roller-bottle system with an
improvement in phalangeal development. Kocklar et al., (1982) in-
corporated a metabolic system in the form of sera from drug-treated mice
and exposed humans. In a novel approach, Agnish and Kochlar (1976)
exposed the whole embryos in culture to bromodeoxyuridine and then
organ cultured the fore-limbs for an additional nine days when structural
development could be accurately assessed. Combining whole embryo and
organ culture satisfies the conditions for exposure during the organo-
genesis period and allows assessment of differentiated structure, but
cannot be considered rapid and inexpensive relative to whole animal
testing.
Diana Anderson 367
The suitability of limb bud culture has been doubted (Wilson, 1978;
Friedman, 1987) because the development of the organ does not closely
parallel in vivo development, in that 6 days of culture is equivalent to
approximately 3 days in utero, and there is organ distortion. The in uitro
differentiation is reproducible and consistent in itself, however, and a
scoring system has been devised to quantitate the extent of limb bud
differentiation. The effect of exposure to test chemicals can be quanti-
tatively assessed (ECETOC, 1989; Hales, 1992).
D Cell Culture
Various assays have been established using isolated cells in culture.
Rat Embryo Limb Cells and Mid-brain Cells in Micromass

Culture
Cultures are started by plating cells at high density in a single small drop.
They adhere together to the dish within about two hours. The cells are
prepared from a thirteen day embryo after trypsinisation before pre-
cartilage condensation (Flint, 1979; 1982; 1983). After cell adhesion,
medium containing the test substance floods the culture for 5 or 6 days.
In the case of limb bud cells, untreated cultures develop foci of
chondrogenic cells which secrete proteoglycan and this can be stained
with alcian blue. Mid-brain (mesencephalon) cells develop into neurons.
In a blind trial, where test chemicals with or without S-9 mix were added
directly to the culture medium, Flint and Orton (1984) measured foci or
neurons and found a concentration-dependent inhibition in the number
of differentiating cell foci following exposure to 25 out of 27 teratogens
and 2 out of 19 non-teratogens. This complementary assay is the best
characterised of the primary cell culture methods. By treating the dams
(intraperitoneally on day 12 of pregnancy) before moving embryonic
tissue, Flint et al. (1984) have shown that of 31 chemicals tested, 17 out of
18 teratogens and one out of 13 non-teratogens were positive when a
number of end-points were examined. The micromass test has been used in
the early stages of pharmacological evaluation to screen out potentially
teratogenic compounds (Flint, 1986). Criteria for a positive in vitro result
have been suggested by Flint ( I 987). There have been several independent
trials of the test (Bacon et a/., 1990; Parsons et al., 1990; Uphill ct a/., 1990;
Newall and Beedles, 1993).
Neural Crest and Neuroblastoma Cells

Primary neural crest cells are explanted from chick embryos after 1: days
incubation, into culture segments of neural tube (Wilk et al., 1979). The
tube is removed after the crest cells have migrated from the tube to form
a monolayer. Cells are replated as small drops as for micromass cultures.
Depending on the type of serum present in the medium, such cells can
develop into melanocytes or neurons and can be assayed for melanin or
acetyl choline transferase, respectively. Only 14 agents have been tested;
10 out of 12 teratogens and none of the 4 non-teratogens were positive
(Greenberg, 1980; Greenberg et al., 1982).
Mummery et al. (1984a and b) using a clone (NIE 115) of the murine
neuroblastoma cell line examined induction and inhibition of cell
differentiation. Of 39 teratogens and 18 non-teratogens, the system
detected as positive 35 teratogens and 4 non-teratogens.
Chinese Hamster Ovary V7 Cells and Human Embryonic

Palatal Mesenchyme (HEPM) Cells
Welsch and Stedman (1984a and b) have used the disruption of cell to
cell communication of these cell lines as a screen for teratogens. For V79
cells a toxic metabolite of 6-thioguanine is transferred from wild type to
mutant cells by metabolic co-operation; for HEPM cells tritiated uridine
is transferred through gap junctions. Treatment with a test compound
reduces the transfer of the toxic metabolite causing an increase in mutant
survival in V79 cells or an increase in incorporation of label as measured
by autoradiography in HEPM cells. Eleven chemicals have been tested for
increased mutant survival and 3 for gap junction transfer. Swierenga and
Yamasaki (1992) have reviewed the disruption of gap-junction intercellu-
lar communication for screening chemicals.
Ascitic Mouse Ovarian Tumour Cells ( M O T )

Braun and colleagues (1979 and 1982) have measured the inhibition of
tumour cell attachment to Concanavalin-A-coated surfaces as an assay
for teratogenic agents, since it is believed that cell adhesions may be
critical to embryonic morphogenesis. MOT cells are grown as an ascites
tumour in the peritoneum of mice and labelled with tritiated thymidine
the day before harvest. After harvest they are treated with the test
compound and then allowed to attach to the Con-A coated discs. In a
study of 102 chemicals, 60 of 74 teratogens and 7 of the 28 non-teratogens
were inhibitory. Cell growth and division are not involved in the cell
attachment assay, so teratogens which affect DNA replication are not
affected.
Thirteen of the teratogenic agents which were false negatives in this
tumour cell attachment assay were tested with the HEPM cell line
(above). Cell proliferation was used as an end-point for teratogenicity
and this was decreased for 12 out of 13 of these agents (Pratt et al.,
1982). A system combining these two assays, known as BRAT, is
being evaluated by the US National Toxicology Programme.
Diana Anderson 369
Other Cell Lines

The measurement of cell death or growth inhibition in cell cultures has
been suggested by Freese (1982) as a means of identifying teratogens,
since differentiating cells may be more vulnerable to treatment. The dose
of compound required to inhibit 50% of cell growth was examined in
various cells (e.g. Hela cell line and rat glial cell line C) but there was no
obvious difference between teratogens and non-teratogens.
BSC 40 cells, a primate line of monkey origin, have been infected with
pox virus and then cultured with the test chemical, after which cells are
frozen. They are subsequently assayed for the number of virus colonies
using a plaque-forming assay, on the assumption that viral replication is
only possible if the host cell is actively proliferating. In a screen of 51
chemicals (Keller and Smith, 1982), measuring the inhibition of plaque
number by 50%, 36 of 42 teratogens and 1 of 9 non-teratogens were
positive.
It is not yet clear whether the cell culture systems described in 1-5 are
able to determine teratogens as opposed to generally toxic chemicals.
4 Usefulness of the In Vitro Methodology

The tests described offer advantages of speed and ease of operation, and
are relatively inexpensive to perform (ECETOC, 1989; Daston and
D’Amato, 1989). In vitro tests permit precise regulation of exposure, and
maternal factors such as the placental ‘barrier’ are excluded. A well-
controlled in vitro test could give a precise estimate of the inherent
teratogenic potential of a chemical under defined conditions. Fabro et al.
(1982), Brown and Freeman (1984), and Brown (1987) have discussed the
use of these systems in terms of developmental toxicity risk assessment,
where concepts of teratogenic potential, potency, and hazard are
described: potential is the ability of an agent under any circumstances to
induce terata, potency is the dose required to induce abnormalities, and
hazard is any estimate of the relationship between teratogenic doses and
general or adult toxic doses. A chemical with high hazard is one which is
teratogenic at doses which are well below maternally toxic doses. Both
potency and hazard are important components of the estimation of risk to
humans. According to Fabro et ul. (1982), hazard is the more important
for an initial screen which should be able to detect those chemicals which
have specific action on development. Comparison of hazard in nitro and in
vivo could determine whether the in vitro assay predicts strong and weak
teratogens as determined by animal studies. (A strong teratogen is one
which will induce a high incidence of malformation at doses well below
maternally toxic doses and a non-teratogen is one which will not induce
fatal abnormalities even up to maternally lethal doses.)
Obviously, the absence of the maternal-conceptus-placental-
relationship reduces the practical value of in vitro screens in predicting
possible mammalian teratogenicity and these tests may never replace the
pregnant mammal as the primary system for teratogenicity testing.
Assessment of an agent’s developmental toxicity is often confounded by
the close relationsip between the maternal and developing systems
(Kimmel et al., 1987). The relative performances of the in vitro tests will
only become apparent when they have been tested with the same group
of chemicals such as those suggested by Smith et al. (1983). To date,
groups of chemicals used for validation purposes vary widely.
The tests do offer promise, however, as pre-screening procedures and
provided such procedures are assessed as to their ability to distinguish
mammalian teratogens from non-teratogens before the outcome of
mammalian studies is known, they may find a role in reducing the burden
of animal testing required (Faustman-Watts, 1988).
Once again, it is unlikely that the various models will be fully accepted
until an understanding of the mechanisms of control of embryonic and
foetal development is achieved and the methods directed towards the
examination of the effect of compounds on these mechanisms.
References
Anderson, D., Jenkinson, P. C., and Gangolli, S. D. (1986). The effect of
chemicals and radical species on rat embryos in culture. Food Chem.
Toxicol., 24, (6.7) 637-638.
Agnish, N. D. and Kochlar, D. M. (1976). Direct exposure of post-implantation
mouse embryos to 5-bromodeoxyuridine in vitro and its effect on subsequent
chondrogenesis in the limbs. J. Embryol. Exp. Morphol., 36, 623-638.
Ames, B. N., McCann, J., and Yamasaki, E. (1975). Methods for detecting
carcinogens and mutagens with the Salmonella/mammalian microsome
Aydelotte, M. B. and Kochlar, D. M. (1972). Development of mouse limb buds
in organ culture: chondrogenesis in the presence of a proline analogue,
~-azetidine-2-carboxylicacid. Dev. Biol., 28, 191-201.
Bacon, W. J., Duffy, P. A., and Jones K. (1990). Studies on variability of the
micromass teratogen test. Toxicol. In Vitro, 4, 577-58 1.
Best, J. B. and Morita, M. (1982). Planarians as a model system for in vitro
teratogenesis studies. Teratogen. Carcinogen. Mutagen., 2, 277-291.
Birge, W. J., Black, J. A., Hudson, J. E., and Bruser, D. M. (1979). In:
‘Aquatic Toxicology ASTM STP 667’, Eds L. L. Marking and R. A.
Kimerle. American Society for Testing Materials, pp. 131-147.
Bournias-Vardiabasis, N. and Teplitz, R. L. (1982). Use of Drosophila embryo
cell cultures as an in vitro teratogen assay. Teratogen. Carcinogen. Mutagen.,
2, NO. 3/4, pp. 333-343.
Bournias-Vardiabasis, N., Teplitz, R. L., Chernoff, G . F., and Seccof, R. L.
(1983). Detection of teratogens in the Drosophila embryonic cell culture test
assay of 100 chemicals. Teratology, 28, 109-122.
Braun, A. G . , Emerson, D. J., and Nichinson, B. B. (1979). Teratogen drugs
inhibit tumour cell attachment to lectin-coated surfaces. Nature (London),
282, 507-509.
Diana Anderson 37 I
Braun, A. G., Nichinson, B. B., and Horowicz, P. B. (1982). Inhibition of

tumour cell attachment to Conconavalin-A coated surfaces as an assay for
teratogenic agents, approaches to validation. Teratogen. Carcinogen.
Mutagen., 2, No. 3/4, pp. 343-355.
Brown, N. A. (1987). Arch. Toxicol., Suppl. 11, 105-114.
Brown, N. A. and Fabro, S. E . (1982). The in vitro approach to teratogenicity
testing. In : ‘Developmental Toxicology,’ Ed. Keith Snell. Croom-Helm,
London, pp. 33-54.
Brown, N. A., and Freeman, S. J. (1984). Alternative tests for teratogenicity.
‘Alternatives to Laboratory Animals’, 12, 7-23.
Brown, N. A., Goulding, E . H., and Fabro, S. E . (1979). Ethanol ernbryotoxi-
city: direct effects on mammalian embryos in vitro. Science, 206, 573-575.
Brown, J., Jorgenson, T. A., and Anderson, D. G. (1983). Validation of an in
vivo pre-screen for the determination of embryo/foetal toxicity in mammals.
Teratology, 27, 34A.
Chatot, D. L., Klein, N. W., Piatek, J., and Pierre, L. J. (1980). Successful
culture of rat embryos on human serum: use in the detection of teratogens.
Science, 207, 1471-1473.
Cicurel, L. and Schmid, B. P. (1988a). Post-implantation embryo culture:
validation with selected compounds for teratogenicity testing. Xenohioticu,
18, 617-624.
Cicurel, L. and Schmid, B. P. (1988b). Post-implantation embryo culture for the
assessment of the teratogenic potential and potency of compounds. E - q w i -
rntia, 44, 833-840.
Collins, T. F. X. (1987). Teratological research using in vitro systems.
V. Nonmammalian model systems. Environ. Health Perspect., 72, 237-249.
Danicl, J . C. (Ed.) (1971). ‘Methods in mammalian embryology’. Frecman and
Company, San Francisco.
Daston, G. P. and D’Amato, R. A. (1989). In vitro techniques in teratology.
Toxicol. Ind. Health, 5, 555-585.
Dawson, D. A., Fort, D. J., Newell, D. L., and Bantle, J. A. (1989). Develop-
mental toxicity testing with FETAX: evaluation of five compounds. Drug
C‘hem. To.vicol., 12, 67-75.
Deepe, U., Schierenberg, E., Cole, T., Krieg, C., Schmitt, D., Yoder, B., and
von Ehrenstein, G. (1978). Cell lineages of the embryo of the nematode
Caenorhabolitis elegans. Proc. Natl. Acad. Sci. ( U S A ) , 75, 376-380.
de Serres, F. J. and Ashby, J. (Eds) (1981). Evaluation of short term tests for
carcinogens. ‘Report of the International Collaborative Programme’.
Elsevier/North-Holland Biomedical Press, Amsterdam, Holland.
Dumont, J. N., Schultz, T. W . , and Newman, S. M. (1982). A frog embryo
teratogenesis assay. Xenopus (FETAX)-a model for teratogen screening.
Teratology, 25, 37A.
Dumont, J. N. and Epler, R. G. (1984). Validation studies of the FETAX
teratogenesis assay (frog embryos). Teratology, 29, 38A.
ECETOC (1 989). ‘Alternative Approaches for the Assessment of Reproductive
Toxicity (with Emphasis on Embryotoxicity/Teratogenicity)’, Monograph
No. 12, 50 pp.
Fabro, S . , Shull, G., and Brown, N. A . (1982). The relative teratogenic index
and teratogenic potency: proposed components of developmental toxicity
risk assessment. Teratogen. Carcinogen. Mutagen., 2, No. 1, pp. 61-77.
Fantel, A. G. (1982). Culture of whole rodent embryos in teratogen screening.

Teratogen. Carcinogen. Mutagen., 2, No. 314, pp. 231-243.
Fantel, A. G., Greenaway, J. C., Juchau, M. R., and Shephard, T. H. (1979).
Teratogenesis bioactivation of cyclophosamide in vitro. Life Sci., 25, 67-73.
Faustman-Watts, E. M. (1988). Mutat. Res. , 205, 355-384.
Flint, 0. P. (1980). The effects of sodium salicylate, cytosine, arabinoside, and
eosine sulphate in rat limb buds in culture. In: ‘Teratology of the Limbs’,
Eds H. J. Merker, H. Nau, and D. Neubert de Gruyter. Berlin, pp. 325-338.
Flint, 0. P. (1982). An in uitro assay for teratogens using embryonic cells
exposed to compound in utero or directly in culture. Teratology, 15, 40A.
Flint, 0. P. (1983). A micromass method for rat embryonic neural cell. J . Cell
Sci., 61, 247-262.
Flint, 0. P. and Orton, T. C. (1984). An in vitro assay for teratogens using
cultures of rat embryo cells, brain, and limb bud cells. Toxicol. Appl.
Pharmacol., 76, 383-395.
Flint, 0. P., Orton, T. C., and Ferguson, R. A. (1984). Differentiation of rat
embryo cells in culture. Response following acute maternal exposure to
teratogens and non-teratogens. J . Appl. Toxicol., 4, No. 2, pp. 109-116.
Flint, 0. P. (1986). An in vitro test for teratogens: its practical application. Food
Chem. Toxicol., 24, 627-63 1.
Flint, 0.P. (1987). An in vitro test for teratogens using cultures of rat embryo cells.
In: ‘In vitro Methods in Toxicology’. Eds C. K. Atterwill and C. E. Steele.
Cambridge University Press, Cambridge, pp. 339-363.
Flynn, T. J. (1987). Teratological research using in vitro systems. I. Mammalian
whole embryo culture. Environ. Health Perspect., 72, 203-210.
Fort, D. J., Dawson, D. A., and Bantle, J. A. (1988). Development of a metabolic
activation system for the Frog Embryo Teratogenesis Assay: Xenopus
(FETAX). Teratogen. Curcinogen. Mutugen., 8, 251-263.
Fort, D. J., James, B. L., and Bantle, J. A. (1 989). Evaluation of the developmental
toxicity of five compounds with the Frog Embryo Teratogcnesis Assay:
Xenopus (FETAX) and a metabolic activation system. J . Appl. Toxicol., 9,
377-388.
Freese, E. (1982). Use of cultured cells in the identification of potential teratogens.
Teratogen. Carcinogen. Mutagen., 2, 355 -360.
Friedman, L. (1987). Teratological research using in vitro systems. 11. Rodent limb
bud culture system. Environ. Health Perspect., 72, 2 1 1-2 19.
Greenberg, J. H. (1980). Normal and Abnormal Differentiation of Neural Crest
Cells in Culture. In : ‘Current Research Trends in Prenatal Craniofacial
Development’. Eds R. M. Pratt and A. Christiansen. ElsevierINew York,
pp. 65-80.
Greenberg, J. H. (1982). Detection of teratogens by differentiating embryonic
neural crest cells in culture-evaluation as a screening system. Teratogen.
Carcinogen. Mutagen., 2, No. 314, pp. 319-325.
Greenaway, J. C., Fantel, A. G., Shephard, T. J., and Juchau, M. R. (1982).
The in uitro teratogenicity of cyclophosphamide on rat embryos. Teratology,
25, 335-343.
Gross, P., Humphreys, T., and Anderson, E. (1972). ‘The Sea Urchin:
developmental urchin’. MSS Information Corporation, New York.
Hales, B. F. (1992). Teratogenicity. In: ‘In Vitro Toxicity Testing. Applications to
Safety Evaluation’. Ed. J. M. Frazier. Marcel Dekker, Inc., New York, Ch. 9,
pp. 205-220.
Diana Anderson 37 3
Hardin , B, D. (1987). Special Issue Teratogen. Carcinogen. Mutagen., 7 , No. 1.

Hsu, Y. C. (1973). Differentiation in witro of mouse embryos in the stage of early
somite. Dew. Biol., 33, 403-411.
Hsu, Y. C. (1980). Embryo growth and differentiation factors in embryonic sera
of mammals. Dew. Biol., 76, 465-474.
Jelinek, R. (1982). Use of Chick embryo in screening for embryotoxicity.
Teratogen. Carcinogen. Mutagen., 2, No. 314, pp. 225-263.
Jelinek, R. and Rychter, Z. (1980). Morphogenetic systems and screening for
embrytoxicity. Arch. Toxicol. Suppl., 4, 267-273.
Jenkinson, P. C., Anderson, D., and Gangolli, S. D. (1986). Malformations
induced in cultured rat embryos by enzymically generated active oxygen
species. Teratogen. Carcinogen. Mutagen., 6, No. 6, 547-554.
Jensen, M., Newman, L. M., and Johnson, E. M. (1989). Re: Problems in
validation of in vitro developmental toxicity assays. Fundarn. Appl. Toxicol.,
13, 863-865.
Johnson, E . M. (1980). Detection of teratogenic potential of drugs and chemicals
by an artificial embryo. Anat. Rec., 196, 89A.
Johnson, E. M., Gorman, R. M., Bradley, E. G., and George, M. E. (1982).
The hydra attenuata system for detection of teratogenic hazards. Teratogen.
Carcinogen. Mutagen., 2, No. 314, 263-277.
Keller, S. J. and Smith, M. K. (1982). Animal virus screen for potential
teratogens. I. Pox virus morphogenesis. Teratogen. Carcinogen. Mutagen., 2,
361-374.
Kimmel, E. L., Kimmel, C. A., and Francis, E. Z . (1987). Special Issue
Teratogen. Carcinogen. Mutagen., 7 , No. 3.
Kochlar, D. M. (1982). Embryonic limb bud culture in assessment of teratogeni-
city of environmental agents. Teratogen. Carcinogen. Mutagen., 2 , No. 314,
303-313.
KuEera, P. and Burnand, M. B. (1987). Teratogen. Carcinogen. Mutagen., 7,
NO. 5, 427-449.
Lallier, R. (1965) J . Embryol. Exp. Morphol., 31, 721-734.
Lyng, R. D. (1989). Test of six chemicals for embryotoxicity using fetal mouse
salivary glands in culture. Teratology, 39, 59 1-599.
Lyng, R. D. (1990). Use of fetal mouse salivary glands in culture to detect
embryotoxicity: evaluation of eight additional chemicals. Toxicol. Lett., 54,
245-25 1.
Maron, D. and Ames, B. N. (1983). Revised methods for the Salmonella
mutagenicity tests. Mutat. Res., 113, 173-215.
Mummery, C., van den Brink, C., van der Saag, P., and de Laat, S. (1984). A
short-term screening test for teratogens using neuroblastoma cells in witro.
Teratology , 29, 271-279, and abstracts from European Teratology Society ,
Holland 5-7 Sept, 1984.
Neubert, D. and Barrach, H. J. (1977). Techniques applicable to study
morphogenetic differentiation of limb buds in organ culture. In: ‘Methods in
Prenatal Toxicology’, Eds D. Neubert, H. J. Merker, and T. E. Kwasigroch.
George Thieme Publishers, Stuttgart, pp. 241-251.
Newall, D. R. and Beedles, K. (1993). The stem-cell test-a novel in vitro assay for
teratogenic potential. In: Book of Abstracts, Practical In Vitro Toxicology
111, 25-29 July. Abstract 3.5, p. 20.
Parsons, J. F., Rockley, J., and Richold, M. (1990). In vitro micromass teratogen
374 In Vitro Methods ,fbr TeratokogJ* Testing
test: interpretation of results from a blind trial of 25 compounds using thrcc

separate criteria. Toxicol. In Vitro, 4, 609-61 1.
Pratt, R. M., Grove, R. I., and Willis, W. D. (1982). Prescreening for
environmental teratogens using cultured mesenchymal cells from the human
embryonic palate. Teratogen. Carcinogen. Mutagen., 2, No. 314, 313-318.
Schuler, R. L., Hardin, B. D., and Niemeier, R. W. (1982). Drosophila as a tool
for the rapid assessment of chemicals for teratogenicity . Teratogen. Car-
cinogen. Mutagen., 2, No. 314, 293-303.
Smith, M. K., Kimmel, G. L., Kochlar, D. M., Shephard, T. H., Spielberg, S.
P., and Wilson, J. G. (1983). A selection of candidate compounds for in vitro
teratogenesis test validation. Teratogen. Carcinogen. Mutagen., 3, 461-480.
Spielmann, H. and Vogel, R. (1989). Unique role of studies on preimplantation
embryos to understand mechanisms of embryotoxicity in early pregnancy.
CRC Crit. Rev. Toxicol., 20, 51-64.
Staples, R. E. (1971). In ‘Methods in Mammalian Embryology’, Ed. J. C. Daniel.
Freeman and Company, San Francisco, pp. 190-204.
Staples, R. E. (1975). In: ‘New Approaches to the Evaluation of Abnormal
Embryonic Development’, Eds D. Neubert and H. J. Merker. Thieme
Edition Publishing Sciences Group, Berlin, pp. 71-81.
Sulston, J. E. and Horvitz, H. R. (1977). Post-embryonic cell lineages of the
nematode, Caenorhabolitis elegans. Dev. Biol., 56, 110-156,
Swierenga, S. H. H. and Yamasaki, H. (1992). Performance of tests for cell
transformation and gap-junction intercellular communication for detecting
non-genotoxic carcinogenic activity. In: ‘Mechanisms of Carcinogenesis in
Risk Identification’. Eds H. Vainio, P. N. Magee, D. B. McGregor, and
A. J . McMichael. Lyon International Agency for Research on Cancer,
pp. 165-193.
Tesh, J. M. (1988). The application of whole-embryo culture to new product
development. Toxicol. In Vitro, 2, 189-194.
Uphill, P. F., Wilkins, S. R., and Allen, J. A. (1990). In vitro micromass teratogen
test: results from a blind trial of 25 compounds. Toxicol. In Vitro,4,623-626.
Welsch, F. and Stedman, D. B. (1984a). Inhibition of metabolic co-operation
between Chinese hamster V79 cells by structurally diverse teratogens.
Teratogen. Carcinogen. Mutagen., 4, 285-302.
Welsch, F. and Stedman, D. B. (1984b). Inhibition of intercellular communica-
tion between normal human embryonic palatal mesenchyme cells by terato-
genic glycol ethers. Environ. Health Perspect. 57, 125-133.
Whitby, K. E. (1987). Teratological research using in vitro systems. 111. Embry-
onic organs in culture. Environ. Health Perspect., 72, 221-223.
Wilk, A. L., Greenberg, J. H., Horigan, E. A., Pratt, R. M., and Martin, G. R.
(1979), Detection of teratogenic compounds using differentiating embryonic
cells in culture. In Vitro, 16, 269-276.
Wilson, J. G. (1977). Current status of teratology. In: ‘Handbook of Teratology’,
Eds J. G. Wilson and F. C. Fraser. Vol. 1, pp. 47-74. Plenum Press, New
York.
Wilson, J. G. (1978). Survey of in vitro systems, their potential use in
teratogenicity screening. In: ‘Handbook of Teratology’, Eds J. G. Wilson
and F. R. Fraser. Vol. 4, pp. 135-153. Plenum Press, New York.
World Health Organisation (1967). Principles for the testing of drugs for
teratogenicity. Technical Report Series 167, No. 364.
Diunu Anderson 375
Wiirgler, F. E., Sobels, F. H., and Vogel, E. (1977). Drosophilu as assay system
for detecting genetic changes. In: ‘Handbook of Mutagenicity Test Proce-
dures’, Eds B. J. Kilbey, M. Legator, W. Nichols, and C. Ramel.
Elsevier/North-Holland Biomedical Press, pp. 335-373.
CHAPTER 17
Assessing Chemical Injury to

the Reproductive System
S. D. GANGOLLI AND J. C. PHILLIPS
1 Introduction
The mammalian reproductive process is long and complex, and thus
uniquely susceptible to disruption by exogenous agents. Foreign com-
pounds can interfere with either the neurophysiological processes in-
volved in the process or in gametogenesis. In the male, for example,
spermatogenesis can be altered directly by an effect on the germ cells at
the various stages in their maturation process or indirectly by disruption
of the endocrine functions essential for regulating their production and
delivery. In the female, oogenesis can be affected directly (oocyte
destruction) or indirectly by interference with the ovarian and uterine
cycles. As the integrity of the reproductive system is of crucial impor-
tance to the survival of the species, the identification of foreign chemicals
which can affect human reproduction is of major importance. Despite a
lack of understanding of the biochemical mechanisms underlying the
toxicity of many agents, a wide variety of compounds including drugs,
pesticides, organic solvents, and heavy metals (Table 1) have been found
to be reproductive toxins in both the male and female (Bernstein, 1984;
Lucier, Lee, and Dixon 1977; Schrag and Dixon, 1985).
In this chapter, the investigation of chemical injury to the mammalian
reproductive system will be discussed from a mechanistic standpoint, first
by describing the essential components of the male and female reproduc-
tive system; secondly by outlining the toxic consequences resulting from
chemical injury to specific male (testicular and extra testicular) and
female (ovarian and uterine) sites by some model toxins, and thirdly by
discussing practical techniques for identifying and assessing toxicity.
To this end, the chapter is divided into the following main sections:
The anatomy and physiology of the male reproductive system.
Anatomy and physiology of the female reproductive system.
Mechanisms regulating testicular function.
Mechanisms regulating ovarian function.
376
Table 1 Some environmental agents associated with reproductive dys-

function in humans.
Male: confirmed effectt
Carbon disulphide; dibromochloropropane (DBCP) , lead; oral contraceptives
: probable
Anaesthetic gases; benzene; carbaryl; chlordecone; chloroprene;
dinitrotoluene; ethylene dibromide; hexane; hexachlorobenzene; glycol ethers;
metals (arsenic, boron, cadmium, manganese); pentachlorophenol; radiation;
dioxins; vinyl chloride monomer.
Female:
Anaesthetic gases; aniline; benzene; carbon disulphide; chloroprene; ethanol;
ethylene oxide; glycol ethers; formaldehyde; metals (lead, organic mercury);
phthalate esters; polychlorinated biphenyls; styrene; tobacco smoke; toluene;
viny chloride monomer.
t Adapted from Dixon (1986) and Schrag and Dixon, (1985).
Site specificity of chemical injury by model toxins.

Methods for studying reproductive toxicity in the male.
Methods for studying reproductive toxicity in the female.
Methods for assessing reproductive function,
2 Anatomy and Physiology of the Male

Reproductive System
The principal components of the male reproductive system are the
primary sex organs (the testes) and the accessory sex organs, which
include the epididymis, vas deferens, prostate, Cowper’s gland, and the
penis (Figure 1). As a number of other (extra testicular) sites are
involved in regulating the reproductive system, they will also be
considered in this context. These sites include the pituitary, adrenals,
thymus, and pineal gland.
The testis consists of two main components, the vascular interstitial
tissue containing Leydig cells and the seminiferous tubules containing
Sertoli and germ cells. The Leydig cells, which occur in clusters in the
sinusoidal spaces between seminiferous tubules, are directly exposed to
the lymph and are often closely associated with blood vessels. They
contain abundant smooth endoplasmic reticulum, mitochondria, and a
prominent Golgi complex. The normal function of the Leydig cell is to
produce testosterone and other steroid hormones de now0 from acetate
via cholesterol (Figure 2). In the rat, the major pathway is the
A4-pathway (Bell, Vinson, Hopkin, and Lacy, 1968), whereas the
A’-pathway is important in other species and in man (Yanahara and
Troen, 1972).
The seminiferous tubules, in which spermatozoa are produced, form
long coiled loops opening at both ends into the rete testis. The number of
tubules, arranged in lobules, vary from approximately 30 in the rat to up
378 Assessing Chemical Injury to the Reproductive System
Ureter
vas de
Epididymis
Figure 1 Diagram of the human male reproductive system
to 1000 in man. Within the tubule are the somatic Sertoli cells, and germ
cells at various stages of maturity.
Spermatogenesis is the process of maturation of the germ cells, the
spermatogonia (derived from the prenatal gonocyte) into sperm. During
the first phase, the so-called type A spermatogonia undergo six mitotic
divisions to form type B spermatogonia. Some spermatogonia remain
undifferentiated to give rise to more stem cells. These may either be type
A4 cells or a separate long-cycling population (Ao). After the last mitotic
stage, the type B spermatogonia divide mitotically to form preleptotene
primary spermatocytes, which are able to pass through the blood-testis
barrier into the more protected environment of the adluminal compart-
ment of the seminiferous tubule. These spermatocytes replicate their
DNA, initiating meiosis, to proceed to the leptotene step. This is
followed by a zygotene step, during which the pairing of the homologous
chromosomes occurs. Cells with completely paired chromosomes are
termed pachytene primary spermatocytes. The progression from early
pachytene through mid- and late-pachytene to form secondary sperm-
atocytes is the longest in the meiotic prophase and also the stage at
-
F?
Cholesterol f--
+
Acetate
Progesterone
Pregnenolone VI
I
17a-OH-pregnenolone --+ 17a-OH-progesterone
Dehydroepiandrosterone + Androst-4-enedione +
IX
- Estrone
Androst-5-enediol --+ + --+ Estradiol
5 a-Dihyd rotest oste rone

~ 1 1 1
Androstandiol
Figure 2 Pathways of steroid synthesis in the testes. The enzyme(s) involved in
the various pathways are as follows:
I Multiple enzyme steps-HMG-CoA reductase is rate limiting
I1 20a-hydroxylase, 22-hydroxylase, 20-22 lyase
I11 17a-hydroxylase
IV 17-20 lyase
V 17P-hydroxysteroid dehydrogenase
VI Isomerase, 3P-hydroxysteroid dehydrogenase
VII Sa-reductase
VIII 3 a-reductase
IX l9a-hydroxylase, 19P-hydroxysteroid dehydrogenase, ‘aromatase’
X 17/3-hydroxysteroid dehydrogenase.
The enzymes mediating the conversion of pregnenolone to androst-5-enediol
(A5-pathway) are the same as those mediating the conversion of progesterone to
testosterone (A4-pathway>. Similarly, the conversion of androst-4-enedione to
oestrone is mediated by the same enzymes that convert testosterone to oestradiol.
Each intermediate in the As-pathway can be converted to its corresponding
A4-pathway intermediate by enzymes VI (isomerase/dehydrogenase)
which the cells are particularly susceptible to damage. The secondary

spermatocytes have a short lifespan, and without duplicating their DNA,
they enter the second maturation division forming haploid spermatids. In
a series of well-defined transformations (spermiogenesis) , spermatids
become spermatozoa. During the second phase of spermiogenesis, the
cytoplasm of the Sertoli cell is gradually attenuated as the sperm are
slowly extruded towards the lumen. The residual cytoplasm is sub-
sequently phagocytosed.
The Sertoli cell, which extends from the base of the seminiferous
epithelium into the lumen of the tubule, contains a large irregular-shaped
nucleus, many spherical and rod shaped mitochondria, a prominent Golgi
complex, and large amounts of smooth endoplasmic reticulum. The tight
junction between Sertoli cells delimits two compartments in the semin-
iferous epithelium and is probably the morphological site of the blood-
testis barrier. As well as participating in the maturation and release of
spermatozoa from the germinal epithelium, Sertoli cells are the site of
synthesis of some testis-specific proteins, such as androgen binding
protein (ABP), transferrin, and ceruloplasmin.
Sperm deposited in the rete testis from the seminiferous tubules leave
the testis through the efferent duct, which empties into the initial segment
of the epididymis. During their passage through the epididymal duct, a
long tortuous tube surrounded by connective tissue and smooth muscle,
the sperm attain functional maturity, emerging with a strong unidirec-
tional motility. The maturation process is androgen-dependent (Orgebin-
Crist, Danzo, and Davies, 1975). The mature sperm travel along the vas
deferens to the ejaculatory duct, which is made up of the proximal
extremities of the seminal vesicles. The epithelium of these vesicles is
packed with rough endoplasmic reticulum and dense excretory granules,
and the secretions of this gland, which contain fructose, prostaglandins,
and various reducing substances, make up a substantial proportion of
the seminal plasma. The prostate, which surrounds the uretha below the
seminal vesicles, also contributes to the ejaculate. The prostatic fluid is
rich in proteolytic enzymes such as fibrinolysin, and also contains high
concentrations of zinc, citric acid, and acid phosphatase. The functional
activity of both the prostate and seminal vesicles is dependent on
testosterone. A detailed description of the morphology of the male
reproductive system can be found in Weiss and Greep (1977), and recent
reviews of aspects of the physiology of the male reproductive system have
been published (Lamb and Foster, 1988; Waller and Nikurs, 1986; Waller,
Killinger, and Zaneveld, 1985).
3 Anatomy and Physiology of the Female

Reproductive System
The female reproductive system consists of the ovaries, fallopian tubes
(oviducts), uterus, and vagina (Figure 3). Hormonal control of the
S. D. Gangolli and J . C. Phillips 38 I
Oviduct
a - i - j : ampullary- isthmic junction

u - t - j : utero - tu bal junction
Figure 3 Reproductive organs of the human female
uterine and ovarian cycles in sexually mature females is mediated by the

hypothalamic-pituitary axis, in particular the pars distalis of the pi-
tuitary. The ovaries, which lie on either side of the uterus contain all of
the eggs (oocytes) that the female will ever have. During the foetal
period, the oogonia which are the stem cells for the oocytes and the
female equivalent of spermatogonia, proliferate and shortly after birth
are arrested in the primary oocytes stage, diplotene. The oocytes, and
several granulosa cells (equivalent to testicular Sertoli cells), are sur-
rounded by a basement membrane forming a follicle. The oocyte remains
in this resting phase until just before ovulation. During each menstrual
cycle a number of follicles increase in size, although only one releases its
oocyte (i.e. ovulates). Following ovulation, the ovum is carried into the
oviduct either by the action of cilia on the epithelial cells of the fimbria in
sub-primates, or by muscular contraction in primates. The ovum travels
down the ampulla to the ampullary-isthmic junction, where fertilisation
takes place. During passage down the ampulla, the ovum progresses to
the metaphase stage of the second meiotic division. If the ovum is
fertilised, it implants in the endometrial lining of the uterus. In the rat,
this occurs approximately 5 to 6 days after fertilisation.
Thus, the essential difference between gametogenesis in males and
females is that in male, from puberty to death, there is a continuous
production of mature germ cells from stem cells, whereas in females,
germ cell maturation essentially ceases at birth, and thereafter there is a
discontinuous release of these cells, which ceases at menopause. For a
detailed description of the female reproductive tract see Weiss and Greep
(1977) or Takizawa and Mattison (1983).
4 Mechanisms Regulating Testicular Function

A Intra-testicular Mechanism
The most important intra-testicular sites of regulatory mechanisms are
the Leydig and Sertoli cells. The function of these cells is mainly
regulated by the principal testicular androgen, testosterone, and by the
gonadotropic hormones, luteinising hormone (LH) and follicle stimulat-
ing hormone (FSH), produced by the anterior pituitary. However, recent
studies have shown that refinements to this ‘classical’ picture are
necessary (see below).
The Leydig cell is the major, if not the only source of de nouo
testosterone production, so that inhibition of Leydig cell activity can have
important consequences for the regulation of testicular function. A
reduction in testosterone synthesis, brought about by inhibition of the
enzymes in the biosynthetic pathway (e.g. 17c~-hydroxylase, 3P-
hydroxysteroid dehydrogenase) may result in lowered plasma testos-
terone levels, reduced testicular and accessory sex organ weights, and
decreased spermatogenesis (Dixon 1986; Willis, Anderson, Oswald, and
Zaneveld, 1983).
The Sertoli cell provides a site for the proliferation and differentiation
of germ cells, and it is thought that many of the effects of circulating
hormones, such as the gonadotropin FSH, are mediated via the Sertoli
cells. The major function of the Sertoli cell is the secretion of tubular
fluid, which contains a number of important proteins, including plas-
minogen activator and androgen binding protein (ABP). Sertoli cells also
secrete large amounts of lactate and pyruvate which appear to be
important in maintainting ATP levels in spermatids and primary sper-
matocytes (LeGac, Attramadal, Horn, et aE., 1982). Other functions of
the Sertoli cell include steroidogenesis, the production of intracellular
proteins such as protein-kinase inhibitor (PKI) and the secretion of the
peptide hormones, luteinising-hormone releasing hormone (LHRH)-like
peptide (Hansson, Jegou, Attramadal, et al., 1983).
Although Sertoli cells are not capable of de nouo synthesis of
androgens, their complement of steroidogenic enzymes can convert
progesterone to testosterone, testosterone to androstenedione and 5a-
diol, and testosterone, androstenedione and their 19-hydroxy derivatives
to oestradiol-17P and oestrone (see Figure 2) (Tcholakian and Stein-
burger, 1978; Dorrington, Fritz, and Armstrong, 1976; Welsh and Wiebe,
1978).
The secretion of the testis- and epididymis-specific high affinity ABP
(French and Ritzen, 1973) is regulated by both FSH and testosterone in
the interstitial tissue, thereby facilitating its entry into the seminiferous
tubules. The secretion rate of ABP in uiuo can be used as an index of
Sertoli cell function (Setchell, 1978; Gray, Beamand, and Gangolli,
1982). In in uitro studies with isolated rat seminiferous tubules, the rate
of ABP secretion was found to vary throughout the spermatogenic cycle,

being maximal at stages VIII to XI and minimal at stages IV to V
(Ritzen, Boitani, Parvinen, et al., 1982).
Other proteins known to be produced in the Sertoli cell include protein
kinase inhibitor (Tash, Welsh, and Means, 1980) and inhibin (Van Thiel,
Sherins, Myers, and DeVita, 1972; Steinberger and Steinberger, 1976).
The function of the former is not clear; the latter is secreted and acts as a
negative feedback control on the release of FSH (Figure 4). It is not yet
known whether the effect of inhibin is a direct effect on the release
mechanism or is a consequence of reduced FSH synthesis.
FSH binds to membrane receptors on Sertoli cells, thereby regulating
the activity of adenylyl cyclase. The intracellular steady-state level of
cyclic AMP, which controls the production of lactate and pyruvate and
the activities of protein kinases, is dependent on the relative activities of
LEYDIG SERTOLI
CELL CELL
Figure 4 Schematic diagram f o r hormonal regulation of testicular steroidogenesis

showing classical ‘long’ feedback loops between testis and hypothalamus-pituitary
axis and hypothetical ‘short’ feedback loops between Leydig and Sertoli cells.
C = cholesterol, T = testosterone, Tp = polar testosterone metabolite, and
GnRH = Gonadotropin releasing hormone. Inhibin is also known as Sertoli cell
factor [Dashed line indicates down -regulation]
adenylyl cyclase and phosphodiesterase. Germ cells may also produce

cyclic AMP via a germ-cell specific adenylyl cyclase, located primarily in
the spermatids. Whether this enzyme is regulated by Sertoli cell factor(s)
is not known.
B Extratesticular Mechanisms
The regulation of testicular function by the endocrine system has been
reviewed by numerous authors (for example, see Steinberger and
Steinberger, 1974; Setchell, 1978; DiZerega and Sherins, 1981; Sharpe,
1982a)- The classical ‘long’ feedback loops between the hypothalamic-
pituitary axis and the testis requires the regulation of Leydig cells by LH
and the seminiferous epithelium by FSH (Steinberger, 1976). These
glycoprotein hormones, which are released from the anterior pituitary by
blood-borne releasing factors produced in the hypothalamus, bind to
specific cell surface receptors to exert their effects. The immediate effect
of LH is to stimulate testosterone synthesis (Rommerts, Bakker, and Van
der Molen, 1983) but it is also believed to be involved in the early
development of Leydig cell function (Chase, Dixon, and Payne, 1982)
and in the maintenance of the integrity of the smooth endoplasmic
reticulum (Ewing, Wing, Cochran, et al., 1983). The major effect of FSH
on the Sertoli cell is a general stimulation of protein synthesis, resulting
in part in the stimulation of testosterone metabolism (Steinberger ,
Browning, and Grotjan, 1983).
Numerous other hormones have been shown to influence testicular
function. Thus, for example, prolactin derived from the pituitary can
increase the number of Leydig cell LH receptors leading to a stimulation
of steroidogenesis and can stimulate cyclic AMP production (Baranao,
Tesone, Calvo, et al., 1983). Luteinising hormone-releasing hormone
(LHRH) derived from the hypothalamus not only stimulates gonadotro-
phin production by the pituitary but has a direct effect on Leydig cells,
initially stimulating steroidogenesis but inhibiting on prolonged exposure
(Auclair, Kelly, Coy, et al., 1977; Auclair, Kelly, Labrie, et al., 1977;
Sharpe, 1983). The effects of these and other hormones have been
reviewed by Sharpe (1982a).
Although the regulatory mechanisms acting via the hypothalamic-
pituitary axis are of major importance, recent studies have suggested that
various ‘short’ loop regulatory mechanisms are involved in the fine tuning
of testicular function. In particular it has been shown that there is
intratesticular regulation of Leydig cell function by Sertoli cells, and vice
versa. Thus, Leydig cell steroidogenesis may be regulated by as yet
unidentified polar testosterone metabolites produced by the Sertoli cells
and many of the processes of the Sertoli cell initiated by FSH can be
maintained by testosterone secreted by the Leydig cells. In addition,
gonadotropin releasing hormone (GnRH)-like peptide, secreted by the
Sertoli cells, has been suggested as an inhibitor of Leydig cell function
S . D . Gangolli and J . C . Phillips 385
through an indirect effect on testosterone synthesis (Sharpe, Frazer,

Cooper, and Rommerts, 1981; Steinberger, Browning, and Grotjan,
1983).
5 Mechanisms Regulating Ovarian Function

The details of the mechanisms underlying the process of follicle growth,
selection, and ovulation (expulsion of the ovum as the gamete) are not
well understood. It is clear, however, that the growth of the dominant
follicle is influenced by FSH secreted by the pituitary, and that as it grows
it produces oestrogen which acts initially as a feed-back inhibitor of FSH
production. Oestrogen also stimulates proliferation of the endometrium
and its appearance in the circulation coincides with an increase in plasma
LH concentration. The continued secretion of oestrogen triggers a
discharge of LH and FSH from the pituitary, the LH being involved in
the preovulatory maturation of the oocyte. LH also stimulates the
granulosa cells to secrete progesterone and the ovaries to synthesise
prostaglandins (PGE2 and PGF2), resulting in the expulsion of the ovum.
The follicle remaining after ovulation undergoes morphological changes
1r-l HYPOTH;LI\MUS -1
FSH/LH-RH
LH-RH
I
LH-RH
Figure 5 Schematic diagram for hormonal regulation of the female reproductive

system. E/P = Oestradiol-17P/Progesterone [Dashed line indicates a pre-
dominantly inhibitory effect]
(luteinisation) forming the corpus luteum, which secretes both oestrogen

and progesterone. These steroids are involved in preparing the en-
dometrium for implantation (Figure 5). Luteinisation appears to be
controlled by prolactin in both primate and sub-primate species and
luteolysis, the breakdown of the corpus luteum, appears to be controlled
by prostaglandins in sub-primates, but not in primates.
LHRH also has a role in the regulation of ovarian function, being a
potent inhibitor of steroidogenesis in both the developing and post-
ovulatory follicle. Because of its inhibitory effects on granulosa and luteal
cell function, it has been postulated to be important in follicular atresia
and luteal regression (Sharpe, 1982b).
6 Site Specificity of Chemical Injury by Model

Toxins
A Testicular sites
Sertoli Cells
Many testicular toxins exert their effects on the Sertoli cell. Administra-
tion of the neurotoxins, n-hexane and methyl n-butyl ketone or their
metabolite, 2,5-hexandione (2,5-HD), results in extensive atrophy of the
germinal epithelium of the testes of experimental animals (Krasavage,
O’Donoghue, DiVincenzo, and Terhaar, 1980). As the CNS controls the
release of gonadotropins it was postulated that the effects of these
compounds were due to changes in testicular homeostasis. However, it
was found that 2,5-hexanedione treatment had no effect on serum
testosterone concentration or on the circulating levels of FSH and LH
(Chapin, Norton, Popp, and Bus, 1982), whereas a significant reduction in
the activities of the Sertoli cell-localised enzymes, P-glucuronidase and
;I-glutamyl transpeptidase, has been seen. Vacuolation of Sertoli cells after
treatment with a subneurotoxic dose of 2,5-HD (Boekelheide, 1988)
confirms a direct toxic effect on the Sertoli cell. The mechanism by which
2,Shexanedione exerts its effects is not clear, although detailed
morphological studies suggest this compound interferes with the micro-
tubule network (Johnson, Hall, and Boekelheide, 1991).
Di-(2-ethylhexyl)phthalate (DEHP), a widely used industrial solvent,
was first reported to cause testicular atrophy in the rat by Schaffer,
Carpenter, and Smyth, (1945). Subsequent studies showed that the
testicular injury produced by this and some other phthalate diesters was
confined to a marked reduction in the diameter of seminiferous tubules
and loss of germ cells (Gangolli, 1982). Sequential morphological studies
of the effects of dipentyl phthalate on pre-pubertal rats showed a rapid
development of vacuolation of the Sertoli cells, and disruption of the basal
tight-junctions between Sertoli cells, which form the basis of the blood-
testis barrier (Creasy, Beech, Gray, and Butler, 1987). Early degenerative
changes were also apparent in germ cells (Creasy, Foster, and Foster,
1983). Other effects seen included a marked reduction in Sertoli cell
mitochondria1 succinate dehydrogenase activity and an inhibition of

seminiferous tubule fluid and ABP secretion (Gray and Gangolli, 1986).
Although the mechanisms underlying phthalate-induced testicular injury
are not fully understood, it is clear that the proximate toxin is the
monoester or a metabolite thereof (Sjoberg, Bondesson, Gray, and Ploen,
1986), and that the Sertoli cell is the principal target site.
Other chemicals targeting the Sertoli cell include nitrobenzenes and
nitrotoluenes (Reader and Foster, 1990; Allenby, Sharpe, and Foster,
1990). The effects produced, such as cell vacuolation and germ cell
sloughing, are similar to those of the toxic phthalate esters, although, as for
phthalate esters, the biochemical mechanisms underlying the toxicity are
not known.
Leydig Cells
Leydig cells are the target for many drugs and environmental chemicals
including cadmium, ethane dimethylsulphonate (EDS), aminoglute-
thimide, and phthalate esters. EDS appears to be uniquely capable of
depleting the testis of Leydig cells in vivo (Kerr, Donachie, and Rommerts,
1985) as a consequence of its cytotoxicity whereas other toxicants inhibit
enzymic reactions in Leydig cells. Thus, for example, cannabinoids,
particularly tetrahydrocannabinol, inhibit steroidogenesis by reducing
cholesterol esterase activity (Burstein, Hunter, and Shoupe, 1979) and
hexachlorocyclohexane inhibits the NAD-requiring AS-3P-hydroxy-
steroid dehydrogenase (A5-3p-HSD)and 17p-hydroxysteroid dehydro-
genase activity (Shivanandappa and Krishnakumari, 1983). LHRH
agonists ketoconazole and spironolactone inhibit steroidogenesis by
reducing both 17a-hydroxylase and 17-20 lyase activities (Penhoat,
Darbeida, Bernier, Saez, et al., 1988; Sharpe, 1982b; Menard, Guenther,
Kon, and Gillette, 1979; Brun, Leonard, Moronvaille, Caillard, P t ul.,
1991).
Ethanol and its principal metabolite, acetaldehyde, induce testicular
atrophy in the rat and mouse (Van Thiel, Gavaler, Cobb, et al., 1979;
Cicero, Bell, Meyers, and Badger, 1980; Willis, Anderson, Oswald, and
Zanveld, 1983). In most studies plasma testosterone levels are depressed
by chronic ethanol treatment, suggesting an effect on the Leydig cells.
Ultrastructural studies have shown intracellular changes in Leydig cells
subsequent to chronic ethanol administration (Gavaler, Perez, and Van
Thiel, 1982). The precise mechanism of ethanol-induced testicular dys-
function remains to be elucidated; however, large doses of ethanol have
been shown to reduce the number of LH binding sites on Leydig cell
membranes and several enzymatic steps in the biosynthesis of testo-
sterone, including A5-3P-HSD appear to be inhibited by ethanol and
acetaldehyde (Boyden and Pamenter, 1983).
Germ Cells
A wide range of chemicals are known to affect spermatogenic cells
directly, resulting in interference with the production of normal sperma-
tozoa. The mechanisms of action vary as does the stage of sper-

matogenesis affected. Mono-functional alkylating agents, such as methyl
methane sulphonate (MMS), react with germ cell DNA causing single-
strand breaks. In both pre- and meiotic cells, including early spermatids
(Lee, 1981), unscheduled DNA synthesis is induced; however, late
spermatids and spermatozoa are unable to repair the DNA damage.
Bifunctional alkylating agents, such as busulphan, stabilise DNA forming
inter- or intrastrand links. These agents mainly affect spermatogonial
cells, as do a number of other drugs such as the DNA-intercalators,
adriamycin and daunomycin, and antimetabolites such as 6-
mercaptopurine and 5-fluorouracil. The relatively prolonged sper-
matocyte development phase renders these cells particularly sensitive to
injury by chemicals. A number of other compounds act during this phase,
including oxygen-, sulphur- and nitrogen-containing heterocyclics, such
as nitrofurans, thiophenes, and dinitropyrroles. Nitrofuran, for example,
arrests primary spermatocyte development at the leptotene stage (Jack-
son, 1966).
Ethylene glycol monomethylether (EGME) , a widely used industrial
solvent, which has been shown to produce testicular atrophy in the rat,
also exerts its effect at the primary spermatocyte stage. Morphological
studies have shown that EGME and the corresponding ethyl ether
(EGEE) are toxic specifically to pachytene spermatocytes, and that cells
in meiotic division (Stage XIV) and in early pachytene (Stages I and 11)
are more susceptible than those in the late pachytene phase (Foster,
Creasy, Foster, et al., 1983; Creasy and Foster, 1984). Recently it has been
shown that disruption of calcium homeostasis may be involved in the
development of pachytene cell damage by E G M E (Ghanayem and
Chapin, 1990).
B Extratesticular sites -Hypothalamus/Pituitary

Hypothalamus / Pituitary
It has been known for a long time that oestrogen can inhibit pituitary
gonadotrophic activity (Moore and Price, 1932). Oestrogen treatment of
male rats results in the inhibition of LH secretion, as gauged by the
extent of androgenic stimulation of the sex accessory organs, but has no
effect on FSH secretion. The testes of oestrogen-treated animals show
atrophied Leydig cells, thickened tunica propria, and a reduction in the
size of the seminiferous tubules. Androgens and progestogens, including
testosterone and 17a-acetoxyprogestone derivatives, can also suppress
the gonadotrophic activity of the pituitary-hypothalmic axis. A role for
oestrogens in the pathophysiological effects of alcohol has also been
suggested, although the evidence is not strong. However, there is
evidence to suggest that ethanol may act directly at the hypothalamic
level, in that chronic alcohol feeding impairs LH response to GnRH in
the rat (Van Thiel, Gavaler, Cobb, et al., 1979).
Sex Accessory Organs

As well as causing profound effects to the testes, phthalate esters also
have marked effects on the prostrate and seminal vesicles. Treatment of
rats with both mono- and di-(2-ethylhexyl) phthalate (DEHP) results in a
decrease in prostate weight and a depletion of endogenous zinc levels
(Thomas, Curto, and Thomas, 1982). Zinc has been shown to interfere
with the binding of dihydrotestosterone to androgen receptors in this
tissue in the mouse (Donovan, Schein, and Thomas, 1980). DEHP also
reduces seminal vesicle weight in immature rats (Gray and Butterworth,
1980). The seminal vesicle/prostate complex in rats is also affected by
chronic ethanol treatments. Willis, Anderson, Oswald, and Zanveld
(1983) noted a diminution in the depth of the epithelial lining of the
seminal vesicles and an associated reduction in the overall weight of the
complex. These effects appeared to be a function of the duration and
level of ethanol exposure.
C Ovaries
Oocytes
Exposure to agents that damage oocytes can lead to reduced fertility in
females. It has been long recognised that ionising radiation can cause
destruction of female germ cells, and that the extent of sensitivity is both
species- and age-dependent . Oocyte destruction has been demonstrated
with X-rays, y-radiation, neutron bombardment , and P-irradiation from
3H20 (see review by Dobson and Felton, 1983). The particular sensitivity
of the immature mouse oocyte appears to be related to excessive
vulnerability of the plasma membrane, rather than to nuclear damage. In
most species, including man, the oocyte is not especially sensitive to
radiation, so that the principal cause of cell death is likely to be damage
to nuclear DNA.
A number of xenobiotics can also cause destruction of oocytes,
although the mechanisms involved are not fully understood. Compounds
known to be ovotoxins include polycyclic aromatic hydrocarbons (PAH) ,
such as 7,12-dimethylbenzanthracene, 3-methylcholanthrene, and
benzo[a]pyrene, and chemotherapeutic agents such as bleomycin,
cyclophosphamide, and actinomycin D. As with radiation, the immature
mouse oocyte is also particularly sensitive to chemical destruction,
although not with respect to actinomycin D which does not attack the
plasma membrane. Metabolic activation in the oocyte appears to be a
necessary factor in the ovotoxicity of PAH (Mattison, Shiromizu, and
Nightingale, 1983).
Uterus
As oestrogens act on the endometrium of the uterus to prepare it for
implantation of the fertilised ovum, compounds that are either oestroge-
nic or anti-oestrogens are likely to affect fertility. Administration of

technical grade DDT to the rat was found to be uterotropic (causing an
increase in the weight of the uterus) and subsequent studies showed
similar effects in both the rat and other species following administration
of the o,p’-isomer. A typical oestrogenic response in the endometrium of
uteri from ovariectomised rats was seen with o,p’-DDT, as well as a
reduction in serum LH but not FSH levels. Studies on the mechanism of
action of DDT showed that o,p’-DDT competitively inhibited the
binding of oestradiol to uterine oestrogen receptor sites and translocated
the oestrogen receptors from cystolic to nuclear compartments in a
manner similar to oestradiol (see review by Bulger and Kupfer, 1983). A
number of other pesticides, such as Methoxychlor (bis-p-methoxy DDT)
and Kepone, have similar oestrogenic effects. The structural diversity of
oestrogenic and antioestrogenic chemicals has been discussed by Kat-
zenellenbogen, Katzenellenbogen, Tatee , et al. (1980).
7 Methods for Studying Reproductive Toxicity in

the Male
In assessing chemically-induced reproductive toxicity in experimental
animals, two aspects of the problem need to be investigated. These are,
first, effects on the reproductive systems in male and female animals, and
second, effects on their reproductive function (Section 9). A brief
description of relevant methods employed for assessing effects on the
male reproductive system follows. Some of the methods are also
applicable in assessing toxicity in humans and these will be indicated in
the text. For a more detailed discussion of methods for detecting
alterations in the reproductive system of the male rat, see Zenick and
Goeden, 1988, and quantitative data on inter-species differences in
spermatogenesis, see Amann, 1986.
In selecting criteria for assessing injury to the male reproductive
system, a number of considerations must be taken into account. The
parameters measured must be objective and generate reproducible
quantitative data amenable to statistical analysis. Furthermore, the tests
must be sensitive and capable of indicating early signs of overt toxicity. In
an evaluation of the usefulness of some potential indicators of reproduc-
tive toxicity in the rat, Blazak, Ernst, and Stewart (1985) examined sperm
production rate, epididymal sperm numbers, transit time, and motility in
groups of F334 strain animals of different ages. Their data suggested that
testes weight and sperm numbers are unreliable indicators of sperm
production rate, but that sperm production rate and sperm movement
characteristics are sensitive indicators of adverse effects on male reproduc-
tion capacity. Recently, it has been suggested that the measurement of
intracellular messenger protein levels in blood or semen may be a more
sensitive method for detecting early effects on spermatogenesis and some
S. D . Gangolli and J . C. Phillips 39 I
encouraging results have been obtained with the Sertoli cell product,
inhibin (Allenby et al., 1991).
A Species Selection
The use of more than one animal species is recommended for assessing
the potential risk of reproductive toxicity in human males. Although a
number of laboratory or domesticated animal species may be used, the
rabbit and rat offer several advantages in comparison to the dog and
sub-human primates, and are the animal models most commonly used.
Their main advantage is that all the phases of the conception process,
such as spermatogenesis, sperm maturation , ejaculation, sperm capacita-
tion in the female, and fertilisation can be objectively evaluated,
quantified, and manipulated throughout the year. Additionally, rabbits
have a high predictable libido that may be useful in assessing effects on
sexual behaviour. Interspecies extrapolation factors have been derived for
a number of reproductive toxicants and have been used to develop a
quantitative model for risk assessment in human males (Meistrich, 1992).
B External Genitalia
Physical examination of the testes provides useful information in animal
models and in man. Measurements of testis size and scrota1 circum-
ference can be taken easily in animals with a pendulous scrotum. The
consistency of the testes, measured by using a tonometer, can provide
objective information on intratesticular pressure and is useful in long-
itudinal surveillance studies to provide data on time- and treatment-
related changes in the testes (Overstreet, 1984). The method has
potential value in monitoring human populations.
C Internal Morphology
The comparison of organ weights and volumes between treated and
control animals can provide a valuable indication of treatment-related
effects. Organs that should be examined are the testis, prostate, seminal
vesicles, epididymis, coagulating glands, pituitary, and adrenals.
Histological examination by light and transmission electron microscopy
of these organs and of the vas deferens is essential for detecting
treatment-related morphological changes. Sequential studies enable both
the identification of the initial target site and intracellular organelles
affected, and the subsequent development of the lesion to be determined.
These procedures have been found to be particularly useful in defining
the stage specificity in toxicity to the germ cells and to the spermatogenic
process (see Section 6). A description of methods commonly employed in
morphological studies on the testes and accessory organs of the re-
productive system is given by Christian (1983).
D Biochemical and Hormonal Parameters

The determination in testicular preparations of enzyme activities gene-
rally associated with intracellular metabolic functions or specifically
associated with reproductive function have been widely used for identify-
ing the initial biochemical lesion and the progression of testicular injury.
Thus, for example, Shen and Lee (1977) measured eight enzyme
activities including hyaluronidase, lactate dehydrogenase-isoenzyme X,
and sorbitol and isocitrate dehydrogenase to distinguish normal from
abnormal testicular tissue.
The levels of circulating androgens in experimental animals and in man
can be used as an index of reproductive function, as can the activities of
specific marker enzymes of steroidogenesis in testicular preparations. The
enzymes mediating the conversion of cholesterol to testosterone via
pregnenolone are indicated in Figure 2. Other biochemical markers of
relevance include the circulating levels of androgen-binding protein as
an indicator of Sertoli cell function, oxygen uptake, and carbon dioxide
production as a measure of sperm respiration, and kinase activity as an
index of phosphorylation. The hormones commonly measured as an
index of testicular injury are FSH and LH. Blood levels of these
hormones, measured by radioimmunoassay , are determined either dir-
ectly or following the administration of gonadotropin-releasing hormone
(the ‘GnRH’ test). This latter test is capable of detecting mild degrees of
primary testicular dysfunction.
E Seminal Evaluation
The examination of seminal plasma can provide valuable information in
assessing injury to the male reproductive tract. Human seminal plasma,
approximately 30% of which is derived from the prostate and 60% from
the seminal vesicles, is rich in acid phosphatase, lysozymes, fructose,
citric acid, prostaglandins, zinc, and magnesium. However, the con-
centrations of these components would appear to be of limited diagnostic
value at present and this indicates the need for further research in the
interpetation of the results.
Sperm analysis provides data on sperm concentration in seminal fluid
and information on spermatozoa1 motility, morphology, and in vitro
function. Spermatozoa1 concentration, as an index of fertility in man,
derives its relevance from the pioneering work of MacLeod in the 1940s,
and the generally accepted lower limit of ‘normal’ sperm concentration at
20 million sperm per ml is based on a survey conducted on 2000 men
(MacLeod and Gold, 1951). However, as a measure of human fertility,
sperm concentration must be interpreted with considerable caution as
wide intra-subject variation has been found (WHO, 1987). Thus, semen
samples obtained from fertile donors over a period of one year contained
between 7 and 170 million sperm per ejaculate. Furthermore, in a survey of
infertile men, MacLeod and Wang (1979) found that less than 20% of the
group had sperm concentrations below the norm of 20 million per ml and
the rest of the patients in the group had well above normal sperm counts.
F Sperm motility
Spermatozoa with adequate motility and fertilising capacity are essential
for the successful outcome of the reproductive act, and it is considered that
motility is an important determinant of fertility (Morrissey et al., 1988).
Whereas early methods for assessing sperm mobility were subjective, and
therefore of limited value, methods have now been developed based on
computer analysed videomicrography which generate objective data on
sperm motility, including swimming speed and direction. At a recent
workshop, standardisation of parameters for assessing sperm motility
using computer-assisted sperm analysis (CASA) was discussed (Chapin et
al., 1992). Changes in motility may indicate abnormal spermatogenesis,
abnormal functions of the epididymis, or the presence of antimotility
factors in the semen. These latter factors may be due to foreign chemicals
or their metabolites entering the semen via the prostate, bulbourethral
glands, or vesicular glands. It should also be noted, however, that even a
substantial reduction in the numbers of motile sperm may not have a
marked effect on reproductive function. Thus, in a recent study, treatment
of rats with nitrobenzene reduced the number of motile sperm by
approximately 80%, but had only a marginal effect on numbers of
implants (Blazak, Rushbrook, Ernst, et al., 1985).
G Spermatozoa1 Morphology
Although semen from normal humans can contain up to 40% of
abnormal sperm cells, sperm morphology can provide valuable informa-
tion on chemical toxicity to the male reproductive system. Unfortunately,
there have been no methods available for generating objective data for
assessing sperm morphology and there is no authoritative system of
classification of overall sperm abnormalities, although the mouse sperm
head morphology test (Wyrobek and Bruce, 1975) is widely used for
screening for mutagenic chemicals. The abnormalities commonly
observed include those related to sperm size and shape, either affecting the
sperm head, midpiece, or tail. Videomicrography appears to be a
potentially valuable tool for assessing sperm morphology (Overstreet,
1984) and the use of morphometric methods holds out promise in
generating computer derived calculations of sperm head length, width,
circumference, and other measurements (Katz, Overstreet, and Pelfrey,
1982). These procedures could be applied to spermatozoa from man and
experimental animals.
Two additional criteria of sperm function have received increasing
attention of late. These are, tests of sperm/cervical mucus interaction and
tests of sperm/oocyte interaction. Sperm/cervical interaction occurs
during the passage of spermatozoa through the mucus of the cervix to the
site of fertilisation. Cervical mucus, a well defined fluid, may interact with
and influence the flagellar activity of spermatozoa and the presence of
anti-sperm antibodies in the cervical mucus has been shown to affect
adversely the motility of sperm (Jager, Kremer, and Van Slochteren-
Draaisma, 1979.) The tests are conducted either on a microscope slide or in
a capillary tube. In the former test, the swimming behaviour of sperm
placed in contact with cervical mucus on a microscope slide is investigated,
and in the latter test, a capillary tube filled with cervical mucus is placed in
contact with semen, and the distance travelled by sperm along the tube
during a prescribed time measured (Katz, Overstreet, and Hanson, 1980).
The disadvantage of the sperm/cervical mucus interaction test stems from
the difficulty of obtaining an adequate supply of human cervical mucus. It
is possible that bovine cervical mucus, available in large quantities, may
prove a suitable substitute for the human material (Moghissi, Segal,
Meinhold, and Agronow, 1982). Additionally, research is in progress to
develop synthetic simulants for cervical mucus (Lorton, Kummerfield,
and Foote, 1981).
The penetration of sperm into the ovum and the subsequent physiolog-
ical changes, collectively known as capacitation, prior to the onset of
morphological changes in the acrosomes form the basis of the
sperm/oocyte interaction studies (Yanagimachi, 1981). Since human in
vitro fertilisation studies are usually precluded for obvious ethical
reasons, substitutes have been used for the human ovum. These have
included zona pellucida-stored human follicular oocytes and the zona-free
golden hamster egg. The removal of the zona pellucida of the hamster
egg by hyaluronidase digestion provides a ready source of oocytes
capable of fusing with sperms from many species including man.
Both capacitation and acrosome reaction are necessary to effect the
fusion of human sperm with the zona-free hamster egg thus providing an
appropriate system for assessing sperm function (Overstreet, 1983). The
interpretation of the sperm/oocyte interaction findings as an index of
chemical injury to the male reproductive system remains to be fully
developed.
8 Methods for Studying Reproductive Toxicity in

the Female
Many of the morphological and biochemical investigations carried out for
assessing chemical injury to the male reproductive system are also
employed in assessing toxicity in the female. A brief description of the
relevant procedures used in the study of female reproductive toxicologi-
cal follows.
S . D. Gangolli and J . C. Phillips 395
A Morphological Studies
In addition to the visual examination of the external genitalia and the
internal tissues of the reproductive tract, weights of the relevant organs,
in particular the ovaries, adrenals, and pituitary, can provide useful
information on the onset of chemical injury in female experimental
animals. Histological examination by light microscopy and by electron
microscopy of sections of the vagina, cervix, uterus, fallopian tubes, and
ovaries may reveal early treatment-related morphological and ultra-
structural changes in the female reproductive tract. An important
indicator of reproductive toxicity is the number of oocytes present in the
ovaries, measured in sections of ovaries fixed in Bouins’s solution. The
procedure provides a quantitative measure of effects on follicular
development and degeneration.
B Biochemical and Hormonal Studies

Determinations of the circulating levels of oestradiol and its metabolites,
oestrone and oestriol, and of the activities of enzymes mediating their
biosynthesis provide useful information on the reproductive competence
of experimental animals. The numbers of steroid receptors in the
cytoplasm and nuclei of target tissues, particularly those for oestradiol
and progesterone, are important indicators of chemical injury to the
female reproductive system. Chemicals may adversely affect the steroid
receptor function either by competing for, and occupying, the receptor
sites in preference to the endogenous hormones, or by interfering with
the biosynthesis of the steroid receptors. The former may lead to
hormonal imbalance and reproductive abnormalities and the latter may
affect the hormonally mediated function of the target tissues. Examples
of chemicals in the former category are DDT, polychlorinated biphenyl,
and dimethylbenzanthracene (DMBA); chemicals in the latter category
include alkylating agents and inhibitors of protein synthesis.
The ovarian hormones, oestrogen and progesterone, exhibit an
inverse relationship with FSH, LH, and LTH (luteotropic hormone), the
three gonadotropins derived from the hypophysis. Additionally, as dis-
cussed earlier, the gonadotropins regulate the oestrous cycle and the
preparation of the endometrium for the accommodation and nourishment
of the fertilised ovum. Thus, the estimation of the three gonadotropic
hormones, and of the two sex hormones provides a comprehensive
picture of the development, maintenance, and function of the female
reproductive system. Radioimmunoassay methods are extensively used
for the determination of these hormones.
9 Methods for Assessing Reproductive Function

The procedures for the assessment of chemical injury to reproductive
function in experimental animals have become relatively standardised
and are designed primarily to meet various regulatory requirements. The

guidelines set by the US Food and Drug Agency in 1966 have been the
basis for official procedures prescribed by the US Environmental Protec-
tion Agency, by Japan, UK, OECD, and more than twenty other
regulatory agencies. The US Food and Drug Agency Guidelines are
currently being assessed with a view to revision (Collins et al., 1991).
The approach commonly adopted in the evaluation of reproductive
hazards is the use of a three segment scheme of studies. The ‘Segment 1’
study is designed to assess the toxic effect of the chemical on gonadal
function and mating behaviour in both male and female (experimental)
animals, usually rats. In addition, the effect on conception rates, the early
and late stages of gestation, parturition, lactation, and the development of
offspring is assessed.
The ‘Segment 2’ study is conducted to evaluate the teratogenic potential
of a test compound. F D A guidelines prescribe both a rodent and
non-rodent species, the rat and rabbit being commonly used. Female
animals only are treated with the test compound post-mating and during
the period of major organogenesis.
The ‘Segment 3’ study is intended to evaluate the effect of treatment on
the peri- and post-natal development of the foetus. The study is usually
performed in the rat. Sexually-mature female animals are mated and then
treated with test compound beginning on day 15 of presumed gestation
and continuing throughout delivery and until 2 1 days post-parturition.
The effects of treatments on the duration of gestation, delivery, number,
and viability of naturally delivered litters are evaluated. For further details
of experimental methods, see Christian (1983).
In addition to the three-segment scheme for evaluating chemically
induced reproductive toxicity of drugs, multigeneration reproduction
studies are carried out on food additives and environmental con-
taminants. Treatment of the Fo generation commences immediately
following weaning and continues through mating, delivery, and nursing
of the F, generation pups. These pups are individually treated after
weaning, usually at 21 days of age, and the treatment continued through
mating and delivery of F2 generation pups. If required, selected F2 pups
are similarly treated through to the F3 generation. In addition to the
criteria previously mentioned, histopathological examination of all organ
systems in the foetuses and post-natal ‘behavioural’ assessment are
employed for evaluating reproductive toxicity in the experimental ani-
mals. A number of indices of reproductive function are used for
comparing treated animals with controls. The indices commonly used are:
Number of copulations x 100

Mating Index =
Number of oestrus cycles required
Number of pregnancies x 100

Fecundity Index =
Number of copulations
Number of pregnant females x 100
Male Fertility Index =
Number of non-pregant females presented
Number of females that conceive x 100
Female Fertility Index =
Number of females exposed to fertile males
Number of viable pups born x 100
Live Birth Index =
Total number of pups born
For details of the procedures mentioned the reader is referred to
official methods presented by the various national regulatory agencies
(e.g. Food and Drugs Administration, 1966; Committee on the Safety of
Medicines, 1974; Environmental Protection Agency, 1979).
Recently, the National Toxicology Program (NTP) in the United States
has proposed a new test procedure, which is claimed to be a reliable and
cost-effective alternative to the multigeneration study. This system,
Reproductive Assessment by Continuous Breeding (RACB) has been
evaluated with a wide range of chemicals and to date over 70 studies
have been completed, almost exclusively with mice (Gulati et ul., 1991).
The experimental design is based on four ‘tasks’, which in mice last over a
period of 35 weeks, and include observations on F,, F,, and F,
generations. Task 1 is a 14 day dose-selection study using conventional
toxicity end points (body weight, clinical signs). Task 2 is a 14 week
continuous breeding phase followed by a 3 week period in which the
co-habiting pair are separated. During this task, data is collected on all
newborn pups. If significant effects on fertility are seen, a cross-over
mating trial (task 3) is carried out to determine which sex is more sensitive.
The last litter of the continuous breeding phase (usually the 5th) is used for
the second generation phase (task 4), the mothers (F,) being dosed
through weaning and the offspring (F,) until mated (approximately 11
weeks). Finally, the litters (F,) are examined. If task 3 is undertaken the
offspring are also analysed as in task 4.
10 Conclusions
This chapter has attempted briefly to discuss chemical injury to the male
and female reproductive systems from a mechanistic standpoint and to
describe the various methods employed for investigating chemical injury,
both in experimental animals and in man. No attempt has been made to
address the question of toxicity expressed in the developing foetus or
subsequently in the offspring (developmental and behavioural toxicology);
many recent reviews have appeared on aspects of this area of toxicology
(see, for example, Hemminki and Vineis, 1985; IPCS, 1984; Kimmel,
1988).
The limitations in the methodology currently employed have been

indicated and the difficulties in interpreting experimental results pointed
out. These difficulties arise principally as a result of the lack of detailed
information on both the pathogenesis of chemically-induced testicular
and ovarian injury and on the biochemical mechanisms underlying injury
at these sites. Progress in understanding reproductive toxicity is likely to
be enhanced by further research in a number of areas. These include the
development of more objective methods for assessing toxic effects (e.g.
the use of morphometric methods for examining sperm), investigations
into the influence of diet, age, genetic, and other social factors (e.g.
alcohol intake and drug interactions) on chemically-induced reproductive
toxicity, and inter-species comparisons to allow the extrapolation of
animal data to man.
An understanding of the biochemical mechanisms underlying re-
production is of central importance, and in this context the development
and application of in vitro model systems may be valuable. Examples of
in vitro systems examined recently include seminiferous tubule prepara-
tions, used in the studies on the toxicity of phthalate esters and glycol
ethers (see Section 6) and the zona-free hamster egg for assessing sperm
function (see Section 7 ) . Alternative approaches for assessing embryo-
toxicity and teratogenicity have recently been reviewed by ECETOC
(1989). Clearly, the realistic assessment of toxicity of chemicals to the
human reproductive system requires further fundamental studies using
both in vivo and in vitro techniques.
References
Allenby, G., Sharpe, R. M., and Foster, P. M. D. (1990). Changes in Sertoli cell
function in vitro induced by nitrobenzene. Fundam. Appl. Toxicol., 14,
364-375.
Allenby, G.,Foster, P. M. D., and Sharpe, R. M. (1991). Evaluation of changes in
the secretion of immunoactive inhibin by adult rat seminiferous tubules in
vitro as an indicator of early toxicant action on spermatogenesis. Fundarn.
Appl. Toxicol., 16, 710-724.
Amann, R. P. (1986). Detection of alterations in testicular and epididymal
function in laboratory animals. Environ. Health Persp., 70, 149-158.
Auclair, C., Kelly, P. A., Coy, D. H., Schally, A. V., and Labrie, F. (1977).
Potent inhibitory activity of [D-leu6, Des-gly-NH,( lO)]LHRH ethylamide on
LH/hCG and PRL testicular receptor levels in the rat. Endocrinology, 101,
1890-1893.
Auclair, C., Kelly, P. A., Labrie, F., Coy, D. H., and Schally, A. V . (1977).
Inhibition of testicular luteinising hormone receptor levels by treatment with
a potent luteinising hormone-releasing hormone agonist or human chorionic
gonadotropin. Biochem. Biophys. Res. Commun., 76, 855-862.
Baranao, J. L. S . , Tesone, M., Calvo, J. C., Luthy, I. A., Gonzalez, S. I.,
Charreau, E. H., and Calandra, R. S. (1983). Prolactin effects on the
prepubertal male rat. In: ‘Recent Advances in Male Reproduction: Molecu-
lar Basis and Clinical Implications’. E h R. D’Agata, M. B. Lipsett, P.

Polosa, and H. J. van der Molen. Raven Press, NY, pp. 305-319.
Bell, J. B. G., Vinson, G. P., Hopkin, D. J., and Lacy, D. (1968). Pathways for
androgen biosynthesis from (7-3H)pregnenolone and (4-‘4C) progesterone by
rat testis interstitium in vitro. Biochem. Biophys. Acta, 164, 412-420.
Bernstein, M. E. (1984). Agents affecting the male reproductive system: effects of
structure on activity. Drug Metab. Rev., 15, 941-996.
Blazak, W. F., Ernst, T. L., and Stewart, B. E. (1985). Potential indicators of
reproductive toxicity: testicular sperm production and epididymal sperm
numbers, transit time and motility in Fischer F344 rats. Fundam. Appl.
Toxicol., 5 , 1097-1103.
Blazak, W. F., Rushbrook, C. J., Ernst, T. L., Steward, B. E., Spak, D.,
DiBiasio-Erwin, D., and Black, V. (1985). Relationship between breeding
performance and testicular/epididymal functioning in male Sprague-Dawley
rats exposed to nitrobenzene. Toxicologist, 5, 121.
Boekelheide, K. (1988). Rat testis during 2,5-hexanedione intoxication and
recovery. I . Dose-response and reversibility of germ cell loss. Toxicol. Appl.
Pharmacol., 92, 18-27.
Boyden, T. W. and Pamenter, R. W. (1983). Effect of ethanol on the male
hypot halmic-pituitary -gonadal axis. Endocrine Rev., 4, 389-395.
Brun, H. P., Leonard, J. F., Moronvaille, V., Cailland, J. M., Melcion, C., and
Cordier, A. (1991). Pig Leydig cell culture: a useful in vitro test for evaluating
the testicular toxicity of compounds. Toxicol. Appl. Pharmacol., 108,
307- 320.
Bulger, W. H. and Kupfer, D. (1983). Oestrogenic action of DDT analogs. A m .
J. Ind. Med., 4, 163-173.
Rurstein, S . , Hunter, S. A., and Shoupe, T. S. (1979). Site of inhibition of
Leydig cell testosterone synthesis by A’-tetrahydrocannabinol. Mol.
Pharmacol., 15, 633-640.
Chapin, R. E., Norton, R. M., Popp, J. A., and Bus, J. S. (1982). The effects of
2,5-hexanedione on reproductive hormones and testicular enzyme activities
in the F344 rat. Toxicol. Appl. Pharmacol., 62, 262-272.
Chapin, R. E., Filler, R. S., Gulati, D., et al. (1992). Methods for assessing rat
sperm motility. Reprod. Toxicol., 6, 267-273.
Chase, D. J., Dixon, G. E. K., and Payne, A. H. (1982). Development of Leydig
cell function. Prog. Clin. Biol. Res., 112, 209-219.
Christian, M. S. (1983). Assessment of reproductivity toxicity. State of the art.
In : ‘Assessment of Reproductive and Teratogenic Hazards. Advances in
Modern Environmental Toxicology’. Vol. 111, Chapter 8. Eds M. S.
Christian, W. N. Galbraith, P. Voytek, and M. A. Mehlman. Princeton
Scientific Publishers Inc., Princeton, USA, pp. 65-76.
Cicero, T. J., Bell, R. D., Meyer, E. R., and Badger, T. M. (1980). Ethanol and
acetaldehyde directly inhibit testicular steroidogenesis, J . Pharmacol. Exp.
Ther., 213, 228-233.
Collins, T. F. X., Black, T. N., Graham, S. L., Jackson, B. A., and Welsh, J. J.
(1 991). Updating developmental toxicity testing guidelines for the safety
assessment of direct food additives and color additives used in food : results of
a survey. J. Am. Coll. Toxicol., 10, 461-475.
Committee on the Safety of Medicines (1974). Notes for guidance on reproduc-
tion studies. Dept. of Health and Social Security, Great Britain.
Creasy, D. M., Beech, L. M., Gray, T. J. B., and Butler, W. H. (1987). The
ultrastructural effects of di-n-pentyl phthalate on the testis of the mature rat.
Exp. Mol. Pathol., 46, 357-371.
Creasy, D. M., Foster, J. R., and Foster, P. M. D. (1983). The morphological
development of di-n-pentyl phthalate-induced testicular atrophy in the rat. J.
Pathol., 139, 309-321.
Creasy, D. M. and Foster, P. M. D. (1984). The morphological development of
glycol ether induced testicular atrophy. Exp. Mol. Pathol., 40, 169-176.
Dixon, R. L. (1986). Toxic responses of the reproductive system. In:
‘Toxicology-the basic science of poisons’, 3rd Edn. Eds C. D. Klaassen,
M. 0. Amdur, and J. Doull. Chapter 16. Macmillan, New York, pp. 432-
477.
DiZerega, G. S. and Sherins, R. J. (1981). Endocrine control of adult testicular
function. In: ‘The Testis’. Eds H. Burger and D. Dekrester. New York
Raven Press, pp. 127-140.
Dobson, R. L. and Felton, J. S. (1983). Female germ cell loss from radiation and
chemical exposures. A m . J . Znd. Med., 4, 175-190.
Donovan, M. P., Schein, L. G., and Thomas, J. A. (1980). Inhibition of
androgen-receptor interaction in mouse prostate gland cytosol by divalent
metal ions. Mol. Pharmacol., 17, 156-162.
Dorrington, J. H., Fritz, I. B., and Armstrong, D. T. (1976). Site at which FSH
regulates estradiol-17P biosynthesis in Sertoli cell preparations in culture.
Mol. Cell Endocrinol., 6 , 117-122.
ECETOC (1 989). Alternative approaches for the assessment of reproductive
toxicity (with emphasis on embryotoxicity/teratogenicity). Monograph
No. 12, Brussels, Belgium.
Environmental Protection Agency (1979). Proposed Health Effects Test Stand-
ards for Toxic Substances Control Act Test Rules and Proposed Good
Laboratory Practice Standards for Health Effects. Fed. Reg. pt. IV 44 (145)
Sub-part F. Teratogenic/Reproductive Health Effects, July 26th.
Ewing, L. L., Wing, T.-Y., Cochran, R. C., Kromann, N., and Zirkin, B. R.
(1983). Effect of luteinising hormone on Leydig cell structure and testo-
sterone secretion. Endocrinology, 112, 1763-1769.
Food and Drugs Administration (1966). Guidelines for reproductive studies for
safety evaluation of drugs for human use. Washington, DC.
Foster, P. M. D., Creasy, D. M., Foster, J. R., Thomas, L. V . , Cook, W. M.,
and Gangolli, S. D. (1983). Testicular toxicity of ethylene glycol monomethyl
and monoethyl ethers in the rat. Toxicol. Appl. Pharmacol., 69, 385-399.
French, F. S. and Ritzen, E. M. (1973). Androgen-binding protein in efferent
duct fluid of rat testis. J . Reprod. Fertil., 32, 479-483.
Gangolli, S. D. (1982). Testicular effects of phthalate esters. Environ. Health
Perspect., 45, 77-84.
Gavaler, J. S., Perez, H., and Van Thiel, D. H. (1982). Ethanol produced
morphologic alterations of Leydig cells. Alcoholism Clin. Exp. Res., 6, 296.
Ghanayem, B. I. and Chapin, R. E. (1990). Calcium channel blockers protect
against ethyleneglycol monomethylether (2-methoxyethanol)-induced tes-
ticular toxicity. Exp. Mol. Pathol., 52, 279-290.
Gray, T. J. B. and Butterworth, K. R. (1980). Testicular atrophy produced by
phthalate esters. Arch. Tox. (Suppl 4), 452-455.
Gray, T. J. B., Beamand, J. A., and Gangolli, S. D. (1982). Effects of phthalate
S . D . Gangolli and J . C . Phillips 40 1
esters on rat testicular cell cultures and on Sertoli cell function in the intact
testis. Toxicologist, 2, 78, Abs 276.
Gray, T. J. B. and Gangolli, S. D. (1986). Aspects of the testicular toxicity of
phthalate esters. Environ. Health Perspect., 65, 229-235.
Gulati, D. K., Hope, E., Teague, J. L., and Chapin, R. E. (1991). Reproductive
toxicity assessment by continuous breeding in Sprague-Dawley rats: a
comparison of two study designs. Fundam. Appl. Toxicol., 17, 270-279.
Hansson, V., Jggou, B., Attramadal, H., Jahnsen, T., Le Gac, F., Tvermyr, M.,
Frflysa, A., and Horn, R. (1983). Regulation of Sertoli cell function and
response. In : ‘Recent Advances in Male Reproduction: Molecular Basis and
Clinical Implications’, Eds R. D’Agata, M. B. Lipsett, P. Polosa, and H. J.
van der Molen. Raven Press, New York.
Hemminki, K. and Vineis, P. (1985). Extrapolation of evidence on teratogenicity
of chemicals between humans and experimental animals: Chemicals other
than drugs. Teratol. Carcinogen. Mutagen., 5 , 251-318.
IPCS: International Programme on Chemical Safety (1984). Principles for
evaluating health risks to progeny associated with exposure to chemicals
during pregnancy. Environ. Health. Criteria, 30. WHO, Geneva.
Jackson, H. (1966). ‘Antifertility compounds in the male and female’, Thomas,
Springfield, Ill.
Jager, S , , Kremer, J., and van Slochteren-Draaisma, T. (1979). Presence of
sperm agglutinating antibodies in infertile men and inhibition of in vitro
sperm penetration into cervical mucus. Znt. J . Androl., 2, 117.
Johnson, K. J., Hall, E. S., and Boekelheide, K. (1991). 2,5-Hexanedione exposure
alters rat Sertoli cell cytoskeleton I. Microtubules and seminiferous tubule
fluid secretion. Toxicol. Appl. Pharrnacol., 11 1, 432-442.
Katz, D. F., Overstreet, J. W . , and Hanson, F. W. (1980). A new quantitative
test for sperm penetration into cervical mucus. Fertil. Steril., 33, 179-186.
Katz, D. F., Overstreet, J. W . , and Pelfrey, R. J. (1982). Integrated assessment
of the motility, morphology and morphometry of human spermatozoa. In :
‘Human Fertility Factors’, Eds P. Spira and P. Jouannet. Paris, Inserm, p.
97.
Katzenellenbogen, J. A., Katzenellenbogen, B. S . , Tatee, I . , Robertson, D. W . ,
and Landvatter, S. W. (1980). The chemistry of oestrogens and anti-
oestrogens: Relationships between structure, receptor binding and biological
activity. In: ‘Oestrogens in the Environment’ Ed. J. A. McLachlan. Vol. 1,
Elsevier/North-Holland, NY. pp. 33-52.
Kerr, J. B., Donachie, K., and Rommerts, F. F. G. (1985). Selective destruction
and regeneration of rat Leydig cells in vivo. A new method for the study of
semin i fer o u s t u bu 1a r - i n t er st i t ial ti s u e in t e r act i o ns . Cel I Tissue Res ., 242,
145-156.
Kimmel, C. A. (1988). Current status of behavioural teratology. Science and
Regulation. C R C Crit. Rev. Toxicol., 19, 1-10.
Krasavage, W. J., O’Donoghue, J. L., DiVincenzo, G. D., and Terhaar, C. J.
(1980). The relative neurotoxicity of methyl n-butyl ketone, n-hexane, and
their metabolites. Toxicol. Appl. Pharmacol., 52, 433-441.
Lamb, J. C. and Foster, P. M. D. (1988). ‘Physiology and Toxicology of Male
Reproduction’, Academic Press Inc., San Diego, CA, USA.
Lee, I. P. (1981). Effect of drugs and chemicals on male reproduction. ZNSERM,
103, 311-331.
Le Gac, F., Attramadal, H., Horn, R., Tvermyr, M., Froysa, A., and Hansson,
V. (1982). Hormone stimulation of lactate/pyruvate secretion by cultured
Sertoli cells and maintenance of ATP levels in primary spermatocytes and
round spermatids. Program 2nd Eur. Workshop Mol. Cell. Endocrin. of the
Testis, Rotterdam, 11-14 May, Abstract C-16.
Lorton, S. P., Kummerfield, H. L., and Foote, R. H. (1981). Polyacrylamide as a
substitute for cervical mucus in sperm migration tests. Fertil. Steril., 35,
222-225.
Lucier, G. W., Lee, I. P., and Dixon, R. L. (1977). Effects of environmental
agents on male reproduction. In: ‘The Testis’ Vol. IV. Ed. R. A. Gomes,
Academic Press, NY, pp. 578-604.
MacLeod, J. and Gold, R. Z . (1951). The male factor in fertility and infertility.
11. Spermatozoon counts in 1000 men of known fertility and in 1000 cases of
infertile marriage. J . Urol., 66, 436-449.
MacLeod, J. and Wang, Y. (1979). Male fertility potential in terms of semen
quality: a review of the past, a study of the present. Fertil. Steril., 31,
103-116.
Mattison, D. R., Shiromizu, K., and Nightingale, M. S. (1983). Oocyte
destruction by polycyclic aromatic hydrocarbons. A m . J . Ind. Med., 4,
191-202.
Meistrich, M. L. ( 1 992). A method for quantitative assessment of reproductive
risks to the human male. Fundarn. Appl. Toxicol., 18, 479-490.
Menard, R. H., Guenthner, T. M., Kon, H., and Gillette, J. R. (1979). Studies
on the destruction of adrenal and testicular cytochrome P450 by spironolac-
tone. J. Biol. Chem., 254, 1726-1733.
Moghissi, K. S . , Segal, S . , Meinhold, D., and Agronow, S. J. (1982). In vitro
sperm cervical mucus penetration: studies in human and bovine cervical
mucus. Fertil. Steril., 37, 823-827.
Moore, C. R. and Price, D. (1932). Gonad hormone functions, and the reciprocal
influence between gonads and hypophysis with its bearing on the problem of
sex antagonism. Am. J. Anat., 50, 13-67.
Morrissey, R. E., Lamb, J. C., Schwetz, B. A., Teague, J. L., and Morris, R. W.
( 1 988). Association of sperm, vaginal cytology, and reproductive organ
weight data with results of continuous breeding reproduction studies in Swiss
CD-1 mice. Fundam. Appl. Toxicol., 11, 359-371.
Orgebin-Crist, M. C., Danzo, B. J., and Davies, J. (1975). Endocrine control of
the development and maintenance of sperm fertilising ability in the epididy-
mis. In: ‘Handbook of Physiology’, Section 7, Volume V. Eds D. W.
Hamilton and R. 0. Greep. American Physiological Society, Washington
DC., pp. 319-335.
Overstreet, J. W. (1983). Evaluation and control of the fertilizing power of
sperm. In: ‘The Sperm Cell’. Ed. J. Andre. Martinus Nijhoff Publishers,
Boston, p. 1.
Overstreet, J. W. (1984). Laboratory tests for human male reproductive risk
assessment. Teratogen. Carcinogen. Mutagen., 4, 67-82.
Penhoat, A., Darbeida, H., Bernier, M., Saez, J . M., and Durand, Ph. (1988).
Inhibition of hormonally induced CAMP and steroid production by inhibi-
tors of pregnenolone metabolism in adrenal and Leydig cells. Mol. Cell.
Endocrind., 60, 55-60.
Reader, S. C . J. and Foster, P. M. D. (1990). The in vitro effects of four isomers of
dinitrotoluene on rat Sertoli and Sertoli-germ cell co-cultures: Germ cell

detachment and lactate and pyruvate production. Toxicol. Appl. Phurmucol.,
106,287-294.
Ritzen, E. M., Boitani, C., Parvinen, M., French, F. C., and Feldman, M.
(1982). Age-dependent secretion of ABP by rat seminiferous tubules. Mol.
Cell Endocrinol., 25, 25-33.
Rommerts, F. F. G., Bakker, G. H., and van der Molen, H. J. (1983). The role
of phosphoproteins and newly synthesised proteins in the normal regulation
of steroidogenesis in Leydig cells. J. Steroid Biochern., 19, 367-373.
Schrag, S. D. and Dixon, R. L. (1985). Occupational exposures associated with
male reproductive dysfunction. Annu. Rev. Pharmacol. Toxicol., 25, 567-
592.
Setchell, B. P. (1978). ‘The Mammalian Testis’. New York, Cornell University.
Shaffer, C. B., Carpenter, C. P . , and Smyth, H. R. (1945). Acute and sub-acute
toxicity of di-(2-ethylhexyl)phthalate with notes upon its metabolism. J . Ind.
Hyg. Toxicol., 27, 130-135.
Sharpe, R. M. (1982a). The homonal regulation of the Leydig cell. I n : ‘Oxford
Reviews of Reproductive Biology’. Vol. 4. Ed. C. A. Finn. pp. 241-317.
Sharpe, R. M. (1982b). Cellular aspects of the inhibitory actions of LHRH on the
ovary and testis. J . Reprod. Fertil., 64, 517-527.
Sharpe, R. M., Frazer, H. M., Cooper, I., and Rommerts, F. F. G. (1981).
Sertoli-Leydig cell communication via an LHRH-like factor. Nature
(London), 290, 785-787.
Sharpe, R. M. (1983). Regulation of the Leydig cell by luteinising hormone-
releasing hormone. In : ‘Recent Advances in Male Reproduction: Molecular
Basis and Clinical Implications’ Eds R. D’Agata, M. B. Lipsett, P. Polosa,
and H. J. van der Molen. Raven Press, NY. pp. 197-204.
Shen, R. S. and Lee, I. P. (1977). Developmental patterns of enzymes in mouse
testis. J. Reprod. Fertil., 48, 301-305.
Shivanandappa, T. and Krishnakumari, M. K. (1983). Hexachlorocyclohexane-
induced testicular dysfunction in rats. Acta Pharmacol. Toxicol., 52, 12-17.
Sjoberg, P., Bondesson, U., Gray, T. J. B., and Ploen, L. (1986). Effects of
di-(2-ethylhexyl)phthalate and five of its metabolites on rat testes in vivo and
in uitro. Acta Pharmacol. Toxicol., 58, 225-233.
Steinberger, E. and Steinberger, A. (1974). Hormonal control of testicular
function in mammals. In: ‘Handbook of Physiology’, 4(2), Eds K. Knobil and
W. H. Sawyer. American Physiological Society, Washington D.C. pp.
325-345.
Steinberger, A. and Steinberger, E. (1976). Secretion of an FSH-inhibiting factor
cultured Sertoli cells. Endocrinology, 99, 918-921.
Steinberger, E., Browning, J. Y . , and Grotjan, H. E. (1983). Another look at
steroidogenesis in testicular cells. In: ‘Recent Advances in Male Reproduc-
tion: Molecular Basis and Clinical Implications’. Eds R. D’Agata, M. B.
Lipsett, P. Polosa, and H. J. van der Molen. Raven Press, NY. pp. 113-120.
Steinberger, E. (1976). Biological action of gonadotropins in the male.
Pharmacol. Ther., B , 2, 771-786.
Takizawa, K. and Mattison, D. A. (1983). Female reproduction. Am. J. Ind.
Med., 4, 17-30.
Tash, J. S . , Welsh, M. J., and Means, A. R. (1980). Protein kinase inhibitor as
an intracellular marker of FSH action in the Sertoli cell. In: ‘Testicular
Development, Structure and Function’. Eds A. Steinberger and E. Steinber-

ger. Raven Press. NY, pp. 159-167.
Tcholakian, R. K. and Steinberger, A. (1978). Progesterone metabolism by
cultured Sertoli cells. Endocrinology, 103, 1335-1343.
Thomas, J. A., Curto, K. A . , and Thomas, M. J. (1982). MEHP/DEHP: Gondal
toxicity and effects on rodent accessory sex organs. Environ. Health
Perspec., 45, 85-88.
Van Thiel, D. H., Gavaler, J. F., Cobb, C. F., Sherins, R. J., and Lester, R.
(1979). Alcohol-induced testicular atrophy in the adult male rat.
Endocrinology, 105, 888-895.
Van Thiel, D. H., Sherins, J. R., Myers, G. H., and DeVita, V. T. (1972).
Evidence for a specific seminiferous tubule factor affecting follicle-
stimulating hormone secretion in man. J . Clin. Invest., 51, 1009-1019.
Waller, D. P., Killinger, J. M., and Zaneveld, L. J. D. (1985). Physiology and
toxicology of the male reproductive tract. In : ‘Endrocrine Toxicology’ Eds J.
A. Thomas et al. New York, Raven Press.
Waller, D. P. and Nikurs, A. R. (1986). Review of the physiology and
biochemistry of the male reproductive tract. J . A m . College Toxicol., 5 ,
209-223.
Weiss, L. and Greep, R. 0. (1977). ‘Histology’ 4th Edn. McGraw Hill Book Co.,
New York.
Welsh, M. J. and Wiebe, J. P. (1978). Sertoli cell capacity to metabolise
C,,-steroids. Variation with age and the effect of follicle-stimulating hor-
mone. Endocrinology, 103, 838-844.
WHO (1987). Laboratory manual for the examination of human semen and
semen-cervical mucous interactions. Cambridge University Press, Cam-
bridge, UK.
Willis, B. R., Anderson, R. A., Oswald, C., and Zaneveld, W. D. (1983).
Ethanol-induced male reproductive tract pathology as a function of ethanol
dose and duration of exposure. J . Pharm Exp. Therap., 225, 470-478.
Wyrobek, A. J. and Bruce, W. R. (1975). Chemical induction of sperm
abnormalities in mice. Proc. Natl. Acad. Sci. USA, 72, 4425-4429.
Yanagimachi, R. (1981). Mechanisms of fertilization in mammals. In: ‘Fertiliza-
tion and Embryonic Development In vitro’. Eds L. Mastroianni and J. D.
Biggers. Plenum Publishing Corp. New York, p. 81.
Yanahara, T. and Troen, P. (1972). Studies of the human testis I. Biosynthetic
pathways for androgen formation in human testicular tissue in vitro. J . Clin.
Endocrinol. Metab., 34, 783-792.
Zenick, H. and Goeden, H. (1988). Evaluation of copulatory behaviour and sperm
in rats: role in reproductive risk assessment. In ‘Physiology and Toxicology
of Male Reproduction’. Chapter 8 . Eds J. C. Lamb and P. M . D. Foster.
Academic Press Inc., San Diego, USA.
CHAPTER 18
Statistics
PETER N. LEE
1 Introduction
This chapter is aimed at providing some insight into statistical aspects of
the design, conduct, analyses, and interpretation of toxicological data and
is intended mainly for the reader not qualified in statistics. In general,
mathematical details are kept to a minimum with emphasis being given to
the principles involved. While the chapter should give the reader
sufficient information to choose and carry out appropriate analyses in a
number of situations, the need to have the advice of an expert statistician
available when dealing with toxicological data cannot be over-
emphasised. Scientific journals are full of papers describing studies where
the authors’ conclusions cannot be supported, owing to deficiencies in the
statistical methodology which would not have occurred had a qualified
statistician been available to the researcher or indeed had one refereed
the paper.
2 General Principles and Terminology
A Sources of Treatment/Control Difference

Essentially, the objective of a toxicological study is to determine whether
a treatment elicits a response, but the observation of a difference in
response between a treated and a control group does not necessarily
mean that the difference is a result of treatment. There may be two other
causes of difference, bias, that is systematic differences other than
treatment between the groups, and chance, or random differences. Good
experimental design and analysis should seek to avoid bias where possible
by ensuring that like is compared with like.
Chance factors cannot be wholly excluded since identically treated
animals will not all respond identically, however carefully the experiment
is conducted. It is impossible to be absolutely certain that even the most
40 5
406 Statistics
extreme treatment/control differences are not due to chance, but the

appropriate statistical analysis will allow the experimenter to assess the
probability of a ‘false positive’, that is of the observed difference having
occurred had there been no effect of treatment at all. The smaller the
probability is, the more the experimenter will be confident of having
found a real effect. In order to improve the likelihood of detecting a true
effect with confidence, it is necessary to try to minimise the role of
chance, Experimental design should aim also at trying to ensure that the
‘signal’ can be recognised above the ‘noise’.
B Hypothesis Testing and Probability Values

Toxicological reports often include statements such as ‘the relationship
between treatment and blood glucose levels was statistically significant
( p = 0.02)’. What does this actually mean? There are three
considerations.
Firstly there is a difference in meaning between biological and
statistical significance. It is quite possible to have a relationship that is
unlikely to have happened by chance and therefore is statistically
significant, but of no biological consequence at all, the animal’s well-
being not affected. Similarly, an observation may be biologically but not
statistically significant, as when one or two of an extremely rare tumour
type are seen in treated animals. Overall judgment of the evidence must
take into account both biological and statistical significance.
Secondly, ‘p = 0.02’ does not mean that the probability that there is no
real treatment effect is 0.02. The true meaning is that, given the
treatment actually had no effect whatsoever (or, to phrase it more
technically, ‘under the null hypothesis’), the probability of observing a
difference as great or greater than actually seen is 0.02.
Thirdly, there are two types of probability value (p-value). A
‘one-tailed (or one-sided) p-value’ is the probability of getting, by chance
alone, a treatment effect in a specified direction as great as or greater
than that observed. A ‘two-tailed p-value’ is the probability of getting, by
chance alone, a treatment difference in either direction -positive or
negative-as great as or greater than that observed. Normally, two-
tailed p-values are appropriate, and by convention p-values are assumed
to be two-tailed unless the contrary is stated. However, when there is
prior reason to expect only a treatment effect in one direction, a
one-tailed p-value is normally used. If a one-tailed p-value is used,
differences in the opposite direction to that assumed are ignored.
While, in an unbiased study, a p-value of 0.001 or less can on its own
provide very convincing evidence of a true treatment effect, less extreme
p-values, such as p = 0.05, should be viewed as providing indicative
evidence of a possible treatment effect, to be reinforced or supported by
other evidence. If the difference is similar to one found in a previous study,
or if the response, based on biochemical considerations, is expected, a
Peter N . Lee 407
larger p-value would be needed than if the response was unexpected 01' not
found a t other dose levels.
Some laboratories, when presenting results of statistical analyses,
assign a magical significance to the 95% confidence level ( p < 0.05) and
simply mark results as significant or not significant at this level. This is
poor practice as it gives insufficient information and does not enable the
distinction to be made between a true effect and one which requires other
confirmatory evidence. While it is not necessary to give p-values exactly,
it is essential to give some idea of the degree of confidence. A useful
method is to use plus signs to indicate positive differences (and minus
signs to indicate negative differences) with +++ meaning p < 0.001, + +
p <0.01 and 3 0.001, + p <0.05 and 20.01, and (optionally) (+)
meaning p < 0.1 and 3 0.05. This makes it easier to assimilate findings
when results for many variables are presented.
C Multiple Comparisons
Toxicology studies frequently involve making treatment/control com-
parisons for large numbers of variables. If no treatment effect exists at
all, it is possible that, purely by chance, one or more variables will show
differences significant at the 95% confidence level. For example, with 100
independent variables at least one variable would show significance 99.4%
of the time. Because of this, some have suggested making the critical values
required to achieve significance more stringent with increasing numbers of
variables studied, so that, in testing a t the 95% confidence level, 19 times in
20 all the variables in the test should show non-significance if the treatment
is without effect. This approach is not to be recommended because it is
frequently the case in toxicological studies that a compound has only one
or two real effects and has no effect on a large number of other variables
studied. Such multiple-comparison tests would make it much more
difficult to demonstrate statistical significance for the real effects.
In any case there is something unsatisfactory about a situation
where the relationship between a treatment and a particular response
depends arbitrarily on which other response happens to be investigated at
the same time. For this reason Section 4, which describes methods of
analysis, does not give details of such procedures.
An alternative multiple-comparison problem relates to the situation
where more than one test can be carried out in relation to the same
response variable, because there is information either under different
experimental situations (e.g. males, females) or from a number of
different dose levels of treatment. This is discussed in some detail in
Section 4A.
D Some Statistical Terms

Before going into details of experimental design and analysis it is helpful
to give brief definitions of a number of the more commonly used
408 Statistics
statistical terms, some of which are used in the text that follows and some
of which the reader may come across in other statistical documents. The
terms are considered in groups rather than alphabetically. For a much
more extensive alphabetical list see Marriott (1990).
Frequency Distribution
The frequency distribution is the relationship between the values of the
data that occur and how often they occur. For observed data the
distribution is usually given as a table with some grouping for continuous
variables. Data are often assumed to satisfy some theoretical underlying
frequency distribution in which the relationship between the data values
and their relative frequency is defined by an equation depending on one
or more parameters. For such distributions the formulae are usually
defined so that the total frequency is equal to 1, the relative frequency for
discrete data then being equivalent to the probability of the value
occurring. An example of a theoretical distribution for discrete data is the
binomial distribution, where the probability of an observed value x
occurring is given by:
n!
prob (x) = p"( 1 - p y x
x! ( n - x ) !
This distribution occurs when an event has a probability p of occurring at
any one trial and n independent trials are carried out, x being the number
of times the event occurs.
For continuous data the most commonly found distribution is the
normal distribution, known otherwise as the Gaussian or Laplacian
distribution. It has a bell-shaped curve with the equation:
where m is the mean and (7 the standard deviation (see below). The
probability of an observation occurring within the range ( x l , x2) can be
obtained by integrating the curve over this range.
Histogram
A histogram is a diagrammatic representation of a frequency distribution
in which rectangles proportional in area to the class frequencies are
erected on sections of the horizontal axis, the width of each section
representing the corresponding class interval of the variate.
Statistic
A statistic is a figure calculated from a set of data intended to summarise
some feature of it. Usually it is an estimate of some parameter of the
underlying population from which the sample is drawn.
Peter N . Lee 409
Average
Statistics are often used to indicate central tendency. Average is a
familiar word, not usually used in technical statistical reports, to describe
such statistics. In common usage, average is understood to be the mean,
but the median and the mode are other types of average.
Mean
The mean, or the arithmetic mean, is calculated by summing the
observations and dividing by their number. For a theoretical frequency
distribution, the mean is calculated by the integral Jrm
x dF(x).
Median
The median is that value which divides the total frequency into two
halves so that as many observations are greater as are less.
Mode
The mode is the value which occurs most frequently.
Variance
For a theoretical frequency distribution, the variance is calculated by the
integral JZm ( x - p)2dF(x) where p is the mean. This is a measure of the
extent to which the data are dispersed about the mean. For a normal
distribution the variance can be estimated from a sample by the formula:
where n is the number of samples and p is the sample mean.
Standard Deviation
The standard deviation is the positive square root of the variance.
Standard Error
The standard error is the positive square root of the variance of the
sampling distribution of a statistic, commonly the sample mean.
Skewness
Skewness is a measure of asymmetry of a frequency distribution.
410 Statistics
Kurtosis
Kurtosis is a measure of ‘peakedness’ of a frequency distribution.
Chi-squared Distribution
The chi-squared distribution is the distribution of the sum of squares of n
independent normal variates each with mean 0 and standard deviation 1.
F -distribution
The F-distribution is the distribution of the ratio of two independent
variates, each distributed like a variance in normally distributed samples.
Student’s t-Distribution
The Student’s t-distribution is the distribution of the ratio of a simple
mean to its simple variance, in samples from a normal population.
Degrees of Freedom
Degrees of freedom are the number of independent comparisons which
can be made between members of a sample. For example, if a sample of
a constant size n is grouped into r intervals, there are r - 1 degrees of
freedom as, given r - 1 of the frequencies, the other one is specified.
Contingency Table
A contingency table is a table giving the breakdown of a sample
according to two or more characteristics, each with two or more possible
values. Thus the number of animals in an experiment may be presented
in a 2 x 5 x 2 contingency table according to sex (male or female), group
(control or four dose levels), and presence of a tumour (yes or no).
Confounding
A study design is said to be confounded if the effects of two or more
possible determinants of a response cannot be separated. A study in
which the slides from the test group are examined by pathologist A and
the slides from the control group are examined by pathologist B is
confounded since any difference seen in response cannot be attributed
with certainty either to an effect of treatment or to a difference between
pathologists.
Interaction
Two factors are said to interact if the difference in response between
levels of one factor depends on the level of the other factor. Thus, if
Peter N . Lee 41 1
treatment increased body weight by 30g in males but only by l o g in

females, there would be a sex-treatment interaction.
Latin Square
A latin square is an experimental design which allows comparison of k
experimental treatments without confounding from two other sources of
variation in response, which are identified with rows and columns of a
k x k square in which each treatment occurs exactly once in each row or
column. Thus 25 animals in 5 treatment groups of 5 animals each may be
assigned to the following design to avoid possible effects of time (row)
and day (column) of injection.
Mon . Tue. Wed. Thurs. Fri.

10.30 am A B C D E
11.30 am D E B A C
12.30 am E C D B A
2.30 pm C D A E B
3.30 pm B A E C D
Regression
A regression is a relationship between a response variable, y , and one or
more predictor variables, x i . In simple linear regression the relationship
+
is of the form y = Po P,x where PO and PI are regression coefficients
(the intercept and the slope) to be estimated. In multiple linear
regression the regression equation can be written as y = Po PIxI + +
p2x2 + * * * + ppxp.
Ejgiciency
One estimator is regarded as more efficient than another if it has smaller
variance.
Power
The power of a statistical test is the probability that it rejects the
hypothesis that there is no effect of treatment where a true effect of
treatment exists.
Sensitivity and Specificity

Consider a 2 x 2 table relating the number of times an event actually
occurs or not to the number of times the recording device claims it
occurs.
412 Statistics
0bserved
situation
Yes No
True Yes a b
situation No C d
+
The sensitivity of the recording device is measured by a / ( a b ) , the
specificity by d / ( c + d ) . High sensitivity implies a low false negative
rate, that is the device (or test) is sensitive to what is being measured.
High specificity implies a low false positive rate, that is the device is
specific for the effect being measured.
The question of experimental design of toxicological studies is considered
in detail in Chapter 4, but in view of the importance of statistical
consideration in experimental design some of the main issues involved
will be discussed briefly here. For convenience, the principles are
illustrated with reference to the design of a long term carcinogenicity
study. Sub-sections A-F examine decisions mainly involved with mini-
mising the role of chance while sub-sections G-J discuss decisions mainly
related to avoidance of bias.
A Choice of Species
While maximising the ‘signal’ means avoiding a species where the
response of interest is very rare, the use of an over-responsive species can
bring penalties also. Thus, to have a 50% chance of achieving statistical
significance at the 95% confidence level in an experiment in which the
treated group has a 5% incidence of the lesion and the control group a
0% incidence, two groups of about 80 animals are required. To achieve
the same level of statistical significance where the incidences are 55% and
50%, respectively, group sizes some 10 times larger are required.
Further, it is not certain that an increased incidence of a lesion that is
common spontaneously in the animal species used (such as pituitary
tumours in Wistar rats) provides biological evidence of an effect which
can be extrapolated to other species.
B Dose Levels
The problem of dose selection is one of the most controversial and
important elements in the development of a protocol for a toxicology
bioassay. On biological grounds it would be ideal to test only at dose
levels comparable to those to which humans are exposed. On statistical
and economic grounds this is not usually practicable because the effect will
Peter N . Lee 41 3
be too small to detect without very large numbers of animals. To avoid

the possibility of an effect which would occur in a small proportion of
millions of exposed humans being missed in a study of hundreds or even
thousands of animals, the solution usually employed is to test in animals at
a dose level many times greater than maximum human exposure. Then,
assuming that any effect that exists is dose related, the demonstration of a
non-significant increase in response at a high dose level, though not
providing evidence of absolute safety (an impossible goal), can give
reasonable grounds for believing that any effects that occur at a dose
level very much lower will be at most very small.
A particular problem with this procedure is deciding the dose level. In
long term carcinogenicity studies the dose should clearly be one which is
not so great that the animals die from toxic effects before they have a
chance to get cancer. Based on these principles, the International Agency
for Research on Cancer (1980) has recommended that the high dose to be
used should be one expected, on the basis of an adequate sub-chronic
study, to produce some toxicity when administered for the duration of
the study but which should not induce the following:
(a) overt toxicity, i.e. appreciable death of cells or organ dysfunction,
as determined by appropriate methods,
(b) toxic manifestations which are predicted materially to reduce the
lifespan of the animals except as a result of neoplastic develop-
ment, or
(c) 10% or greater retardation of body weight gain compared with
control animals.
If the substance seems completely non-toxic, the high dose may
represent about 5% of the diet or even more for substances such as food
ingredients.
For a number of reasons it is important to have more than one dose
level. One is to cater for the possibility that a misjudgement has occurred
and the highest dose proves toxic. A second purpose is to cater for the
possibility that the metabolic pathways at the high dose may differ from
those at lower doses. A third is that the whole point of the study may be
to obtain dose-response information. Finally, it may be necessary to
ensure that no large effect occurs at dose levels in the range to be used by
man.
C Number of Animals
The number of animals to be used is clearly an important determinant of
the precision of the findings. The calculation of the appropriate number
depends on the following:
(a) the critical difference, i.e. the size of the effect to be detected,
(b) the false positive rate, i.e. the probability of an effect being
detected when none exists (known also as the ‘type I error’ or the
‘a-level’), and
414 Statistics
(c) the false negative rate, i.e. the probability of no effect being
detected when one of exactly the critical size exists (known as the
‘type I1 error’ or the ‘p-level7).
Reducing any of these increases the number of animals required.
The method of calculation of the number depends on the experimental
design and the type of statistical analysis envisaged. Tables are available
for a number of standard situations. To give an idea how the numbers
depend on the critical difference and on the a- and p-levels, Tables 1 and
2 give examples for two common situations, both of which relate to an
experiment where there is a control and a treated group. In the first there
is a normally distributed variable and the critical difference is expressed
in terms of the number of standard deviations (6) by which the treated
group differs from the control group. Thus, given a control response
known from past experience to have a mean value of 50 units with a
standard deviation of 20 units, one would need 2 groups of 36 animals to
have a 90% chance ( p = 0.10) of detecting a difference in response of 10
units (6 = 10/20 = 0.5) at the 95% confidence level ( a = 0.05).
In the second situation two proportions are compared. Here the
numbers of animals depend not only on the ratio of proportions but also
on the assumed proportion in the controls. Thus, where the control
response is expected to be l o % , the number of animals required in each
group to detect an increase response by a factor of 1.5 ( r ) is 920,
assuming again an a-level of 0.05 and a p-level of 0.1, whereas, if the
control response is expected to be 50%, the number required would be
79 per group.
For more complex situations the advice of a professional statistician
should be sought, though a general rule is that, to increase precision (i.e.
decrease the size of the critical difference) by a factor n, the number of
animals required will have to be increased by a factor of about n2.
Where there are a number of treatments to be tested in an experiment,
each to be compared with a single, untreated control group, it is
advisable that more animals be included in the control group than in each
of the treated groups, since the precision of the control results is relatively
more important, A useful rule in many situations is to have approxi-
Table 1 Number of animals required in each of a control and a treated

group in order to have a probability (1 - p ) of picking up a difference of o
standard deviations as significant at the 100 (1 - a ) % confidence level for
a normally distributed variable
Single-sided test a = 0.005 a = 0.025 a = 0.05
Double-sided test a = 0.01 a = 0.05 a = 0.01
P= 0.01 0.1 0.5 0.01 0.1 0.5 0.01 0.1 0.5
0 = 0.5 100 63 30 76 44 18 65 36 13
0.75 47 30 16 35 21 9 30 17 7
1.0 28 19 10 21 13 6 18 11 5
1.5 15 11 7 11 7 9 6
2.0 1 0 8 5 7 5 6
Peter N . Lee 415
Table 2 Number of animals required in each of a control and a treated

group in order to have a probability (1 - p ) of picking up a proportional
increase by a factor r as signijkant at the I00 (1 - a()% confidence level for
a binomially distributed variable
Single-sided test a = 0.005 a = 0.025 a = 0.05
Double-sided test a = 0.01 a = 0.05 ff =0.1
P= 0.01 0.1 0.5 0.01 0.1 0.5 0.01 0.1 0.5
Control level = 10%
r = 1.25 7679 4754 2120 5871 3358 1228 5039 2737 865
1.5 2103 1302 581 1608 920 337 1380 750 237
2.0 613 380 170 469 268 98 403 219 69
Control level = 20%
r = 1.25 3353 2076 926 2563 1466 536 2200 1195 378
1.5 902 558 249 689 395 145 592 322 102
2.0 253 157 70 193 111 41 166 90 29
Control level = 50%
r = 1.25 757 469 209 579 331 122 497 270 86
1.5 181 112 50 138 79 29 119 65 21
2.0 37 23 10 28 16 6 24 13 5
-
mately v k times as many animals in the control group as in each of the k
treated groups.
One point that is frequently misunderstood by experimentalists relates
to the number of animals required in studies where more than one
treatment is investigated in a crossed design. If, for example, one is
comparing compounds A and B, each dissolved in two different solvents,
in a 2 x 2 design with 4 groups, calculations of sample size to gain an
overall verdict on the difference between the two compounds should
generally be based on the overall numbers of animals on each compound
for both solvents combined unless there is reason to expect
compound/solvent interaction, i.e. that the compound A/compound B
difference depends on which solvent is used. Conversely, if it has been
decided that 2 groups of 100 animals are sufficient for attaining a given
level of precision concerning the difference in effects of a treatment,
additional information can be obtained on another factor (or factors) of
interest, without requiring any additional animals.
D Duration of the Experiment

Duration of the experiment can also affect the sensitivity of the tests
markedly. This is particularly so in long term carcinogenicity studies
where the great majority of cancers are seen in the latter half of an
animal’s lifetime. While studies should therefore not be terminated too
early, it is also important not to go on too long. This is because the last
few weeks or months may produce relatively little extra data at a
disproportionate cost, and diseases of extreme old age may be of little
interest in themselves but render it more difficult to detect tumours and
other conditions that are of interest. Where the study is of the prevalence
of an age-related, non-lethal condition that ultimately occurs in all or
416 Statistics
nearly all of the animals, early termination is required. In this situation

the greatest sensitivity is obtained when the average prevalence is around
50%.
E Accuracy of Determinations
Accuracy of observation is clearly important in minimising error. The
advent of Good Laboratory Practice and Quality Control Units has done
much to improve the quality regarding observations, but the quality
depends still on interested and diligent personnel.
F Stratification
To detect a treatment difference with accuracy, the groups being
compared should be as homogeneous as possible with respect to other
known causes of the response of interest. For example, a set of animals
thought to be homogeneous, but which in fact consisted of two
genetically different sub-strains, might provide the following measure-
ments of body weight in groups of 10 treated and 10 control animals, the
underlined readings relating to the first of the two sub-strains:
Control: 181 192 217 290 321 292 307 347 276 256
Treated: 222
- 249 - 232 - 284 - 270 - 215 - 265 - 378 328 391
Ignoring sub-strain increases the variability of the data so that more
controls are required to detect a treatment effect and, if unequal numbers
of each sub-strain are present in each group, may bias the comparison. In
the example given the means are as follows:
Sub-strain 1 Sub-strain 2 Total
Control 196.7 298.4 267.9
Treated 248.1 365.7 283.4
Diflerence 51.5 67.2 15.5
Although in each sub-strain the treatment results in an increase in body
weight by over 50 units, the greater number of the lower weight strain in
the treated groups results in the difference observed in the totals being
much less.
There are methods to take account of the sub-strain as a ‘stratifying
variable’ at the analysis stage. As described further in Section 4A, these
involve carrying out separate analyses at each level of the variable
considered and combining the results for an overall conclusion about the
treatment effect. This does not, however, preclude the possibility that the
proportions of each sub-strain in each group are so different that the data
provide substantially less comparative information than might otherwise
be achieved. In the extreme case where all control animals are sub-strain
1 and all treated animals sub-strain 2, the experiment would be worthless
to determine whether differences were due to treatment or sub-strain. To
Peter N . Lee 41 7
obviate this possibility, sub-strain can be used as a ‘blocking factor’ in the

design. In this case animals in each sub-strain are allocated equally to
control and treated groups. Although this removes bias, it is still
necessary to treat strain as a stratifying variable in the analysis.
Where more than one known factor affects the response, all can be
taken into account simultaneously. At the design stage, or retrospectively
in the analysis, the results are treated at each combination of levels of the
factors. Thus, to block for sub-strain, sex, and room where 3 experimen-
tal rooms were needed to house the animals, 12 mini-experiments, one
for each of the 2 (sub-strains) x 2 (sexes) x 3 (rooms) combinations,
would be set up.
G Randomisation
Random allocation, or randomisation, of animals to treatment groups is
an essential of good experimental design. If not carried out, it cannot be
ascertained whether a difference between groups is a result of differences
in the treatment applied or is due to some other relevant factor. It is a
fundamental on which the statistical methodology is based that the
probability of a particular response occurring is equal for each animal,
regardless of group. The ability to randomise easily is one great
advantage that animal experiments have over epidemiology.
The process of randomisation eliminates bias, so that the statistical
analysis is concerned only with assessing the probability of an observed
difference happening by chance. The smaller the probability is, the more
it suggests a true treatment effect. The procedure used for randomisation
should genuinely ensure that all possible assignments of animals to
treatment groups are equally probable. Such equal probabilities are best
achieved with pseudo-random numbers as found in tables or produced by
computer, it being difficult to ensure that apparently random devices such
as dice or playing cards really are random. Randomisation should never
be based on a system of testing animals haphazardly as they come and
assigning them to successive treatment groups. Not only do humans find
it virtually impossible to generate random sequences unaided, but it is
well known that the first animals selected may differ markedly from the
last, who are more active and avoid being caught.
In many experimental situations it is adequate to allocate randomly all
the animals to treatment groups, but in some the technique of stratified
random sampling is to be preferred. In this technique the animals are
divided into sub-groups (‘strata’) according to known factors believed to
be strongly related to the response, and random allocation to treatment
groups is carried out within each stratum. Sex is normally treated as a
stratifying variable. In a large experiment where animals are delivered in
batches, a batch could also be treated in this way, each batch forming a
smaller experiment, the results of which can be combined in the analysis.
This discussion on randomisation and stratification has been concerned
418 Statistics
primarily with the allocation of animals to treatment groups. The same

principles apply to anything that can affect the recorded response. Thus,
in a two-group experiment, measurements of some biochemical para-
meter should not be made for the first group in the morning and for the
second group in the afternoon. While the major part of such potential
bias can be averted fairly easily by various simple procedures, such as
doing alternate measurements on treated and control animals, ran-
domisation is preferable. Although many different procedures throughout
an experiment (feeding, weighing, observation, clinical chemistry, patho-
logy) require consideration in this way, the same random number can
usually be applied to all the procedures. Thus, if the cage position of the
animals is randomly allocated and does not depend on treatment, the
animals can always be handled in the same cage sequence.
H Adequacy of Control Group

The principle of comparing like with like implies that the control groups
should be randomly allocated from the same source as the treatment
groups. While historical control data can occasionally be of value in
interpretation of findings in treated animals, there is so much evidence of
quite large systematic differences in response between apparently identi-
cal untreated control groups tested at different times that it is often
impossible to be sure whether a difference seen between a treated group
and a historic group is really due to treatment at all.
It is also essential to be sure that the treated group differs from the
control group only in respect of the treatment of interest. Thus, having a
treated group where animals are exposed to cigarette smoke in a smoking
machine and an untreated control group does not allow evaluation of the
effect of cigarette smoke itself. The stress of being placed in a smoking
machine regardless of whether smoke or air is passed through to the
animal may have a quite marked effect on response. The appropriate
control group to use in this case would be one in which animals are
placed in a machine for the same length of time as the treated animals
but are not exposed to smoke.
I Animal Placement
The general underlying requirement to avoid systematic differences
between groups other than their treatment also demands that attention
be given to the question of animal placement. If all treated animals
are placed on the highest racks or at one end of the room, differ-
ences in heating, lighting, or ventilation may produce effects which
are erroneously attributed to treatment. Such systematic differences
should be avoided, and in many cases randomisation of cage posi-
tions is desirable. It may not be possible in some circumstances, for
example in studies involving volatile agents where cross-contamination
can occur.
Peter N . Lee 419
An example from my personal files highlights the reality of the dangers

of ignoring such problems. In a carcinogenicity study on albino rats the
incidence of severe retinal atrophy among animals housed in cages
located on the top shelf was 100% whilst among rats located on the
bottom shelf it was only about 5%, with intermediate incidences seen in
the central shelves. It was fortunate in this study that all groups were
represented on each shelf. A design and analysis that ignored animal
placement could easily have produced a false-positive effect of treatment.
J Data Recording
The application of the principle of comparing like with like means the
avoidance of systematic bias in data-recording practices. Two distinctly
different types of bias can occur. The first is a systematic shift in the
standard of measurement with time, coupled with a tendency for the time
of measurements to vary from treatment to treatment. The second is that
awareness of the treatment may affect the values recorded by the
measurer, consciously or subconsciously. The second bias is circumvented
by the animal’s treatment not being known to the measurer, i.e. the
readings being carried out ‘blind’. Although not always practical (that an
animal is treated may be obvious from its appearance), it is recom-
mended that laboratories organise their data-recording practices so that,
at least for subjective measurements, the observations are made ‘blind’.
The problem of avoidance of bias due to differences in time of
observation is a particularly important one in histopathological assess-
ment, especially for the recording of lesions of a graded severity, and in
large experiments where the slides may take the pathologist more than a
year to read. Where more than one pathologist reads the slides, there
should be discussion between them as to standardisation of terminology
and data to be recorded, and each should read a random or a stratified
sub-set of the slides to avoid bias.
4 Statistical Analysis
A Introduction
In the simplest situation animals are randomly assigned to a treated or a
control group and one observation is made for each animal, the objective
of the statistical analysis being to determine whether the distribution of
responses in the treated group differs from that in the control group.
There are five characteristics of this situation which require discussion
before detailed methods are described, as all should be clearly under-
stood before an appropriate technique of analysis can be selected.
Univariate or Multivariate Data?

The analyses described in Sections B to D are appropriate for univariate
data, i.e. where one analysis concerns a single response measure per
420 Statistics
animal. Methods for multivariate data, involving more than one response
measure per animal, are discussed more briefly in Section E, which also
covers some other methods not considered elsewhere.
Experimental and Observational Units

In the simple example cited the animal is both the ‘experimental unit’ and
the ‘observational unit’. This is not always so. In the case of feeding
studies, the cage, rather than the animal, is usually the experimental unit
in that it is the cage, rather than the animal, that is assigned to the
treatment. In the case of histopathology, observations are often made
from multiple sections per animal and here the section rather than the
animal is the observational unit. For the purposes of determining
treatment effects by the methods described below it is important that
each experimental unit provides only one item of data for analysis, as the
methods are all based on the assumption that individual data items are
statistically independent. If there are multiple observations per ex-
perimental unit, these observations should be combined in some suitable
way into an overall observation for that experimental unit before
analysis. Thus, in an experiment in which 20 animals were assigned into 2
treatment groups of 10 animals and in which measurements of the weight
of both kidneys were individually made, it would be wrong to carry out
an analysis in which the 20 weights in group 1 were compared with the 20
weights in group 2 since the individual kidney weights are not independ-
ent observations. A valid method would be to carry out an analysis
comparing the 10 average kidney weights in group 1 with the 10 average
kidney weights in group 2.
Type of Response Variable

Responses measured in toxicological studies can normally be classified as
being one of the following three types:
(a) Presencelabsence, in which a condition either occurs or it does not.
(b) Ranked, in which a condition may be present in various degrees.
For example, severity may be classed as minimal, slight, moderate,
severe or very severe.
(c) Continuous, in which a condition may take any value, at least
within a given range.
Each type of response demands a different sort of statistical technique,
described in Sections B to D respectively. It should be noted that by
using the methods of Section B it is possible to analyse ranked or
continuous data by defining values above a given cut-off point as present
and those below as absent. In general, this is not to be recommended; it
is wasteful of information and is less precise. Continuous data may also
be analysed by the methods described in section C for ranked data. While
again this may be somewhat wasteful, it may have advantages in avoiding
Peter N . Lee 42 1
possibly unjustified assumptions about the form of statistical distribution

underlying the response variable. This is explained more fully in Section
D.
Number of Groups Being Compared

Sections B to D each start with methods appropriate for the comparison
of two groups. They then go on to consider methods appropriate for the
comparison of k (more than two) groups. While only one comparison is
possible between two groups, more than one comparison is possible
between k groups. Two particularly important methods which are
described for each type of response are the test for heterogeneity and the
test for trend. The test for heterogeneity determines whether, taken as a
whole, there is significant evidence of departure from the (null) hypothe-
sis that the groups do not differ in their effect. The test for trend is only
applicable in experiments where the groups receive different doses of the
same substance (or have some other natural ordering). It determines
whether there is a tendency for a response to rise in relation to the dose
of the test substance applied. It can be a much more sensitive method of
picking up a true effect of treatment than an overall test for heterogeneity
or the comparison of individual treated groups with the control.
Graphically, the test for trend can be seen as determining whether a
sloped straight line through the dose-response (x-and y-axis, respec-
tively) relationship fits the data significantly better than a horizontal
straight line. That the trend statistic is significantly positive does not
necessarily imply that the treatment increases response at all dose levels,
though it is a particularly good test if the true situation is a linear
non-threshold model. In a true situation in which treatment increases
response at higher dose levels but has no effect at lower levels, it is easy
to construct data (see Section B) where only the trend test is able to
detect any significant effect.
Stratification
In the simplest situation all the animals in the experiment differ
systematically only in respect of the treatment applied. It is often the
case, however, that there are a number of sets of animals, each of which
differs systematically only in respect of treatment, but where the
characteristics of the sets differ. The commonest situation relates to male
and female animals, but there are many other possibilities, such as
different treatment rooms, different secondary treatments or even
different conditions under which the response variable was measured.
While it is often useful to look within each set of animals, or ‘stratum’, as
it is often called, to determine the effects of treatment in each situation, it
is also often useful to determine whether, based on the data from all the
strata, an effect of treatment can be seen overall. In some situations a
422 Statistics
relatively small number of animals in each stratum can make it difficult to

pick up an effect of treatment as significant within individual strata. The
essence of stratification lies in making comparisons within strata and then
accumulating treatment differences over strata. As shown in Section B,
the incorrect alternative of combining the data over strata and then
making a single analysis can lead to erroneous conclusions. Methods for
analysing stratified data are described in Sections B to D for the different
types of response data.
In the sections that follow, three levels of statistical detail are given.
Where the method is likely to be of frequent application, both the
mathematics and an example are given. Where the method is of less
frequent application or is more lengthy to perform, only the mathematics
are given. Where the method is rarer still, or likely to need the assistance of
a statistical expert, only a brief verbal description with references is given.
B Presence/Absence Data
UnstratGed analysis for 2 groups
The data can be expressed in the form of a 2 x 2 table as follows:
Treated
Control Total
Number with condition present a b n1
Number with condition absent C d n0
Total number at risk m, m0 N
A test of whether the proportion with the condition in the treated
group ( a / m l )differs from that in the control group ( b / m o )is given by the
corrected chi-squared statistic: (Breslow and Day, 1980, p. 131).
(lad - bcl - $ N ) 2 ( N- 1)
non1mom1
Referring this statistic to tables of percentiles of the chi-squared
distribution (see Beyer, 1966) with one degree of freedom gives an
give
Treated Control Total

Tumour 22 15 37
No tumour 62 59 121
Total 84 74 158
I
[22(59) - 15(62) - :(158)12157 - o.47
x 2 = -
121(37)74(84)
The chi-squared table value for one degree of freedom is 3.84 at
the 0.05 probability level. There is therefore no significant association
between liver tumour incidence and treatment (at p = 0.05).
Peter N . Lee 423
The corrected form of the chi-squared statistic yields a closer ap-

proximation to the proper p-value than that given by the uncorrected
chi-squared statistic in which the -$N in the above formula is not
included (Mantel and Greenhouse, 1968). Where the sample is small the
corrected chi-squared statistic yields probabilities which are poor ap-
proximations to reality, and Fisher’s exact test should be used instead.
Armitage (1971) notes that approximations to significance levels of about
0.05 are reasonably good, provided that the ‘expected’ numbers for all
four cells in the 2 X 2 tables are at least 5. [Here and subsequently the
expected numbers are calculated by multiplying the proportion present
(or absent) in the total sample by the total number examined in the group
in question. Thus, the expected number of controls with the condition
present is given by rn,n,/N.]
Given that there is no relationsip between treatment and the condition,
and given the marginal totals m,,mo, nl, and no, the exact probability of
an observed 2 x 2 table occurring is given by the formula:
m,!m,! n,! no!
P(a, 6, c, d ) =
a! b !c !d ! N !
This formula can then be used to calculate the exact probability of a
result the same as or more extreme than that actually observed (see
Example 2) In the past exact calculations were only feasible with small
2 x 2 tables but modern high speed computers now allow them to be
conducted in all cases.
EXAMPLE 2
In the same study the incidence of brain tumours was as follows:
Treated Control Total
Tumour 5 1 6
No tumour 79 73 152
Total 84 74 158
The exact probability for this table is given by:
84! 74! 6! 152!
P(5, 1, 79, 73) =
5! l! 79! 73! 158!
84(83)82(81)80(74)6
which by cancellation = = 0.1 164
158(157)156(155)154(153)
Given the marginal totals, the only table showing a more extreme
difference is the one where all 6 tumours occurred in the treated
group. The exact probability for this table is given by:
P(6, 0,78, 74) = 0.0207
Thus the overall (one-tailed) probability of observing a difference as
large as or larger than that seen, given no true effect of treatment, is
0.1164 + 0.0207 = 0.1371.
Note that the calculation of the second probability, 0.0207, would
not normally be necessary in this situation as the first probability,
0.1164 was already larger than needed for significance.
Had a corrected chi-squared statistic been calculated, a value of
1.19 would have been obtained, with a corresponding probability of
0.2758, equivalent to a very similar one-tailed probability of 0.1379.
424 Statistics
UnstratiJiedAnalysis for k Groups

The data can be expressed in the form of a 2 x k table as follows:
Group
1 2 k Total
Present a1 a2 ... ai ... ak fll
Absent c1 c2 ... ci ... ck n0
At risk m, m2 ... mi ... mk N
Dose applied dl d2 ... di ... dk
An overall test of heterogeneity is given by
where ei is the expected number with the condition in group j

(= m j n l / N ) , the associated probability being given by reference to tables
of the chi-squared distribution on k - 1 degrees of freedom (Armitage,
1971).
A test for trend is given by:
which should be referred to tables of chi-squared on one degree of

freedom (Armitage, 1955; Mantel, 1963) (see Example 3).
EXAMPLE 3
In a short term toxicity study the numbers of rats showing evidence of
pigment deposition in the kidney were as follows:
0 p.p.m. 10 p.p.m. 30 p.p.m. 100p.p.m. Total
Deposition 5 8 10 11 34
No deposition 15 12 9 5 41
Total 20 20 19 16 75
Given no relationship between treatment and pigment deposition,
the expected numbers with deposition and the difference between
observed and expected numbers are as follows:
Expected 9.067 9.067 8.613 7.253 34
Observed - expected -4.067 - 1.067 + 1.387 +3.747 0
An overall test of heterogeneity is given by

(-1.067)2+ (1.387)2
20 19 16
which equals 7.42 on three degrees of freedom. From tables the
associated probability is between 0.05 and 0.10 so that there is some
weak evidence of between-treatment variation.
Peter N . Lee 425
I T o test for trend we note that:
ck
j=l
dj(uj - e j ) = lo(-1.067) + 30(1.387) + lOO(3.747) = 405.64
k
2 d:mj = 102(20)+ 302(19)+ 1002(16)= 179,100
j= 1
k
dimj = lO(20) + 30(19) + lOO(16) = 2370
j=l
so that
7S2(74)405.642
= 6.29
x2 =41(34)[75(179, 100) - (2370)2]
The two-tailed probability for this single degree of freedom chi
squared is 0.01, so the trend statistic gives much stronger evidence of
an effect of treatment than does the test for heterogeneity.
Where expected values are low this x2 may be poorly approximated by

a chi-squared statistic. In this case it may be useful to carry out an exact
trend test noting that, given there is no relationship between treatment
and condition, and given the marginal totals m l , m2, . . . , mk, n l , and
no, the exact probability of an observed 2 x k table occurring is given by
the formula:
The formula can then be used to calculate the exact probability of a result
the same as or more extreme than that actually observed. The problem
with this test is not only that the computing effort can become large, even
with tables with small numbers, but also that the definition of what is ‘the
same or more extreme’ may be difficult. One method is to combine the
2 x k tables with the given marginal totals which yield values of
Cf=, d,(a, - e,) equal to or greater than the observed table. The method is
illustrated in Example 4. It is interesting to note that with these data a
significant trend is seen, although none of the individual differences
between any treatment group and control are significant.
EXAMPLE 4
In a small short term study the numbers of dogs showing evidence of
congestion in the lung were as follows:
Omg 1mg 3mg 5mg 10mg Total
Congestion 0 0 0 1 3 4
No congestion 4 4 4 3 1 16
Total 4 4 4 4 4 20
Given the number of cases of congestion that were seen, the exact
probability of observing the table actually seen is given by:
4! 4! 4! 4! 4! 16! 4!
P=
4! 4! 4! 3! 3! 20!
426 Statistics
(omitting terms of O! and l ! which equal 1). Cancellation yields:
The only other table which could have demonstrated a trend that
was at least as strong is the one in which all 4 cases occurred in the
top dose group. This has a probability
p=--16! 4! - 0.0002
20!
The overall one-tailed exact probability for trend is therefore
0.0035, which gives quite strong support to the belief that a true
treatment effect occurred. It is noteworthy that the simple com-
parison of top dose with control only gives a one-tailed exact
probability of 0.07, which would not normally be considered sig-
nificant. The trend test is a much more sensitive test as it takes into
account all the data and knowledge about the rank ordering of the
doses applied.
Stratified Analysis
The basis of stratification methods lies in accumulating over the strata
both the number observed to have the condition and that expected.
Summed over all treatment groups, the total observed and total expected
numbers are mathematically constrained to be the same, so that an excess
of observed over expected numbers in a particular treatment group
indicates a positive effect in that group, whilst a deficiency indicates the
effect is below average.
In the two-group situation the test of whether the treated group differs
from the control group is given by referring:
x’ = (10- El - 3)’
ZV
to tables of chi squared with one degree of freedom. This is the analogue
of the corrected chi-squared statistic for unstratified data. In this formula
0 (the total observed) = Ca and E (the total expected) = C rn,n,/N,
(summation without subscripts being over strata). V is the variance of the
expected number present in the treated group in a stratum and is given
by
nQnlmOrnl
V=
N’(N - I)
Example 5 illustrates the method and also shows how totally mislead-
ing conclusions can be gained by ignoring the stratifying variable.
EXAMPLE 5
In a long-term carcinogenicity study the data on numbers of rats
showing evidence of testicular atrophy (an age-related change which
does not cause death) can be laid out as follows:
Peter N . Lee 427
Variance
Period of Treated Control of
deathlweeks a c m, b d m, Expected expected
1-52 0 4 4 0 10 10 0.0000 0.0000
53-78 1 8 9 2 36 38 0.5745 0.4443
79-104 1 11 12 5 29 34 1.5652 1.0284
Terminal kill 23 52 75 5 13 18 22.5806 3.0878
Total 25 75 100 12 88 100 24.7203 4.5604
For period 53-78 weeks the expected (number of treated rats with
atrophy) is given by 9(1+ 2)/(9 + 38) = 0.5745 while its variance is
given by 9(38)3(44)/(47)246 = 0.4443. Similar calculations for each
period allow completion of the right-hand columns of the table:
((25- 24.72031 - 0.5)*
x2 = = 0.01
4.5604
There is thus no evidence of an association between testicular
atrophy and treatment, after adjusting for period of death. Had an
unstratified analysis been carried out on the totals, a corrected
chi-squared statistic value of 4.75 which is significant at the 95%
confidence level would have been obtained. This would have re-
flected the fact that survival was vastly better in the treated group so
that more rats reached termination where the frequency of atrophy
(about 30% in each group) was much higher than in rats dying
earlier.
In the k-group situation an overall test for trend is given by referring:
to tables of chi squared with one degree of freedom. In this formula, for
each stratum: k
C
T = dj(aj - ej)
j= 1
and its variance
An overall test of homogeneity involves matrix manipulations and will

be beyond the less mathematical reader. It is described by Breslow and
Day (1980).
Age Adjustment
One important application of stratified analysis for presence/absence data
relates to the use of age as a stratifier. For many conditions, such as
tumour incidence, frequency increases markedly with age (and con-
comitantly with length of exposure to the agent), and the overall
frequency in a treatment group can depend as much on the proportion of
animals surviving a long time as on the actual ability of the treatment to
428 Statistics
cause the condition. A simple, unstratified analysis ignoring differences in

survival between treatments can therefore give a false impression as to
the relative effect of treatment on the condition. Stratifying for time at
death in the context of long term carcinogenicity studies is discussed at
length by Pet0 et al. (1980). They point out that the way time of death
should be used as a stratifying variable depends on the circumstances in
which the tumour is seen. There are three different situations:
(a) Visible tumours. When the time of onset (ideally to some defined
size) can be observed exactly, the experimental time period is
divided into small time periods (usually of a week). Then within
each period the ‘number at risk’ (referring back to the terminology
used above) is defined by the number of animals alive and
tumour-free at the beginning of the period while the ‘number with
condition present’ is defined by the number for which a tumour
was first seen during the period.
(b) Fatal turnours. Other tumours are only observable at necropsy, and
in order to be able to analyse them properly one has to be able to
distinguish between those which (directly or indirectly) caused the
death of the animal and those which were seen only because the
animal died of an unrelated cause. In the case of the former group,
‘fatal tumours’, the analysis is similar to that for visible tumours,
with the ‘number at risk’ being the number alive at the beginning
of the period and the ‘number with condition present’ being the
number dying because of the tumour in the period.
(c) Incidental turnours. For tumours only observable because the
animal died of an unrelated cause, the methodology is somewhat
different because one does not know whether living animals have
or do not have a tumour at any given point. Here the ‘total
number at risk’ is defined as the number dying in the period, while
the ‘number with condition present’ is the number dying with the
tumour in the period. In order to maintain adequate numbers
examined in each time period, longer intervals must be used,
perhaps 20 weeks early in a 2-year rat study, reducing to 8 weeks
later. Interim or terminal kills are treated as special time periods in
this analysis.
While there can be problems in defining individual tumours as fatal or
incidental, it is important to realise that the conclusions can, on occasion,
depend critically on the definitions used, as is clearly demonstrated by
Pet0 et al. (1980).
C Ranked Data
Unstrat@ed analysis
An appropriate method to use when comparing observations in two or
more groups suitable for analysis by ranking methods is the Kruskal-
Peter N . Lee 429
Wallis one-way analysis of variance. This test is equivalent in the

two-group situation to the Mann-Whitney U-test or the Wilcoxon rank
sum test. In this method the N observations in all the groups combined
are ranked in a single series from 1 to N. Observations which are tied are
given the average of the ranks, so that if there are four equal lowest
observations each scores (1 + 2 + 3 + 4)/4 = 2.5. For each group the sum
of the ranks R j is then computed. From this a statistic H is calculated as
follows:
1L
H=
N(N + 1 ) C (R:/N,) - 3(N + 1)
j=1
where Nj is the number of observations in group j .

Where, as is usually the case, there are tied observations a correction
factor C is computed as follows:
2 (t3- t )
C=l-
N3-N
where t is the total number of observations in all groups combined with
the same score and summation is over the number of scores where such
ties occur.
The statistic H / C , or simply H if there are no (or just a few) ties, can
be referred to tables of chi squared on k - 1 degrees of freedom,
provided that the sample sizes are sufficiently large. Where sample sizes
are small, reference may be made to tables giving the exact probabilities
which are given in Iman et al. (1975).
As noted by Conover (1980), there should no longer be any hesitation
to apply this rank test to situations that have many ties, the Kruskal-
Wallis test being an excellent test to use in a contingency table, where the
data consists of numbers of animals in each group with one of a limited
range of values, e.g. when observations are scored on a five-point scale.
A measure of response in a group can be calculated by the difference,
Dj, between the scaled rank sum in that group, 2Rj/N, and its expected
+
value, Nj(N l ) / N . Using this, a one degree of freedom chi-squared
statistic for trend can be computed from the formula:
x2= T2/var T
where the trend T is given by:
k
T= djDj
j=l
and the variance is given by:

430 Statistics
Stratified Analysis
For some reason little attention has been given to the use of rank tests in
the analysis of stratified data. However, there seems little reason why
methods analogous to those for presence/absence data cannot be carried
out. Thus a stratified test for trend can be calculated by accumulating T
and var T over strata and referring:
x2 = (C T ) ~ / Cvar T
to tables of chi squared with one degree of freedom. In the two-group
situation also, noting that within a stratum:
var Di = C(N + 1)(N - N,)N,/3N2
one can refer:
x2= (C Di)'/C var Dj
to tables of chi squared with one degree of freedom as an overall test of

difference between treated and control groups.
As for presence/absence data, a stratified test for homogeneity
involves matrix manipulation and is beyond the scope of this chapter.
EXAMPLE 6
Times to death (in hours) in a rat study of three dose levels of an
antidote to a given dose of toxin might be recorded as follows:
3 mg kg-' 2 mg kg-' 1 mg kg-'
18,48,22,5,14,48 10,2,1,14,14,48 12,1,4,4,3,8
To analyse these data by the Kruskal-Wallis one-way analysis of
variance the times are first replaced by their ranks yielding:
3 mg kg-' 2 mg kg-' 1 mg kg-'
14,17,15,7,12,17 9 , 3 , l i , 12,12,17 10, l i , 5 $ , 5 $ , 4 , 8
Totalling these gives rank totals of 82, 54i, and 34;. As a check,
the overall total of ranks should be computed to see that it equals
+
N ( N 1)/2 = 18(19)/2 = 171, which it does. H is then calculated as
follows:
H=-
12 [(82)2 (54i)2 (34:)2
+- +
-1
- 3(19) = 6.65
18(19) 6 6 6
Noting that there are 2 rats with a tied time of 1 hour, 2 with a tied
time of 4 hours, 3 with a tied time of 14 hours, and 3 with a tied time
of 48 hours, the correction factor C is calculated:
C=l-
(23- 2) + (23- 2 ) + (33 - 3) + (33- 3) = 0.9897
183- 18
An overall test for heterogeneity is therefore given by referring
H / C = 6.72 to tables of chi squared on 2 degrees of freedom, which is
significant at the 95% confidence level.
Peter N. Lee 43 I
Testing for trend:

2x82
D1=---- 6 x 1 4 - 2.7778, D2 = -0.2778, D, = -2.50
18 18
T = (3 X 2.7778) + (2 X -0.2778) + (1 X -2.50) = 5.2778
var T = 0*9897 l9 x [18[(32 x 6) + (22 x 6) + (1*x 6)] - (18 + 12 + 6)2]
3 x 18 x 18
= 4.1787
x2 = (5.2778)2/4. 1787 = 6.666
which, on 1 degree of freedom, is significant at the 99% confidence
level.
Continuous Data
Assumptions
Whereas analysis of presence/absence or ranked data makes no assump-
tion concerning the frequency distribution of the underlying data,
analysis of continuous data does of necessity involve some assumptions.
It is not the intention here to discuss the various sorts of distribution
mentioned in the literature. Suffice it to say much biological data can be
analysed by methods which assume that the observations in each group
are normally distributed with a common variance. The methods outlined
below are all based on these assumptions but are ‘robust’ in the sense
that they are not highly contingent on minor departures from them.
Outliers
One common situation where the assumptions are not met is where there
are ‘outliers’, i.e. one or two values in the data which are very different
from the rest. While these have relatively little effect on the results for
presence/absence or ranked data analysis, a single outlier can dramati-
cally distort the conclusions when continuous data are analysed, not only
by affecting the mean value of the group in which the outlier occurs but
also by increasing the overall estimate of the variability of the data. This
makes it important to screen the data with the possibility of outliers in
mind. Choice of an appropriate formal test for outliers is a difficult
subject to which Barnett and Lewis (1978) have devoted a whole book.
Of the tests they discuss, that based on the sample kurtosis is one of the
most useful, and the interested reader should refer to the book for the
methodology.
Homogeneity of Variance
The second common situation where the assumptions are not met relates
to the situation where the larger the value, the more inherently variable it
432 Statistics
tends to be. Thus animals in one group, with a mean response of 100,
may have values varying in the range 80-120, while animals in another
group, with a mean response of 300, may have values varying over the
wider range of 240-360. Where the standard deviation of the data is
approximately directly proportional to the mean, it is appropriate to take
logarithms of the individual data items before carrying out the analysis.
This not only makes the variability independent of the mean but also
tends to ensure that the distribution of logx more closely resembles the
symmetrical bell-shaped normal distribution than does the distribution of
x . Where the variability increases with the mean but not so markedly, a
square-root rather than a logarithmic transformation may tend to be
better in making the variability more independent of the mean. A formal
test as to whether results in different groups are equally variable, often
referred to as a test of homogeneity of variance or of homoscedasticity, is
Bartlett’s test. Where , even after logarithmic or square-root transfor-
mation, there is still evidence of marked heterogeneity of variance, it will
often be preferable to use methods for ranked data as described in
Section 4C. However, it should be noted that it is in general preferable to
use the same method of analysis when analysing similar data at different
times, e.g. routine blood chemistry measurements. This is partly to make
comparisons of different data sets easy and partly because whether a
variable, or its logarithm or square root, is distributed normally is to
some extent a property of an underlying biological process and therefore
likely to be inherent. If experience suggests a variable is generally
normally distributed, it is not justified to take logarithms just because on
one specific occasion it happens to ‘cure’ significant heterogeneity of
variance shown up in Bartlett’s test. Remember that some significant
results are expected by chance and that the analysis of variance
procedures to be described below are robust against slight departures
from the assumptions.
Bartlett ’s Test
Suppose we have k groups of animals and within each group of animals
there are Ni observations xii, the subscript i referring to group and j to
observations within group.
In order to test for homogeneity of variance, the procedure is as
follows. For each group calculate the sum of the observations:
c
?I =
i
xij
the sum of squares of the observations:

si c = x?.
i
and hence the within-group variance:
02 =
NJi - (z)2
Ni(Ni - 1)
Peter N . Lee 433
Defining the degrees of freedom in each group by dJ = Ni-1, the

sta tist ic:
can then be looked up in tables of the chi-squared distribution on k - 1

degrees of freedom. Instead of using logarithms to base e (natural
logarithms), logarithms to base 10 may also be used in the above
formula, the resultant statistic then being multiplied by 2.306 (see
Example 7).
EXAMPLE 7
In a study of 18 rats, body weights at one point in time were recorded
as follows:
Group 1 ( N = 5 ) 124, 161, 143, 182, 202
Group 2 ( N = 6) 182, 304, 194, 262, 135, 188
Group 3 ( N = 7 ) 221, 204, 197, 283, 230, 250, 270
From these data calculate:
T s, of dLd
Group 1 4 812 135,674 951.30 3805.20
Group 2 5 1265 285,389 3736.97 18684.83
Group 3 6 1655 397,655 1060.95 6365.71
Total 15 28855.74
(dL In of)= 110.363
I
7(&) = + + $ = 0.61667
substituting into the formula:
= -
3.0672
x 2 = 2.81
1 + 0.09167
1+ (0.61667 - 1/15)
Referring this to tables of chi squared on 2 degrees of free-

dom, it can be seen that this is not significant even at the 90%
confidence level, and the variances can therefore be accepted as
homogeneous.
One-way Analysis of Variance

Given that there is no significant heterogeneity of variance, one-way
analysis of variance can then be used to test whether there is evidence of
any difference between the group means. Using the same notation as
before, ?;, S j , and o? are calculated as for Bartlett's test and the group
434 Statistics
means are calculated by:

p i = 7;/Ni
Then, summing over groups calculate the total number of observations:
the sum of the observations:
the sum of squares of the observations:
and the sum of squares of the group total divided by the number of
observations:
G= (T;/Ni)
1
Then calculate the ‘correction for the mean’:

c=T ~ I N
and form the analysis of variance table as follows:
Source of Degrees of Sum of Mean square
variation freedom squares
Between groups k-1 G-C (G - C ) / ( k - 1)
Within groups N-k S-G ( S - G ) / ( N- k )
The ratio (between-groups mean square)/(within-groups mean square)
is then distributed on the null hypothesis as an F-statistic on k - 1 and
N - k degrees of freedom and the associated probability can be looked
up using standard statistical tables. When there are only two groups
( k - 1 = 1 ) the square root of this statistic can also be looked up in tables
of Student’s t-statistic on N - k degrees of freedom-this is equivalent to
the well known ‘t-test’.
When there are data from more than two groups, but results from two
particular groups ( i l and i2) are to be compared, the t-test can be used.
involving data from these groups only. In general, however, it is better to
use the following formula:
t = MDWD
where MD is the difference between the two means and V D , the variance
of the difference, is given by:
as this uses information about variability of the data from other groups.
Peter N . Lee 43 5
Given that dose levels, d j , are applied in each group, and given an
observed difference, Dj, between the mean in each group and the overall
mean, an approximate one degree of freedom Student’s t-statistic testing
for trend can be calculated by the formula:
k
T = djN,Dj
j= 1
and its variance is given by:
Alternatively, linear regression analysis (see, for example, Johnson and

Leone, 1964) may be used to test for a linear relationship between dose
and response. See Example 8 for an example of a one-way analysis of
variance with test for trend.
EXAMPLE 8
To compare percentage body weight changes over a defined period in
a study in 4 groups of dogs, the following calculations would be
carried out. First, based on the individual data items, calculate:
Control 50 p.p,m. 100 p.p.m. 150 p.p.m.
21.2 19.5 17.4 13.4
22.8 23.1 19.0 16.5
23.5 22.0 20.2 18.1
20.7 18.4 16.5 12.3
18.9 13.4
Ni 4 4 5 5
T, 88.2 83.0 92.0 73.7
pi 22.05 20.75 18.40 14.74
and summing over groups: N = 18, T = 336.9
and hence the grand mean: p = 18.717
and the correction factor: C = (336.9)2/18 = 6305.645
(88.2)2 (83.0)2 (92. 0)2 --
G=
4
+ 4 + 5 + (73’5 7)2- 6446.198
Next, based on the individual data items squared, calculate
Control 50 p.p.m. 100 p.p.m. 150 p.p.m.
449.44 380.25 302.76 179.56
519.84 533.61 361.OO 272.25
552.25 484.00 408.04 327.61
428.49 338.56 272.25 151.29
357.21 179.56
Si 1950.02 1736.42 1701.26 1110.27
and hence, summing over groups, S = 6497.97
Source of Degrees of Sum of Mean
variation freedom squares square F
Between groups 3 140.553 46.851 2.67
Within groups 14 51.772 3.698
436 Statistics
Using tables of the F-distribution on 3 and 14 degrees of freedom

the value of 12.67 obtained can be seen to be significant at the 99.9%
confidence level. There is thus evidence of variation between the
group means.
An idea of where the variation comes from can be seen by
comparing each group mean with that for the controls, using the
overall estimate of within group variance (mean square). The
difference between high dose mean and the control is (14.74 -
22.05) = -7.31 and its standard error is given by [($+ $)3.698]4=
1-290 yielding a t-value of -5.67 significant at the 99.9% confidence
level, Thus a table can be set up:
Difference from control
Dose level Mean Standard error t
50 p.p.m. -1.30 1.360 -0.96
100 p.p.m. -3.65 1.290 -2.83
150 p.p.m. -7.31 1.290 -5.67
Alternatively a test for trend can be carried out. Taking dJ as
dose/50 for convenience, calculate:
Group
Control
N,
4
dJ
0
D,
+3.333 0
dfv
0
‘JqDl
0
30 p.p.m. 4 1 +2.033 4 4 +8.133
100 p.p.m. 5 2 -0.317 10 20 -3.167
150 p.p.m. 5 3 -3.977 15 45 -59.650
Total 29 69 -54.683
T = -54.683
3.698
var T = -[(18 x 69) - (29 x 29)] = 82.383
18
x2 = T2/var T = 36.30 on 1 degree of freedom
which is again significant at the 99.9% confidence level.
Stratified Analysis
Though, when analysing continuous data that is stratified, it is possible to
devise global tests for homogeneity and trend in a manner similar to that
used for presence/absence and ranked data, i.e. combining results of
independent analyses carried out in each stratum, this technique is not
normally employed in this situation. Where, as is often the case,
variability of response is similar in different strata, it is more efficient,
especially where numbers of observations may be small in particular
strata, to use a method of analysis in which an overall estimate is made of
variability rather than one in which separate estimates are made with
strata. This brings us into the area of more complex analysis of variance,
or more generally into the area of multiple regression and linear models,
details of which are beyond the scope of this chapter but are discussed in
numerous statistical textbooks. These methods normally require exten-
sive calculations and the user will need the availability of appropriate
software such as that provided by the BMD statistical package or the
Peter N . Lee 437
General Linear Interactive Modelling (GLIM) program of Baker and

Nelder (1978).
E Other Methods
Relationships Between T w o Variables
In the methods described in Sections 4B-4D interest has centred on
whether there is evidence of variation between groups in respect of a
single variable. When other variables have appeared, hese have only
been nuisance variables to be taken account of by stratif ;ation. In many
situations, however, interest is centred on relationships between two
variables. For a single group, relationships between two variables can be
investigated by a 2 x 2 chi-squared test or Fisher's exact test in the case of
presence/absence data. In the case of ranked data Spearman's rank
correlation coeficient can be used. This is calculated by:
where n animals are ranked separately in respect of two variables and Di

is the difference between the ranks for the ith animal. In the case of
continuous data the product-moment correlation coeficient can be used.
This is calculated by:
where xi and yi are the values of the two variables for the ith animal. For
both the Spearman rank and the product-moment correlation coefficient
observed values must lie between + 1 (perfect positive relationship) and
-1 (perfect negative relationship) with a value of 0 implying no
association between the variables. Associated probability values for given
values of r and n can be looked up in standard tables (e.g. Beyer, 1966).
Techniques are also available for determining whether these relation-
ships differ for different treatment groups. Breslow and Day (1980)
described methods appropriate for testing constancy of association over
multiple 2 X 2 tables, whilst analysis of covariance (Bennett and Franklin,
1954) can be used for continuous data.
Multivariate Methods
Methods in which the inter-relationship of a large number of variables
are studied will normally require the advice of a professional statistician
and are not considered here in detail. Such techniques are most
commonly applied to continuous data and include the following:
43 8 Statistics
(a) Multiple regression analysis. In this, variation in a dependent

variable, y , is related simultaneously to the values of a number of
known predictor variables, x k , and the significance of each is
assessed.
(b) Discriminant analysis. In this, an individual is classified into one of
g groups, depending on the values recorded of each of a number of
known variables, x k . The relative importance of the variables in
discriminating individuals can be assessed.
(c) Component analysis. In this, the major part of the variation in a
dependent variable? y , can be related to a reduced number of
variables, formed from the original set of known variables, x k .
This technique is closely related to factor analysis. Where a
number of variables are closely related, perhaps by having a
common underlying cause these techniques can simplify complex
data, though in some cases the results can be difficult to interpret.
Multiple Observations on the Same Animal

For body weight and clinical-chemistry data it is common to take
measurements on the same animals at regular intervals throughout a long
term study. Clearly, data at any individual time point can be considered
and a between-group comparison undertaken, but this only gives a
limited amount of information in any one analysis. Often the interest is in
how the measurements are changing rather than in their absolute level, as
in the case of weight gain. An intuitively obvious approach is to use
differences in measurements at two defined time points, but this is in fact
not the best method. The problem is that, for any variable measured with
some error, low values measured at one time point will tend to increase
and high values decrease simply because of the ‘regression to the mean’
effect, even when no true change occurs. It follows that if there is
group-to-group variation between the values measured at the first time
point, estimation of between group differences in the change that has
occurred owing to treatment will be contaminated by this effect. The
correct technique to use is analysis of covariance where differences in
mean values at the second time point after adjusting for the values
observed at the first time point are sought.
In some circumstances a single analysis involving the data measured at
all the time points may be worth attempting. If, for example, the growth
curve for an animal can be expressed in terms of an equation depending
on a limited number of parameters, analyses comparing the groups in
respect of these parameters can be informative. However, such tech-
niques tend to be complex.
Pa ired Data
It is not uncommon for an experiment to consist of a number of pairs of
animals, one animal of each pair being a control and one a treated
Peter N . Lee 439
animal. Thus, to minimise the effects of between-litter variation, it can be

advantageous to treat pairs of animals from the same litter in this way.
Where the criterion on which pairing is carried out can affect the
response, it is important to use appropriate statistical methods that take
the pairing into account.
For presence/absence data the appropriate method to use is
McNemar’s test, in which the number of pairs in which the event occurs
in the test animal and not in the control animal ( N , ) is compared with the
number of pairs in which the event occurs in the control animal and not
in the test animal (N2), the statistic:
being approximately distributed as chi squared on 1 degree of freedom.

The same formula can also be used for continuous or ranked data where
Nl is the number of pairs in which the observation was greater in the test
than in the control animal and N2 the reverse. The exact probability
associated with the observation can also be calculated, when the test is
known as the sign test. For numbers of pairs up to 24, this is available in
tabular form (e.g. Siegel, 1956).
This test uses information only about the direction of the differences
within pairs. A technique which uses information about the relative
magnitude as well as the direction of the difference is the Wilcoxon
matched-pair signed-ranks test. In this method the difference in value
between each pair is calculated (di) and these differences are then ranked
regardless of sign (ri). Then to each rank is affixed the sign of the
difference to form ri. Finally, the sums of the positive and the negative
ranks are formed, which under the null hypothesis are equal to
N ( N + 1)/4, the average of the ranks ( N being the total number of
non-zero differences). The smaller of these, T, is then either tested by
taking:
x 2 =
+
24[ T - N(N 1)/412
+
N(N 1)(2N 1) +
as an approximate chi-squared distribution or, in the case of less than 25
pairs, by looking up the exact probability in tables (e.g. Siegel, 1956).
Alternatively, when the data are normally distributed, the matched-
pair t-test can be used. Here the differences are used as raw data and
their mean ( p ) and variance (a2)calculated as in Section 4D. p / a is then
looked up in tables of the t-distribution.
Pro bit Analysis

The objective of some toxicological experiments is to calculate the dose
of a substance at which 50% of the animals will respond. This is known as
the LDS0 when the response is death (LD =lethal dose) and, more
generally, as the median-effective dose. Typical data consist of results at
440 Statistics
a range of dose levels in which the number of animals exposed and the
number responding are given. A single plot of dose against percentage
response typically gives a relationship in the form of an elongated S. By
plotting the logarithm of dose on the x-axis and the ‘probit’ of response
on the y-axis, as can be easily done with special graph paper, this
relationship usually becomes approximately linear and it is reasonably
easy to get a rough estimate of the median-effective dose. The probit is
defined by Y +5, where Y is the normal equivalent deviate which is
related to the proportion responding ( P ) by the formula:
Techniques for formal estimation of the median-effective dose and its

95% confidence limits are laborious by hand. The reader is referred to
Finney (1977) or for a more rapid, reasonably accurate procedure to Gad
and Weil (1982).
Some Further Reading

This chapter has not given a detailed discussion of all the types of
observation that the toxicologist may have to deal with and the
appropriate methods for each. The interested reader may refer to Gad
and Weil (1982) for some useful material, covering inter aka body and
organ weights, clinical chemistry, haematology , reproduction, teratology ,
dominant lethal assay, and mutagenesis. For further details on methods
for analysis of presence/absence data, Breslow and Day (1980) is useful,
while both Siege1 (1956) and Conover (1980) are valuable for the analysis
of non-parametric data including rank methods. Many books are avail-
able on analysis of continuous data; two of these, Johnson and Leone
(1964) and Bennett and Franklin (1954), though both rather old, are very
clear. All these references contain numerous worked examples.
References
Armitage, P. (1955). Test for linear trend in proportions and frequencies.
Biostatistics, 11, 375-386.
Armitage, P. (1971). ‘Statistical Methods in Medical Research’. Blackwell
Scientific Publications, Oxford.
Baker, R. J. and Nelder, J. A . (1978). ‘Generalised Linear Interactive Modell-
ing’. Release 3. Numerical Algorithms Group, Oxford.
Barnett, V. and Lewis, T. (1978). ‘Outliers in Statistical Data’. John Wiley and
Sons, New York.
Bennett, C. A. and Franklin, N . L. (1954). ‘Statistical Analysis in Chemistry and
the Chemical Industry’. John Wiley and Sons, New York.
Beyer, W. H. ( E d . ) (1966). ‘CRC Handbook of Tables for Probability and
Statistics’. The Chemical Rubber Company, Cleveland, Ohio.
Peter N . Lee 44 1
Breslow, N. E. and Day, N. E. (1980). ‘Statistical Methods in Cancer Research.

Volume 1-The Analysis of Case-Control Studies’. IARC Scient. Publ. No.
32. International Agency for Research on Cancer, Lyon.
Conover, W. J. (1980). ‘Practical Nonparametric Statistics’. 2nd Edn. John Wiley
and Sons, New York.
Feron, V. J. et al. (1980). Basic requirements for long-term assays for
carcinogenicity. In : ‘Long-term and Short-term Screening Assays for Car-
cinogens: A Critical Appraisal’. IARC Monographs on the Evaluation of the
Carcinogenic Risk of Chemicals to Humans. Suppl. 2. International Agency
for Research on Cancer, Lyon.
Finney, D. J. (1977). ‘Probit Analysis’. 3rd Edn. Cambridge University Press,
Cambridge, England.
Gad, S. C. and Weil, C. S. (1982). Statistics for toxicologists. In: ‘Principles and
Methods of Toxicology’. Ed. A. W. Hayes. Raven Press, New York.
Iman, R. L., Quade, D., and Alexander, A. (1975). Exact probability levels for
the Kruskal-Wallis test statistic. Selected Tables in Mathematical Statktics, 3,
329-384 (5.2 Appendix).
Johnson, N. L. and Leone, F. C. (1964). ‘Statistics and Experimental Design in
Engineering and the Physical Sciences’. Vol. 1. John Wiley and Sons, New
York.
Mantel, N. (1963). Chi-squared tests with one degree of freedom: extensions of
the Mantel-Haenszel procedure. J . A m . Stat. Assoc., 58, 690-700.
Mantel, N. and Greenhouse, S. W. (1968). What is the continuity correction?
A m . Stat., 22, 27-30.
Marriott, F. H. C. (1990). ‘A Dictionary of Statistical Terms’. 5th Edn. Longman
Scientific and Technical, Harlow.
Peto, R. et al. (1980). Guidelines for simple, sensitive significance tests for
carcinogenic effects in long term animal experiments. In: ‘Long-term and
Short-term Screening Assays for Carcinogens: A Critical Appraisal’. IARC
Monographs on the Evaluation of the Carcinogenic Risks of Chemicals to
Humans. Suppl. 2. International Agency for Research on Cancer, Lyon.
Siegel, S. (1956). ‘Nonparametric Statistics for the Behavioural Sciences’.
McGraw-Hill Book Co., New York.
CHAPTER 19
Risk Assessment of
Chemicals
DAVID P. LOVELL
1 Introduction
The basic objective of experimental toxicology is the identification and
understanding of the properties of chemicals which are detrimental to the
human population. These effects may be seen directly as adverse
consequences on health, and indirectly through damage to economically
important crops, livestock, the wider environment, and ecosystems. Such
an objective is, of course, ambitious. Chemicals have a spectrum of
effects, some considered beneficial to the human population, others
deleterious. For instance, a pesticide may kill mosquitoes and reduce the
burden of malaria but may leave residues which some consider unwel-.
come contaminants of food; or a drug may protect against heart disease
but may cause damage to the patient’s kidneys or eyes.
The procedures for determining the potential benefits and side effects
of exposure to chemicals in the environment form the basis of risk
assessment.
A What is Risk Assessment?

Unfortunately for a young discipline like risk assessment, the word ‘risk’
has a number of unspecific meanings. A more precise definition is
necessary otherwise concepts such as risk estimation become nebulous.
A clear distinction is made between the words ‘risk’ and ‘hazard’. This
is not so in everyday speech where they are synonymous with
chance, danger, peril, and related to concepts like harm.
The definitions below derive from those developed by a Royal Society
Study Group (1983) and have subsequently been used by Love11 (1986)
and Richardson (1986) in reviews of the Toxic Hazards of Chemicals.
442
443
David P. Love11
B Definitions of Hazard, Risk, and Risk Assessment
A hazard is the circumstance where a particular incident could lead to
harm, where harm is death, injury, or loss to an individual or society.
A more specific definition relevant to chemicals is the set of inherent
properties of a chemical substance or mixture which makes it capable of
causing adverse effects in man or the environment when a particular
degree of exposure occurs.
Risk is the chance that a particular adverse effect actually occurs in a
particular time period or after a specific challenge.
Relating this to chemicals: risk is the predicted or actual frequency of
occurrence of an adverse effect of a chemical substance or mixture from a
given exposure to humans or the environment.
Some examples should clarify the distinction. Ionising radiation is a
known hazard to the human population causing harm in the form of cell
death, genetic damage, and cancer in exposed individuals. The risk of
such injuries actually occurring depends upon factors such as the level of
exposure and individual susceptibility. Similarly, from epidemiological
investigations, medical case reports, and experiments using volunteers,
many chemicals are known to be toxic [for instance, TCDD
(dioxin) ,methylisocyanate, and vinyl chloride] in various ways and
degrees.
The risk that individuals will suffer toxic effects from exposure to such
chemicals depends upon a range of variables. Hazard is therefore the
potential for harm, risk the chance that harm will actually occur.
Determining the chance, or probability, that harm will occur and the
subsequent consequences form the basis of Risk Assessment.
Risk assessment is defined as the process of decision making applied to
problems where there are a variety of possible outcomes and it is
uncertain which event will happen. Risk assessment can be broken down
into stages (Table 1). The problem has to be defined and the different
risks estimated. These risks are evaluated in relation to other factors or
interests involved and choices are made between the possible course of
action. The total process of hazard identification, risk estimation, risk
evaluation, decision making, and implementation makes up what is called
risk assessment.
This definition is not the only one that has been developed. Confusion
can arise because not all definitions use the same names or categories.
However, the general ideas and principles are similar. For instance, the
Table 1 The stages of risk

assessment
Hazard Identification
Risk Estimation
Risk Evaluation
Deciding on Acceptable Risk
Managing the Risk
444 Risk Assessment of Chemicals
US Environmental Protection Agency (EPA) have produced guidelines

for risk assessment for carcinogenicity and mutagenicity. They make a
clear demarcation between what they call risk assessment and risk
management. The former consists of hazard identification, dose-response
assessment, exposure assessment and risk characterisation; it defines the
adverse health consequence using scientific methods. Risk management
combines technical, socioeconomic, political, and other considerations in
a decision making process (EPA, 1986).
This chapter will mainly concentrate on the methods available to the
toxicologist to assist in risk estimation. The determination of toxic effects
and identification of hazards is dealt with in other chapters in this book,
while the more general aspects of evaluation, decision making and risk
management are reviewed elsewhere (Lovell, 1986).
2 Estimating the Risks

There are four broad areas from which toxicological data can be used for
estimating risk: human studies, theoretical considerations, short term
testing, and animal testing.
A Human Studies
Measures of the consequences of exposure of the human population to
chemicals provide the most direct information available for risk estima-
tion. Such exposures may be unintentional, such as a result of an
industrial accident; deliberate, as in the case of a volunteer study; or a
consequence of incidental exposure during the course of normal activity.
Risk estimates are based upon a calibration of the observed ill-health
associated with an estimate of the exposure to the chemical to provide a
dose-response relationship. Such studies are limited for chemicals but a
major ‘model’ for the type of investigation is the prediction of the effects
of radiation from various accidental or deliberate exposures. A major
problem in studies of chemicals is obtaining reliable estimates of the
exposures received under non-experimental conditons.
Estimates of the risks associated with exposure to a chemical can be
drawn up based upon the results of these studies (sometimes in
conjunction with the results of animal studies). They are then used to
help manage the risks by the setting of occupational exposure levels such
as ‘Threshold Limit Values’ (TLV).
These levels are often based upon the assumption that the effects of
long term, low level exposure can be estimated from the results of short
term, high concentration exposure. The main limitation with human
studies are that they are, in general, retrospective as exposure has
already occurred. Epidemiological studies of possible carcinogenic effects
are particularly limited by the low statistical power of the designs, which
David P . Lowell 445
make it difficult to detect any effects other than increases in rare

diseases, the long term nature of the studies, and the consequences of the
sub-group being exposed to a chemical before the risks associated with it
can be evaluated.
B Theoretical Approaches and Short Term Tests

There is an increasing interest in the use of surrogates of either human
exposure or animal studies to try to identify toxicological hazards. The
results of such methods could then be used in the risk assessment process.
Theoretical approaches include molecular modelling and quantitative
structure-activity relationships (QSAR). Compounds with specific chem-
ical structures which may pose a hazard can either be considered to need
specific further testing or rejected as possible therapeutic or industrial
substances.
Similarly, short term tests for toxicological activity such as potential
genotoxicity or teratogenicity are used to try to predict potential hazards
to the human population. The results of these tests also determine
whether future testing is necessary, or whether the tested product will be
developed further.
Theoretical approaches and short term tests are mainly of value in
hazard assessment because of the difficulties of obtaining quantitative
measures of risk applicable to man. This limits the replacement of animal
studies by short term in vitro testing, which may be desirable on ethical
grounds. Consequently, short tests are mainly used to supplement the
results of animal and human studies.
C Animal Studies
Most new chemicals likely to be introduced into the environment have no
relevant data on human exposure for assessing risks. Risk estimation is
therefore based upon animal studies. These are also used to augment the
results from studies in humans of chemicals already present in the
environment.
Animal studies are characterised by an attempt to identify compound-
related toxic effects following the administration of high doses of the
compound to small numbers of animals. Such studies have two major
complications when used for risk assessment. Firstly, the difficulty of
estimating an effect at a low dose from the effects observed at high doses
(a complication also of studies based on accidental exposures of groups of
humans) and secondly, the transfer of conclusions from one species,
often a rodent, to another, man.
The choice of high doses and low numbers of animals is forced on
toxicologists because the obvious experiment of testing for effects at
doses comparable with likely human exposures would require very large
numbers of animals. Such experiments would be very expensive (the
standard US National Toxicology Program assay using 800 animals has
been costed at $1,000,000 at 1986 prices, Lave and Omenn, 1986). There
are also increasing ethical objections to using large numbers of animals
and such studies would probably be impractical to conduct. The largest
experiment of this type carried out was, the EDol (colloquially called the
‘megamouse’ experiment) study. This was a considerable strain on the
resources and facilities of the US National Center for Toxicological
Research.
Even with a design which has relatively large numbers of animals and.
high doses, the statistical power, or ability to detect successfully a true
effect, of the experiment is limited. Figure 1 illustrates the probability of
finding a condition induced by the test chemical at frequencies of
between 1 and 5% for various different sample sizes. Over a third of
samples of 100 animals would not include an animal with a particular
condition even when the induced incidence was 1%. This incidence is
considerably higher than many adverse drug reactions noted in patients
who had been administered pharmaceutical products.
0 10 20 30 40 50 60 70 80 90 100
Sample size ( n )
Figure 1 Graph showing the probability of one or more animals carrying an

abnormality (such as a tumour) being found in a group of animals for various
samples sizes (n) and background incidences. The lines show the probabilities
associated with background incidences of 1% (0), 2% (A),3% (O), 4% (0) and
5% (0). Note that with a sample size of 100 ouer a third (probability of ~ 0 . 3 3 of
)
samples will not contain a tumour-carrying animal when the background incidence
k 1%, while over a tenth will not contain a tumour-carrying animal when the
background incidence is 2%
David P. Love11 447
The use of high doses also has complications. The methods of

deactivation occurring at high doses may be different from those
operating at lower, more realistic dose levels while tissue distribution and
pharmacokinetics may also vary. Criticism is particularly strong of the
high dose levels used in the long term rodent carcinogenicity bioassay
where the highest dose is called the maximum tolerated dose (MTD).
This dose is based upon previous smaller experiments and is chosen so
that no lethality is produced and body weight in the treated group is not
reduced by more than 10% of the weight of the control animals. Critics
have argued that toxic effects, particularly tumours, may be artefacts
based upon chronic administration of the compound in high doses. This
results in metabolic overload and pathological changes resulting from
chronic repair processes.
A further complication of animal studies is that the route of ad-
ministration may be different from how the human population is exposed.
Dosing animals in a way that mimics human exposure is not always
practicable and much of the safety evaluation of chemicals is done using
either intraperitoneal or oral administration of the compound.
3 Approaches to Risk Estimation from Animal

Studies
Two general approaches have been developed to use animal data to
estimate the risk to human populations of a chemical.
A Safety Factor Approach

This derives from methods developed in the 1950s when an attempt was
made to provide a legislative framework for the control of food additives.
The Delaney amendment passed by Congress in the the USA, which
states that no substance carcinogenic to animals should be deliberately
added to human food, resulted in the application of safety factor methods
only for non-carcinogenic findings. The most widely used method is the
determination of an acceptable (sometimes referred to as allowable) daily
intake (ADI) of a compound. This is obtained by dividing a value called
the No Observable Effect Level (NOEL) by a safety factor (SF), i.e.
NOEL
ADI=-
SF
The NOEL is obtained from a study where groups of animals are exposed
to different doses of the compound. The NOEL is the highest dose at
which no observable effects are detected. The safety factor is tradition-
ally, but not necessarily, 100. This is supposed to reflect the possibility that
man may be 10-fold more sensitive than the test species and that there
may be a 10-fold range in susceptibility within the human population.
448 Risk Assessment of Chernicu1.r
The NOEL is not as objective a value as it might first appear as the

following two descriptions quoted by Crump (1986) show. The first is a
circular argument with NOEL defined ‘as the level at which no effects
were observed’, while the second, used by the US Environmental
Protection Agency (EPA), is ‘the exposure level at which there are no
statistically significant increases in frequency or severity of effects
between the exposed population and its appropriate control’. Defining
whether a statistically significant increase is a biologically important
finding is a subjective decision partly determined by the choice of
statistical test used. The NOEL could, for instance, be affected by the
sample sizes at each dose. Table 2 illustrates how smaller sample sizes
can produce apparently higher doses which are the NOEL. One
objection, therefore, has been that the investigator who uses insensitive
designs with too few animals is ‘rewarded’ with a higher safe dose.
In practice, the results of the statistical analysis of such tests is only one
input into the assessment of the results. Other biological and toxicologi-
cal considerations are considered in the evaluation. In this way the
decision on the NOEL is based upon expert but nevertheless subjective
decisions.
Use of a NOEL might seem to give support to the concept that there is
a threshold dose, below which a chemical will not produce toxic effects.
However, increasing the number of animals sampled at each dose and
improving the sensitivity of the methods for detecting toxicity may find
effects at lower doses. The choice of the NOEL is thus a reflection of
relative rather than absolute safety. Others have suggested it should be
called the no adverse effect level or that the lowest dose at which effects
occur should be used in conjunction with a larger safety factor.
The size of the safety factor is also debated. Proposals have been made
Table 2 Effect of varying sample size on the deter-

mination of a N O E L using a statistical test
Data Set 1
Control Low Medium High NOEL
0 3
- - -
A -
10 10
6
10 l”0 Low
NS ** ***
0 6 12 18
B -
20
-
20
-
20 20 None
* *** ***
Data Set 2
0 1 2
A -
10
-
10 10
- High
NS NS r;”S
B 0
-
20
-
20 -
20 &, Medium
NS NS *
Comparison between dose levels and control by one-sided Fisher
exact test (NS Not significant; * P < 0.05; **P< 0.01; ***P<
0.001).
Note that in both data sets results A are based upon sample sizes
of 10 and results B on sample sizes of 20. The proportion of
affected animals is the same in both A and B at each dose level
with each data set.
David P . Love11 449
for a factor of 10 to be used when chronic human exposure data are

available, or a factor of a 1000 when there is no good human data (Hogan
and Hoel, 1982). Weil (1972) has proposed that a safety factor of 5000
could be used when carcinogenic effects were found in the experiment.
Proponents of the safety factor approach argue that the method is
satisfactory despite the lack of a real biological justification for a
particular safety factor. They contend that the method has an added
safety factor because the true dose level at which toxic effects would
occur in practice may be considerably greater than the NOEL.
The approach incorporates both the estimation of low dose effects and
the inter-species comparison in a single stage. However, the failure to
take dose-response relationships into account, the arbitrary nature of the
safety factor, and its unsuitability for the risk assessment of carcinogens
have stimulated other approaches.
B Mathematical Modelling
The second approach has been to attempt to model mathematically the
dose-response relationship in toxicological data. The aim is to try to
provide quantitative risk estimates (this is confusingly called quantitative
risk assessment by some US researchers).
The methods used attempt to extrapolate from the results of high doses
to predict effects at low doses. (This should strictly be called interpolation
as there is often an estimate of the incidence of the condition in the
control group.)
The use of low-dose extrapolation is controversial. There are some
statisticians who believe that ‘mathematical models should be used to
project results only within the range of observed data’ (Salsburg, 1986).
Others have developed a series of models for use in low-dose extrapola-
tion. The problem has interested many leading biostatisticians and
resulted in a large number of scientific publications spread through a
range of journals. Such a large and disparate literature is impossible to
summarise, reference, or review adequately in a chapter such as this.
Much of the work depends upon a series of complex assumptions and is
argued using statistical concepts and nomenclature. This makes the
subject difficult for many toxicologists to follow. Worked examples of the
various methods are rare and some of the models require access to
specialised computer programs before they can be fitted to the data.
In this chapter an attempt will be made to determine the underlying
concepts behind some of the methods, briefly discuss in non-statistical
terms the more commonly encountered models, and then review the
performance of the various approaches.
Virtually Safe Dose ( V S D )

Mathematical modelling of carcinogenicity data arose out of the concerns
described previously about the safety factor approach, together with an
attempt to provide a rational alternative to the 'no safe dose of a

carcinogen' concept behind the Delaney amendment. Mantel and Bryan
(1961) proposed the use of a method of linear extrapolation based upon
carcinogenicity data such that the dose of carcinogen was identified which
caused a lo-* increase in the risk of developing cancer. Such an increased
risk was considered to be effectively zero and the dose causing the
increased risk was called the virtual safe dose (VSD). Mantel and Bryan's
method involved fitting the log-probit model developed for bioassay data
to the data on the dose-response results from the chemical. The upper
99% confidence limits were calculated for the different doses and lines
with a probit response of one drawn downwards (Figure 2). The most
2
m
a,
3 1
$ 0
4-
7
QJ
Maximum llkellhood regrpssion line
TY
6 4 normal deviates per log)
d -1
Extrapolated llne with slope
5
2 -2
-3
-4
-5
-5.612
1 1
-7 -6 -5 -4 -3 -2 -1 0 1
Log dose of methylcholanthrene ( m g mouse-')
Figure 2 Illustration of Mantel- Bryan method for low-dose extrapolation.
Observed data are fitted using maximum likelihood regression line. The most
conservative (i.e. gives lowest safe dose) extrapolated line with a slope of 1 normal
deviate per unit log dose is drawn from the 99% upper confidence levels (A)of the
observed proportion of tumour-carrying animals (e). The virtual safe dose
b calculated by reading off the dose which produces an increased tumour risk
of (lop8) or -5.162 normal deviates. This is 10p7."s ( - 9 x lo-') mg
methylcholanthrene per mouse (as shown in bottom left hand corner of graph).
The graph and calculations are modified from those presented by Mantel and
Bryan in their original 1961 paper. The arrows ?& indicate 100% and 0%
respectively of animals were carrying tumours
David P. Love11 45 1
conservative of these lines was then used to estimate the dose leading to a
lo-’ extra risk.
Although the concepts of different individual tolerances underlying the
probit model had some biological appeal, this method was shown not to
be as conservative as the authors had initially expected. The probit
method also proved to be a poor fit to many experimental data sets and
the method had to be modified to include cases when there was a
background incidence of the condition under study.
The level of acceptable risk chosen by Mantel and Bryan of was
also subsequently raised to lop6by the Food and Drug Administration,
based upon consideration of the types of risk readily accepted by the
population in general. The VSD which leads to a lifetime risk of death of
is equivalent to an effect which would cause about 3 extra deaths per
year in USA based upon an average life span of 73 years and a
population of 220 million or about one extra death per year in the United
Kingdom.
Linear Extrapolation Method

A somewhat similar approach for low dose extrapolation is to find an
upper confidence level for the lowest dose effect and then draw a straight
line back to the spontaneous or control incidence (Figure 3). Such an
approach has been popular because it is both simple and conservative.
There has been some criticism that it may be too conservative and also
does not reflect dose-response relationships found in multiple dose
studies.
The Range of Mathematical Models

The failure of the Mantel-Bryan log-probit method to fit some of the
observed sets of data stimulated other biomathematicians to try to find
functions which were better fits to the observed dose-response relation-
ships. The mathematical modelling has two main purposes. Firstly, to
describe the observed data by some equation relating the incidence to the
dose administered. Secondly, to try to estimate effects, using the
equation rather than the linear extrapolation method, at doses not tested,
particularly low doses. The types of models used fall into two categories:
those based upon an assumed distribution of tolerances such as the probit
used by Mantel and Bryan or the somewhat similar logit models, and the
mechanistic models which are supposedly simulating biological events
which lead to carcinogenicity. These include the one-hit, multi-hit, and
multistage models.
One-hit models. This is one of the easier models to use. In its simplest form
it assumes that cancer can result from a ‘single hit’, such as irreversible
in 100 c o n t r o l animals
0 1
l o p 6 x 50000
dose associated with a risk of is or 0.1843 p.p.m. (The
0.27 1256
calculations upon which this graph is based can be.found in Gad and Weil, 1986,
p . 209)
genetic damage or interaction with a receptor. Fitting the model does not
require sophisticated computing facilities.
A consequence of the model is that in the low dose region it is
approximately a simple linear model. This assumption of low dose
linearity results in extremely low dose or conservative estimates of the
virtual safe dose (VSD). In practical terms, using the one-hit model may
be the same as applying the Delaney amendment as the VSD will be
extremely small.
Although appealing in its relative simplicity, the one-hit model has
only a single parameter to model what may be a complex shape of
dose-response; the model, therefore, is often a poor fit to the data
compared with the other models. The 'one-hit' concept of the mechanism
of carcinogenicity may also be oversimple in view of developments in the
theory of carcinogenicity.
David P . Lovell 453
Multi-hit models. This is considered to be a more realistic mechanistic

model than the one-hit. It assumes that there are multiple hits (it is
sometimes called the Gamma multi-hit model) required at the cellular
levels to initiate carcinogenesis. Although this model has been used by a
number of regulatory bodies and recommended by the Scientific Com-
mittee of the Food Safety Council (1980), there are critics who have
questioned its use because of statistical problems and the production of
inconsistent results. They have suggested that its use may lead to
misleading conclusions in some circumstances.
Multistage models. This is another mechanistic model assuming that the
development of cancer is a multistage process. This makes it biologically
appealing but difficulties arise because the estimated parameters do not
have any direct biological interpretation. It, too, has been criticised
because assumptions implicit in its formulation may not actually apply in
general. Various variants and modifications of the model have been
developed but Park and Snee (1983) point to the potential of the model
to overestimate risk and give misleading results under certain
circumstances.
Nevertheless, the EPA (1986) suggest that in the absence of adequate
information to the contrary, the linearised multistage procedure should
be used. This is a variant of the multistage model which uses linear
extrapolation. Extrapolation is from an upper 95% confidence bound
calculated for a low dose down to the control levels (Crump, 1981). The
EPA will allow results using other models to be presented but a rationale
must be included to justify use of any model chosen.
Pharmacokinetic models. There is some concern that the mechanistic
models developed for carcinogenicity oversimplify the pharmacological
processing and metabolism of toxic compounds. Models have, therefore,
been developed to account for such factors as: the dose reaching the
target organ compared with that administered, thresholds, and inter-
species differences in response. Their practical use is limited by the
increased numbers of parameters needed to specify the models, which
mean they can only be applied to studies with many doses. Development
of models incorporating pharmacokinetic information is likely to
continue.
The Weibull Model. The Weibull model is a tolerance distribution
increasingly fitted to toxicological data. It is widely used in industry to
study product reliability, such as the life span of electronic components.
The Weibull model can be made more complex by the addition of extra
parameters or simplified by specifying fewer (it eventually becomes a
special case of the one-hit model). Models can be developed including
parameters to take into account the length of time before a tumour
develops or the length of administration of the treatment. Its use is
limited by needing sufficient numbers of treated groups to allow
parameters to be estimated. The use of the Weibull has increased in part

because it can incorporate information on the time up to when a tumour
develops. The inclusion of these data into the hypothesis testing
procedure for carcinogenic effects (Peto et al., 1980) was an important
advance in the analysis of rodent carcinogenicity bioassay studies.
The performance of diferent mathematical models. Fitting of models to
toxicological data is a complex and controversial subject with a volumi-
nous literature. Most toxicologists do not need to know the underlying
theoretical assumptions for the various methods. However, an apprecia-
tion of the practical results of applying various models is useful.
In general, many of the models fit the observed data satisfactorily, but
they differ in their estimates of lower dose effects such as the VSD or the
e
0 - 2
I
4 6 8
o/o Sodium saccharin in diet
(ii)
-7 -5 -3 -1 1
Dose of sodium saccharin
Figure 4 Graphs to illustrate: (i) fit of mathematical models to observed data on
the incidence of tumours in animals fed diet containing various percentages of
sodium saccharin (a) is one-hit model; (b) is Weibull; (ii) low dose extrapolation
of fitted lines for different models (a) one hit; (b) multi-stage; ( c ) Weibull; (d)
multi-hit. Both axes on logarithmic scale. (Graphs reproduced by permission from
Food Cosmet. Toxicol., 1980, 18, 725).
David P. Love11 455
extra risk. The Food Safety Council (1980) compared the results of 4
models on 14 different data sets. Three of the models, the
multistage, multi-hit, and Weibull, provided reasonable fits to the data
while the other, the one-hit model, has a poorer fit. Estimates of the
VSDs differ quite appreciably between models, with the one-hit, in
general, giving the smallest safe doses, the multistage, next lowest, then
the Weibull, followed by the multi-hit. Figure 4 illustrates both the
failure of the one-hit model to fit data for saccharin compared with the
Weibull, and the differences in low dose extrapolations for the four
models. Other studies have suggested that the probit and logit models
give higher VSDs than the multi-hit, they are, in other words, less
conservative.
Variability in the estimates of the results produced by the different
models has resulted in the models having proponents. Considerable
debate has ensued over their respective merits. In practice, the actual
choice of a model must depend on the professional judgement of the
investigator. The choice of method for low dose extrapolation must
always be made with an appreciation of the statistical limitations of any
such methods, as well as the limited biological applicability of any of the
models.
4 Interspecies Comparisons
The problem of extrapolating from the test species to the human
population remains, even if a satisfactory low dose extrapolation is
achieved. This species scale-up has to take into account possible
differences in distribution, metabolism, and excretion of a compound,
together with factors such as differences in size and life span.
The three most frequently used approaches to the interspecies scale-up
are: the conversion of results on the basis of the fraction of the diet, a
body weight basis, and differences in surface area.
A Fraction of Diet Scaling Factor

This converts intake in mg kg- (diet) per day or parts per million (p.p.m.)
or billion (p.p.b.) in the rat or mouse diet to the human intake on a similar
basis. The human diet is estimated to weigh between 600 and 700 g day-'
while that of the rat is 21 g day-' and the mouse 4 g day-'. The effect of
using mg kg-' per lifetime is to produce a risk which is 35 times greater as
it is based upon an average lifespan of 70 years for man and 2 years for
rodents.
B Scaling Based Upon Body Weight

The administered dose is expressed in terms of mg kg-' body weight of
the test animal. Human exposure is then considered on the same basis of
body weight. A conversion factor can be calculated which is a constant

based on the ratio of average body weight of a rat or mouse to the
average body weight of a human. The average body weights of a
mouse, rate, and human being are assumed to be 20 g, 400 g, and 70 kg,
respectively, giving a ratio of 3500 for the mouse and 175 for the rat.
Clearly not everybody (or every mouse) conforms to these average
weights.
C Scaling Based upon Surface Area

Dosage in terms of mg kg-’ body weight in the test animal is converted to
mg per square metre of body surface area. Human exposure is then also
expressed in terms of mg per square metre surface area.
Two alternative methods are used. The first calculates a term called the
surface factor as the cubed root of the mass of human (Mh) divided by the
mass of the test animal ( M J , e.g. (Mh/Mt)t’3.In the case of the rat this
factor is(x)
70000 ‘ I 3
(7)
= 5.59 and for mice it is
70000 113
= 15.18.
The second method is based upon calculating the surface area by using
the formula of the weight of the animal ( W ) to the power 2/3 multiplied
by a species constant ( K ) provided by Spector (1956). [Algebraically the
surface area is = K(W)2’3].This results in a man: rat surface area
ratio of 55.4 and for man: mouse, a ratio of 387.9.
D Choice of Method of Scaling

The choice of which of these methods should be used is a matter of
debate. Hogan and Hoe1 (1982) have concluded that estimates based
upon either mgkg-’ body weight day-’ or mgm-2 body surface area
day-’ seem to lead to better projections of risk in man from carcino-
genicity data than those based upon lifetime estimates. However, such
methods have to be considered somewhat crude and experience of their
use is limited. Considerable care is needed, therefore, in using such
species scale-ups and consideration needs to be taken of species
differences which could complicate such attempts.
The EPA (1986) suggest that extrapolation on the basis of surface area
is considered appropriate mainly because some pharmacological effects
appear to be scaled according to surface area.
Ames, Magaw, and Gold (1987) have taken an alternative approach for
interspecies comparison. They have used rodent carcinogenicity data to
try to rank possible carcinogenic hazards to the human population, rather
than to predict absolute human risks. They used an index called HERP
(the Human Exposure/Rodent Potency dose) to indicate the possible
David P. Love11 457
hazard. Human exposures were calculated for populations at risk and the
Rodent Potency dose calculated using a method developed by Sawyer et
al. (1984). They concluded that the carcinogenic hazards resulting from
pesticides or water pollution are likely to be minimal compared with
those from background levels of natural substances.
5 An Example of Risk Estimation in Practice

An example of the application of mathematical modelling, low dose
extrapolation and species scale-up is illustrated by the approach used by
Carlborg (1985) for estimating the risk associated with saccharin from the
results of a long term carcinogenicity bioassay in rats.
Carlborg (1985) estimated the excess risk of a bladder tumour resulting
from exposure to a dose of 0.01% saccharin fed in the solid food of rats
for over two years. He converted this dose in rats to its human equivalent
using a scaling based on body weights. The average rat body weight was
taken as 742g (see Table 3), and average daily intake of diet was 27.2g.
This resulted in an average daily intake per kg body weight for each rat
of 3.666 mg of saccharin.
Table 3 Steps used by Carlborg (1985) in estimating the risk of human

consumption of saccharin f r o m a long term rat feeding study
( 1 ) Estimated from the rat study
Average body weight of rat 742 g
Average daily intake of diet day ' rat -' 27.2 g
Estimated excess risk of bladder cancer in rats from addition of 0.01% saccharin
to diet based upon Weilbull model is 4.1 X lo-' to 2.5 x 10 '()
(2) Estimated human exposure based upon body weight

Average human body weight 70 kg
Average intake day-' rat-' of saccharin
on 0.01% of diet = 2.72mg day-' rat-'
Average intake day-' kg-' of rat = 2.72 X 1000
742
= 3.666mg day-' kg rat-'
Converting intake for a 70 kg man = 70 X 3.666 mg day-' man-'
= 256.6 mg day-' man-'
(3) Estimating human exposure

1 can (12fl. ozs) of 'diet-drink' contains
110.4 mg saccharin
256.6 mg saccharin is equivalent to 2.3 cans
The 90 percentile for consumption of diet drinks in the USA is 2.3 cans per
day. Therefore only 10% of the US population drink more than 2.3 cans
per day or consume more than 0.01% saccharin in their diet.
Converting this to a daily intake for a 70 kg man gives a total dose of

256.6mg per day for each person. A 12fl. oz can of ‘diet’ soft drink
contains 110.4mg of saccharin, so a daily intake of 256.6mg would be
equivalent to 2.3 such cans. This is about the 90 percentile for the intake
of saccharin within the USA. Carlborg used a series of mathematical
models to try to fit the observed bladder cancer incidence in a long term
rat carcinogenicity bioassay. The estimates of the risk obtained by
extrapolation using the Weibull model for the 0.01% dose of saccharin
were from 4.1 x to 2.5 x lo-’’ suggesting a very low and acceptable
risk from saccharin if directly applied to comparable consumption in the
human population.
Clearly such risk estimation involves a considerable number of
assumptions which may be wrong to some degree. One check on the
approach is to alter some of the estimates to see if changing the variables
in the equations could alter the conclusions. This is called a sensitivity
analysis.
It is important to realise that such estimates of risk are dependent upon
the quality of data included in the equations. This is illustrated by
another example.
6 An Example of Risk Estimation using Mouse

Dominant Skeletal Mutations to Estimate
Human Risk from Radiation
Dominant mutations in the offspring of individuals exposed to radiation
or chemical exposure are of special concern because their effects are seen
in this subsequent generation. Data from experiments on mice has been
used to try to estimate the genetic risk to the human population from
radiation. The model used was the production of dominant skeletal
mutations in the mouse (Selby, 1982).
Estimates of the mutation frequencies of serious disorders per million
livebirths for each 0.01 Gy (1Rad) of radiation in the human population
have been produced. The central estimate is about 20 per million with a
range from about 5 to 66 per million, depending upon various assump-
tions made (Table 4). This compares with a current or spontaneous
incidence of about 10,000 per million livebirths. The only direct
experimental data used in the estimate are Selby’s finding of 37 dominant
skeletal mutations out of 2646 offspring of exposed male mice. The
remaining 6 calculations are based upon corrections of the doses
administered to make them applicable for low dose estimates, assump-
tions about the relative frequency and severity of skeletal mutations and
the relative sensitivities of male and females to radiation. These represent
the expert considerations of the UNSCEAR and BEIR I11 committees
(UNSCEAR, 1977; BEIR 111, 1980).
A sensitivity analysis was carried out in effect, because ranges of
Table 4 Steps in the estimation of genetic risk in humans from induced

skeletal mutations in mice
mutations/l0' live births
(1) Dominant skeletal mutations after 1 Gy + 5 Gy

at 0.60 Gy min-' gamma radiation (37/2646) 13983.4
(2) Rate per 0.01 Gy (division by 600) 23.3056
(3) Dose-rate effect (division by 3) 7.7685
(4) Dose-fractionation effect (division by 1.9) 4.08871
(5) Total dominant mutations assuming skeletal
mutations are a tenth (10%) of all 40.887 1
dominant mutations
(multiplication by 10)
[range 5-15%] [20.4435 to 61.33061
(6) Proportion of mutations that are serious
handicaps 50% (divide by 2) 20.4435
[range 25-75%] [5.1109 to 45.99791
(7) Relative sensitivity of maternal to paternal 20.4435
exposure. Assume same.
[range: equal* risk to 44% greater than
risk in male] [5.1109 to 66.23701
* The BEIR Committee (1977) gave a negligible risk as their lower band to the risk of
protracted maternal irradiation
estimates were given for the proportion of dominant mutations that are
skeletal, the proportion that produce severe disabilities, and the
differential sensitivities of the two sexes. These give estimates of between
5 and 66 mutations per million livebirths.
These estimates are, of course, dependent upon the quality of the
experimental data used in the equations. This may be open to interpreta-
tion. In this case, of the 37 presumed mutations, 6 were sterile and 3
others only identified by their small size. Subsequent confirmation and
characterisation of the mutants was hindered by the loss of 13 of the
mutants when the stocks were transferred from West Germany to the
USA, while studies have shown that one of the strains used in these
experiments have at some time become genetically contaminated
(West, Peters, and Lyon, 1984).
These difficulties in the interpretation of the biological data do not
necessarily alter the conclusions of the risk estimation process ap-
preciably. However, they do highlight the need to evaluate critically the
assumptions made in such methods and the potential dangers from
over-interpreting data. Risk estimation is an important and useful
procedure but must remain closely linked to data derived from good
scientific experimentation.
7 Risk Evaluation and Management

The identification and estimation of the risks associated with particular
hazards is followed by the stage of risk evaluation. This involves deciding
on what levels of risk are acceptable and whether there are any offsetting
benefits. Risk management is the series of procedures for controlling risk
to this acceptable level. The steps of the risk estimation, evaluation, and
management overlap. A more detailed discussion of the issues surround-
ing these concepts can be found in Love11 (1986).
One method which tries to evaluate the relative benefits and risks of
actions is risk-benefit analysis. This is a special case of cost-benefit
analysis in which the relative costs of the advantageous and disadvan-
tageous outcomes are compared. Lave (1987) has identified two different
types of such analyses. One, he termed ‘risk-risk’, is when both
alternative actions have risks and benefits which have to be compared;
the other, called ‘how safe is safe enough’, results in increasing cost for
each successive reduction in risk. In the latter case a decision has to be
made as to whether the costs of extra safety are worth the sacrifice of
other desired actions. There are, of course, considerable difficulties in
such analyses in providing monetary values to injury, illness, or death.
Lave and Omenn (1986) used a cost-benefit approach in an assessment
Carcinogen Non-carcinogen
True positive False positive

Prediction from Carcinogen
in vitro tests 900 4900
False negative True negative
Non-carcinogen
100 44100
1000 49000
(2%) (98%)
Carcinogen Non-carcinogen
True positive False positive
Prediction from Carcinogen
in vitro tests 4500 4500
False negative True negative
Non-carcinogen
500 40500
5000 45000
(10%) (90%)
David P. Love11 46 1
of the effectiveness of short term testing of chemicals for the prediction of

carcinogenicity.
They illustrated the case (Table 5) of a series of 50,000 new chemicals
of which either 2% or 10% were carcinogenic. They estimated the
relative costs of incorrectly predicting the carcinogenicity of the chemical
using in vitro short term tests. They assumed the cost of testing each
chemical in the battery of tests was $8,000 and the cost of accidentally
allowing a carcinogen (false negative) into the environment was $10 m,
while the cost of unnecessary controls on a non-carcinogen incorrectly
called a carcinogen (false positive) was $1m. The balance sheets
comparing the costs of testing against no testing, shown in Table 6, give
net benefits of $3.7 bn if 2% of chemicals are carcinogens and $40.1 bn if
10% are carcinogens for short term testing.
Obviously such conclusions can be dramatically altered by changing the
assumptions (such as that the short term tests have 90% accuracy) or the
costs ascribed to different outcomes. However, although such approaches
can be criticised, particularly when dealing with non-monetary outcomes,
they do provide a framework for exploring the implications of various
strategies.
Table 6 Balance sheet of social costs of short term testing for car-
cinogenicity compared with no testing
Percentage chemicals that
are carcinogens
2% 10%
Cost of testing 50,000 chemicals

at $8,000 per chemical 0.4 0.4
Cost to society of false positives
($1 m per chemical) 4.9 4.5
Cost to society of false negatives
($10 m per chemical) 1.o 5.0
Total cost to society of test strategy 6.3 9.9
Cost to society of not testing

($10 m per carcinogen) 10.0 50.0
Net saving to society by test strategy 3.7 40.1
(Tables 5 and 6 based upon Lave and Omenn, 1986)

8 Conclusions
Risk assessment is an imprecise procedure despite the statistical and
mathematical aspects incorporated in it. The use of mathematical models
and derivation of various estimates of risks from the data must not be
used to provide a spurious precision or to create an appearance of
scientific sophistication. Instead? such approaches can provide a frame-
work to help and illustrate the implications of various evaluations and
decision processes.
The multitude of conflicting interests of the producers, regulators, and
consumers of chemicals is too complex a pattern of behaviour to be
reduced to simple mathematical models. Individuals may not always act
in a manner which, based upon an assessment, appears to be in their best
interests. However, actions and ideas which may appear irrational have
to be accommodated within the risk assessment processes. Risk assess-
ment must always reflect the changing pattern of society’s perception of
its ideas and values.
References
Ames, B. N., Magaw, R., and Gold, L. S. (1987). Ranking possible carcinogenic
hazards. Science, 236, 271-280.
BElR 111 Committee. Advisory Committee on the Biological Effects of Ionizing
Radiation of the United States National Academy of Sciences, (1980). The
effects on populations of exposure to low levels of ionising radiation,
National Academic Press, Washington, DC.
Carlborg, F. W. (1985). A cancer risk assessment for saccharin. Food Chem.
Toxicol., 23, 499-506.
Crump, K. S. (1981). An improved procedure for low-dose carcinogenic risk
assessment from animal data. J . Enuiron. Pathol. Toxicol., 5, 339-348.
Crump, M. S. (1986). Letter to editor. Fundam. Appl. Toxicol., 6, 183-184.
Environmental Protection Agency (1986). Guidelines for carcinogen risk assess-
ment. Federal Register, 51, 33992-34003.
Food Safety Council (1980). Quantitative Risk Assessment. Food Cosmet.
Toxicol., 18, 711-734.
Gad, S. C. and Weil, C. S. (1986). ‘Statistics and Experimental Design for
Toxicologists~.Telford Press, Caldwell, New Jersey.
Hogan, M. D. and Hoel, D. G. (1982). Extrapolation to Man. In: ‘Principles and
Methods of Toxicology’. Ed. A. Wallace Hayes. Raven Press, New York,
pp. 711-731.
Lave, L. B. (1987). Health and Safety Risk Analyses: Information €or better
decisions. Science, 226, 291-295.
Lave, L. B. and Omenn, G. S. (1986). Cost-effectiveness of short term tests for
carcinogenicity. Nature (London), 324, 29-34.
Lovell, D. P. (1986). Risk assessment-general principles. In: ‘Toxic Hazard
Assessment of Chemicals’, Ed. M. L. Richardson. Royal Society of
Chemistry, London pp. 207-222.
Mantel, N. and Bryan, W. R. (1961). ‘Safety’ testing of carcinogenic agents. J .

Natl. Cancer Inst., 27, 455-470.
Park, C. N. and Snee, R. D. (1983). Quantitative risk assessment: state of the art
for carcinogenesis. Fundam. Appl. Toxicol., 3, 320-333.
Peto, R., Pike, M. C., Day, N. E., Gray, R. G., Lee, P. N., Parish, S . , Peto, J.,
Richards, S . , and Wahrendorf, J. (1980). Guidelines for simple, sensitive
significance tests for carcinogenic effects in long term animal experiments.
In: ‘Long Term and Short Term Screening Assays for Carcinogens’. A
Critical Appraisal (IARC Monograph No. 2), Lyon, International Agency
for Research on Cancer, pp. 311-426.
Richardson, M. L. (1986). Preface to ‘Toxic Hazard Assessment of Chemicals’.
Ed. M. L. Richardson. The Royal Society of Chemistry, London.
The Royal Society (1983). ‘Risk Assessment-A Study Group Report’. The
Royal Society.
Salsburg, D. S. (1986). ‘Statistics for Toxicologists’. Marcel Dekker Inc., New
York.
Sawyer, C., Peto, R., Bernstein, L., and Pike, M. C. (1984). Calculation of
carcinogenic potency from long term animal carcinogenesis experiments.
Biometrics, 40, 27-40.
Selby, P. B. (1982). Induced mutations in mice and genetic risk assessment in
humans. ‘Progress in Mutation Research’. Vol. 3. Ed. K. C. Bora et al.
Elsevier Biomedical Press, pp. 275-288.
Spector, W. S. (1956). ‘Handbook of Biological Data’. W. B. Saunders,
Philadelphia.
UNSCEAR Committee (United Nations Scientific Committee on the Effects of
Atomic Radiation) (1977). Sources and effects of ionising radiation, Report
to the General Assembly, with annexes, United Nations, Sales No. E77.1x.l
pp. 425-564.
Weil, C. S. (1972). Statistics us. Safety Factors and Scientific Judgement in the
Evaluation of safety for man. Toxicol. Appl. Pharmacol., 21, 454-463.
West, J. D., Peters, J., and Lyon, M. F. (1984). Genetic differences between two
substrains of the inbred 101 mouse strain. Genet. Res., 44, 343-346.
CHAPTER 20
Epidemiology
G. M. PADDLE
1 Introduction
For present purposes, epidemiology will be defined as the study of
patterns of health in groups of people. Besides brevity, this definition has
the virtue that the key words all have colloquial meanings, which convey
a picture of simplicity, and more esoteric meanings, which convey an air
of profundity. Epidemiology is an observational, rather than experimen-
tal, pursuit, and the contrast between the pragmatism and opportunism
of the barefoot epidemiologist and the purity of approach of the
academics is one of its many fascinations. Many other definitions of
epidemiology have been proposed (Lillienfield, 1978) but none is
universally accepted. Indeed, as the range of epidemiological applications
expands, there is a temptation to argue that, just as toxicology is what
toxicologists do (Casarett and Doull, 1980), so epidemiology is what
epidemiologists do. The aim of this chapter is to describe what
epidemiologists do, and how they do it. The aim of this introduction is to
link ‘the study of patterns of health in groups of people’ to the activities
of other relevant disciplines.
Epidemiologists have to work closely with clinicians, who could be said
to study the patterns of health of individuals. However, an
epidemiologist’s concern with the health of groups of people requires a
different approach from that of a clinician concerned with the ill-health of
an individual, and the goal of restoring it to normality. The information
collected by epidemiologists and clinicians may often be similar. Each
may record that John Smith was born in 1937, is 175 centimetres tall,
weighs 86 kilograms, has worked for 20 years in the chemical industry,
smokes 40 cigarettes a day, drinks 15litres of beer a week, and is
suffering from bronchitis. The clinician will use this information to assess
John Smith’s state of ill-health, to recommend a cure, and to counsel him
about his future welfare. These consultations between the clinician and
his patient usually occur only when the patient is unwell. The epidemiol-
464
G. M . Paddle 465
ogist will use the same information to classify John Smith alongside other
men of similar age, with similar occupational histories, and similar
smoking and drinking habits, in order, perhaps, to gain a better
understanding of bronchitis patterns, which for some reason John Smith
is suffering from, and hypertension patterns which for some reason he is
not suffering from. This accumulation of information by the epidemiolo-
gist about the subject in his study may occur when the subject is fit or
unfit, in his presence or in his absence, or even after his death.
Epidemiologists and toxicologists, particularly those toxicologists who
could be said to study patterns of health in groups of laboratory animals,
often have common objectives. An epidemiologist and a toxicologist may
have data about the effects of the same compound on the same target
organ. The toxicologist may have conducted a long term rodent study
with the aim of predicting the health effects that people will experience at
the exposure levels prevailing in the workplace or in the environment.
The epidemiologist may have studied the health patterns actually
experienced by people exposed to the compound occupationally or at
home. The total evidence about the risk to humans of this compound will
consist of the toxicologist’s precise, experimental data about the wrong
species at the wrong exposure, and the epidemiologist’s imprecise,
observational data about the right species at the right exposure. The
moulding of these data into a formal risk assessment for regulatory
purposes is a problem receiving much attention currently (Shore et al.,
1992). At the development stage of a compound, and during the early
days of its use, it is logical for the laboratory data to dominate the
assessment, but, after several decades of use, it seems more logical for
the human data to hold sway (Doll, 1987).
The definition of the exposure levels in an epidemiological study can be
improved by the involvement of an industrial hygienist, whose role could
be described as the study of patterns of exposure to hazards in groups of
people. The ultimate aim of epidemiology is to establish the dose-
response curve for human health effects versus exposure level of
compound. Success in achieving this objective depends upon the collec-
tion of appropriate exposure data (Landrigan, 1982). Whereas it would
be inconceivable to conduct a laboratory experiment without painstak-
ingly specifying and controlling the conditions of exposure, it is the
exception for an industrial hygienist to collect exposure data for
epidemiological purposes, rather than to ensure compliance with
regulations.
An alternative to the exposure data collected by a hygienist is the
measurement of delivered dose by the analysis of body fluids, such as
blood and urine. Collaboration with a biochemical laboratory is increas-
ingly an important element in an epidemiological study. If the same
analyses have been done on experimental animals, perhaps by the same
laboratory using the same techniques, it becomes possible to compare the
metabolic pathways and pharmacokinetics of different species, and hence
466 Epidemiology
arrive at a more rational and scientific risk assessment (Anderson et al.,

1987).
The aim of this introduction has been to show the extent to which
epidemiology is reliant upon, and complementary to, other areas of
activity. Epidemiology can make a substantial contribution to the
improvement of health patterns in groups of people on its own. Greater
improvements will be achieved with greater efficiency if epidemiology is
appreciated and supported by other health science specialists.
2 Sources of Problems
Several of the chapters in this book have been devoted to the techniques
available to the toxicologist, and to the systems which he studies when
examining the toxicity of chemicals. The studies an epidemiologist can
mount are similar in range, but they are conducted in a totally different
fashion. The reasons for the differences are obvious. Whereas the
toxicologist can contain his animals within a specified environment , and
ensure that the food eaten and the air breathed by them is of a
predetermined and standardised form; the epidemiologist has no such
control over the people in his studies. Therefore, although their studies
may be similar in basic design, the inability to impose an exposure
regime, the absence of an ideal control group, and the need to analyse
the data for factors additional to the one being studied, mean that the
epidemiologist can never claim that his study approaches the rigour of
that of the toxicologist.
The technique an epidemiologist applies will depend upon the source
of the problem and the urgency attached to it. Some of the ways in which
problems arise are as follows:
(a) An animal study points to a hazard to a human population
(b) A clinician has noticed a pattern of ill-health
(c) An epidemiological study prompts further work
(d) A body of health data has been accumulated
(e) Records for a large population are available
(f) People have expressed concern.
An example from each of these sources will help to introduce many of
the principles and methods to be discussed in later sections.
A An Animal Study Points to a Hazard in a Human

Population
Whenever a laboratory study of a compound to which people are already
exposed demonstrates an excess of disease in animals, it is a natural
consequence to study the exposed people to protect them against the
same or related health problems.
G. M . Paddle 467
In 1981 it was reported that formaldehyde had been found to cause

nasal squamous cell carcinomas in a chronic inhalation study in both rats
and mice (C.I.I.T., 1981). Formaldehyde is a ubiquitous substance in the
environment and it has a long history of widespread use in industry.
There was obviously an urgent requirement to study the patterns of
health in groups of people exposed to formaldehyde. The first epidemio-
logical reports to appear were of case-control studies of nasal cancer.
These are the easiest studies to complete quickly but their interpretation
is always open to debate. Negative outcomes from these nasal cancer
studies could not have been considered completely reassuring, because
they left open the possibility that the target organ was different in people.
Subsequently, two retrospective cohort studies have been published
involving occupationally exposed people in the UK (Acheson et al.,
1984), and occupationally exposed people in the USA (Blair et al., 1986).
These studies examined mortality rates for all causes of death, not just
nasal cancer, and have provided a sound basis on which to assess the risk
attributable to formaldehyde exposure.
B A Clinician has Noticed a Pattern of 111-health

Clinicians, through experience, become aware of the normal pattern of
health of their patients, and are able, without sophisticated calculations,
to detect departures from this pattern.
A cluster of cases of adenocarcinoma of the nose and sinuses was
observed by two specialists in Buckinghamshire, and they were able to
associate it with employment in the local furniture industry (Macbeth,
1965). The initial discovery that smoking is causally related to lung cancer
was due to several clinicians independently reporting that a large
proportion of the lung cancer cases referred to them were heavy smokers
(e.g. Doll and Hill, 1950). Subsequent studies of workers have confirmed
an excess risk of nasal cancer in woodworkers and progress has been
made towards identifying the causative agent (MRC, 1981). Further
studies of the patterns of health in smokers have confirmed the link
between smoking and lung cancer, and helped to estimate the dose-
response curve. One of the smoking studies was a prospective study
which became an intervention study, inasmuch as many British doctors
were persuaded to give up smoking, and their health patterns were
compared with those of continuing smokers and non-smokers, (Doll and
Peto, 1976). This study showed that cessation of smoking gave an
immediate beneficial effect.
The observation by an astute clinician of an unusual pattern of health
in his patients has been the single source of almost all the major
epidemiological discoveries to date. When the excess risk is high, and
particularly when it is concentrated in a small community, there is no
reason to think that modern epidemiological techniques will prove to be a
more successful method of detection.
468 Epidemiology
C An Epidemiological Study Prompts Further Work

This is a major source of epidemiological projects, despite the fact that
very few major health risks have been discovered by formal epidemiolog-
ical studies.
An excess of a syndrome of such non-specific nature that it is often
referred to as ‘Danish painters disease’ was reported, (Axelson et al.,
1976) and it was postulated that the syndrome was associated with
chronic exposure to solvents. This led to the conduct of cross-sectional
studies in Germany to test the hypothesis in a different environment and
conferences in Copenhagen (WHO, 1985) and North Carolina (Cranmer
and Golberg, 1986) at which the various aspects of the syndrome were
discussed, so that a protocol could be established for further studies.
It is in the nature of observational studies to be inconclusive, and to
pose more questions than they answer. The quest for indisputable
conclusions, and the desire to resolve all the related issues necessitates a
sequence of studies on the same topic. The time consumed by such
activities, without any major decisions being reached, is justifiable
because the excess risk is often small, or the health effect is not
debilitating or mortal.
D A Body of Health Data has been Accumulated

Records collected on a routine basis by national and regional authorities
permit patterns of health to be displayed, and variations to be identified
and explored. The death records for England and Wales have been a
source for epidemiological studies for several centuries (Graunt , 1662)
and statistical methods for their analysis were first propounded over it
century ago (Farr, 1839).
The Medical Research Council Environmental Epidemiology Unit has
published maps of cancer mortality based on the England and Wales
mortality data for 1968-1978 (Gardner et al., 1983). These maps, an
example of descriptive epidemiology, illustrate regional variations in
cancer mortality and serve to target epidemiological effort into finding
causes for the highest rates. As the maps clearly depict the excess cancer
rates of known origin, it is reasonable to assume that there are underlying
reasons for the so far unexplained peak figures. The same Unit has
conducted a case-control study of cancers in young and middle-aged men
in the North of England (Coggon et al., 1986), because the British
chemical industry is located there, and cancer rates are higher there than
in the South of England.
E Records for a Large Population have Accumulated

Large chemical Companies that have provided a pension fund for several
decades collect mortality records for deaths in service and deaths of
469
G. M . Paddle
pensioners in order to manage the fund. For many compounds, these

records will relate to the most highly exposed and longest exposed
members of the population.
It is, therefore, reasonable to suppose that analysis of these records
will provide evidence about the health patterns generated by these
compounds, and it would be uncaring of the Companies not to analyse
them routinely.
Several Companies have published papers about their accumulated
death records (e.g. O’Berg et al., 1987). A standard analysis for a
mortality study is used, and statistically significant clusters are detected
and followed up, either by retrospective cohort studies, or nested
case-control studies.
F People have Expressed Concern

Just as individuals have access to a clinician to express concern about
their health, so do groups of people have avenues they can explore when
concerned about their mutual health pattern. All such expressions of
concern have to be evaluated and resolved, even though a report from the
responsible American authority (Schulte et al., 1986) has shown that in
their experience the concern is usually scientifically unfounded.
As heart disease is very common, and is increasing in incidence in men
of working age in the UK, it is frequently a health concern among
employed groups (Paddle and Bennett, 1983). In this example the
employees were those of a power station run as part of a large chemical
company, but separated from the neighbouring plants by a boundary
fence. Through the normal consultative procedure, the men had ex-
pressed anxiety about a sequence of coronary illnesses, some of them
fatal, and had inquired whether there was an occupational explanation.
The Medical Officer felt that the frequency of the incidents may have
been out of the ordinary, and initiated a detailed investigation. Careful
scrutiny of the medical records for the workforce disclosed eleven
coronary events in a period of eight years. Further research into this
small cluster involved a variety of techniques. The frequency was found
to be out of the ordinary, but it did not appear to be of occupational
origin.
3 Guidelines
One consequence of epidemiological studies having a variety of back-
grounds, being observational in character, and inconclusive in outcome,
is that they are difficult to report clearly and concisely. Any two readers
of a typical epidemiological report will perceive different real or
imaginary flaws, will experience different levels of concern, and will
recommend different courses of action. I:f a decision maker in industry,
470 Epidemiology
government, or academia is to do his job by co-ordinating the opinions of

such readers, it is necessary for standards of reporting to be established.
Guidelines written by the Inter-agency Regulatory Liaison Group have
been published (Epid. Work Group of the IRLG, 1981) that can be used to
ensure that all important aspects of a study are reported. The guidelines
cover the headings: Background and Objectives, Study Design, Study
Subjects, Comparison Subjects, Data Collection Procedures, Analytical
Methods and Statistical Procedures, Interpretation, Limitations and
Inferences, and Supportive Documentation. Most of these headings are
self-explanatory and clearly essential. Although these guidelines were
written with reporting in mind, they are equally of value when planning
studies. Perhaps the most frequently met failing in an epidemiological
report is a lack of clear objectives and a consequential lack of justification
for the interpretation and conclusions. If a study is mounted to determine
whether smoking habits are related to baldness, it is inappropriate to
conclude from it that alcohol intake is related to flat feet. The latter is at
best a chance finding worthy of further study, and is at worst a
predictable outcome of a study, that was designed for a different purpose.
4 Cohort Studies
Influential epidemiological output is dominated by two types of study; the
cohort study and the case-control study. To a toxicologist, the former
may look like an attempt to mimic a laboratory experiment, while the
latter is more like an attempt to recreate the history of a laboratory from
a study of the pathology reports. Let us look first at the most prestigious
and most transparently valid of epidemiological techniques. It is known
as the cohort study, where ‘cohort’ means a number of people grouped
together by a common feature such as exposure to a chemical,
occupation, domicile, or social habit. A classical cohort study is
conducted prospectively, that is to say, the cohort is formed at the outset
of the study and is followed for a period of time during which its health
pattern will be recorded until such time as the entire cohort has died, or
until sufficient data has been accumulated. This raises one obvious
disadvantage which the epidemiologist faces compared with the toxicolo-
gist, namely the longer lifespan of humans compared to rodents. The
epidemiologist runs the risk of not achieving results in his own lifetime.
To overcome this difficulty, and to provide the scientific community with
the data it needs to resolve today’s problems, the epidemiologist can
resort to a retrospective or historical cohort study. Using this technique,
the epidemiologist recreates a cohort from the past and follows it towards
the present day.
An advantage of the cohort study is that the cohort is usually known to
be free of disease at the time of formation, or recruitment. As disease
occurs, it can be assumed that it is either the result of everyday risks or
G. M . Paddle 47 1
the ageing process, and therefore equivalent to the incidence in a control

group, or that it is due to the factor that exemplifies the cohort.
Examples of retrospective cohort studies were referenced in the discus-
sion of formaldehyde. A further example is that of employees in the
United Kingdom exposed to vinyl chloride monomer (VCM) during the
course of their work (Fox and Collier, 1977). In that study, an excess
mortality from liver cancer was demonstrated, but, unlike studies of
employees in other countries exposed to VCM, no excess mortality from
cancer at other sites, or excess deaths due to other forms of disease, were
found.
In that study, as in all such studies, it is possible to find ways in which
the epidemiological work is of less validity than the equivalent rodent
toxicology (Maltoni et al., 1981). The humans were indeed exposed at
work to VCM, but that was not the only potentially harmful substance to
which they were exposed, either at work, or elsewhere. Many of them
will have worked with other chemicals at the same time as they were
exposed to VCM, many will have worked in different industries before
and after employment in the plastics industry, and many will have
hobbies or a social style which brought them into contact with other
potentially harmful substances. Even the extent of their exposure to
VCM had been sparsely recorded. Prior to the rnid-l970s, when VCM
was first shown to be harmful to people at high levels, the exposure had
been measured very infrequently.
These drawbacks have tended to make the regulatory authorities
dismiss epidemiological results as invalid evidence, despite the fact that
epidemiology reflects the real risk at real exposures to real people.
The data from a cohort study is summarised by two tables. Table 1 sets
out the deaths by age at death and year of death. Age at death is
important because death rates increase rapidly with increasing age, and
because the major causes of death of young people differ from the major
causes of death of old people. Year of death is important because death
rates have changed from one decade to the next. Some causes of death,
such as lung cancer, have increased in frequency while others have
Table 1 Deaths from all causes by age

and calendar year (Segment from a
hypothetical study)
I Calendar year
50-54
55-59
60-64 47 62
65-69 68 69 73
47 2 Epidemiology
Table 2 Person years at risk by

age and calendar year (Segment
from a hypothetical study)
I Calendar Year
Age 1951-5 1956-60 1961-5

50-54 2940 3410 4143
55-59 2407 2895 3290
60-64 2310 2380 2810
65-69 1976 2198 2304
decreased. Changes in diagnostic fashions and classification systems also

have to be allowed for. Table 2 sets out the person-years at risk by age
and calendar year. Each person, whether dead or alive at the end of the
study, contributes to this table, according to his year of birth, his date of
entry to the study, and his date of exit from the study. A person can
leave the study because he dies, because the study ends or, unlike a
laboratory study, because it becomes impossible to trace him. In the
simplest report of a cohort study, age and calendar year, as displayed in
Table 2, are sufficient to complete the analysis. If, however, there is an
excess risk of cancer in the cohort due to exposure, any reasonable model
of cancer in man will specify that there is no excess risk until a latent
period has elapsed since first exposure, and that there is a higher risk at
higher exposures. Latent periods for cancer are typically 15-30 years or
more, which means that the first 15 or so years of exposure should be
treated as if the people were unexposed. Several computer packages are
available for partitioning the person-years at risk table into latency
periods and exposure periods, so that the estimate of the excess risk due
to exposure is compatible with a simple model of tumour development.
5 Case-control Studies
The other ubiquitous type of epidemiological study is the case-control or
case-referent study. This is always a retrospective study, and will,
therefore, produce results more quickly than a prospective approach. It is
also a very efficient type of study because it only uses people who are
known to be able to contribute valuable information. It can therefore
produce results more rapidly and with less resources than a retrospective
cohort study. It has a reputation of being a ‘quick and dirty’ technique,
but reassessments and improvements to its design and analysis have led
to a growing acceptance that it can be the most informative as well as the
cheapest approach to a problem (Breslow and Day, 1980).
Typically, a case-control study will test the hypothesis that the presence
of factor A in a person’s life is related to his risk of contracting disease X.
G. M . Paddle 473
Table 3 Typical outcome of case-control study
I Disease X
cases 1 No disease X
controls
Total
Factor A 34 40 74
present
Factor A 62 152 214
not present
Total I 96 I 192 1 288
For example, factor A could be exposure to asbestos at work,chewing

tobacco, or taking part in amateur dramatics, and disease X could be
pleural mesothelioma, cancer of the larynx, or suicide. The first
requirement to be met is a list of cases of X about whom the requisite
information about factor A can be collected. The second requirement is a
set of controls, or referents, for each case, who differ from the case only
inasmuch as they have not contracted disease X. The outcome of the
study is a table like Table 3.
In Table 3 the matching of controls to cases has been ignored. If a
second factor is also important, for instance age, then a matched analysis
can be much more powerful. The overwhelming advantage of this design
is that the incidence of the disease is almost irrelevant. For a very rare
disease, a prospective cohort study may require a thirty year follow-up of
several million persons to produce worthwhile results. An equivalent
case-control study can be productive even if there have only been forty
cases, say, in England and Wales in the past twenty years.
An example of a case-control study is the survey of cancer and
occupation in young and middle-aged men referred to in Section 2. In
this project, the investigation of several diseases and many occupations
was encompassed in a single study. Occupational and smoking histories
were collected for all the cancer patients aged 18-54 in the area of the
survey. When respiratory cancer was the disease specifying the cases, all
the other patients were used as controls. A disadvantage of this approach
is that so many hypothetical dose-response relationships are being tested
simultaneously that some of them are bound to be statistically significant,
even when no real relationships exist. This portmanteau approach to
epidemiology can be very cost-effective if practised judiciously, but if
applied without thought it can justifiably be described as a fishing
expedition.
The case-control study is an approach that toxicologists can find
counter-intuitive. This is because each animal they handle is taking part
in a single organised experiment, and its demise is attributed to the
presence or absence of the substance under test. There is little point in
analysing all the liver tumours to determine what caused them, because
474 Epidemiology
the answer is already known. This is far from true of humans, each of
whom is taking part, unwittingly, in a multitude of epidemiological tests
by being, for example, a smoker, a VCM worker, a painter and
decorator, and a heavy drinker.
The checkered history of case-control studies can be attributed to two
features which prompt criticism. Firstly, the cases of disease X are often
to be found in hospital or in the graveyard. This can make it difficult to
obtain information about their exposure to factor A, and it can prevent
them or their surrogate from being dispassionate about factor A. If factor
A is to any extent suggested as a possible cause of disease X, the
possibility that they admit to exposure can be increased. Secondly, if the
cases are hospitalised or deceased, it is not obvious who to use as
controls, differing only to the extent that they have not contracted disease
X. Should the controls be correspondingly hospitalised or dead, and, if
so, hospitalised with what or dead from what? If the controls are chosen
from healthy persons, however, it is obvious that they differ in at least
one important respect from the controls.
The rigour of a cohort study and the efficiency of a case-control study
can both be obtained by using a case-control study nested within a cohort
study (Kupper et al., 1975). The cohort study is conducted in standard
fashion but without reference to any exposure data. If it detects any
categories of disease worthy of further investigation, a case-control study
is mounted for each of these diseases. This approach is very economical if
the exposure data for each individual is complicated by mixtures of
chemicals, changes of job and changes in processes, as is usually the case.
6 Other Types of Study

This section describes the other types of study alluded to in Section 2.
A mortality study can be conducted when the results of a cohort study
are required but either only details of deaths are available or there is not
time or resource for a full study. Its disadvantage is that the results are
not as valid as those of a cohort study and the extent of invalidity cannot
be deduced.
The epidemiologist collects details of all the deaths that he can muster
in the population of interest. The files of decedents collected by large
chemical firms are examples. He extracts from the death certificates the
underlying cause of death for each one, classifying them according to the
International Classification of Diseases, (WHO, 1977) and tabulates them
according to age at death. A hypothetical data set is shown in Table 4.
A difficulty in conducting this type of study is that some of the death
records required will be easy to acquire, some less easy, and some will
prove to be very elusive. The temptation will be to analyse the more
accessible records, but this can mean that the study is biased towards
some outcomes rather than others. As with cohort studies, the data in
G. M . Paddle 475
Table 4 Mortality data grouped b y age and cause

Cause of ICD
death codes 25-34 35-44 45-54 55-64 65-74 74- All
Cancer 140-239 2 8 45 114 51 224
Coronary 393-429 5 28 67 124 106 332
Cerebra1 430-439 2 2 7 42 60 113
Respiratory 460-519 0 1 13 62 61 137
Accidents 800-999 3 3 2 7 5 21
Other Other 1 6 9 36 36 90
All 001-999 13 48 143 385 319 917
Table 4 can be partitioned in various ways to increase the validity of the

analysis.
A cross-sectional study can be used when past records do not exist or
cannot be accessed and a group of people available at the time have to be
used.
A typical cross-sectional study is a morbidity prevalence study. This
may be appropriate if a report is received of a possible link between a
substance and a non-fatal disease. The epidemiologist will identify all
those in the group in which he is interested, such as the painters in the
Danish and German studies, and will have each individual assessed to
decide whether or not he is suffering from the disease in question. He will
categorise the sufferers and the non-sufferers according to their age at the
time of examination. A hypothetical set of data is shown in Table 5.
A disadvantage of cross-sectional studies is that the health effect being
studied must not be such as to affect who will and who will not remain in
the exposed group, otherwise only fit people will be available for study.
This may be true of most blood and urine parameters, and minor
respiratory or coronary ailments, but it will not be true for debilitating or
lethal diseases. The technical problem is that cross-sectional studies
measure the prevalence of a disease, that is, the accumulation of cases
that have not been cured, rather than the incidence, which is the
frequency with which new cases occur.
I Table 5 Prevalence of disease grouped by age

I
15-24 2 137 139
25-34 7 209 216
35-44 17 172 189
45-54 30 169 199
55-64 15 65 80
476 Epidemiology
In the power station example, a cross-sectional study of coronary risk

factors was conducted on the present employees at the power station and
employees drawn from the neighbouring factor. Controls were chosen at
random from those employees at the factory who matched each case for
age,length of service, and staff category. The disease in question was
‘suspect ischaemia’ measured by the answers to a questionnaire or the
results of an ECG. The study demonstrated no difference in prevalence
of the condition in the two groups.
Descriptive studies can be either an end in themselves or they can form
part of the background data for another form of study. The essence of a
descriptive study is to describe the data as simply as possible but without
introducing bias. Descriptive data can be presented as tables with
summary statistics, as tables in which the results have been sequenced in
descending order for easy reference, or as maps, in which shading or
colours can be used to highlight the abnormal values.
An intervention study, such as the smoking study on British doctors,
involves changing the exposure of some of the people in a study and
observing whether their health patterns change as a result. In the
industrial setting this could mean changing a regulatory exposure limit or
ceasing using a compound altogether. This is likely to demonstrate a
beneficial effect, because a chance finding will go away of its own accord,
or a real excess risk will be reduced or eliminated. It can, however, take
a long time for the benefit to become clear because the latency effect will
continue to produce illness from past exposure. Dietary epidemiology is a
particularly fertile area for intervention studies.
7 Control Groups
Having taken his study as far as a tabulation of the results for his exposed
group, the epidemiologist will be faced by a problem which is met with
equanimity by the toxicologist. He will need data for a control group with
which to compare his figures. The toxicologist, in accordance with the
principles laid out in previous chapters, takes another group of animals as
his control group. These will be identical to those in the exposed group
and, indeed, will have been allocated to the exposed and unexposed
groups at random. This is far from being the way in which people conduct
their lives. Where is the epidemiologist to find a control group identical
to his exposed group except only that they do not live in the same village,
or have not been employed in the same industry? The intricacies of this
problem, and the measures that have been taken to alleviate it were
discussed at a conference solely devoted to this topic (MRC, 1984).
Suppose that he chooses the population of the country in which the
village or industry is situated. One major advantage of this apparently
inappropriate choice may be that equivalent death, morbidity, or coro-
nary statistics are available in published form. The disease pattern in his
G. M. Paddle 477
Table 6 Expected deaths from all causes by

age and calendar year (Segment from a
hypothetical calculation)
I Calendar year
Age of death I 1951-5 I 1956-60 I 1961-5

50-54 29.1 32.6 34.3
55-59 36.7 43.8 45.9
60-64 49.6 53.4 67.8
65-69 62.4 73.7 80.2
study can now be compared with the pattern in the control population. In
the common jargon, he will be able to calculate an ‘expected’ pattern of
ill-health which he can compare statistically with the pattern ‘observed’.
In a cohort study, the expected figures are calculated by multiplying
the figures in the man-years at risk table (Table 2) by the age and
calendar year death rates for the control group to arrive at the figures in
Table 6. Although expected ‘deaths’, these figures will not be whole
numbers.
The figures in Table 1 are summed to give a total ‘observed’ figure and
those in Table 6 are summed to give a total ‘expected’ figure, and the
ratio (total observed)/(total expected) is multiplied by 100 to give a
standardised mortality ratio (or SMR) for all causes of death. The
calculation can be repeated for any group of diagnoses, such as all
cancers, or any specific diagnosis, such as chronic lymphatic leukaemia.
A similar calculation, employing proportions of deaths rather than
man-years at risk, converts the figures in a mortality study into
proportional mortality ratios (or PMRs).
From the observed and expected figures, it is simple to calculate
whether the health of the exposed population is statistically significantly
different from that of the control group.
Any ‘observed’ figures that are significantly in excess of the ‘expected’
figures are ‘evidence’ of an ‘exposure’ effect. Unfortunately, as there is
no single control group which is obviously ‘best’, any choice of control
group can be criticised. In the heart disease example, the population of
England and Wales, the population of North West England, the
employees of power stations, the employees of the company, and the
employees of the neighbouring factory were all plausible control groups.
If alternatives generate similar ‘expected’ values then the choice among
them is largely academic, but if, as can happen, the expected values have
a wide range, then the choice of control group will determine the
outcome of the study. In the example, no sensible control group was
478 Epidemiology
found which gave an expected figure of more than about four, so that the
observed figure of eleven clearly warranted further investigation. Had the
‘expected’ values spanned the range from two to twelve, however, it
would first have been necessary to agree upon the most appropriate
control group.
Whilst criticisms are always possible of any control group, it is unwise
to dismiss all epidemiological studies on the grounds that the choice of
control population is suspect. The study may still contain an important
message about people, health, and the toxicity of chemicals.
Confounding Factors
In a laboratory study of formaldehyde, for example, the only difference
between the exposed and control animals is their exposure to form-
aldehyde. In an epidemiological study of formaldehyde, the health pattern
of the exposure group will be affected by their age, by their smoking
habits, by their exposure to other hazards at work, such as wood dust, and
by their state of health at recruitment to the cohort, and a host of other
factors besides exposure to formaldehyde. It would be absurd to assume
that these factors will not affect the study, and wildly optimistic to assume
that they will have the same effects in whatever control group is chosen.
Techniques are needed that will allow for the effects of these other factors
and thus permit the effect of formaldehyde to be estimated. Unfortunately
a single technique is not sufficient. Age is a factor that everyone has and
knows. It can be allowed for by matching the control group to the exposed
group, or by an age standardisation calculation. Smoking habits can be
built into the analysis for both exposed and control groups, but as there
may be a synergistic or antagonistic relationship between formaldehyde
exposure and smoking, a simple linear correction may not be sufficient.
Wood dust exposure may be so closely linked to formaldehyde exposure
that there separate effects cannot be disentangled. Wood dust would then
be a confounding factor and the reliability of any separate estimates of the
effects of formaldehyde and wood dust exposure would remain in some
doubt.
The most demanding aspect of the statistical design and analysis of
epidemiological data is the effort that has to be made to allow for these
confounding, or nuisance, variables. Even in the best planned studies,
there will be some need for sophisticated statistical analysis rather than
the straightforward analysis that was originally planned.
Data Collection
At the data collection stage of a project, the epidemiologist has a great
advantage over the laboratory experimenter; he is able to communicate
with his subjects. The options of asking questions, recording answers,
probing for additional information, and noting opinions are elements of
most epidemiological studies. Data can be collected by telephone,
G. M. Paddle 479
correspondence, and direct computer interaction as well as by face to

face interview. The questionnaire is a very versatile tool. It is also such a
simple tool that it can easily be misused, and its complexities can easily
be overlooked.
When collecting data the epidemiologist will be largely reliant upon
three techniques-incident reporting, questionnaires, and quantitative
measurement. Whichever of these techniques is used, it is important to
avoid biased results by defining and following a rigorous protocol.
Typically, incident reporting will cover birth and death, for which the
data are usually adequate, but also ill-health, as measured perhaps by
sickness absence. Before attaching too much importance to routine
absence data, it should be realised that a period of sickness absence
might be due to a football match taking place in the neighbourhood
rather than to ill-health; and that a period of presence at work, even
when sickness exists, may be due to a bonus being paid prior to a holiday
period. It may be necessary to audit the routine sickness data before
using it to assess patterns of health.
When a questionnaire has been completed, it is possible that the
responder has not understood the questions, or is not providing truthful
answers. When quantitative measurement is involved, it is possible that
the patient has biased the result by not fasting before the sample is taken.
The data then are subject to some doubt. It is possible to include
validation techniques but these will usually involve further costs of one
sort or another.
Recall for re-examination is hard to justify if management are
requesting that the men are not withdrawn from work too frequently, or
the budget for a study may be insufficient to cover repeat visits to check
questionnaire responses. Nonetheless, the study should have been
conducted according to a strict protocol and each item of information
collected, (whether a response to a question, or an incident report, or a
quantitative measurement) should have been checked for accuracy. If this
has been the case, then the statistical analysis of the data should be a
straightforward exercise. In the power station example all three methods
of data collection were used at various stages. Initially, incident reports
of coronary disease had to be counted. This may have been no easy task
if the information had needed to be extracted from the medical records.
Each record would have had to be examined, and there would have been
no reason to expect the clinicians involved to have used consistent
terminology. Fortunately the company had been operating a com-
puterised medical records system for the period involved and the data for
deaths and long term absences had been coded in a standardised manner.
When a questionnaire was required for the latter part of the study, it was
sensible to use the one devised and validated by the London School of
Hygiene and Tropical Medicine. When ECGs were collected, it was
preferable to rely on an internationally accepted method of classification
rather than use the more detailed, yet less formalised, descriptions of the
local clinicians.
480 Epidemiology
10 Data Banks
One difference between epidemiology and toxicology is highlighted by a
discussion of data banks.
It seems rather inconsistent that, with the benefit of huge data banks,
epidemiology has managed to demonstrate few definite relationships
between factors and diseases in humans, whereas toxicology with
relatively small data banks has established many more relationships
between factors and diseases in laboratory animals. A defence of
epidemiology is that its data banks have only recently been made
sufficiently detailed and comprehensive to contain the necessary evidence
for relationships to be harvested in twenty years time. Previously the
records had been kept in manual, unstandardised form. Occasional
attempts to analyse toxicological data accumulated over a series of
studies have revealed that animal data can be just as inconsistent and
subject to just as many temporal variations as human data. The
computerised files of health and occupational data kept in the Scandi-
navian countries, and linked by personal identification numbers are
enabling epidemiological studies to be done almost wholly by modern
technology. Whether the increase in sensitivity of toxicological studies and
the more rapid access to epidemiological results will lead to a safer
environment for mankind remains to be seen.
11 Interpretation
It is at the interpretation stage that arguments about the validity of an
epidemiological study will arise. It is a straightforward statement in an
animal study to say that exposure to a certain chemical has caused a
certain incidence of cancer in a particular species of rodent; that is only
bad news for the chemical species and for the rodents.
Indeed the experiment has been planned and conducted in such a way
that very few interpretations are plausible. An excess of disease in the
exposed group can only be attributed to exposure or to chance.
If the same sort of statement is made about the same chemical in
humans, then one needs to be extremely sure that the statement is
correct. Otherwise the unhappiness caused to those people exposed to
that chemical but still alive, and to the best of their knowledge in good
health, is considerable.
Thus the interpretation of the power station study was that the
employees had been the subjects of a long run of bad luck, and the
cross-sectional study gave no reason to suppose that the sequence would
continue. The long history of healthy employees in power stations in all
parts of the world makes this a more plausible explanation than the
existence of some malevolent factor in this particular building. Of course,
there could be a complicated multi-factorial explanation of the excess
incidence, but all plausible simple explanations have been exhausted.
G. M. Paddle 48 1
One difficulty in the interpretation of an epidemiological study is that

very few studies stand on their own two feet. Usually, similar studies of
similar groups of people exposed to similar chemicals in similar cir-
cumstances will already be in the literature.
It is extremely difficult in this situation to piece together the various
bits of epidemiological information to reach a conclusion about the
harmfulness of the chemical in question (Beaumont and Breslow, 1981).
A technique called meta-analysis can be used to combine the results of
clinical trials to obtain a more precise estimate of the efficacy of a new
drug. Whether meta-analysis can be used to combine the results of
epidemiological studies, when the protocols are usually quite varied,
remains a controversial issue (Spitzer, 1991). The extent to which the
interpretation of a single study can be generalised to other situations, in
which the epidemiology has yet to be done, is always uncertain,
This issue has been addressed in detail in Bradford Hill’s fundamental
book (Bradford Hill, 1971) but we may note that the requirements laid
down by Bradford Hill are very rarely found in a single study. One of the
remarkable features of the relationship in people between exposure to
vinyl chloride monomer and the occurrence of angiosarcoma of the liver
is that all of the requirements are met. In most cases only one or two of
Bradford Hill’s requirements obtain.
In the reporting of many epidemiological studies, no greater claim is
made than that the study has managed to generate a hypothesis about a
chemical and a disease that is worthy of further study. This manner of
reporting may be justified by the extent of the data, or it may be that the
author does not wish to risk subsequent discredit. Nonetheless, the
difference between hypothesis generating studies and hypothesis testing
studies is becoming part of the folklore. If the extent of hypothesis
generating studies is of some embarrassment to the manufacturer of the
chemicals, then the plethora of negative studies is becoming an equal
embarrassment to the environmentalist in his quest for a world free of
toxic chemicals. Many epidemiological studies are almost certainly
negative from the day they commence.
In many cases, the population available for study, the degree of
exposure within that population, and the background incidence of the
disease being studied are all so small that the relative risk due to
exposure would have to be astronomical for a statistically significant
result to be achieved.
It may be asked whether these negative studies, of very low statistical
power, can be justified on any grounds whatsoever. One justification for
them is the extent to which they set at rest the minds of those who have
to work with the chemicals and who have been subjected to disquieting
reports based on studies of laboratory animals. Nonetheless, it has to be
admitted that a great deal of epidemiological work doomed to be
negative is reported as if there is an equal chance of it being positive or
negative. It is of doubtful sense to rely simply on the statistical analysis of
482 Epidemiology
the data and the p value which results from that analysis. The other
considerations listed by Bradford Hill should always be taken into
account and in particular the biological plausibility of the findings must
be considered, in the light of the available toxicological information.
12 Essential Skills
From a sample of the techniques that an epidemiologist will use, it is
possible to draw up a list of skills that he will need to be able to draw
upon. Ideally an epidemiologist will have been trained in both medicine
and statistics, and many leading epidemiologists do have qualifications in
both subjects. More commonly, a medically qualified person will acquire
the specialised statistical knowledge necessary for the role of epidemiolo-
gist, or a statistician will perform the role by using medically qualified
colleagues as consultants on the clinical aspects. All epidemiologists will,
however, need to consult a higher level of expertise than their own on
several facets of their work.
Consider again a mortality study. Suppose that the aim of the study is
to test the hypothesis that the age corrected proportions of deaths due to
respiratory disease in two towns are equal after allowing for smoking
habits, despite the presence of more heavy industry in one town than the
other. Clinical skill will be required to derive a tight definition of
‘respiratory disease’, and statistical skill will be required for the analysis
and the hypothesis testing. In addition, the classification of disease is
itself a science, called nosology, which requires training and experience.
The collection of data on smoking habits and domicile will involve a
questionnaire, which, as it clearly cannot be answered by the decedents,
must be entrusted to a relative or neighbour. The extent to which
ambiguities, leading questions and misunderstandings can invalidate
questionnaire data is itself a topic that has generated books and papers
(Kalton and Schuman, 1982). If the study is large and concerned with
many factors it will prove essential to use a computer for the storage and
analysis of the data. It then requires systems analysis skills to prevent
incorrect data, imprecise programming, and inappropriate statistical
packages from destroying what had so far been an immaculate study.
The danger is that so many skills have to be harnessed for a successful
epidemiological study that the task becomes too daunting to be under-
taken. Fortunately, such defeatism is not warranted. Although a perfect
epidemiological study may never take place, it is possible to set out
procedures which, if followed, will guarantee an acceptable study that
will not deserve to be torn to shreds when reported.
13 Conclusion
That then is what epidemiologists, or at least some epidemiologists, do,
and how they do it. They face many obstacles and they never achieve
G. M . Puddle 483
perfection. Are their efforts always in vain, and always to be discounted,

or do they produce the most valid evidence about human toxicology?
What can be said about the relative weight that can be attached to
epidemiological information? On the credit side, the species is certainly
the most appropriate one to study, and the dose levels are certainly those
which are appropriate to that species. On the negative side are all the
deficiencies admitted to in the course of this chapter. In many cases,
those deficiencies outweigh the advantages. It does seem, however, that
where epidemiological evidence exists that can rightly be called substan-
tial, in terms of population size, latency, and dose, then the two prime
advantages of that evidence must to a great extent outweigh conflicting
findings in animal or bacterial species.
References
Acheson, E. D., Barnes, H. R., Gardner, M. J., Osmond, C., Pannett, B., and
Taylor, C. D. (1984). Formaldehyde in the British chemical industry. An
occupational cohort study. Lancet, 1, 611-616.
Anderson, M. E., Clewell 111, H. J., Gargas, M. L., Smith, F. A., and Reitz, R.
H. (1987). Physiologically based pharmacokinetics and the risk assessment
process for MeCl. Toxicol. Appl. Pharmucol., 87, 185-205.
Axelson, O., Hane, M., and Hogstedt, C. (1976). A case-reference study of
neuropsychiatric disorders among workers exposed to solvents. Scan. J .
Work Environ. Health, 2, 14-20.
Beaumont, J. J. and Breslow, N. E. (1981). Power considerations in epidemiolo-
gical studies of vinyl chloride workers. A m . J . Epidemiol., 114, 725-734.
Blair, A., Stewart, P., O’Berg, M., Gaffey, W., Walrath, J., Ward, J., Bales, R.,
Kaplan, S . , and Cubit, D. (1986). Mortality among industrial workers
exposed to formaldehyde. J . Nutl. Cancer Inst., 76(6), 1071-1084.
Bradford Hill, A. (1971). ‘Principles of Medical Statistics’. 9th Edn. Lancet.
Breslow, N. E. and Day, N. E. (1980). Statistical methods in cancer research.
Volume 1 -The analysis of case-control studies. International Agency for
Casarett, L. J. and Doull, J. (1980). ‘Toxicology. The Basic Science of Poisons’.
Macmillan. pp. 3-10. 2nd Edn.
C.I.I.T. (1981). Battelle Columbus Laboratories, Final Report, Docket No.
10922. A chronic inhalation study in rats and mice exposed to formaldehyde.
C.I.I.T. Research Triangle Park N.C.
Coggon, D., Pannett, B . , Osmond, C., and Acheson, E. D. (1985). A survey of
cancer and occupation in young and middle-aged men. I. Cancers of the
Respiratory Tract. Br. J . Ind. Med., 43, 332-338.
Cranmer, J. M. and Goldberg, L. (1986). Proceedings of the Workshop on
Neurobehavioural Effects of Solvents. Neurotox., 7(4).
Doll, R. (1987). Environmental chemicals and cancer. Chem. Br., (Sept)
847-850.
Doll, R. and Hill, A. B. (1950). Smoking and carcinoma of the lung-
preliminary report. Br. Med. J . , 2 , 739-748.
Doll, R. and Peto, R. (1976). Mortality in relation to smoking: 20 years’
484 Epidemiology
observation on male British doctors. Brit. Med. J . , 4, 1525-1536.

Epid. Work Group of the IRLG. (1981). Guidelines for documentation of
epidemiological Studies. Am. J . Epidemiol., 114, 609-613.
Farr, W. (1839). Letter to Registrar General, in first annual report of the
Register General of Births, Deaths and Marriages in England. HMSO.
London.
Fox, A. J. and Collier, P. F. (1977). Mortality experience of workers exposed to
vinyl chloride monomer in the manufacture of polyvinylchloride in Great
Britain. Br. J Ind. Med., 34, 1-10.
Gardner, M . J., Winter, P. D., Taylor, C. P., and Acheson, E. D. (1983). ‘Atlas
of Cancer Mortality in England and Wales 1968-1978’, John Wiley and Sons.
Graunt, J. (1662). Preface to ‘Observations upon the bills of mortality, 1662’ in
the Economic Writings of Sir William Petty. Vol. 2. Ed. C. H. Hull. p. 333.
C.U.P. London, 1899.
Kalton, G. and Schuman, H. (1982). The effect of the question on survey
responses: a review. J . R. Stat. SOC., (A) 145, 42-75.
Kupper, L. L., McMichael, A. J., and Spirtas, R. (1975). A hybrid epidemiologi-
cal study design useful in estimating relative risk. J . Am. Stat. Assoc., 70,
NO. 351, 524-528.
Landrigan, P. J. (1982) Recent advances in assessment of workplace exposure-
epidemiologic linkage of medical and environmental data. J . Environ. Sci.
Health, A17, 4, 499-513.
Lillienfield, D. E. (1978). Definitions of epidemiology. Am. J . Epidemiol., 107,
87-90.
Macbeth, R. G. (1965). Malignant disease of the paransal sinuses. J . Laryngol.,
79, 592.
Maltoni, C., Lefemine, G., Ciliberti, A., Cotti, G . , and Carretti, D. (1981).
Carcinogenicity bioassays of vinyl chloride monomer: a model of risk
assessment on an experimental basis. Envir. Health Perspect., 41, 3-29.
MRC Environmental Epidemiology Unit (1981) The carcinogenicity and mutage-
nicity of wood dust. Scientific Report No. 1. MRC, Southampton.
MRC Environmental Epidemiology Unit (1984) Expected numbers in cohort
studies. Scientific Report No. 6. MRC, Southampton.
O’Berg, M.T., Burke, C. A., Chen, J. L., Walrath, J., Pell, S . , and Gallie, C. R.
(1987). Cancer incidence and mortality in the Du Pont Company: an update.
J . Occup. Med., 29(3), 245-252.
Paddle, G. M. and Bennett, B. (1983). An investigation of an excess incidence of
heart disease. Stat. Med., 2, 477-484.
Schulte, P. A., Ehrenberg, R. L., and Singal, M. Investigation of occupational
cancer clusters: theory and practice. Am. J . Public Health, 77(1), 52-55.
Shore, R. E., Iyer, V., Altshuler, B., and Pasternack, B. S. (1992). Use of human
data in quantitative risk assessment of carcinogens : Impact on epidemiologic
practice and the regulatory process. Reg. Toxicol. Pharmucol., 15(2),
190-22 1.
Spitzer, W. 0. (1991). Meta-meta analysis: Unanswered questions about
aggregating data. J . Clin. Epidemiol., 44(2), 103-107.
World Health Organisation (1977). ‘Manual of the International Statistical
Classification of Diseases, Injuries, and Causes of Death’. WHO.
World Health Organisation (1985). Neurobehavioural methods in occupational
and environmental health. WHO, Copenhagen.
CHAPTER 21
Information and Consultancy

Services in Toxicology
D. M. CONNING
1 Introduction
Toxicology holds a unique position among the natural sciences. Not only
does the solution to a toxicological problem usually depend on the
assembly of data from several disciplines, the solution is itself often used
to resolve a practical problem concerning human health or safety. Such
problems may arise in relation to the general public with respect to
community health or consumer safety; they may be confined to the
industrial setting and relate to occupational health; or they may be
concerned with a potential environment hazard. Toxicology supplies the
data with which such problems are resolved and thus contributes directly
to decisions affecting human health.
Information Services in toxicology, therefore, have to meet two
important obligations-to ensure first that available data are assimilated,
and second that the quality of the data is subject to expert judgement.
These obligations apply, of course, to any issue where scientific judge-
ment is involved but are particularly onerous for the information
consultant in toxicology because of the wide spectrum of technical data to
be utilised and the added importance of the potential human danger.
A third important aspect of a toxicological information service are the
possible vested interests of those using the resultant data or opinion. In a
society that places a high value on the quality of the lives of its citizens,
three types of vested interest can be identified. The first is the
manufacturer eager to sell his wares. A great many obstacles in relation
to health and safety are imposed on anyone attempting to bring a new
chemical product to the market. These, increasingly, exist on an
international as well as a national basis and, although great strides have
been made by international regulatory authorities to ‘harmonise’ their
regulations, there are still many areas where conflict exists and possible
barriers to trade occur. There is always the possibility that a manufac-
485
Table 1 Examples of the wide range of databases available on toxicology and related subjects
Hostslsystems providing
Examples of databases access include:
Toxicology : TOXLINE includes Environmental Mutagen Information

Centre file, Environmental Teratology
Information Centre file, Pesticide Abstracts [ll
subfiles]
RTECS NIOSH Registry of Toxic Effects of Chemical
Substances z
TDB
SPHERE
Toxicology Data Bank
Scientific Parameters for Health and the
BLAISE-LINK
[NLM computer, Bethesda] 8g.
Environment 0
CTCP Clinical Toxicology of Commercial Products BRS [BRS, New York] 3
a
ECDIN Environmental Chemicals Data and 3
A
Information Network
OHM-TADS Oil and Hazards Materials Technical Assistance 2
E
Carcinogenicity : CANCERLIT
Data System
Carcinogenesis Abstracts and Cancer Therapy
Abstracts
DATA STAR [Radio-Suisse, Berne] '2
CANCERPROJ Cancer research projects in progress Q
h
CCRIS Chemical Carcinogenesis Research Information 2
System s.
c,
Biomedicine : MEDLINE includes Index Medicus DIALOG [Lockheed, Palo Alto] 2
h.
EMBASE Excerpta Medica 3
BIOSIS PREVIEWS Biological Abstracts and Bioresearch Index DIALTECH [IRS, Frascati, Italy] y
Chemical Information : CHEMLINE, CHEMNAME, etc. Dictionary files $:
ECOIN K
inventories DIMDI [Koln, W. Germany] 2
TOSCA
CA SEARCH Chemical Abstracts 6
Pharmaceuticals : IPA International Pharmaceutical Abstracts P
MERCK INDEX ECDIN [Datacentralen,
Copenhagen] is
Agriculture:
MARTINDALE ONLINE
AGRICOLA represents holdings of US National Agricultural EPA/NIH-CIS[ISC, Washington]
s
Library 2.
0s
CAB Abstracts Commonwealth Agriculture Bureau
AGRIS agriculture, forestry, fisheries, food
Food: FSTA Food Science and Technology Abstracts INFOLINE [Pergamon -Infoline,
London]
FOODS ADLIBRA developments in food technology, packaging, efc.
Environment : ENVIROLINE pollution, contamination, efc. ORBIT [SDC, Santa Monica]
POLLUTION ABSTRACTS
Packaging : RAPRA ABSTRACTS commercial, technological and research aspects QUESTEL [Telesystemes, Paris]
of rubber and plastics
PIRA ABSTRACTS Paper and Board Abstracts, Printing Abstracts,
Packaging Abstracts
PSTA Packaging Sciences and Technology Abstracts
Legislation /Guidelines: CRGS Chemical Regulations and Guidelines System; and others
USA
FEDREG Federal Register Abstracts
Paklegis national, international and EEC regulations
affecting packaging
HSE-line legislation, industrial health and safety; UK
and many more
488 Information and Consultancy Services in Toxicology
turer will be impatient of what he sees as unnecessary bureaucracy and be

tempted to diminish the value or importance of having his product
properly evaluated for potential hazards. He must be protected against
this very understandable tendency and one line of protection is the
information service.
The second interest arises in that government departments charged
with the protection of public health cannot afford to ignore any data that
might have a bearing on the problem. The personnel involved must
acquaint themselves with all aspects of the issues-no easy task when few
have any formal training in the subject. In these circumstances there has
been a tendency to establish the ‘checklist’ approach in which a series of
items of data are devised and applied to any particular issue. Although
this is very satisfactory administratively, it undermines the requirement
for expert judgement of difficult biological issues and there has been a
tendency in some departments for the ‘checklists’ to be sustained when
the ‘science’ has left them far behind, because of the urge to protect the
system of regulatory control.
The third vested interest is that of the research scientist in toxicology
who sees a number of exciting biological problems which have to be
resolved before definitive decisions on product-safety can be made. Such
a man, dismissive or impatient of the need for pragmatism in resolving
the issues in a reasonable time, may, if unrestrained, impose many
difficulties in terms of unavailable or unexplained data.
An information service has a duty to serve all of these groups and has
the added difficulty of remaining independent of each. The independent
but accurate assessment of toxicological information is thus the main
objective of any information service and this chapter outlines the basic
considerations needed in the attempt to achieve it.
2 Data Acquisition
The data to be used by the information section should relate not only to
available data but to data that is likely to accrue from on-going
experiments-and that may become available by the time a given product
is in use. Therefore, it is of great value if the information personnel are in
touch with toxicology research workers who operate in the various
disciplines concerned, and who are aware of the activities in their field.
With the advent of computerised systems, the acquisition of data has
been revolutionised in the last 10 years. There are now a large number of
data bases which can be tapped by anyone willing to purchase the access
(Table 1). There are, in addition, sources of information on on-going
experiments in toxicology being run by or on behalf of regulatory
authorities. These sources are not yet widespread and there is still a
problem of duplication of toxicity studies, but the position is likely to
improve as regulatory requirements become more international in their
scope.
D. M. Conning 489
In addition to computerised data, there is still a large amount of

toxicological information in conventional form in the older literature and
this cannot be ignored in any serious appraisal. Here too, it is now easier
to acquire copies of older documents from centralised systems (such as
National libraries), but the costs can be substantial.
3 Data Interpretation
Ideally, information personnel researching the data on a particular
compound should be able to obtain papers reporting high-quality
experimental work on biological effects pertinent to the intended use and
thus to the expected human exposures, and to define safe dosage and
conditions. In many respects this ideal was nearly achieved before the
advent of computerised data bases, because information personnel then
had to search personally to acquire relevant papers and consequently the
objectives were more narrowly defined and the information was more
restricted. Now it is easy to obtain a list of all the papers that might
conceivably be connected and the problem is to sift them for their true
pertinence. This task can assume massive proportions because of the
considerable expansion of the data now considered relevant to toxicolog-
ical assessments and the tendency for papers merely to record technical
data while presenting a diminishing amount of truly scientific informa-
tion. It remains of paramount importance to interpret results in the light
of human need and, for this, good science is essential.
4 Data Dissemination
Most assessments of toxicological data are in response to specific queries
on certain compounds or groups of compounds, or on certain functional
effects. Apart from the accurate resolution of potential hazards, there is a
need to ensure that the recipient of the report understands its content
and implications.
Two factors must be considered in the latter connection. First, the
recipient may not be a trained toxicologist and the data must be
expressed in terms that he understands, i.e. in ‘lay’ language if necessary.
This often requires considerable skill and is helped by a thorough
understanding of the second factor-that the information provided must
relate to the problem giving rise to the request. The information
consultant must seek to relate the data he acquires to the intended
application. Thus, a problem concerning the hazards of exposure by
inhalation in the work setting will usually not benefit from a treatise on
teratogenicity or on whether or why the compound shows greater
carcinogenicity towards one tissue than to another.
Such considerations give added impetus to the assertion that a proper
understanding of toxicology data in relation to the actual problem is
imperative; hence the need for interpretive expertise to be available.
490 Information and Consultancy Seruices in Toxicology
Apart from the response to specific enquiries, an information service is

much enhanced in value by the provision of a service in current
awareness. Regular bulletins covering contemporary developments in
specific fields in respect of both new science and new regulatory
developments can be very useful. It is helpful if these, too, are written in
‘lay’ language but, again, the emphasis should be on quality and the
identification of data that allow real conclusions to be drawn on health
and safety problems for man and do not merely stimulate academic
debate about interesting biological problems.
Finally, the very nature of information work leads to the compilation
of surveys of available data about particular compounds. Such surveys,
particularly if conducted in critical fashion, represent invaluable compen-
dia of available data and are too often allowed to languish in archives.
This results in duplication of effort and the loss of opportunity for each
survey to build on the last. Wide dissemination has the added value that
surveys undertaken for a particular reason (and thus from a restricted
viewpoint) can often stimulate a cross-fertilisation of ideas when the topic
is approached from another direction. Thus, the evaluation of a
compound for its genotoxic potential may present data of value to the
toxicologists, to the expert in carcinogenicity, or to the enzymologist.
Wherever possible, such surveys should be made available by publication
or other means.
5 Summary and Conclusions

This chapter has attempted to enumerate the principles on which an
information service should be based. The most important principle is that
the scientific work generating the information should be of high quality
and should be interpreted in relation to the human hazards envisaged.
This requires a knowledge of the circumstances in which the opinion is to
be used, and access to toxicological expertise which may be of a specialist
type. It is essential that papers in the literature are read carefully and that
opinions are not based mainly on available abstracts.
Every effort should be made, within the bounds of confidentiality, to
disseminate the results of surveys as widely as possible, and advice and
consultancy should be couched in non-specialist language where possible.
Current awareness surveys , especially involving data from different
sub-specialists, are to be encouraged if resources allow. The acquisition
of many references now possible through computerised data banks is
useless if the results are not vetted scientifically.
CHAPTER 22
Regulations and Advisory

Requirements in Relation to
Food
D. M. CONNING
1 Introduction
The development of toxicology as a separate biological science has been
dependent largely on its role in underpinning the laws and regulations
designed to safeguard human health and safety. Such regulations are now
commonplace throughout the world and cover most aspects of human
wealth-generating activity. Thus, there are regulations designed to
safeguard the health of the individual in industry, in agriculture, and in
nearly all aspects of consumer affairs.
That branch of toxicology dealing with this aspect-namely Regulatory
Toxicology-is beyond the scope of this book but it is important that any
student of toxicology is aware of the ramifications of the regulatory
system. What follows is an account of the principles and development of
regulations as applied to the safety of foodstuffs, the most developed part
of the regulatory system, which will serve as an illustration of the
complexities involved.
A The Origin of Food Laws

The laws on food safety differ considerably from one country to another,
reflecting their own particular needs and experiences, The basic prin-
ciples of food law, however, are the same in all countries, and are
designed to protect public health and to prevent fraud.
The countries of Western Europe and North America have the most
comprehensive and complicated food laws, many of which have been
widely adopted by other countries. Developing countries are showing an
increasing tendency to develop specific food laws that are more in
keeping with their ethnic requirements.
49 1
492 Regulations and Advisory Requirements
This chapter concentrates on the laws in developed countries with

particular emphasis on the laws applicable in the USA and United
Kingdom. The underlying principles of the laws are explained. Although
food laws are subject to frequent change, it is unlikely that the broad
principles will differ appreciably.
The present food laws evolved from attempts to legislate against the
adulteration of foodstuffs, a practice that was fairly common in the
nineteenth century. The slow progress between the first food acts of the
mid-18th century and those of the modern era was largely due to the lack
of analytical expertise with which to enforce the various acts. As such
methods have been developed, so has the extent of the regulations been
increased.
2 Evolution of Food Law Exemplified by the

United Kingdom
The evolution of food law in the United Kingdom resembles that seen in
most other developed countries. Early in the nineteenth century, no laws
existed to protect a consumer from toxic substances that were present in
their food. Fraudulent trade practices, such as the dilution of cocoa with
farina, coffee with chicory, or milk with water, were not unusual.
Formaldehyde and boric acid were permitted as food preservatives and
lead compounds were used as food colours. The increasing demand for
laws to prohibit the adulteration of food carried with it the need for
government inspectors and analysts, and was dependent on the availabil-
ity of analytical techniques.
As reliable methods became available medical practitioners, scientists,
and social reformers began to press for governmental legislation, but
there was little activity until the so-called ‘Bradford Incident’ of 1858
when as many as 200 people suffered an outbreak of poisoning.
The first General Act to Prevent the Adulteration of Food and Drink
was passed in 1860 and several more parliamentary bills and acts
culminated in the ‘Sale of Food and Drugs Act, 1875’.
The first laws relating to food colouring materials were of limited
value. A list of those substances which were not allowed in food included
lead compounds, arsenic, picric acid, and Manchester yellow, which
from experience were known to be hazardous. A list of substances
permitted for use in food was not available so that compounds such as
artificial dyes, which had been developed for the textile industry, could
be used in food. In due course, a list of permitted food colours was
produced but the list of prohibited additives was retained.
Significant factors in the decline of food adulteration were the
improved means for enforcing government legislation, more stringent
analytical methods, and better qualified personnel. The increase in
scientific knowledge enabled specific laws on food additives to be made,
D. M. Conning 493
relating the level of use of an additive to that which is unlikely to be

associated with human risk.
By the time the Food and Drugs Act was passed in 1955, the attitudes
of food manufacturers, wholesalers, and retailers had changed ap-
preciably. As a consequence of improved means of enforcing food laws,
and continuing surveillance by consumer protection groups, detailed and
comprehensive food legislation was included in the Food Act, 1984. Since
then, however, a number of problems relating to animal health and food
hygiene, together with the increasing requirement to ensure compatibility
throughout the European Community, resulted in the Food Safety Act,
1990. This is an enabling Act which allows the government to make
regulations governing any part of the food chain.
3 Structure of Food Laws

A Food Additives
In general, two approaches have been used for the control of additives.
Most common is the ‘positive list’, that is, a list of those substances which
may be used in food for the purpose of processing, preserving, or
colouring. In principle, this system is adopted by all developed countries
for most additive classes.
To a lesser extent, the ‘negative list’ system, which specifies only those
substances which may not be added to food, has been used and is
increasingly being replaced by the positive list.
B Standard of Composition
In most countries where food quality is controlled by law, the staple
foods and certain luxury items are frequently subject to a standard of
composition. This normally specifies minimum levels of important
constituents (for example, the meat or fruit content) and the additives
that may be used. Standards of composition are intended usually to
prevent fraud rather than to protect health. In recent times, with the
advent of comprehensive food and nutrition labelling, composition
standards are being phased out.
C Codes of Practice
To cover all aspects of food quality by specific laws would be a
formidable task. In those areas where some common standards are
required but specific legislation is not justified, Codes of Practice have
often been adopted which, while not legally binding, are widely accepted
by the food industry. They specify an agreed standard and if food of a
lower standard is marketed it may be construed as an attempt to defraud
the customer. Codes of Practice are popular with industry because they
are flexible, less expensive to enforce, and are voluntary.
D Enforcement
To be effective, any law should not only be readily understood by those
to whom it applies, but should be enforcible.
Specific methods relating to food additives are frequently dependent on
the technical ability of the industry (for example, the accuracy by which a
curer can control the level of nitrite present in a finished product, or how
evenly can the nitrite be distributed throughout the meat). The accuracy
and reliability of the analytical methods available are also important,
since the maximum permitted levels of a substance may be directly
related to the limits of detection by the analytical methods available at
the time. Purity criteria for food additives are usually decided in
consultation with manufacturers, since they are in the best position to
know what level of purity is achievable. Maximum levels of permitted
food additives depend on the levels necessary to achieve the desired
technological effect, within the constraints of safety.
Usually, the government department responsible for enforcing food
law (which may be the local authority) will appoint analysts and
inspectors who take samples at all places where the food is prepared,
manufactured, or offered for sale. Failure to comply with food law
usually constitutes a criminal offence; this is the case in the United
Kingdom.
E Amending Legislation
Food laws throughout the world are subject to frequent changes due to
the addition of new additives, the restriction in the use list of substances
which, on the basis of new information, are found to be toxic, and the
revision of regulations on package labelling. Food manufacturers them-
selves may initiate changes by the process of petition of the government
department concerned.
F Analytical Methods
These may be specified but often this is not the case, as laws cannot be
changed with sufficient speed to keep pace with technicological advances.
If an analytical method is not specified by law, an analyst may use the
best method available at the time.
4 Naturally Occurring Hazards in Food

Although most developed countries agree on the need to have laws to
control the use of additives in food, there are few laws relating to
naturally occurring hazards. Such laws that do exist usually relate to
heavy metal contamination which, though unavoidable, it may be
possible to reduce. For example, in the United Kingdom there is a
maximum level of lead permitted in foods.
The presence of other contaminants is sometimes controlled by
D.M. Conning 495
specifying the ‘level’ above which action will be taken on the basis of
hazard to health. The FDA, for example, have listed a number of food
groups with action levels. Amongst the food groups covered are: fish and
shellfish; flours and cornmeals; fruit chocolate, and chocolate products;
nuts and spices; vegetables. Some examples of the defect action levels are
insect fragments in cornmeal and Drosophilu fly eggs in canned tomatoes.
Countries rarely specify microbiological standards for use in the food
industry. (Exceptions include France and the USA, where certain foods
are controlled by individual standards.) The principal reason for this is
that methods by which microbiological contaminations are determined
tend to vary greatly and they are insufficiently precise for legislative
purposes. In practice, government inspectors concerned with food-quality
adopt their own standards for regulating food. As a consequence,
advantage may be taken of new microbiological assays as they become
available without restriction by legislative processes. With continually
changing methodology it would be counter-productive if the regulations
specified a method that might quickly become out-dated.
The absence of microbiological standards in food law does not diminish
the protection of the consumer, given that the basic Food Acts in all
countries state that food must not endanger human health. Most
countries have regulations controlling food hygiene and officials are
authorised to inspect food premises to ensure that good hygienic practices
are maintained.
The number of countries introducing legislation setting maximum
tolerances in food for the common mycotoxins, particularly aflatoxins, is
growing steadily. Countries without any legislation in this area do not
necessarily tolerate mycotoxins; their presence at toxicologically un-
acceptable levels would be an offence under the general provisions of
Food Acts which forbid the sale of any food that constitutes a danger to
human health.
5 Food Additive Legislation

In the USA a food additive has been defined as any substance, the
intended use of which results or may reasonably be expected to result,
directly or indirectly, either in their becoming a component of food or
otherwise affecting the characteristics of food. A material used in the
production of containers and packages is subject to the definition if it
may reasonably be expected to become a component? or to affect the
characteristics, directly or indirectly, of food packed in the container.
A Legislation on Food Additives in the USA

All comments relating to food legislation in the USA refer only to the
Code of Federal Regulations. The separate food laws which exist in the
individual states are not considered here. If the food complies with the
federal law then it should be accepted in the individual states.
The Federal Food, Drug and Cosmetic Act specifies that ‘no additive
shall be deemed to be safe it is found to induce cancer when ingested by
man or animal, . .
This is the so-called ‘Delaney clause,’ which has been the cause of
much debate since it was enacted in 1958. Also in 1958, the food additive
amendment was passed and a regulation was made listing specified food
additives with ‘Generally Recognised as Safe’ (GRAS) status. If an
additive appeared on the list then it was considered suitable for food use.
The GRAS list consisted of about 500 food additives divided into
categories, for example chemical preservatives, emulsifying agents,
nutrient and/or dietary supplements, sequestrants, stabilizers, anticaking
agents, spices and other natural seasonings and flavourings, essential oils,
oleoresins, and natural extractives.
The food additives appearing in the GRAS list were those which had
been in common use. Although most of the GRAS-listed substances were
without maximum levels of use, other than those in accordance with good
manufacturing practice, there were restrictions laid down for some, such
as 0.1% maximum for benzoic acid or the prohibition of the use of
sulphurous acid and its salts in meats or foods recognised as a source of
vitamin B1.
After the GRAS list regulation came into force, any food additives
which were not prior sanctioned, that is in common use before 1958 and
which were not on the GRAS list, could not be used until a specific
regulation was passed permitting their use. Often such regulatory
approval could only be achieved by providing detailed results from a
comprehensive programme of toxicity testing.
A number of regulations were passed subsequently permitting a further
large number of food additives, including emulsifiers (tweens and spans) .,
stabilisers (carrageenan, furcelleran) , chewing gum base , modified
starches, and indirect additives including boiler water additives, chemi-
cals used in the washing or lye peeling of fruits and vegetables, and
chemicals for controlling micro-organisms in cane and beet-sugar mills.
A food ingredient of natural biological origin, which has been widely
consumed for its nutrient properties in the United States prior to 1
January 1958, without known detrimental effcts, and which is processed
by conventional means as practised prior to 1 January 1958 and for which
no known safety hazard exists, is regarded as GRAS without it being
specifically listed. (However, a food which has been widely consumed in
another country may not necessarily be accepted as GRAS in the United
States.)
In 1969, the FDA decided that all the substances on the 1958 GRAS
list should be re-evaluated because they were originally added to the list
without a detailed examination of all available information. The review
was conducted by the Select Committee on GRAS Substances of the Life
Sciences Research Office, Federation of American Societies for Experi-
D.M . Conning 497
mental Biology (FASEB). Its reports on individual GRAS substances

were made available for comment at public hearings.
It is inevitable that food ingredients which have been in use for some
time, and hitherto considered safe, should sometimes have their safety
questioned when additional information becomes available. New safety
evidence may not be conclusive but once an additive has been banned, it
is difficult to reinstate it. To cover this situation, the US authorities
invoke the concept of an interim food additive regulation. Additives for
which new information indicates a possible hazard, but where there is still
reasonable certainty that the substance is not harmful, are given interim
status and may continue to be used pending further study. An additive in
this situation will eventually be deleted from the permitted list unless
toxicological work is initiated to prove its safety. When the work is
finished the additive will either be reinstated or deleted.
B Purity Criteria for Food Additives

The introduction of statutory purity criteria for food additives is
increasing. In the USA, all GRAS listed additives must comply with the
purity criteria laid down in the Food Chemicals Codex. Other additives
often have purity criteria specified in the regulation permitting their use.
In the UK there are purity criteria for all the permitted additives and
often these are the Food Chemicals Codex criteria or those specified in
EEC Directives on food additives. The EEC has introduced criteria for
all of the permitted additives covered by Directives and these have been
adopted by the individual Member States.
One of the most important contributions on purity criteria has come
from the Joint FAO/WHO Expert Committee on Food Additives
(JECFA), a committee that has evaluated the acceptability of a wide
range of food additives. Its reports are published annually, and are
accompanied by monographs and recommended specifications. These
purity criteria are used by many countries as guidelines in the absence of
national legislation.
C Regulations on Specific Additives

Antioxid ants
The regulations on the addition of antioxidants to foods are similar to
those on food preservatives, in that it is usual to limit the use of
antioxidants to a short list of specified foods and to set maximum levels of
addition of each antioxidant in each food. The restriction on the ‘natural’
antioxidants, such as ascorbic acid, are usually less than those for the
synthetic antioxidants.
The most commonly permitted antioxidants are: butylated hydroxy-
anisole (BHA); butylated hydroxytoluene (BHT); propyl, octyl, and
dodecyl gallates (gallates); ascorbic acid; ascorbyl palmitate;
tocopherols.
49 8 Regulations and Advisory Requirements
A rtijicial Sweeteners
Legislation on artificial sweeteners has received more publicity than any
other food additive. The prohibition of cyclamates in the USA in 1969
was swiftly followed by a world wide prohibition. Even countries which
previously had no food additive legislation introduced laws banning
cyclamates. The subsequent doubts about the validity of the USA ban led
national governments to be more circumspect in their consideration of
permitted lists. Consequently when, in 1971, the USA authorities
threatened to introduce severe restrictions on the use of saccharin
amounting to a virtual ban, other countries were reluctant to follow suit
and most of those which did permit saccharin still continue to do so.
In most European countries, the use of saccharin as a food additive is
limited to soft drinks and slimming foods. In the United Kingdom, there
is no such restriction and saccharin is generally permitted in foods, except
those covered by a standard of composition; its use in soft drinks is
subject to maximum permitted levels. Other artificial sweeteners, such as
acesulfame and aspartame, have recently been added to the UK
permitted list, a reflection of the more liberal regulatory stance to this
additive class being seen world-wide.
The legislative history of sweeteners indicates the dilemma governments
experience trying to balance the risk and benefits of additives. In theory,
as saccharin has been shown to cause cancer, the Delaney clause should
have ensured a ban under USA law. In practice, the American Congress
recognises that the dangers from overweight are real and the dangers
from saccharin, whilst not easy to define, are probably very small; they
therefore wish to keep saccharin as a permitted additive until the position
is clarified.
Colouring Material
Most countries have lists of permitted colouring materials, even if they
have very little else in the way of other food legislation. The addition of
colourings to food has long been a source of concern to consumer
pressure groups in several countries. Safety testing has been carried out
to a greater extent on food colours than any other class of additive with
the consequent steady reduction in the numbers of permitted substances.
The most commonly permitted synthetic colourings are Ponceau 4R,
Carmorsine, Amaranth, Erythrosine, Tartrazine, Sunset Yellow, Quino-
line Yellow, Greens, Indigo Carmine, Patent Blue V, and Black DN.
The ‘natural’ colourings often subject to legislative control include
annatto, caramel, vegetable carbon, carmine, carenlenoids, iron oxide,
and titanium dioxide.
Caramel. The name caramel covers a wide range of different products,
most of which cannot be chemically defined. Some countries have
attempted to distinguish between the different types of caramels and to
D.M. Conning 499
introduce specific maximum levels for them. The Danish and Swedish
colouring regulations‘ differentiate between caramels made by the am-
monia process and non-ammonia caramels, setting higher maximum
permitted levels for the latter. Both countries also specify maximum
levels of 4-methyl imidazole in the caramels themselves.
In the United Kingdom, there have been suggestions over the past few
years from the Ministry of Agriculture, Fisheries and Food that caramels
should be divided into four types for the purpose of regulations: burnt
sugar; caustic caramel (catalysed by sodium hydroxide) ; ammonia cara-
mel; and ammonium sulphite caramel. Within these types, the ‘catalysed’
caramels should fall within six basic specifications. Once the numbers of
caramels available have been reduced in this manner, it will be possible
to carry out suitable safety testing.
EmulsiJiers and Stabilisers

The number of permitted emulsifiers and stabilisers is greater than any
other group of food additives. They include: sorbitan esters; polyoxy-
ethylated sorbitan esters; cellulose ethers; alginic acid and its salts;
and phosphates.
Flavourings and Flavour Enhancers

When considering the legislations on flavourings, it is necessary to
recognise a few commonly accepted definitions:
Natural flavourings are obtained from plant or animal sources and are
used as such or extracted and concentrated by purely physical
methods.
Nature-identical flavourings are substances obtained by synthesis or
isolated by chemical processes from raw materials but are chemically
identical to substances present in natural products.
Artificial flavourings are substances which have not yet been identified
in nature.
The problem legislators face in this area is the very large number of
substances involved. Many natural flavourings are complex mixtures and
the composition varies according to the source, time of harvest, and other
factors. Artificial flavourings and nature-identical synthetics can usually
be precisely defined chemically, and purity criteria established.
The International Organisation of the Flavour Industry (IOFI) has
recommended that flavourings should be controlled by a ‘mixed list’
system of legislation. This would consist of permitting all natural
flavourings except those shown on a list of prohibited substances or on a
restricted list (e.g. maximum tolerance in food or restricted to certain
foods only). Nature-identical flavourings would be treated in the same
way as natural flavourings. Artificial flavourings would be controlled by a
positive list and only those substances on the list would be permitted,
with any necessary restrictions. It is suggested by IOFI that this would
solve the problem arising from the enormous number of flavourings. It

would offer some legislative control of flavourings but would still permit
freedom to use a wide variety of substances in food to suit ethnic tastes.
JECFA has evaluated a number of flavourings for food safety. The
Council of Europe (which is also part of the United Nations) published in
1970 a Partial Agreement on Flavourings, drawn up by a Working Party
supported by the Member States of the European Community and
Switzerland. This lengthy document, which was revised and reissued in
1973 and is still under review, contains a list of artificial flavourings
considered suitable for use in foodstuffs without hazard to health; a list of
natural flavourings admissible for use in foodstuffs, based on their
sources; a list of flavourings considered harmful; and criteria for the
toxicological evaluation of new flavourings. The lists include natural and
artificial flavourings which may be temporarily added to foodstuffs
without hazard to health, and a list of artificial and natural flavouring
substances not fully evaluated.
The 1976 United Kingdom recommendations for flavourings legislation
are based largely on the Council of Europe document. They inspired
much opposition and no move has been made to adopt them, partly due
to the fact that the European Community is also working on proposals
for flavourings legislation.
In Belgium, Denmark, and West Germany there are lists of flavouring
substances which may not be used in food. These include: coumarin,
tonka bean, safrole, sassafras wood, birch tar oil, bitter almond oil
containing hydrocyanic acid, thujone, cade oil, and calamus oil contain-
ing more than 10% asarone.
In the USA there are a number of natural and artificial flavouring
substances which are considered ‘GRAS’. The extension of the original
GRAS list was due in no small way to the Flavouring Extract Manufac-
turers Association of the USA (FEMA), which evaluated the safety of
flavourings required by industry and petitioned the FDA to have them
included in the regulations. Not all the flavourings considered to be safe
by FEMA feature in the appropriate section of the Code of Federal
Regulations; FEMA publish their list.
The most widely known and widely permitted flavour enhancer is
monosodium glutamate (MSG). Repeated investigations into the safety
of MSG do not appear to have greatly restricted its use, although in many
countries it is not permitted in infant foods. In some cases, manufacturers
have voluntarily ceased to use MSG.
Hydrolysed vegetable protein and, to a lesser extent , hydrolysed
animal protein are used widely but they are usually exempt from food
additive legislation, although their use may be controlled by standards of
composition of the food. The lesser used flavour enhancers are the
sodium salts of 5‘4nosinic acid and 5‘-guanylic acid.
D . M . Conning 50 1
Miscellaneous Food Additives

Food acids. The most commonly permitted food acids are: acetic, adipic,
citric, fumaric, lactic, malic, tartaric, succinic, hydrochloric, phosphoric,
and sulphuric.
The laws on food acids do not usually specify maximum levels and the
use of food acids is often not restricted to specific foods.
Bases. The most commonly permitted bases are: carbonates, hydroxides,
phosphates, citrates, lactates, tartrates, acetates, and malates.
These are treated in the same way as food acids.
Anticaking Agents. The most commonly permitted are: silicon dioxide,
magnesium silicate, sodium silicoaluminate, calcium silicate, ferric am-
monium citrate, sodium and potassium ferrocyanide, tricalcium phosph-
ate, silica gel, and magnesium oxide. They are restricted usually to such
as powder mixes for vending machines, instant desserts, and salt.
Antifoaming Agents. The most commonly permitted are: dimethylpoly-
siloxane and vegetable oils.
Firming agents. Calcium chloride and calcium lactate are often pemitted
as firming agents in pickles and canned vegetables, for example.
Glazing agents. The most common use of these is in sugar confectionery;
the typical glazing agents permitted are: beeswax, spermaceti, carnauba
wax, magnesium and calcium stearates, paraffin wax, and shellac.
Release agents. Some countries do not permit release agents other than
vegetable oils, others permit mineral hydrocarbons (UK and USA),
oxystearin, and sperm oil.
Humectants. Sorbitol and mannitol are widely permitted for this purpose.
Clarifying or filtering agents. Diatomaceous earth, bentonite, kaolin, and
tannic acid are commonly permitted in a few countries which allow
polyvin y lpolypyrrolidone .
Flour improvers and bleaches. Of the few countries which do permit flour
treatment (e.g. USA, UK, and Canada) the permitted substances
include: acetone peroxide, benzoyl peroxide, potassium bromate, am-
monium or potassium persulphate , chloride dioxide, azodicarbonamide ,
L-cysteine hydrochloride monohydrate, sodium and calcium stearoyl-2-
lactylate, oxides of nitrogen, nitrosyl chloride, and sulphur dioxide.
Wetting agents. Dioctyl sodium sulphosuccinate is a permitted wetting
agent in a few countries.
Clouding agents for soft drinks. Brominated vegetable oils have largely
fallen from grace world-wide. Glycerol ester of wood rosin, glyceryl ester
of colophony , and sucrose diacetate hexaisobutyrate are permitted in a
few countries.
Propellants. Carbon dioxide, nitrous oxide, chloropentafluoroethane, and

octafluorocyclobutane are the most widely permitted propellants for
food aerosols (mostly whipped cream); butane, propane, and chloro-
fluorocarbon propellants are also occasionally permitted.
Packaging gases. Nitrogen and oxygen are the usual gases permitted.
Sequestrants. Ethylenediamine tetraacetic acid (EDTA) and its salts, and
calcium phytate sequestrants are permitted under specified conditions in
a few countries. Phosphates are very widely permitted in many foods in
most countries.
Liquid freezants in contact with food. Dichlorodifluoromethane is per-
mitted in a few countries only.
Preservatives .
The legislation on preservatives in food is similar in most countries. The
most commonly permitted preservatives are: benzoic, sorbic, and propi-
onic acids and their sodium, potassium, and calcium salts; ethyl, propyl,
and methyl para-hydroxybenzoic acid esters and their sodium salts;
sulphur dioxide and various sulphur dioxide-generating salts; sodium and
potassium nitrate and nitrite; nisin; thiabendazole; hexamine; ortho-
phenylphenol; and pimaricin.
As with the antioxidants, the use of preservatives is usually restricted
to a narrow range of specific foods.
Nitrite. Since the demonstration of the possible formation of carcinogenic
nitrosamines in nitrite-cured meat, governments have been trying to find
ways of reducing the maximum levels of nitrite that can be safely
permitted in cured meat, without incurring the danger of the growth of
Clostridium botulinum whose toxin is fatal to man. The search for a safe
processing method using minimal levels of nitrite is continuing.
Ethylene oxide. Ethylene oxide is used for the fumigation of certain foods
such as spices. It is not a permitted preservative but rather an insecticidal
substance. Doubts over safety have resulted in a recent reduction in
legislative approval.
Processing Aids
There is no clear distinction made between a processing aid and a food
additive. Processing aids are an unpredictable area as regards the
pressure of legislation; some countries do not legislate for them at all and
others, such as the USA attempt to legislate in considerable detail.
Specific usage additives. The following list of ‘specific usage additives’
taken from the USA regulations provides a picture of the type of
substance which would fit the concept of processing aid: boiler water
additives used in the preparation of steam which will contact food;
D. M . Conning 503
chemicals used in washing or to assist in the lye peeling of fruit and

vegetables; chemicals for controlling micro-organisms in cane sugar and
beet-sugar mills; defoaming agents; sanitising solutions; flocculents for
the clarification of beet sugar and cane sugar juice and liquor; ion-
exchange resins; molecular sieve resins; extraction solvents; for the
extraction of spice oleoresins; for decaffeination of coffee; and enzymes.
6 Pesticides
A large number of insecticides, fungicides, and herbicides are used for
treating crops in many parts of the world. Because the residues of these
substances persist in human food, it is necessary for countries to control
their use. It is essential to keep the level of residues as low as possible,
therefore maximum tolerances are laid down.
International organisations such as the FAO/WHO Codex Alimen-
tarius Commission, JECFA, and the E E C have produced lists of
tolerances for pesticide residues in or on food. The maximum permitted
residue levels generally lie within the range 0.1 mg kg-' to 20 mg kg-',
with the majority in the 1mg kg-' zone.
7 Packaging Materials
This term includes any material or equipment intended to come into
contact with food. Many countries do not have detailed regulations on
packaging materials; the laws in the USA and the Netherlands and the
recommendation of the BGA in West Germany often act as guidelines
for those countries without their own packaging laws.
The first EEC Directive o n materials arid articles intended to come into
contact with food was brief and merely specified the basic concept that
nothing which is harmful to human health or detrimental to the food shall
migrate from the packaging material into the food. It is intended that
subsequent directives will specify details of constituents of plastics and
regenerated cellulose, glass, ceramics, and stainless steel. A limit for the
tolerance of vinyl chloride monomer in materials and articles in contact
with food to and in foods has been agreed. Global limits for the
migration of harmful substances from materials into food will probably be
set. Maximum tolerances for migration of certain heavy metals from glass
and ceramics have been set. Discussions are underway to produce E E C
lists of approved monomers and additives for food-contact plastics.
A The Netherlands
The Dutch have a sophisticated system of regulations broadly similar in
form to the BGA lists; a series of positive lists has been drawn up for
each commonly used plastic as well as other packaging surfaces. Many of
the additives are controlled by specific migration criteria.
B United Kingdom
The regulations reflect the existing EEC directives, and as such generally
lack specificity. There is a Code of Practice drawn up by the British
Plastics Federation in conjunction with the British Industrial Biological
Research Association. This code, first published in 1969 and revised
periodically, provides recommendations on the safe use of ingredients of
plastics for food contact applications. In drawing up some of the
recommendations, the legislation in other countries is taken into account.
The Code includes specifications for a number of polymer classes.
C USA
Several types of food-contact materials (i. e. indirect food additives) are
specifically covered by US regulations, including plastics, rubbers, paper
and paperboard, cotton, and cotton fabrics. The indirect food additive
regulations comprise positive lists which are published in Title 21 of the
US Code of Federal Regulations. These regulations cover the two main
components of packaging materials (adjuvants and base materials), which
are either listed separately or as part of composite regulations on end
use. Additional qualification (for example regarding maximum permitted
concentration , restricted end uses, and migrational requirements) are
also often stipulated.
D West Germany
Although the German Food Law gives Ministers the power to legislate
for control of the migration of substances from packaging and utensils,
little detailed legislation has yet been enacted. At present ‘recommenda-
tions, from the Bundesgesundheitsamt (BGA) , whilst not having the
status of law, are strictly followed by industry. The various BGA
Recommendations over the years have resulted in a series of positive lists
of ingredients, subdivided according to the different types of polymers or
other packaging materials, such as paper and paperboard.
Acknowledgement
We gratefully acknowledge the contribution of Mrs Carol Burke, who
prepared the data on which this chapter is, in part, based.
References
This chapter has not been referenced but is merely a guide to the legislative
approach adopted by a number of countries.
CHAPTER 23
The Influence of a Growing

Environmental Awareness on
Laboratory Design
D. ANDERSON AND B . COPELAND
1 Introduction
Within the space of little more than a generation, the manufacture of all
kinds of chemical products has become surrounded by an array of
legislative and advisory requirements. These have been instituted by
Governments or advisory bodies in many countries and have stemmed
from a growing awareness of the possible hazards which attend some of
the chemical compounds we use today.
Since the Food and Drug Association in the USA published its
guidelines on toxicity testing in 1959 (Association of Food and Drug
Officials, 1959) followed in 1965 in the UK (Ministry of Agriculture,
Fisheries and Food, 1965), many authorities have devised schemes for
assessing the safety of chemicals based on studies of structure, acute toxic
effects, etc. All of these schemes until the last decade or so depended for
toxicological information on animal studies. More recently schemes such
as that of the EEC 6th amendment (Sixth Amendment to Directive,
1979) and the Health and Safety Executive in the U K on the Notification
of New Substances (Health and Safety Commission, 1981) have included
non-animal studies. The EEC Directive 86/609/EEC requires a reduction
wherever possible in the numbers of animals used and encourages research
into the development and validation of alternative techniques. The EEC
intends to ban the use of animals for testing cosmetics in the Community
from 1 January, 1988, but only if alternative methods of testing have been
scientifically validated (CEC, 1993).
Because of the expanding numbers of new chemicals and the anti-
vivisection movement, it is becoming increasingly impractical to use
animals for all the necessary test procedures. Two year carcinogenicity
studies, for example, are very demanding of resources, both in time and
505
506 Influence of a Growing Environmental Awareness
money. There has thus been a need to develop more rapid and
cost-effective screening procedures. In response, new methods have
emerged which involve less work with animals. IARC (the International
Agency for Research into Cancer) is currently using data from non-
animal studies, i.e. short term tests, in its evaluation of carcinogens.
Earlier requirements were more concerned with standards of safety
and accuracy within the laboratory than with the compounds them-
selves. The general concern for safety of workers was given legal status in
the UK with the wide ranging Act of 1974 (Health and Safety at Work
etc. Act, 1974) and for their offspring with the Act of 1976 [Congenital
Disabilities (Civil Liability) Act, 19761. To augment the intentions of this
legislation, various specialist advisory groups have established recom-
mendations regarding standards and methods within particular types of
laboratory. Thus, regulations or recommendations have been made to
protect the workers when handling dangerous organisms (Howie Report,
1978), against possible hazards from genetic engineering experimentation
[Reports of the Genetic Manipulation Advisory Group, 1978, 1979;
Health and Safety (Genetic Manipulation) regulations, 1978; consultative
review of proposals to amend regulations, 19871, when handling car-
cinogens (Precautions for laboratory workers, 1966, Factories-HMSO,
1967; Department of Labour, 1974), against over-exposure to chemicals
in the workplace (American Conference Governmental Hygienist from
1938; Guidance notes from Health and Safety Executive-to date; The
Control of Substances Hazardous to Health Regulations-pending
COSHH regulations), and against radiation (Workplace Health and
Safety Service-TUC Proposals, 1981). The two aspects of regulatory
controls, namely chemical and worker safety, have continued to evolve in
parallel (Figure 1) and have substantially influenced optimal laboratory
design.
The development of laboratories carrying out the new type of
toxicological experiments has been relatively slow. Whereas the impor-
tance of good environmental control for animal accommodation (mainly
with a view to keeping the animals comfortable and disease free) has long
been recognised, laboratories have been given little attention. Many
existing buildings equipped with low grade fume cupboards and ‘open
window’ ventilation have had to be adapted as well as possible to the
greater scale of activity and the more critical standards of hazard
containment. The development of Genetic Toxicology (e.g. Ames et al.,
1973; Department of Health and Social Security, 1981; Department of
Health, 1989) has in particular precipitated onerous environmental
requirements for a whole range of procedures, from the microbial level to
the complex animal studies, such as long term carcinogenicity tests. In
carrying out these studies to accepted codes of safety practice, the need for
controlled environmental conditions, which has been highlighted in the
genotoxic field, has been recognised. The main criteria for correct
laboratory design are:
(a) The provision of the correct conditions for the experiment.
Pre-1959
Generally
1959
More complex
Dermal and --+ Animal Studies
LD, Studies
,/in Animals
for Food, Drugs
and Cosmetics I
1979
6th Amendment
to EEC Directive
(Stringent
Toxicology
- \
P
?
8-
1973
Requirements)
1981
,Continued
Activity
- R
R
The Birth DHSS (UK)
ENVIRON- of Short term + Guidelines for
P
MENTAL In V / t r o Assays Mutagenicity 2
AWARE N E SS (Ames Test) Testing
-Updated 2
Updated Threshold Limit Values (TLVs)

(ACGIH and guidance note from HSE)
Figure 1 Some events influencing toxicology testing in the last generation
Abbreviations for Figure 1 HSE-Health and Safety Executive
ACGIH-American Conference Governmental Industrial Hygienists LD-Lethal dose
COSHH-Control of Substances Hazardous to Health TLV-Threshold limit value
DHSS-Department of Health and Social Security TUC-Trade Union Congress
EEC-European Economic Community UK-United Kingdom
GMAG-Genetic Manipulation Advisory Group
(b) The protection of the laboratory worker against hazard from the
materials or organisms used.
(c) The protection of the external environment against noxious
laboratory waste products.
Of these three aspects, (c) has least effect on the internal laboratory
environment itself and will not be considered here.
Within the toxicology laboratory today a wide range of scientific
disciplines are practised. These include;
Acute Toxicity/Long Term Toxicity Immunology
Bacteriology Pharmacology
Biochemistry Pathology
Genetic Toxicology Reproductive Toxicity
Haematology Ter atology
Inhalation Toxicology
Their requisites in terms of laboratory conditions separate basically into
facilities for in uivo and for in uitro work.
2 Animal Laboratories
It has long been recognised that artificial, highly engineered environ-
ments, maintained at constant humidity and temperature, are desirable
for animal accommodation. This is because each animal species has
preferred environments and where long term tests are concerned, there is
a need to keep animals free from disease. High efficiency air filtration
and non-recirculatory ventilation systems are essential.
The subject of staff well-being has been of secondary concern and this
has had consequences. For example, the incidence of allergies in humans
caused by contact with animals or animal wastes, such as faeces or
dander, is substantial. Proper attention to detailed design will eliminate
many of the problems. Thus, (1) Careful integration of the airflow
pattern with the room design, utilising a linear room layout, ensures that
the operator (as well as each animal cage) has a continual supply of clean
air (Figure 2) and eliminates ‘Dead Spots’ and low level extract grilles
(where harmful particulates may collect in the room). (2) Particulates are
removed by introducing a filter at the air extract point if required. (3)
Surfaces should be constructed of materials and finishes which avoid
ledges and joints, to facilitate cleaning. (4) Where protective clothing is
worn within the animal room (gown, mask, boots, gloves) a strict
operating protocol for personnel washing/changing and for
storage/disposal of protective wear is necessary, and facilities for this
procedure should be planned en suite with the laboratory accommoda-
tion. ( 5 ) To give adequate protection against the test substance,
procedures such as dispensing or dosing of compounds should be carried
out in a separate enclosure within the room. A simple negative pressure
isolator design which can be used in such circumstances (Figure 3) has the
advantage that, unlike a fume cupboard, it passes relatively small
quantities of air and thus does not disrupt the air flow pattern (Figure 2).
P
5
R
2
0
3
ROOM SUPPLY AIR
SUSPENDED VALANCE TO PREVENT SHORT- s
Q.
ClRCUlTlNG OF AIR FROM SUPPLY TO EXTRACT 9
FRESH AIR FLOW DOWN FACE OF OPERATIVE
AIR PASSES BETWEEN EACH LEVEL OF CAGES 9
APPROXIMATING TO LAMINA FLOW CONDITIONS %
MOBILE RACK CONTAINING CAGES SPECIALLY a"
3
DESIGNED TO MEET AIR FLOW REQUIREMENTS
AIR EXTRACT
SAFE CHANGE FILTER ACCESSIBLE FROM
OUTSIDE ROOM
BENCH
Figure 2
2
n
(1) SIDE PORT FOR INTRODUCTION OF SPECIMENS,

E TC.
(2) FLEXIBLE EXTRACT DUCT (CONNECTED TO
CEILING OUTLET)
(3) PORT IN WORK TOP FllTED WITH LID
(4) DRAWER IN WHICH SPECIMENS ARE DEPOSITED
FOR REMOVAL AFTER PROCEDURE
(5) WIDE DIAMETER FLEXIBLE PLASTIC GLOVE
PORTS
(6) SEMI-FLEXIBLE CLEAR PLASTIC FRONT
Figure 3
D . Anderson and B. Copeland 51 1
3 General Laboratories
The most common and familiar item of furniture within the laboratory is
the bench and at one time-with great convenience but perhaps less
accuracy and safety-virtually all laboratory work was carried out on the
bench top. In order to achieve open bench top working to satisfy today’s
requirements, it would be necessary to guarantee a stable environment
for the experiment and that dangerous byproducts are removed from the
occupants of the room. This would require the installation of a
comprehensive and costly mechanical ventilation system to constantly
purge the whole room in a controlled way.
Another alternative, used in hospital operating theatres and manufac-
turing clean rooms where very specific activities may be undertaken in
permanent locations, has been to install sophisticated whole-room air
systems (using, for instance, vertical laminar flow techniques). The
complexity and variety of activities within the toxicology laboratory
generally preclude this solution, however, and in most instances a more
flexible and cost-effective approach must be found. One method is to
provide each worker with a ‘space suit’ complete with its own life support
system, and leave the laboratory as a simple untreated room. This idea
works in terms of building economics, but involves a great deal of
inconvenience and manipulative difficulty for the worker and has the
added risk of cross contamination between experiments.
The strategy generally adopted is to use an additional enclosure within
the laboratory, either on the bench or instead of the bench, to provide a
localised environment for each individual experiment.
A Fume Cupboards
The common general purpose enclosure is the fume cupboard. In its most
basic form this is a box fitted on the bench in which an experiment is
carried out, and through which air is drawn from the room and exhausted
via a duct to the outside. The basic design has been greatly developed in
recent years. The chief problem has been in the eddy currents set up by
items of equipment inside the cupboard and by the operator’s hand
and body movements. In modern cupboards the surrounds have been
streamlined to enable a smoother air entry, thus reducing the risk of
turbulence. To ensure efficient purging, air speeds have been increased
and, for the most demanding work, face velocities (of the air entering
across the sash) normally fall within the range 0.4-0.8 m sec-’ (Figure 4).
The quantities of air to be removed from the room to satisfy this
requirement are considerable. For instance, in a general purpose
toxicology laboratory a ventilation rate of 25-30 air changes per hour
may be necessary in order to provide a 1.5 m wide fume cupboard for
each pair of workers.
Despite the usefulness of fume cupboards in the laboratory, there are a
number of short-comings associated with them:
(1) EXTRACT DUCT NOTE THAT FAN IS LOCATED
REMOTELY AS PART OF THE BUILDING
VENTILATION SYSTEM)
(2) VERTICALLY SLIDING LAMINATED GLASS SASH
(3) FLARED SIDES ACCOMMODATING SERVICES
3 CONTROLS 2
(4) AEROFOIL SECTION AT SILL ENSURES GOOD %
SWEEP OF AIR ACROSS WORK SURFACE AND a
FORMS AN AUTOMATIC BYPASS TO PREVENT
FACE VELOCITIES AT LOWER SASH OPENINGS
FROM BECOMING TOO HIGH s*
@Q
Figure 4
E
b
F
a
$
2
D . Anderson and B. Copeland 513
(a) The air movement in a fume cupboard can upset the accuracy of
some experiments, for instance in the microbiological area.
(b) Despite modern fume cupboard design, personnel movement may
cause eddy currents which allow fumes to spill from the front
opening.
(c) The high extraction rate of non-recirculated room air causes high
energy demands for heating and ventilation, with consequent effect
on running costs.
B Safety Cabinets
A great many activities demanding special precautions within the
laboratory may be better undertaken in the safety cabinet. Laminar flow
cabinets have been used in tissue culture for some time. Their use,
primarily, was to prevent the work from being contaminated by the
environment, using horizontal laminar flow systems which did not
provide the worker with any particular protection. Designs of cabinet are
now available incorporating more sophisticated air flow arrangements,
based on a vertical laminar flow system that reduces turbulence caused by
the operator’s hand movements to a minimum, whilst at the same time
keeping the working environment at bench top level constant (Figure 5).
Some safety cabinets are designed primarily to protect the worker and
make no provision for constant experimental conditions (Figure 6). The
impetus for the development of these units in Britain, for example,
has been created by the Howie Report (1978) which suggested basic
safety standards for work with micro-organisms. This report has
been followed by a new British Standard (BS 5726) which lays down
specific requirements for safety cabinets. One general advantage of these
cabinets is that they use a great deal less air than fume cupboards, either
recycling all air back into the room via powerful (HEPA) filters, or
exhausting only 10-20% of the air passing through the cabinet with
consequent saving on energy consumption and heat loss.
C Isolators
A third type of enclosure is the isolator, which completely separates the
operator from the workplace by a continuous membrane. The use of
isolators is usually associated with work with highly dangerous pathogens
in the medical field, and they range in size from a small glove box for the
bench work, to a complete tent for the enclosure of a patient. Whilst
offering extremely good protection to the operator, there is the additional
advantage that the air inside is relatively still and for this reason, isolators
are particularly useful for the weighing and dispensing of dangerous
chemicals. One toxicological application for an isolator has been illus-
trated previously for use in an animal room (Figure 4). Another (Figure
7) is a proposed design for a combined isolator and fume hood for use in
a laboratory involved in the handling of hazardous chemicals.
(1) FAN
(2) HEPA FILTERS
(3) DOWNWARD LAMINAR AIRFLOW
(4) AIRFLOW FROM ROOM 3
(5) HEPA FILTERED EXHAUST AIR RELEASED BACK
INTO ROOM (AN EXHAUST SYSTEM COULD BE $
FITTED FOR DISCHARGE TO OUTSIDE IF g
REQUIRED) F)
(6) PERFORATED WORK SURFACE

(7) FLUORESCENT LIGHTS 2
F
09
Ps.
53
3
9
s.b
Figure 5 z
F)
2
3
2
?
n
( I ) FAN
(2) HEPA FILTER
(3) LAMINATED GLASS SCREEN
(4) TURBULENCE-FREE AIRFLOW
(5) EXHAUST DUCT TO OUTSIDE
ATM 0s PHERE
Figure 6
1
4
( 1 ) EXHAUST DUCTS
(2) LARGE WALK-IN FUME CUPBOARD WITH
HORIZONTAL SLIDING GLAZED DOORS
(3) SMALL GLASS SLIDING PANELS FllTED TO EACH
DOOR AT BENCH HEIGHT
(4) SAFETY CABINET FOR WEIGHING
(5) VERTICALLY SLIDING SASH FllTED WITH GLOVES
(6) VERTICALLY SLIDING SASH BETWEEN FUME
CUPBOARD AND SAFETY CABINET z5
SAFETY CABINET IS USED FOR STORAGE, WEIGHING, ?
cb
AND DISPENSING OF DANGEROUS CHEMICALS. %
NORMAL OPERATION IS WITH SASH UP, OPERATING P
LIKE A CLASS 1 CABINET. FOR WEIGHING, IN ORDER
\ TO ENSURE MINIMUM AIR MOVEMENT SASH COMES
DOWN AND WORK IS CARRIED OUT THROUGH
?
T
5-
GLOVES. CHEMICALS CAN BE PASSED THROUGH TO 00
FUME CUPBOARD ADJACENT, AVOIDING CARRYING
DANGEROUS COMPOUNDS THROUGH THE
9
s.
LABORATORY. s
\ 3
3
s"
Figure 7 E
b
s
3
3
2
i
tl
(1) EXHAUST DUCT TO OUTSIDE

(2) HEADER BOX
(3) BENCH SERVICES SPINE
(4) PERFORATED STAINLESS STEEL PLATE
(5) FILTER PAPER CIRCLE
Figure 8
Notwithstanding the various enclosures now available, there are still

some tasks which, although hazardous, can only be carried out on the
bench because of the dexterity required. In these instances there may be
no alternative but to provide adequate personal protection, though it is
still sometimes possible to devise apparatus to satisfy particular problems.
An example is the simple bench equipment developed for use in the
preparation of histology slides, where formaldehyde fumes are kept from
affecting the worker by extracting them from below through a perforated
metal plate (Figure 8).
D Ventilation
These special environments for individual experiments depend entirely
on the precise control of air flow and it is essential that the laboratory
itself has a predictable environment in terms of temperature, humidity,
filtration and air movement. Ventilating the laboratory by natural means
(i.e. opening windows) cannot give an adequate degree of control; rather
it is necessary to seal the room and install air conditioning to provide
comfort conditions for the workers and equilibration of air for fume
cupboards and other enclosures.
Not all the spaces in a laboratory building need to be air conditoned,
and much cruder, and therefore cheaper, environmental systems can be
conditions for
Figure 9
a
Figure 10
5 20 Influence of a Growing Environmental Awareness
used for such areas as corridors, engineering workshops, stores, and

offices. (Figure 9). The hierarchical principle illustrated has resulted in a
design (Figure 10) in which comparatively expensive air conditioned
laboratory space has been accommodated in a box inside a larger,
comparatively inexpensive ‘shed’ building. The ancilliary accommodation
is arranged within a buffer space between the inner laboratory box and
the outside walls of the building. This arrangement significantly reduces
the load on the laboratory air conditioning due to variations in outside
weather conditions.
In such a building the costs of running the air conditioning can be
reduced by installing a heat recovery system which utilises the cooling
plant as a heat pump during the winter period. The design requires a
large amount of space for the engineering installation. The quantities of
plant and distribution systems installed are considerable and additional
space is needed for accessibility for maintenance and replacement. To
avoid disruption to laboratory work and to reduce potential hazards to
nonscientific staff, it is desirable that maintenance access be from outside
the laboratory space. It is possible to accommodate all these engineering
requirements successfully but it would, in most cases, be extremely
difficult to achieve the same standards by attempting to refurbish an
existing laboratory building of a multi-storey layout with inadequate
space between floors to accommodate the ventilation ductwork.
It is clear that the complexities of providing the correct scientific
environments, together with adequate safety, has a radical effect on the
building design required. Substantial investment in more controllable
environments is necessary, together with a responsibility, placed on the
scientist, to ensure that each laboratory task is carried out safely. The
scientist will inevitably need to be aware of the standards achievable
within his particular building, and what they cost.
In many cases where standards have to be improved, the cheapest and
most effective solution in the long run may well be to build anew rather
than refurbish. Keeping in mind the disruption caused to ongoing work
by construction activities in an occupied building, the apparent additional
costs of a new structure may be theoretical rather than real. Its
advantages in terms of a well-laid out and efficient engineering system, as
well as an internal environment designed flexibly to accommodate the
many types of enclosures and other laboratory equipment now necessary,
are self evident.
References
American Conference Governmental Industrial Hygienists (1983-to date).
Threshold Limit Values for Chemical Substances and Physical Agents in the
Workroom Environment with Intended Changes for a Particular Year.
Ames, B. N., Durston, W. E., Yamasaki, E., and Lee, F. D. (1973).
Carcinogens and mutagens: a simple test system combining liver homogen-
D . Anderson and B. Copeland 52 1
ates for activation and bacteria for detection. Proc. Nutl. Acud. Sci. (USA),
70, 2281.
Association of Food and Drug Officials of the United States (1959). Appraisal of
the Safety of Chemicals in Foods, Drugs and Cosmetics. Topeka, Kansas.
CEC (1993). Council Directive 93/35/EEC of 14 June, 1993 amending for the sixth
time Directive 76/768/EEC on the approximation of the laws of the member
states relating to cosmetic products. Qfl J. European Comniiriiities, 36,
(L151), 32-37.
Congenital Disabilities (Civil Liability) Act, HMSO, London, (1976).
Department of Labour (1974). Occupational Safety and Health Standards,
Subpart G.-Regulations on the Handling of Carcinogens. Part 1910 of Title
29. Fed. Regist., 39, 23502.
Department of Health and Social Security (1 98 1). No, 24. Report on Health and
Social Subjects. Guidelines for the Testing of Chemicals for Mutagenicity.
Committee on Mutagenicity of Chemicals in Food, Consumer Products and
the Environment. HMSO, London.
Department of Health (1989). No. 35. Report on Health and Social Subjects.
Guidelines for the testing of Chemicals for Mutagenicity. Committee on
Mutagenicity of Chemicals in Food Consumer Products and the Environ-
ment. HMSO. London.
European Economic Community (1 986). Council Directive 86/609/EEC on the
approximation of laws, regulations and administrative provisions regarding
the protection of animals used for experimental and other scientific purposes.
O.J. No. L358, 18.12.1986.
Factories, The Carcinogenic Substances Regulations. HMSO, London, ( 1967).
First Report of the Genetic Manipulation Advisory Group. (7215). HMSO,
London, 1978 and 1979.
Guidance Note from the Health and Safety Executive. Threshold Limit Values
for a Particular Year (to date) HMSO, London,
Health and Safety at Work Etc. Act. (1974) Chapter 37. Health Safety and
Welfare in Connection with Work and Control of Dangerous Substances and
Certain Emissions into the Atmsophere. HMSO, London.
Health and Safety Commission (1981). Consultative Document. Notification of
New Substances. Draft Regulations and Approved Codes of Practice.
HMSO, London.
Health and Safety (Genetic Manipulation) Regulations 1978 and Review of
Health and Safety (Genetic Manipulation) Regulations 1978, consultative
paper (1987).
Health and Safety Annex 1, 21 July, 1987. The Control of Substances Hazardous
to Health Regulations 198-(Draft).
Health and Safety at Work (1987). What the COSHH Regs will mean in practice.
9(11), 21-22.
Howie, J. (1978). Department of Health and Social Security-Working Party to
Formulate a Code of Practice for the Prevention of Infection in Clinical
Laboratories. HMSO, London.
Ministry of Agriculture Fisheries and Food. (1965). Memorandum on Procedure
for Submissions of Food Additives and on Methods of Toxicity Testing.
HMSO, London.
Precautions for Laboratory Workers who Handle Carcinogenic Aromatic Am-
ines. Chester Beatty Research Institute, 1966.
Sixth amendment to Directive 67/548/EEC on the Approximation of the Laws,
Regulations and Administrative Provisions Relating to the Classification,
Packing and Labelling of Dangerous Substances. (1979). Annex VII. Off. J .
European Communities (Sept .).
Second Report of the Genetic Manipulation Advisory Group. (1979). (7785).
HMSO London.
Workplace Health and Safety Service (1981). TUC Proposals for an Integrated
Approach, TUC Publications.
CHAPTER 24
Good Laboratory Practice

ROSEMARY I. HAWKINS
1 Introduction and Background

Good Laboratory Practice (GLP) Compliance is a legislative requirement
which has completely altered in concept since its initiation in 1979, from
being guidelines applicable only to American Food and Drug Admin-
istration (FDA)-related non-clinical studies, to a universal requirement
for the International acceptance of toxicological data.
It was introduced by the FDA because this agency was dissatisfied with
the quality of data submitted for safety-in-use evaluation. The legislation
proposed in 1976 for GLPs was finalised in the Federal Register in
December 1978 (Federal Register, 1978) with an enforcement date of 20
June 1979. While not applicable in retrospect, the ruling made life
initially difficult because of its applicability to that part of any study
underway at the time. Although aimed primarily at toxicological labora-
tories, the GLPs regulate any laboratory contributing data to non-clinical
safety tests of FDA-related products (e.g. food and colour additives,
drugs, medical devices, etc. ), the exemption being method development
and efficacy tests.
The legislation, which will be considered in detail later in this chapter,
embodies those routinely expected standards of good scientific practice,
together with certain additional requirements hitherto unpractised in
toxicological laboratories. Any such legislation requires both implemen-
tation, and the FDA opted for a mechanism of both laboratory and study
data inspection, including foreign as well as US facilities, and enforce-
ment operating by eventual disqualification.
Following the FDA initiative, however, other legislative authorities in
the US felt the need for drawing up similar guidelines, the most notable
being the draft proposal from the Environmental Protection Agency
(EPA). This agency proposed not one, but several sets of GLPs
regulating for the various legislation under its jurisdiction, and thus the
proposed Guidelines for the Toxic Substances Control Act (TSCA)
(Federal Register, 1979), Federal Insecticide, Fungicide and Rodenticide
523
524 Good Laboratory Practice
Act (FIFRA) (Federal Register, 1980) and the Physical, Chemical,

Persistence and Ecological Effects Testing (Federal Register, 1980), all
purported to include GLP standards. These multiple GLP requirements
posed a potential threat to the toxicology laboratory. Firstly, the draft
proposals were all sufficiently different to cause compliance problems for
the testing facility, more especially for the laboratories conducting
contract safety evaluation studies on clients’ behalf, as the study
submission agency had to be first clarified from the sponsor and it was
inevitable that compliance of these laboratories must be to those most
stringent requirements. More importantly, there was the confusion of the
generic principles of laboratory practice with those specific requirements
pertaining to a particular type of study. Furthermore, the draft GLPs
were completely inflexible (in contrast to those of the FDA), being
drafted in terms of absolute requirement rather than guidelines, and they
were not internationally applicable, demanding for example adherence to
American international legislation and personnel qualification. Such
documents certainly did not fulfil the uniformity of interagency legislation
as promised by the Interagency Regulatory Liaison Group (IRLG)
established at that time to standardise American legislation.
Together with the multiplicity of GLPs there was, of course, the added
burden of the multiple inspections by agency teams for their enforce-
ment, and European laboratories were faced with a prospect of the
disruption and subsequent cost burden of multinational and multiagency
inspections.
In Stockholm in 1978, therefore, the Organisations for Economic
Co-operation and Development (OECD) initiated the formation of an
expert group to consider the whole question of GLP guidelines for its
member nations and to formulate an international document. The
requests of such a document are that it is mutually acceptable to all
member states and that it is internationally applicable.
Any requirements for GLP must embody three aspects, namely to
establish the guidelines themselves, a mechanism for the assurance of
compliance, and a mechanism for enforcement. Initiated after the
legislation from America, it seemed logical that the OECD requirements
should be formulated along similar lines; industry had spent considerable
resources both in time and capital expenditure to comply with the FDA
and the acceptance by this country of any European system was
paramount to the exercise.
2 Current Situation
The OECD Expert Committee has now finalised three documents
(OECD, 1982):
(1) OECD Principles of Good Laboratory Practice.
(2) Implementation of OECD Principles of Good Laboratory Practice.
(3) Guidelines of national GLP inspection and study audits.
Rosemary I . Hawkins 525
It has been agreed that study data generated under conditions of the
OECD Principles of GLP will be internationally acceptable.
The OECD has therefore fulfilled its remit in this area and produced
generic, internationally acceptable guidelines, from which it was intended
that member nations would prepare their own national GLPs, so
formulated to take into account internal legislation and needs. However,
it was agreed that it was the National government’s own responsibility to
set up a suitable and mutually acceptable mechanism for the implementa-
tion and for the enforcement of compliance.
The World Health Organisation (WHO) considered GLP as part of the
International Programme on Chemical Safety, and agreed to adopt the
OECD approach, thus embracing a still wider range of countries.
OECD guidelines carry no member enforcement requirement and the
GLP Principles were mostly used directly translated as recommendations
by European governments to their national industries. However, an EEC
Directive must be implemented by Member States and following the 6th
Amendment to the Dangerous Substance Directive, 1981, the UK Health
and Safety Executive (HSE) published a set of GLPs (HMSO, 1983) to
regulate for those tests carried out under the Notification of New
Substances Regulation, 1982. The authority also set up an implementa-
tion mechanism of faulty inspection and study audit similar to that of the
FDA, and a GLP Monitoring Unit was established.
It must be remembered, however, that this was only operative in the
UK for a limited area of work and not for other toxicological studies. A
step in this direction was precipitated when the Japanese Pharmaceutical
Affairs Bureau, Ministry of Health and Welfare (MHW), established
‘standards for Conducting Safety Evaluation Tests on Drugs’ (Ministry of
Health and Welfare, 1981) which became fully applicable as from 1 April
1983, for those non-clinical studies conducted for new drug applications,
for manufacturing and importing, and for application for re-examination.
Studies submitted from foreign laboratories had to be accompanied by
government assurance of compliance to national GLPs of similar or
higher standards compared with the Japanese GLP requirements. Even-
tually there was agreement to accept FDA or OECD regulations as this
equivalent, but the need for a government Statement of Compliance
acted as a form of trade barrier to those countries where no such
government’s inspection mechanism had been established.
As the initial inspection mechanism operative in the UK was that for
the HSE for monitoring compliance of organisations generating data in
respect of the Notification of New Substances Regulations, the Depart-
ment of Health and Social Security (DHSS) established, as an interim
measure, a GLP Unit located in, but distinct from, the Medical Division
concerned with Toxicology and Environmental Protection to ensure
compliance with OECD Principles for the Pharmaceutical and Cosmetics
Industry. Distinct from the HSE Monitoring Unit, the authority assured
industry that it would avoid duplicate inspection, but this was clearly an
unsatisfactory situation. However, as a result of discussions with the

DHSS, an arrangement has now been made with the Japanese MHW as
being the first step to producing a full agreement on GLP, by exchange of
notes verbal in March 1984.
For UK industry there followed another problematical Japanese
requirement for the GLP compliance of agrochemicals, issued by the
Ministry of Agriculture, Forestry and Fisheries and made known to UK
industry with little or no forewarning, with an effective date of 1 October
1984. As compliance to OECD GLPs was accepted by the Japanese
Government as the equivalent to its national requirements, there was no
problem with study conduct, but the preliminary registration of the
testing facility with the Director of the Agricultural Production Bureau
was a requirement fraught with difficulties. Details of retrospective GLP
compliance over a period of 3 years, plus facility description and floor
plans, instrument information and location, personnel curriculum vitae,
and even financial aspects of the organisation were required-clearly a
topic for discussion between governments, and one which is currently
under careful review.
In October 1984, the UK Government finally announced the decision
to base the long term comprehensive GLP monitoring scheme in the
Toxicological and Environmental Protection Division of the DHSS. The
Department arrangements for pharmaceuticals and cosmetics were then
extended to include all laboratories involved in the health and environ-
mental safety testing of agrochemicals and food additives, together with
the HSE responsibility for GLP monitoring for the notification of new
chemicals. In 1986 the DHSS published ‘Good Laboratory Practice, The
United Kingdom Compliance Programme’ , which was the explanatory
document in the principles and operational procedures adopted by the
UK GLP Monitoring Unit. To ensure maximum uniformity at an
international level, the GLP principles are those requirements as
developed by the OECD and reproducing and extending the provisions
published initially by the HSE. The Monitoring Unit is structured to
carry out laboratory inspection and study audits and is currently
negotiating with the regulatory authorities, notably Japan and USA, for
agreement of mutual acceptance of inspection programmes. The inspec-
tions are of two categories, regular monitoring every 2 to 3 years, and
monitoring at the specific request of a regulatory authority, either
national or foreign. The former is initially conducted at the request of the
facility itself, but once inspected it is included in the Unit’s on-going
monitoring programme in order to maintain the knowledge of the
laboratories compliance status. The operating costs are recovered from
the participating laboratories which are allocated to one of three
categories of scale of charge.
European countries are similarly in the position of requiring to
formulate GLPs from the OECD guidelines and, more important, to
establish a mechanism of government assurance that these have been
Rosemary 1. Hawkins 527
adhered to and which will be acceptable to all authorities. Most national

governments are in the position of publishing a set of GLPs, although it is
not always clear as to the status. The situation is obviously one of
continual change and the following resum6 must, when at the time of
going to press, be out of date. In France for example, the Ministere des
Affairs Sociales et de la Solidarite Nationale Sante published an official
text ‘Bonnes Pratiques De Laboratoire’ (BPL), instructions regarding
principles of BPL (May, 1983) and Laboratory Inspection (September,
1984), but this applies at the moment only to pharmaceuticals and
medicines. A memorandum of understanding is being negotiated between
the FDA and the French Directorate for Pharmacies and Medicaments
(Food Chemical News, 1984). Currently the separate inspectorate for
studies regulated under the EEC 6th Amendment is not yet in place.
In Germany it is necessary to differentiate between chemicals and
pharmaceuticals. While no GLP regulations have yet been issued, the
Chemicals Act following the EEC Directive on Dangerous Substances
allows for governmental action and the OECD principles have been
published reflecting that non-compliance under the German legal system
could be regarded as culpable negligence and failure to exercise
reasonable care in study conduct. There is no requirement for GLP-
compliance in the generation of data on pharmaceutical products but the
local health authorities are prepared, on request and a fee-paying basis,
to inspect and issue compliance certificates for the Japanese requirement.
The Italian Minister0 della Sainta has started a GLP programme by
sending questionnaires to industry, but meanwhile recommending com-
pliance to either FDA or OECD GLPs.
In Scandinavia, the Swedish National Board of Health and Welfare
published some guidelines in the summer of 1983 which apply to
pharmaceuticals. In Denmark, public laboratories concerned with control
of industrial chemicals, pesticides, and food additives are working
towards preparation of determined guidelines and their implementation;
private laboratories comply with OECD requirements by request, but so
far there is no requirement by the National Board of Health for
Pharmaceuticals.
When considering the European situation, its complexity must be
viewed with the realisation that each country not only has established
mechanisms of government agency authorities with responsibility for
distinct areas, but also in some cases regional authorities. Thus, the
OECD principles needed to be generic in order to allow flexibility of
existing national regulations and mechanisms.
Further change in the GLP status of various countries has been
initiated, led by new legislation, namely the EEC Pharmaceuticals
Directive, December 1984, covering the requirement for GLPs for
pharmaceuticals , and the proposed Directive on Good Laboratory
Practice designed to ensure that all facilities in the Community comply
with G L P by 1 July 1988. The draft Directive requires member states to
ensure compliance by inspection and also outlines the need for authorisa-
tion of GLP requirements within the community.
The American scenario is also changing. While there is yet no update
of the FDA GLPs as promised (for example, with the onerous require-
ments of sample retention), the final EPA GLPs have now been
published (Federal Register, 1983), both from TSCA (applicable to both
Section 4, the 6th Amendment of the Dangerous Substances Directive,
and Section 5, the testing of existing chemicals if so requested), and from
FIFRA. These requirements apply to any study conducted, initiated, or
supported on or after 29 December 1983. Thankfully there is harmonisa-
tion with both the FDA requirements and with the OECD, and within
the EPA itself the only differences in the 2 sets of laws relate to different
statutory requirements. The most significant of these differences is in the
scope; the TSCA regulations contain a special section to cover environ-
mental toxicology as well as health effects.
The concept regarding GLP requirements has thus changed radically
since 1978, when only those laboratories involved with US FDA
submitted safety-in-use studies were concerned. Now these requirements
are universal for the generation of toxicological data in non-clinical
studies and for this data to be internationally acceptable, and this, of
course, applies to all facilities conducting such studies, academic,
industrial, or contract laboratories.
More and more countries are recognising the need for establishing
National GLP principles and a mechanism of compliance assurance
outside of OECD, EEC, and EFTA groups, an example being South
Korea, thus allowing for the confidence in, and exchange of, reliable data
on an ever-increasing scale.
Following these fragmented beginnings progress since the first edition
of this book has happily unified the issue for all Member States of the
European Economic Community (EEC). Requirements mutually agreed
by all Member States finally became embodied in a Council Directive to be
implemented into the National legislation of each State. Negotiations in
Brussels, where finalised when the EEC accepted the OECD principles of
GLP on behalf of all Member States and requirements harmonised in
Directives 87/118/EEC of 17 January 1987 and 88/320/EEC of 11 June
1988 on the application of GLP principles to chemical testing and the
inspection and verification, respectively. The latter was amended in
90/ 18/EEC to replace the reference and OECD guidance documents in
order to facilitate interpretation throughout the Community, Annex A
and B contain the guides for monitoring procedures for GLP and the
conduct of laboratory inspection and study audits.
Now established in Community requirements, GLP has become an
integral part of numerous other Directives on toxicological and safety
issues (e.g. cosmetics, industrial chemicals, medicinal products, food
additives, animal feed additives, pesticides, genetic engineering). All
laboratories have a need to ensure all such relevant texts are part of its
Rosemary I. Hawkins 529
testing programme and be alert to all subsequent amendments which may

follow.
3 Specific Requirements
The requirements for GLP compliance can be separated into those
practices which are standard in any facility fulfilling good standards of
study conduct, and those which are a new addition to the routine of the
toxicological laboratory. However, as with most guidelines it is all a
question of interpretation and of rational judgement as to the best way
the requirements of each particular organisation can be served, so as to
maintain compliance with the maximum cost-effectiveness. The FDA
have given sufficient scope for company policy to be operative and to
allow for managerial thought along these lines.
A Special Requirements
Standard Operating Procedures
Standard Operating Procedures (SOPs) are defined as ‘those procedures
in writing setting forth non-clinical laboratory study methods that
management is satisfied are adequate to ensure the quality and integrity
of the data generated in the course of a study’.
The FDA suggest areas which require SOPs to be established, but in
practice every department needs to have SOPs for each and every
procedure which is performed, from cleaning glassware and laboratory
equipment, reagent and media preparation, to carrying out specific test
methods. In initiating GLP in any laboratory it is the writing of SOPs that
becomes the time-consuming tedium against which all scientists rebel.
However, carefully constructed, they can become valuable tools as
laboratory training documents and reference manuals.
Authorisation by management is required; the easiest technique is to
have the signature of the scientist in charge to assure technical authen-
ticity, and the signature of the manager responsible for facility com-
pliance, both of which signatures must be dated. A sufficient number of
copies of each SOP must be available to ensure that members of staff
conducting each procedure have ready access to the SOP. With tech-
nological change, method development, acquisition of new equipment,
progress with new techniques, etc., continual updating of SOPs is
required. A historical file of all SOPs must be maintained, thus providing
a procedural document as reference for past studies.
In practice, in laboratories conducting studies on behalf of a sponsor,
procedures are often developed which are specific to the nature of
material under test. In these circumstances study-specific SOPs are
written, but confidentiality must be respected so that they are not
included in the departmental manuals but retained separately with other
study data.
The format of SOPS depends upon the laboratory but it is best to keep
them straightforward and easy to follow. Additions of technique can
easily be made, and use of standard text book methods or maintained
manuals are permissible either as a direct photocopied inclusion, or as a
reference providing the location of the original is clearly indicated and
adequate for the procedure to be carried out without detriment.
Quality Assurance Unit

The Quality Assurance Unit (QAU) is the mechanism whereby each
study is monitored to assure management that it is in compliance with
GLP requirements. Thus, it must monitor the facilities, equipment,
personnel, methods, records, quality control, and performance. The
QAU may be a separate department staffed by several individuals if the
organisation is sufficiently large to demand this (and in practice most are:)
but in smaller laboratories it can be any member of staff, providing they
are independent of the personnel engaged in the direction and conduct of
the study in question.
Industry was familiar with quality control, but the Q A U is distin-
guished from this by being the overall monitor of all those actions
necessary to ensure the proper conduct of the study and the integrity of
the data generated. Quality control is the responsibility of the head of the
department; quality assurance ensures there is adequate quality control
as part of its function, the other functions are specifically designated by
the FDA as follows.
The Q A U must maintain a master schedule identifying all the
non-clinical studies being carried out, the nature of each study including
documentation of the test material and test system, sponsor, study
director, start date, current status, and final report status. The rationale
behind such demands is to ensure that the facility is not overloaded and
can give correct application to each individual study, but in reality
maintenance of confidentiality only allows for an agency to be permitted
access to information on studies relating to its own legislative remit, so
that this requisite plus the retaining of past schedules loses some of its
impetus in practice.
The Q A U must maintain a copy of the protocol for each study for
which it is responsible. The QAU must inspect each critical phase of each
study, and report this inspection, identified with the date and phase
indicated and any action required, recorded together with the date when
it is reported to management. In effect, the definition of what constitutes
a critical phase is a matter for decision for each organisation. For most
safety evaluation studies, there will be certain standard phases but QAU
inspections must be related to all individual studies. It is usually most
helpful to designate the inspection phases on finalisation of the author-
ised protocol and at this time to determine any unusual or non-routine
steps. The inspection programme can then be scheduled from the study
Rosemary I . Hawkins 53 1
plan. The actual Q A U inspection should be conducted without prior

knowledge of the staff carrying out the study whenever possible, and
should be performed with the minimum of disturbance. In order to
achieve this it is useful to design a check list for each inspection phase
which can be completed simply and quietly by the QAU staff, and who
then undertake a full inspection report on returning to the Unit. For
studies lasting longer than 6 months, as inspection is required every 3
months, usually this can be designed to coincide with a phase inspection,
for example, in a long term study involving animal parameters at the time
of general observations or weight measurements, and thus reduce the
time involved for QAU staff. Similarly, the requisite report on study
status can be so timed.
Apart from the study inspection, the audit of the final report is one of
the most important and time-consuming areas of the QAU requisite. The
audit is to assure that the report correctly describes the methods used and
that it accurately reflects the raw data. It must be remembered that in no
way is the Q A U involved with the scientific interpretation of the raw
data, which is the responsibility of the study director, but it is responsible
for reducing the number of errors reproduced in the final report from the
raw data to the minimum possible level.
Final report audits can best be divided into two sections, raw data and
the actual report.
Raw data. The raw data includes all the original observations, work-
sheets, records, memoranda, and notes generated during the conduct of
the study, and may include photographs, microfilm or microfiche copies,
computer printouts, magnetic media, dictated observations, and records
from automated instruments. The raw data from any one study must be
clearly and correctly annotated and authorised and should be in such a
condition that the study report and conclusions could be reconstructed
from the data alone in future years without the original report or
personnel being available.
The raw data are typically included in the final report in the form of
appendices. These must be checked for transcription and omission errors.
In most toxicological studies this will include hundreds or even thousands
of numerical information and descriptions, including the laboratory
(including animal house environmental data), animal, and pathological
data. A complete check on every item, while desirable, and I believe is
the only sure way if the report size and staff permit, can become
impractical and a sampling procedure must be utilised. The procedure
used will depend upon the choice of each laboratory, but the two most
frequently employed are the technique of sequential analysis (Moroney ,
1956) and the British Standards Institution method (British Standards
Institution, 1972).
Apart from the straightforward errors of transcription, one of the most
frequent problems encountered with raw data audits is the error of
omission. All the raw data must be included in the final report unless
there is a specific reason for this exclusion. Often this will arise,
especially with certain biochemical parameters or if not required by the
sponsor, but the difficulty is to train the scientist responsible to annotate
the raw data clearly and authoritatively as to what has been omitted and
the reason why. Other common problems are the re-identification of
groups of values which have not been clearly indicated, the non-inclusion
of the derivation of manipulated data (especially where this has been
generated using a hand-held calculator) and the different phraseology
employed in pathology reporting. This latter is a matter of education of
the pathologists to provide a code or key to those comparable diagnoses
in the raw data which have been grouped together under a common term
in the final report. While the FDA accept the final diagnosis as the study
raw data, thus allowing the pathologist the necessary flexibility of
thought, it is impossible for an auditor to understand or accept
differences in terminology and this must be regarded as an error.
It is usual to check 10% of the calculations, but machine-derived data
is of course not subject to re-calculation, although the data input must be
checked.
The error level acceptable to the QAU auditor must depend upon the
nature of the errors involved. Usually once an error level of greater than
10% is apparent the complete set of raw data is returned to the study
director to carry out a 100% data check. If only a few errors are
detected, and of a magnitude and nature not to affect the final outcome
of the report, then these are identified and returned for correction. An
audit report is always made, identifying fully the study, project number,
study director, identity of raw data, the result of the audit, and any action
taken or test required. This is authorised and dated by the auditor and by
management.
The audit of computer-driven data is the most recent problem for the
QAU with the increasing development of this instrumentation in toxicol-
ogy. Direct on-line capture of data must be conceived with the advice of
the unit staff so that some form of assurance will be built into the
programme of the identification of the individual responsible (together
with prevention of unauthorised data entry by personnel or study
identification); date of entry; return capacity of incomplete or erroneous
input; and any changes made, together with the data, reason, and
personnel responsible for the change.
When observations are so entered, directly as input to a computer data
base (either on-line, or as batch records) which is maintained on
magnetic media (disc or tape), the computer data base is considered as
raw data, or if preferred for economic reasons the resultant printout
which contains the collected data (providing these are clearly authorised).
When observations are captured in the form of the written record, it is
this written record which forms the raw data and any subsequent steps,
such as transfer of a computer data base, are not considered as raw data.
Report. The actual typed text of the report must be checked for
procedural details, for completeness, and for accurate reflection of the
data in a non-scientific context, such as dates, times, frequencies, values,
and numerical conclusions. The tables and figures must also be checked
for completeness and the accuracy of any annotation, and to ensure that
this tallies with any references made to them in the text.
It is usually more expedient to check the final draft of the written
report, as this is more cost-effective if changes have to be made. Re-audit
of both text and appendices should take little time, the check for final
editorial and typing errors is the responsibility of the study director.
Each final report must contain a compliance statement, prepared and
signed by the QAU, which specifies the dates inspections were made and
the findings reported to management. It should be noted that the
inspection findings themselves are confidential to the laboratory and the
FDA cannot ask to see them, but a laboratory may be requested to
certify that inspections were implemented, performed, documented, and
any action required fulfilled.
All the procedures and responsibilities of the QAU must be fully
documented, SOPS are required for all aspects of the units remit, and all
records maintained by the unit are required to be kept in one location.
Study management. This third requisite caused some laboratories to
undergo an organisational re-think, while for others it was merely an
adjustment of an already existing managerial system.
This requirement was for the establishment of a study director before
the initiation of a study, to have overall responsibility for the technical
conduct of the study, as well as for the interpretation, analysis,
documentation, and reporting of results. In this way, a single point of
study control is established and becomes the contact for external and
internal liaison. It is the organisation’s responsibility to ensure that the
study director designated has the appropriate education, training and/or
experience for such a position, and that replacement, if it beomes
necessary to do so during the conduct of the study, is carried out
promptly and with the required documentation. There are certain
activities which have been specifically regulated for the study director: to
assure that the protocol is approved and followed; that all experimental
data are accurately recorded and verified; that unforeseen circumstances
which may affect the study quality and integrity are noted and corrected
if needed; that all applicable GLP regulations are followed and that all
raw data, documentation, protocols, specimens, and final reports are
eventually deposited in the archives.
B General Requirements
The remaining requirements can be considered as those which would be
normal practice for the correct conduct of a non-clinical study in any
properly constructed facility, but now with the necessity of the documen-
tation of previously accepted conditions and the maintenance of records

and observations in a specific way.
Personnel
Personnel conducting or supervising non-clinical studies should have the
correct education/training/experience for the function performed, and
the job description and training/experience records must be documented.
Assurance is required that there are sufficient number of personnel,
correctly clothed for the study conduct and taking the appropriate
precautions against personal contamination. Any ill-health that may
adversely affect the quality or integrity of the study should be docu-
mented and the necessary action taken.
Facilities
Organisations have now to be able to demonstrate that they are of
suitable size, construction, and location to undertake non-clinical studies,
and that there is an adequate degree of separation of various facilities to
prevent any inter-contamination. This includes adequate space for
administration and supervision and sufficient locker and toilet facilities
for all of the staff, as well as the necessary general and specialised
laboratory areas and space for cleaning and sterilisation. Specific require-
ments for animal care and facilities are designated, and it must be
remembered that these requirements were stuctured in conditions exist-
ing in the USA and that in the UK, for example, such specifications
appear superfluous in light of the existing inspections carried out by the
Home Office. Thus, the facilities are required to be constructed and
located to minimise any disturbances to the studies and there must be
adequate space for separation of different species and studies. Disposal of
animal waste and refuse must be such as to prevent any environmental
contamination, and disease hazard. Separate and adequately furnished
areas are required for storage of bedding, supplies, equipment, and
animal feeds, which are protected from infestation from wild animals and
contamination. Similarly, in order to prevent contamination and possible
mix-up of materials for different studies, separate areas are required for
receipt and storage of the test article (material under test) and any
control article, for the mixing of the test and control article with a vehicle
if required, and the storage of these mixtures.
Equipment
Equipment employed in the conduct of the studies needs to be
demonstrated to be of appropriate design and adequate in capacity. The
location needs to be that for correct operation, inspection, cleaning and
maintenance. While the necessary SOPS for these functions are located
by the instrument, written records of all inspection, maintenance, testing,

calibration, and/or standardisation operations, together with the date and
staff involved are needed, and any non-routine repairs and details of
default are recorded.
In order to be cost-effective, compliance to any requirements should be
carried out with the specifics of the individual organisation in mind. Thus,
the combination of these first three requirements to form a departmental
manual can be met with some success. Each department has specific
facility and operational requirements, and these are detailed together
with a ground-plan to demonstrate adequacy of that department.
Personnel details are confidential and as such should be restricted to
location in the Personnel Department. However, each manual contains a
list of the staff with their general background, current education details,
date of organisation membership, signature, and initials-most important
as they play a vital role on raw data sheets and identification is essential.
Staff changes are indicated by a single line drawn through the resumk,
together with the reason and date of change, thus maintaining a historical
record for past studies. Equipment is listed with description, serial
number, and date of purchase? together with the details of manufacturer
and/or in-house maintenance of relevant SOPs detailing its cases. The
in-house training is the responsibility of the head of that department and,
as such, designation of a member of staff to carry out each department’s
and other department’s SOPs is the indication that training is judged
fulfilled and the training record for that member of staff is complete.
Thus a check-board listing of SOPs against staff proceeds a list and copy
of each SOP and completes the departmental manual.
Such a document, held by the QAU whose senioral staff as required
fulfils requirements, and assures compliance, ensures confidentiality, and
allows for rapid managerial analysis and decision making.
Reagents and Solutions

All reagents and solutions are required to be labelled for identity, titre or
concentration, storage requirements, and expiration date. The rationale
for this is pefectly clear, it ensures deteriorated or outdated material will
not be inadvertently employed during the study, but it has not been fully
thought out. It is impossible, for example, to place an expiry date for a
stock bottle of analar grade sodium chloride. It is preferable that, for
such reagents ordered after the GLP compliance date, the date of receipt
from the suppliers is stamped onto the label as an indication of age, and
any storage requirements, other than those of normal conditions, clearly
indicated. Shelves in laboratories can easily be marked as being for those
reagents requiring normal storage conditions only.
Animal Husbandry
The requirements set out in this section again must be considered in view
of conditions and existing legislation in the USA, and the need for the
FDA to incorporate certain requirements to make good the deficits which

do not necessarily exist in other countries.
Separate areas for quarantine and the isolation and treatment of sick
animals are not usually provided in such organisations in the UK.
Animals designated for a study are immediately housed in the room
identified for that study alone, and any animal so ill as to need treatment
would be killed, preventing possible spread of infection and/or removal
from the test for medication.
However, other details of animal care are applicable: the appropriate
identification of each animal with full details on the cage, and the
adequate separation of different studies and animals species.
Cleaning of cages and accessory equipment, the use of adequate
non-contaminating bedding, and recording of pest control are all routine.
However, initially in the UK a quite problematical requirement was that
of the analysis of animal feed and water to ensure contaminants capable
of interfering with the study are not present. Animal feed manufacturers
are now well informed of this requirement and provide adequate
analytical data with each batch of foodstuffs. The FDA have agreed that
the manufacturer batch analysis is acceptable, as most laboratories would
not have the necessary equipment or expertise, but do require analysis if
the test article is such that a possible food contaminant might interfere
with the study. It should be remembered, however, that each batch of
animal food requires an analysis document (and for each study supplied
from this batch) and that these analyses must be kept together with the
rest of the raw data for each study. A more difficult requirement for the
UK was that of drinking water analysis; the water supply is more
standardised than in America where private well, bore hole, or river
supply is common. By this time most water authorities are used to the
approach from laboratories and will offer 3 monthly analysis of random
site samples.
Test and Control Articles

Many laboratories, which conduct safety evaluation studies on behalf of
sponsors, are not able to specifically fulfil the FDA requirements for
complete documentation on the test-and sometimes the control-
article(s). The requirements for definition and documentation of identity,
strength, purity, and composition (or characteristics to allow for defini-
tion) may well be confidential information and even if it can be made
available, the laboratory may not have the equipment or expertise to
ensure the necessary characterisation. One method of complying to this
request with the limited means available is to define the sponsor’s
responsibility for having this information and for assurance that the
material given to the facility to test fulfils this analysis. On receipt of the
material from the sponsor, the laboratory should check the container and
ensure that it is visibly undamaged in any way; the receipt form once
signed should then be returned to the sponsor for his signature that the
undamaged material so received fulfils the specification analysis and/or

identity requirements. It is the laboratory’s responsibility to obtain from
the sponsor all knowledge of stability, storage requirements, and possible
hazard handling precautions and whether these will be altered in any way
following mixture with the vehicle, where applicable. Each sample must
be correctly identified-and this must include the lid as well as the body
of any large container, and for a study lasting longer than 4 week’s
duration, reserve samples should be kept for as long as they afford
analysis-if mixed with the diet as vehicle this will be for the duration of
the viability of that diet batch as specified by the manufacturer. When
mixed with such a vehicle, the uniformity and concentration of the
article/vehicle mix is determined.
Study Conduct
The requirements specified for the conduct of the non-clinical studies are
not unusual in themselves but did require, for many laboratories, the
redesign of worksheets for data capture.
Thus, the study is required to be in accordance with the protocol,
specimens adequately identified, and all experimental procedures carried
out as requested. All data generated is recorded directly, legibly, and in
ink, or at least in an unerasible medium. All data entries are to be dated
and authorised by the member of staff responsible for the entry, hence
the importance of the recognition of initials. Changes in the entry must
not obliterate the original entry, and must be dated and signed and the
reason for the change given. In practice, no data sheets can be that large,
so it is normal for a code of errors to be keyed for each set of data and
this to be used to keep laboratory worksheets clear, simple, and legible.
Computer data needs the data input to be identified and the same
requirements for data change.
Protocol
The requirements set out by the FDA for the information to be contained
in the study protocol are those which it considers minimal. These are as
follows:
(1) Descriptive title and study purpose.
(2) Test and control article identification.
(3) Name of the sponsor and the name and address of the testing
facility .
(4) Proposed starting and completion dates.
( 5 ) Justification of choice of test system.
(6) Details of test system where applicable, such as number, body
weight, sex, supply source, species, strain.
(7) Method of identification of test system.
(8) Description of experimental design (including randomisation).
(9) Description and/or identification of diet and any materials

employed with the test and control article.
(10) Route of administration and its justification.
(11) Dosage level, method, and frequency of administration of test
and control articles.
(12) If necessary, the method used to determine the degree of
absorption of the test and control article by the test system.
(13) The type of frequency of the tests, analyses, and measurements to
be made.
(14) The records to be maintained.
(15) The date of protocol approval by the sponsor, and the signature
of the study director.
(16) The statistical methods to be used.
Any changes and revisions to the protocol are issued as a protocol
amendment and signed and dated and maintained with the protocol.
It is often useful to include with the protocol the signature of the head
of the QAU which assures both the management of the testing facility
and the sponsor that the protocol is in compliance with GLP require-
ments. While some organisations list the SOPs involved in the study
protocol this seems unnecessary and time-consuming , and a general
statement that all procedures will be carried out according to the
organisation’s SOPs would seem to be sufficient.
Reports
Similarly to the protocol contents, the FDA also sets out the minimum
contents that should be included in the final report of each study:
(1) Name and address of the facility conducting the study and the start
and completion dates.
(2) The objectives and procedures as stated in the protocol (together
with any amendments).
The archives are limited by requirement to authorised personnel only
and require indexing for expedient retrieval by test article, date of study,
test system, and nature of study. Most facilities will produce the system
best suited for their organisation, but colour-coding is often applicable.
In small organisations, the employment of a specialist archivist is not
cost-effective and the librarian or member of the QAU staff can just as
efficiently be designated for this responsibility. Material withdrawn from
the archives at any time should be well documented and with the date
and signature of the person responsible together with its temporary
location, and countersigned on return.
4 Conclusions and Summary

Since its inception in 1978, the concept of GLP has altered considerably.
No longer an American regulatory-agency-orientated exercise, GLP has
become an established code of conduct of non-clinical studies on an

international level. As far as industry was concerned, the original
exercise to ensure compliance was extremely costly, both in real terms
and in terms of time and manpower; it is estimated that technician time,
for example has been increased by at least 30%. Continual monitoring
and updating is necessary and costly, and the auditing of reports has
necessarily increased the length of time taken, especially for long term
studies. Animal suppliers and diet manufacturers have had to improve
their records and documentation and improve analytical techniques all of
which have added to greatly increasing their costs, which in turn have
been passed on to the customer. However, against this must be balanced
the greatly improved quality of the generated data and study, integrity
and, more recently, the importance of the international acceptance of
data.
Any mechanism which improves the quality of a toxicology study can
only be a good thing; laboratories have certainly had to look to their
housekeeping, and even the burdensome chores of SOPS can be proved
useful in the prevention of slipshod habits being passed on to newly
recruited staff, and be valuable training manuals. It must be remembered
that these guidelines are sufficientlyflexible to allow for company policy to
operate. It is not a question of blindly following a set of rules but of
reasoned thought of the rationale underlying the requirements. Surely the
critics of GLP should be rather bemoaning the reason why these
requirements were necessary, and strive to improve the standards of
technique and scientific integrity of their laboratory.
This ending to the first edition has been retained simply to illustrate the
problems, that as the original instigators of the GLP principle, we
encountered both with staff and management.
It seems especially pertinent that the second edition should be published
in 1993, a year which heralds the completion of the internal market and the
removal of trade barriers. In such an environment it is essential for the free
movement of goods to have a harmonised system which allows for the
mutual recognition of test data. This can only be achieved by ensuring
data is generated using standard and recognised test methods and that
there are recognised procedures for monitoring compliance to these
throughout the Community. Thus test data generated by one Member
State are recognised by all other Member States.
One issue of great concern to all testing laboratories was data
confidentiality. Reassurance can be found in Directive 88/320/EEC,
Article 4, that all commercially sensitive and confidential information
which might become available during inspection is so respected.
Lastly, but by no means of the least importance, is the reduction in
manpower and resources of the testing laboratories by negating the need
for duplicating effort. Most important of these is the consequent reduction
of experiments on animals which has always been an issue of great concern
to all involved, and is embodied in a separate Council Directive
86/609/EEC.
540 Good Luborutory Pructict:
Any system that is new meets with certain opposition. Much hard work
and lengthy negotiations with governments, both National and Inter-
national, has produced a reasonable framework acceptable to both
scientists and legislators which should enhance and not hinder the
progress of toxicology.
References
British Standards Institution (1972). Specification for Sampling Procedures and
Tables for Inspection by Attributes (BS.6001). Published by the British
Standards Institution.
British Standards Institution (1972). Guide to the Use of BS.6001 (BS.6000).
Published by the British Standards Institution.
European Economic Community (1987). Council Directive 87/118/EEC on the
harmonisation of laws, regulations and administrative provisions of relating
to the application of the principles of good laboratory practice and the
verification of their applications for tests on chemical substances. O.J. No.
L15, 17.1.1987.
inspection and verification of good laboratory practice. O.J. No. L145,
1 1.6.1988.
European Economic Community (1990). Council Directive 90/18/EEC adapting
to technical progress the Annex to Council Directive 88/320/EEC. O.J. No.
L11, 13.1.1990.
approximation of laws, regulations and administrative provisions regarding
the protection of animals used for experimental and other scientific purposes.
O.J. No. L358, 18.12.1986.
Fed. Regist. (1978), 43, (247) 22 December.
Fed. Regist. (1979), 44, (91), 9 May.
Fed. Regist. (1980), 45, (77), 18 April.
Fed. Regist. (1980), 45, (117), 1 November.
Fed. Regist. (1983), 48, (230), 29 November.
Food Chern. News (1982), 24, (12), 31 May.
Food Chern. News (1982), 17 October.
Food Chern. News (1984), 26, (22), 6 August.
Ministry of Health and Welfare (1981). Notification No. 313 of Pharmaceutical
Affairs Bureau.
Moroney, H. J. (1956). ‘Fact from Figures.’ Penguin Books, 3rd edn; reprint with
revisions, 1969.
OECD (1982). ‘Good Laboratory Practice in the Testing of Chemicals’, Final
Report of the OECD Expert Group on Good Laboratory Practice, Paris.
(ISBN 92-69-12367-9).
HMSO (1983). Health and Safety Commission Approved Code of Practice
COP7, Principles of Good Laboratory Practice. HMSO (ISBN 011 8836587).
HSE (1983). Health and Safety Executive. Notification of New Substances
Regulations, 1982. Establishment of a good laboratory practice compliance
programme (ISBN 0717601250).
DHSS (1983). Department of Health and Social Security. Arrangements to
monitor compliance with Good Laboratory Practice, NS/14 March 1983.
DHSS (1986). ‘Good Laboratory Practice’. The United Kingdom Compliance

Programme. Published by DHSS, Hannibal House, Elephant and Castle,
London.
Ministere Des Affaires Sonales Et De La Solidante Nationale, Emnes Pratignes
De Laboratoire (B.P.L.) ISSN 0758-1998.
CHAPTER 25
Ethics in Experiments on
Animals*
DOUGLAS BROWN MCGREGOR
1 Introduction
Ethics is moral philosophy. It is a subject which is constantly reviewed,
analysed, and synthesised, reflecting the spirit of the time. It attempts to
identify the standards by which one ‘ought’ to behave.
Ethics is fraught with difficulties? as was recognised by Hume when he
stated that there is no deductive relation between an ‘is’ and an ‘ought’
(MacNabb, 1962). Thus, there is no way of deducing how one ought to
behave from observation of the world. But, if this view is correct, then it
puts in jeopardy all claims to moral knowledge, leaving ethics at the
mercy of subjective whim, against which no arguments will stand
(Scruton, 1984). The difficulties have been restated many times, for
example in this century by the pragmatist, C. I. Lewis (1946), ‘Valuation
is always a matter of empirical knowledge. But what is right and what is
just, can never be determined by empirical facts alone.’
Kant (Abbott, 1909) believed that the ‘is-ought’ problem could be
answered by practical reason, based upon the exercise of the Categorical
Imperative: ‘Act only on the maxim which you can at the same time will
as a universal law.’ Practical reason guides action and aims at rightness,
while theoretical reason guides belief and aims at truth. Practical reason
is nonsense without freedom of action and only rational beings can be
free in the sense required of morality. It is against this background that
we can attempt to answer the question, ‘is it morally right to involve
non-rational creatures (animals) in our efforts to understand the Universe
and to safeguard our position in it?’
Philosophy, unlike science, does not progress as new results are
established? replacing existing ones. Much of philosophy is restatement in
a contemporary context of ideas and arguments which may have existed
* Modified from Ethics of Animal Experimentation. Drug Metab. Rev.,1986, 17, 349-361.
542
Douglas Brown MeGregor 543
for centuries. For example, Judaism is the background from which

Christianity developed and Christianity is at the foundation of our
society, no matter how eroded this may have become today. Perhaps our
concern for animals has its basis in these religions. In the Old Testament,
ethics means conformity of human activity to the will of God and it is
stated in the Old Testament that man shall have dominion over all living
things. However-and it cannot be stressed too strongly-to have
dominion over something is not the same as having no consideration for
its welfare.
Of all the many ethical theories, the Judaeo-Christian ethic of love is
one of the very few that might have anything at all to say about animal
welfare. The ethic of love holds that there is only one basic ethical
imperative-love-and that all others are to be derived from it. Thus,
“‘Love the Lord your God with all your heart, with all your soul, with all
your mind”. It comes first. The second is like it: ‘love your neighbour as
yourself‘. Everything in the Law and the prophets hangs on these two
commandments. ” (N. E.B .) .
The second commandment would seem to set ethics of love on a
teleological basis, the ultimate standard of what is morally right or wrong
being the non-moral value that is brought about. The final appeal,
directly or indirectly, must be to the comparative balance of good over
evil produced. Thus,’an act should be done only if it produces more
good than evil than would result from not acting or from acting in an
alternative manner. Teleologists may hold very different views about
what is good in a non-moral sense. For example, a teleologist who is a
hedonist will identify good with pleasure, evil with pain. Other tele-
ologists may identify good with power, knowledge, self-realisation, or
perfection. In the current context, the moral question is whether an
experiment is right or wrong; the non-moral standard against which such
a question is set is whether the knowledge gained from that experiment
serves a purpose. The second commandment suggests that the purpose is
benefit to Man.
The first commandment, on the other hand, is not a teleological
principle. One can hardly say that loving God means promoting His
good. Only if one identifies loving God with loving His creatures, and
loving them by promoting for them the greatest balance of good over
evil, can one interpret the Judaeo-Christian ethic of love as a teleological
theory, It has a bearing on the use of animals in experiments only if one
includes non-human as well as human animals in one’s description of
God’s creatures. If such a broad interpretation is allowed, then the
ethical basis for objection to animal experiments requires absolute
obedience to God. For the person who rejects obedience to God, an
alternative basis for an ethical theory decrying animal experiments could
be the ancient reverence for a so-called life-force.
In the ancient world, people were fascinated by the ‘life force’ that
made the corn grow and the olives and grapes ripen. This awe-inspiring
544 Ethics in Experiments on Animals
‘life force’ was also responsible for the generation of Man, so it is readily
understandable that such an important concept should be worshipped.
An anti-vivisectionist view might be an appeal to a life-force, or some
modern version of it. If such a thing or concept is to be placed in high
authority, commanding reverence (and one might question why this
should be so), then all life is to be held in reverence, be it human, non-
human, vegetable, or microbial. With such a principle, actions may be
judged on the basis of those which are to bring about the greatest
expression of the life-force. But it leaves open to question whether or
not, for example, that expression is size, number, or level of
consciousness.
It very quickly becomes obvious that reverence for life-force cannot be
a basis for determining our actions. Many different living things inflict
harm on other living things simply so that they can survive. All animal
life is dependent upon this activity and evolution is a description of
changes where some individuals thrive better than others and result in
population imbalances, leading to either success of some or, with time,
disappearance of others. Life as such is not imbued with qualities which
can serve as a basis for our relationships with each other and with the rest
of the living world.
There may be some characteristic of certain forms of life which may be
a basis for a morality, in which case this should be identified. One such
characteristic is that we can see something of ourselves in animals. But
this characteristic is a qualitative, subjective one based upon human
perception which stirs feelings of guardianship and is far from the
objectivity sought in ethics. In the same way, a fit and able man can
adopt a position of guardianship over a sick man in whom he sees a much
more accurate reflection of himself than he would in any other animal.
Consequently, if a situation arises whereby the sick man can benefit if the
fit and able man causes harm to an animal, then the animal should be
harmed. Appeal to a life-force without qualification will result in the
animal remaining unharmed and the sick man remaining sick.
2 Human Experiments
The issue of experiments with animals could be avoided if human
subjects were used. It is argued that non-human animals make poor
models for Man anyway, therefore experiments with rodents, dogs, or
primates are a waste of time and money. This argument does have some
currency but experiments are performed on people also and it is
necessary to consider whether these are ethical if supporting information
is not already available from another source. Most human experiments
are ethical, in the judgment of those involved with their sanctioning.
Household products are tested upon human volunteers prior to market-
ing, but only after at least some animal testing has been completed.
Douglas Brown McGregor 545
Other, nonpharmaceutical, products such as pesticides are sometimes

tested in human volunteers; not necessarily to demonstrate some adverse
response, but to gain important information on metabolism to help
identify better non-human animal models with which to study the
toxicology of the substance. Pharmaceutical products entering Phase 1
clinical (tolerance) trials are subject to lay and expert review-Ethics
Review Committees in the UK, Institutional Review Boards in the
US-and the healthy volunteers are not expected to derive any direct
benefit from participation, or experience any significant health risk. The
Phase 1 experiments initially involve rising dose exposures to the
compound until the appearance of side effects. They are conducted under
careful clinical supervision, the volunteers are well informed about the
compound and the results of earlier animal tests, and these people are free
to withdraw at any time.
Any surgical advance is an experiment. Thus, the transplantation of
adrenal medullary tissues into the brains of two very advanced
Parkinson’s disease patients in Sweden, in an attempt to promote
regeneration of dopamine secreting cells, was certainly an experiment
and the result was far from certain at the outset (Backlund et al., 1985).
There was a risk of very serious side effects because the adrenal tissue
secretes a whole range of highly active pharmacological substances in
addition to dopamine, e.g. adrenalin, noradrenalin, opiates, etc. The
effects of such ectopically secreted substances were essentially unknown
prior to the experiment and there was virtually no improvement in the
patients’ condition. More recent experiments using a different technique
were more successful in alleviating the symptoms of Parkinsonism,
although the results are being met with skepticism (Madrazo et al., 1987).
It is unlikely that either experiment would have been undertaken without
extensive investigation upon an animal model: 6-hydroxydopamine-
denervated rats (e.g. Freed et al., 1981).
Any patient given a new chemotherapeutic cocktail is a human
guinea-pig. Heart and heart-lung transplantations are called ‘routine’
these days, but this is not so; they are still experimental, involving
combinations of surgical skill (probably gained on pigs), immunosuppres-
sive regimes, tissue-typing expertise, and live organ storage capability.
Unfortunately, some experiments upon people have not been either of
this benign nature or patient benefit orientated.
In Germany, under Nazi domination, atrocities were perpetrated
against prisoners of race and conscience, sometimes in circumstances
where wily criminal prisoners had ingratiated themselves with the guards
and were in ‘trusted’ positions. Little work of medical importance
resulted from these experiments. One example, chosen because it was by
no means the worst, and because it could be argued that benefits did
result, was the study on typhus immunisation by Dr. E. Ding (Mellanby,
1973). He prepared various anti-sera for the purpose of protecting people
from louse-borne typhus, which is caused by a rickettsia. He then
546 Ethics in Experiments on Animah
deliberately infected groups of people with the organisms causing the

disease; some of these had been protected by his different anti-sera while
some were unprotected controls. Scientifically, it was a well-conducted
experiment, and although many of those taking part died, including most
of the controls of course, Ding’s work formed the basis not only of
German, but also of British and Allied anti-typhus vaccine policy. Ding
appeared to think that the deaths caused by his experiments-about
250-were justified; about 5 million people died of typhus during and
after World War 11. It is to be hoped that those days are behind us.
Ding’s utilitarian view is indefensible since people have inviolable rights;
there are certain actions which should never to taken against them. All
innocent human beings have a right to life; it is wrong to kill innocent
human beings; and it is wrong to deny innocent human beings reasonable
straightforward protection against life threatening conditions (Gillon,
1985).
In contrast to this de-ontological position, another utilitarian view
states that pains of the same intensity and duration are equally bad
whether felt by humans or animals (Singer, 1976). In his arguments
against ‘speciesism’, Singer claims the difference between men and other
animals is our greater capacity for self-awareness, abstraction, complex
communication, and extended social relationships. This claim has been
taken up by a portion of the animal rights lobby-and unambiguously
stated by Singer himself-as a position which must grant equal or greater
right to life to animals than to retarded or senile human beings. Since
these words were uttered, the prospect of experiments upon human
embryos has become a reality, as a consequence of excess in vitro
fertilisations and of abortions. Embryos in research have again focused
attention upon our definition of humanness. Is there any fundamental
change which occurs during development to permit us to say the
conceptus is not human at one stage, but is human at another? It is
unlikely that any absolute change will be identified to support our
convenient definitions. One definition suggested has been the possession
of a rudimentary sense of awareness which might be equated to
sensitivity to pain. However, this has not prevented legal abortions at
ages where this ‘rudimentary sense of awareness’ is almost certainly
present. Such a definition would permit experiments upon suitably
anaesthetised people of any age. While permitting these acts, are we to
prohibit work with protozoans which do have sensory perception and will
react to stimuli?
3 Animal Experiments
A Their Necessity in Principle
The machinery of life is complex, Any electron micrograph of a cell
demonstrates something of the extent of that complexity. Nevertheless,
the picture is only a frozen moment in the death of that cell. It is a fossil.
Look at another cell and one can see that it is different. Each cell is an
individual representation of life, unique, in spite of the readily recog-
nisable sub-cellular organisation within it. Attempted definitions of Life
are irrelevant in this context and only detract from the essence of the
concept, Knowledge and truth are different.
Microanatomists have identified the structures, distorted though they
may be, in which chemists have recognised an abundance of reactions
working in a co-ordinated way, permitting survival and propagation.
Even the simplest cells are complex. In multi-cellular organisms, size
dictates that there should be specialisation since some cells are physically
better placed for some functions than others. The advantages of size are
not immediately obvious, but clearly they were sufficient to allow, for
example, Voluox to survive along side Chlorella.
The organisational step from Chlorella to Volvox is small, but the step
becomes greater as the phylogenetic tree is climbed until, in a very short
time, there is no single cell which is ‘typical’ of an entire organism. This is
all so obvious that it hardly needs to be mentioned and, when the effects
of a chemical upon any animal need to be known, it is also obvious that
there is ultimately no substitute for the exposure of that animal. Animals
are needed for research because it is currently impossible to model the
complex interactions of the biological machinery used by mammalian
systems by other than experimental animals or man. Perhaps the more
fundamental issue is do we need the research?
The form which the research may take varies greatly. Research using
animals may be in nutrition, e.g., the study of energy utilisation, amino
acid balance, the requirements of man or domesticated animals for
vitamins and minerals; in physiology, e.g., the study of renal or hepatic
function; in pharmacology, e.g., the study of muscular activity, neural
transmission, CNS function, cardiac activity of drugs used in medicine or
components of our diet; virology; bacteriology; parasitology, e.g., to
study host-parasite relationships and the relative responses of the host
and parasite to various treatment regimes; psychology, e.g., enabling us
to understand components of complex behaviour patterns; surgery, e.g.,
for the development of new techniques; and toxicology for the investiga-
tion of adverse responses to chemicals and the establishment of safe
levels of exposure.
For all of these purposes combined, enormous numbers of animals are
used, mainly in the industrialised, developed countries. There are
strengthening voices which say that these numbers are too high and that
they should be cut drastically. Some of the people involved in animal
liberation groups are more willing to countenance experiments upon
people who have been defined as mentally sub-normal (by how much is
not made clear) than experiments upon animals, and there is a much
larger proportion who are willing for experiments to be conducted on
legally aborted foetuses. These appalling suggestions are unethical on the
basis of most moral philosophies and dependent upon a background of

knowledge gained from non-human sources.
B The Cruelty to Animals Act, 1876 and The Animals

(Scientific Procedures) Act, 1986
One of the first pieces of legislation relating to animal experiments was
created more than 100 years ago, in the Cruelty to Animals Act, 1876. At
that time, perhaps a few hundred animals were used each year for
experiments. When legislation was first introduced, the main use of
animals was to extend the scientific basis for medical and veterinary
work.
Public opinion, however, had been aroused by description of cruel and
unnecessary experiments conducted in certain countries. Legislation was
passed through parliament without serious opposition and has remained
on the statute books ever since, without any substantial amendments. It
has, however, been repealed and replaced by new legislation, The
Animals (Scientific Procedures) Act, 1986. Only time will tell whether
moderate opinions are contented with the changes encompassed by the
new legislation; activist opinions are unlikely to be satisfied, but
appeasement is an ephemeral process and should not form the basis for a
document which may remain operative for many years.
The 1876 Act controlled all experiments which were likely to cause
pain in living animals. The purpose of an experiment had to be for the
advancement, by new discovery, of physiological knowledge or the
acquisition of knowledge useful for saving or prolonging life or alleviating
suffering. Experiments were absolutely prohibited in which the purpose
was to attain manual skills. Within these constraints, the Act authorised
the Home Secretary to license ‘suitable qualified’ persons to perform
experiments. But there were safeguards. No experimental animal was
allowed to suffer severe pain which was likely to endure.
In any experiment performed under licence alone, the animal had to be
under the influence of an anaesthetic to prevent it feeling pain. If pain
was likely to continue once the effect of the anaesthetic had worn off, or
if any serious injury had occurred, the animal had to be killed before it
recovered from the influence of the anaesthetic. Where these restrictions
frustrated the objectives of the experiment, then certificates could be
granted at the discretion of the Home Secretary which allowed the
requirement for anaesthetic to be relaxed. In effect the Act was exercised
through Cruelty to Animals Inspectors, who advised on all applications
for authority to conduct animal experiments and visited the laboratories
to ensure that the law was observed by the licensees.
Problems arose because the Act contained no definitions of terms and
because the conditions and scope of animal experimentation today are
very different from what they were in 1876. The Act had the object of
controlling experiments, but the term ‘experiments’ was not defined and
no definition was made in a court of law. Also, the definition of pain was
a difficulty because this is subjective and it cannot be measured even on
this basis in animals, except from an anthropocentric view point. This
problem persists in the new legislation.
The new Act requires licensing of people, places, projects, and
procedures involving ‘protected animals’ which may cause pain, suffering,
distress, or lasting harm for scientific purposes. The definition of a
‘protected animal’ is any living vertebrate, other than man, which has
developed beyond the mid-point of their foetal, embryonic, or larval
stages. The Act also controls the breeding and use of animals with
harmful genetic defects in scientific procedures, the production of
antisera and other blood products, the maintenance and passage of
tumours and parasites and, for purposes of scientific investigation, the
administration of any substance to dull perception. Breeding and
supplying establishments are also regulated for certain commonly used
mammalian species.
The change in conditions which has occurred since 1876 is the vast
increase in the tests on products from the pharmaceutical, cosmetic,
food, pesticide, and chemical industries to meet statutory obligations. In
the United Kingdom these may be in compliance with the Agriculture
(Poisonous Substances) Act, 1952, the Food and Drug Act, 1955, the
Medicines Act, 1968, Consumer Protection Act, 1971, and the Health
and Safety at Work Act, 1974. Equivalent legislation is to be found in
other countries. International bodies, such as the European Economic
Community (EEC) and the Organisation for Economic Cooperation and
Development (OECD) have developed chemical safety testing require-
ments or guidelines which have become law or need to be ratified by
legislation within the member countries.
This type of legislation is mentioned to remind us that manufacturers
would be breaking various laws if they did not test their products for
hazardous properties and that such testing does require animals.
C Number of Animals Used

Moderate and widely held opinion asks that toxicologists use fewer
animals and restrict the type of chemical which is tested on them; they
should not be subjected at all to cosmetics. There is also concern about
the type of animal used. There is certainly an emotional attraction about
such views, but they appear to embrace conflicting ideas.
If the argument for fewer animals is based on a belief that Man abuses
his privileged position by using other living creatures to serve his own
ends then the number of animals used is irrelevant. Each animal is an
individual and it does not diminish the individual animal’s plight if the
total is reduced. Man may persuade himself, by virtue of his reasoning
empathy, that the ethical status of an experiment is improved by smaller
numbers but this form of rationalisation involves a negation of the
5 50 Ethics in Experiments on Animals
perceived ethical principle and is forgetful of the relative isolation in an

animal’s experience, because of its inability to communicate at a distance,
and its lack of a highly developed language.
D Species
People are generally much more concerned about experiments involving
animals with which they can develop an emotional rapport, than those
involving animals where such relationships seldom occur. Thus, experi-
ments on a species which constitute domestic pets provoke dispropor-
tionate reactions, when compared with rodents or even certain non-
human primates. Yet, empathic behaviour is important among
primates-particularly apes-while it is probably of little or no sig-
nificance in rodents and is at best poorly developed in common domestic
pets. So, if one of these animals in an experiment experiences well-being,
discomfort, or suffering, these feelings are not shared by others unless
there is close proximity of the individuals in the experiment and empathy
is a possible experience. Even Man generally shows little concern for the
plight of others of his own kind. And yet, no matter what uncertainties
exist in our understanding of behaviour, Man’s ability to contemplate is
far higher than in any other animal. Nevertheless, because of these
uncertainties, we should probably refrain totally from experimental
toxicology on higher apes. Rarely is there argument for toxicology or
metabolism involving these creatures, in spite of there being certain
aspects of their physiology with which there are substantial human
similarities.
Judgement of whether experiments are required with these highly
intelligent animals or with domesticated animals should be made on a
case-by-case basis. This is already the usual situation with large or
expensive animals, but these considerations are made on practical or
economic grounds. There is a formidable price barrier to the indiscrimin-
ate use of dogs, primates, or horses, but this hurdle is much lower for
most other domesticated animals used in laboratories. Nevertheless,
scientists would be foolish to disregard the strong emotional bonds which
can develop between Man and domesticated animals. It is arguable that
the real issue is not the ethics of experiments on animals, but the
psychology of relationships between animals and men. Because they
accept obligations, men have rights; animals also would have rights if
they could accept responsibilities, but they cannot, living as they do
outside the sphere of rights and in a state of innocence. Perhaps it is this
innocence which is so beguiling to men’s emotions, or is it an underlying
gratitude for some received benefit, which may be as simple, but as
welcome, as unquestioning affection? A person having experienced this
relationship may then feel obliged to care for all individuals of a species.
Douglas Brown McGregor 55 1
E Chemicals
Safety testing of cosmetics and toiletries is the topic of much anti-
vivisection rhetoric. The arguments are: (a) these are trivial uses of
chemicals and (b) there are enough available anyway. There is less
criticism of testing pharmaceutical preparations. Here again the anthro-
pocentric view is expressed. The animal experiencing discomfort, pain, or
well-being is unconcerned about the human objectives of the experiment.
It is irrelevant for the animal whether the chemical is a cosmetic, a
pesticide, or a life-saving drug.
F Pain
Various states of suffering can be recognised, ranging from discomfort
(poor condition, reduced appetite, inactivity, avoidance behaviour)
through stress to pain (struggling, screaming, convulsions, severe palpita-
tions). Physical pain results from an interaction of neurological response
to injury with mind. The perception of pain, then, is subjective and
without some physical response it is not possible for another person to
know whether pain is being experienced. The special problem with
animal pain is that our projections of experience are even less likely to be
correct, when made into animals, than when they are made into other
people’s lives. However, this problem extends beyond pain to all animal
perception, be this visual, auditory, tactile, etc. It has been said that there
might be circumstances in which it is legitimate to restrict the freedom of
scientific investigation (Huxley , 1984). The position stated is that,
‘Provided a proposal is scientifically promising, it should normally be
permitted if the pain and distress involved is only ‘mild’, but if
‘substantial’, then the proposal should be scrutinised with care before
being approved. Severe, but short lived pain should be permitted only
exceptionally, e.g., perhaps occasionally for the investigation of pain
itself; severe and enduring pain, or excruciating pain, never.’ While this
opinion will have many supporters, one has still to determine the basis,
from the animal’s point of view, why it is occasionally, but not always,
permissible to inflict ‘substantial’ pain.
If the argument for never inflicting pain prevails, then it is possible that
the good of animals is the guiding principle; but to argue for inconsis-
tency suggests the presence of some other, usually unstated, Frinciple.
4 Conclusions
It is important that we afford the same rights to all members of our species.
If rights are to be gained only by the acceptance of responsibility, then
rational thought is a requisite. This attribute develops as one progresses
from infanthood to adulthood and may be lost with encroaching senility
or mental defectiveness, be this either congenital/hereditary or the result of
accident. Are animal rights to be acknowledged on a similar scale? T o d o
so would present a situation open to all kinds of abuse.
The fundamental question is whether experiments on non-human
animals are or are not to be done. The basis for this choice has been
inadequately explored here. But, if they are to be done, it is argued here
that objective concern for the animals used is not based upon numbers,
value of the experiment to Man, or the type of chemicals being tested.
Concern for the use of animals is widespread, the husbandry of our
fellow earth-dwellers being a matter to be treated seriously and respon-
sibly. Members of animal welfare groups have significant contributions to
make to these discussions of ethics in biological science, but some fringe
members of these groups are only helping to antagonise moderate
support when their morality allows them to harrass, frighten, or threaten
the families of people engaged in biological research. The values of these
demonstrators are not clear; they should take another look at the
principles upon which their own behaviour is based, then ask whether
this is how their fellow human beings should be treated. Is it right or
justifiable for a human being to cause pain, discomfort, or distress to
another human in the name of protecting animals from pain, discomfort,
or distress?
We cannot deny the historical framework in which we live. We cannot
dismantle the results of the industrial revolution. In that revolution,
which had its origins in Man’s curiosity about the physical universe, there
was a great deal of human suffering and it has been a slow process to
obtain general acceptance of the idea as a guiding principle that neither a
company nor an individual should present to the public a product which
will cause harm. Gewirth (1986) concludes that, ‘the rights of workers to
health and safety are of paramount importance and must override all
other considerations that may be adduced to remove or limit them.’ This
is a restatement of the de-ontological moral principle referred to earlier.
The choices open to us are:
(1) Abandon our industrialised society, leaving our future generations
without its bad aspects or its benefits, or
(2) ensure as best we can that we do not harm, knowingly or through
negligence, ourselves or our unborn children. If one accepts the
second of these choices, then we must take action to bring about
that situation. Toxicologists are not wedded to the use of animals
in their work. Their objectives are to identify hazard and to
quantify risk. If these objectives cannot be met, or can be obtained
in some other way which is better (i.e. more accurate, or more
efficient), then animals would be abandoned, but it is less likely
that animals, as experimental subjects, could becomes redundant
in many other branches of biology.
Acknowledgement
The author is grateful for the helpful suggestions made by Dr Roger
Scruton, Department of Philosophy, Birkbeck College, University of
London.
References
Abbott, T. K. (1909). Translator of I. Kant: ‘Groundwork of the Metaphysic of
Morals’, 6th edn. London.
Backlund, E.-O., Grandberg, P.-O., Hamberger, B. er al. (1985). Trans-
plantation of adrenal medullary tissue to straitum in Parkinsonism: first
clinical trials. J . Neurosurg., 62, 169-173.
Freed, W. J., Morihisa, J. M . , Spoor, H. E. et al. (1981). Transplanted adrenal
chrornaffin cells in rat brain reduce lesion-induced rotational behaviour.
Nature (London), 292, 351-352.
Gewirth, A. (1986). Human rights and the workplace. Am. J . Znd. Med., 9,
31-40.
Gillon, R. (1985). An introduction to philosophical medical ethics: the Arthur
case. Br. Med. J . , 290, 1117-1119.
Huxley, A. (1984). Anniversary address by the President. Proc. R. Soc. A . , 391,
215-230.
Lewis, C. I. (1964) ‘An Analysis of Knowledge and Valuation’. London.
MacNabb, D. G. C. (1962). Ed. of D. Hume: ‘A Treatise of Human Nature’,
1740. Book 3, Part 1, Section 1. Collins, Glasgow.
Madrazo, I., Drucker-Colin, R., Diaz, V., er al. (1987). Open microsurgical
autograft of adrenal medulla to the right caudate nucleus in two patients with
intractable Parkinson’s disease. New Engl. J. Med., 316, 831-834.
Mellanby, K. (1973). ‘Human Guinea Pigs’. Merlin Press, London.
New English Bible (1970). St. Matthew Chapter 22, verses 37-40. Oxford
University and Cambridge University Press.
Scruton, R. (1983). ‘A Short History of Modern Philosophy from Descartes to
Wittgenstein’. Ark, London.
tndex
Abnormality, 218 lathyrogenic, 227

Absorption, stathmokinetic, 225
neonatal, I34 A h locus, 205
of lead, 25 Ah receptor, 205, 206, 328
rate of, 6, 9 Air chamber, 363
sites of, 131 Air,
Accommodation, of animals, 88 conditioning, 5 I8
Accreditation schemes, 85 pollution, 108
Accuracy, 416 Alcohol, 93, 221, 232
Acetaldehyde, 387 Alcohol dehydrogenase, 142
Acetylation, 149 Alcoholism, 217, 220
Acetylcholinesterase, 72 Aldehyde oxidase, 142
Acid fuchsin, 121 Aldehyde reductase, 144
Act, Aldrin, 224
Animal (Scientific Procedures), 83, Alizarin red, 230, 237
103 Alkyl tin, 203
Cruelty to Animals, 83 Allantois, 363
Dangerous Wild Animals, 84 Allergy, 508
Health and Safety at Work, 506 Alveolar cells, 156
Additives. in food, 221, 229 Aveoli, 1 I5
ADI, 447 Amaranth, 191
Administration Ames test, 224, 247
dietary, 101 Amines, vasoactive, 21 1
oral, 100 Amino acid conjugates, 148
Adrenal medullary hyperplasia/neoplasia, 29 Amino acids, 319
Adrenal medullary neoplasia, 29 P-Aminopropionitrile, 227
Adrenals, 159 Ammonia, 88
Adulteration, of foodstuffs, 492 Amnion, 363
Aerosol, 12, 108 Amphibia, 361
liquid, 1 1 1 Amplification, 317, 318
solid, I l l Amyloid degeneration, 28
Aflatoxin B , . 68 Amyloid, deposits of, 126
Age, 166 Anaemias, 21 5
Age adjustment, 427 Anaesthetic gases, 220
Ageing, 23 Analysis,
Agent Orange, 215 component, 438
Agents, cost-benefit, 460
alkylating, 388 discriminant, 438
anti-cancer, 219 factor, 438
environmental, 243, 377 for 2 groups, 422
infectious, 2 I5 linear regression, 435
554
Irides 555
Analysis (cont.) Asphyxiation, 214

multiple regression, 438 Assay, plaque-forming, 369
of covariance. 438 Association, Food and Drug, 505
of variance, Wallis one-way, 429 Atmospheres,
of variance, one-way, 433 generation of, 10, 11
probit, 439 test, concentration of, 112
risk-benefit, 460 test, generation of, 1 1 1
sensitivity, 458 Atomisation, 1 1 1
statistical, 405, 406 Atrial thrombosis, 28
Anaphylatoxins, 21 1 Atrophy,
Anaphylaxis, clinical, 210 splenic, 126
Animal experiments, 546 testicular, 126, 386
Animal laboratory, 508 Autoradiography, 237
Animals, low temperature, 194
breeding. 93 Auxotrophy, 247
caging of, 339 Availability, apparent, 185
Chinese hamster, 20 Aves, 361
cynomolgus monkey, 19 Axis, hypothalamic-pituitary, 384, 388
defined, 83 Axon degeneration. 126
diet, 339 Azodye, 191
dog, 18, 95 Azo-reductase, 144
examination of, 101
facilities, 87
ferret, 20 B 16F10 melanoma cells, 207
gnotobiotic, 97 Bacteriophage, 248, 3 15
guinea pig, 18 Barrier,
hamster, 28 blood-testis. 378, 380
health, 103 maintenance of, 99
hen, 20 Bartlett’s test. 432
husbandry, 535 Bastocyst, 364
intact, 192 Bcdding, 89
laboratory, 16 Bench, laboratory, 5 1 1
maintenance, 93 Benlate, 224
marmoset, 19 Beryllium, 63
models in, immunotoxological, 206 Bezoar, 95
mouse, 17, 93 Bias, 405
multiple observations on, 438 Bile, 157, 195
numbers of, 26, 43, 413, 549 Binding covalent, 193
placement, 418 Bioactivation, 164
primates, 19 Bioassay, rodent carcinogenicity, 447
rabbit, 18, 94 Bioavailability, 185
rat, 16, 93 Biotransformation, 190
research, 83 Biphenyls,
species, 43 polybrominated, 204
supply, 85 polychlorinated, 205, 252
Syrian hamster, 20 Bisbiguanide, 210
transgenic, 327 Blindness, 92
transportation, 86 Blood samples, 44
Antagonists, folate, 225 Blot,
Antibiotic resistance, 315 Northern, 290
Antibiotics, beta-lactam, 208 Southern, 290
Antibody production, T cell dependent, 205 Bodies, YY, 285
x,-Antitrypsin deficiency, 24 ‘Brain growth spurt,’ 217
Application, topical, 342 Breath test, I4CO2. 193
Araldite, 121 Breeding, 91
Aromatic compounds, tricyclic, 205 continuous, reproductive assessment by,
Articles, test and control, 537 397
Aryl hydrocarbon hydroxylase, 167 of animals, 93
Aryl hydrocarbon responsiveness, 205 Breeding stock, 102
Ascorbic acid, 92 Bronchiolar tree, 114
Ash fly, 205 Bronchiolcs, terminal, I14
556 Index
C,, direct fixation of, 21 1 T suppressor, 203

Cabinets, safety, 513 Centrifugation, differential, 197
Cadmium, 225 Chambers,
accumulation in kidney, 72 exposure, 109
accumulation in liver, 72 inhalation, 109
compounds, 91 Chance, 405
ions, 219, 234 Changes, hormonal, 22 I , 236
Caffeine, 234 Chemicals, 309, 551
Cage, 88, 93, 94 dangerous, dispensing, 5 13
Caging, of animals, 88 dangerous, weighing, 5 13
Calcification, dystrophilic, 125 Chernoff test, 361
Calcium, 227 Chinese hamster hprt, 327
Cancer, somatic theory of, 243 Chi-squared statistic, corrected, 422
Cannabinoids, 387 Chloramide, 210
Capaci t a t ion, 394 Chlordane, 224
Captan, 224 Chlorfenvinphos, 67
Carbaryl, 224 Chlorhexidine, 210
Carbohydrates, 167 Chloridine, 224
Carbon disulphide, 68 Chlorination, 89
Carbon monoxide, 70 Chlorine, 210
Carbon tetrachloride, 60 Chondrodystrophy, 227
Carcinogenesis, 320, 334 Christianity, 543
dose threshold, 73 Chromosome,
Carcinogenicity, 10, 26 damage, 245
Carcinogenicity bioassay, rodent, 447 Philadelphia, 322
Cardiovascular disease, 125 X, 249
Cartilaginous, 366 Circulation, entero-hepatic, 157, 165
Cascade Impactor, 1 13 Cirrhosis, of liver, 61
Case-control study, 470 Clearance,
Case-referent study, 472 inulin, 184
Categorical Imperative, 542 renal, 184
Cavity, whole body, 184, 189
nasal, I17 Cleft palate, 226. 233, 234
sub-germinal, 363 Clinical chemistry, 49, 50
Cell death, 226 Clothing, protective, 508
Cell proliferation, 280 Co-operation. metabolic, 263
Cell transformation, 265 Co-oxidation, 143
Cell suspension, 195 Codes of Practice, 493
Cells, 246 Codon, triplet, 245
alveolar, 156 Coelenterates, 361
alveolar, type 11, 196 Cohort study, 470
Clara, 156, 196 prospective, 470
germ, 377, 387 retrospective, 470
glial, 369 Collagen, 60
granulosa, 381 Colon cancer, 323
in culture, 192 COMET assay, 259
kidney, 196 Comparative metabolism, 25
Leydig, 377, 382, 387 Comparison,
melanoma, B16F10, 207 in terspecies, 449
musclc, striated, 60 of two groups, 421
natural killer, 207 Comparisons, multiple, 407
nerve, 196 Compartment,
neural crest, 361 central, 187
null, 204 peripheral, 187
ovarian tumour, 361 ‘Compartments,’ 182
pulmonary, 196 Component analysis, 438
sarcoma, PYB6, 207 Components, sub-cellular, I50
Sertoli, 197, 377, 382, 386 Compounds,
T cytotoxic effector, 206 aromatic, tricyclic, 205
testicular, 197 organophosphorus, 71, 72
transformation of, 249 Concentrations, atmospheric, 107
557
Concordance, 270 univariate, 419

Conduct, DDC, 68
of study, 537 DDT. 61. 224
statistical, 405 Debrisoquine hydroxylase, 319
Confounding, 410 Defects. enzymic, 85
Conjugation, 145 Degeneration,
Contamination, 52 amyloid, 28
genetical, 98 axon, 126
Contingency table, 410 cellular, 123
Continuous data, myxoid, 125
stratified analysis, 436, 437 Degradation, 208
unstratified analysis, 433-436 Degrees of freedom, 4 I0
Control group, adequacy of, 418 Delaney amendment, 447
Control groups, 40, 476 Deletion, 253
Control, historical, 418 Delivery, 66
Conversion, metabolic, 209 Demyelination, 126
Copper, 227 Denmark, 527
Coronary events, 469 Dermal, 9
‘Correction for the mean.’ 434 Dermatosparaxis. 227
Correlation coeficient, Dermis, 155
Spearman’s rank, 437 Design,
product-moment, 437 of laboratory, 505
Cortex, 159 statistical, 405
Corticosteroids, 203 Detoxification, 160
Cortisone, 225, 227, 234 by metabolism, 67
Cosmids, 315 Development, 366
Cost-benefit analysis, 460 DFP, 71
Countries, European, 526 DHSS, 525, 526
Coxsackievirus, 228 Diabetes, 215, 225
‘Critical mass,’ 225 Diameter, aerodynamic, 1 1 1
Cupboards, fume. 5 1 1 Diazinon, 224
Curve, Di benzo-p-dioxins,
concentration-time, 184 accidents involving, 205
dose-response, 465 polychlorinated (TCCD), 205, 328
plasma concentration -time, 190 Dibenzofurans, polychlorinated, 205
Cyanide, 69 Dichlorvos, 224
Cycle. reproductive, 360 Dicoumarol, 92
Cyclophosphamide, 203 Dieldrin, 224
Cynomolgus monkey, 19 Diet, 8, 9, 89, 91, 93, 167
CYPlAI, 318 analysis of, 347
Cytochrome P-450, 136 animal, 90, 9 1, 339
Cytochrome P-450 mono-oxygenase system, fraction of, scaling factor, 455
67 restriction of, 31
Cytochrome P-450 superfamily, 137, 143 O,O-Diethyl-O-2-chloro-1 -( 2,4-dichloro-
Cytochrome oxidase, 70 phenyl)vinyl phosphate, 67
Cytolysis. depressed, 207 Di-(2-ethylhexyl)phthalate,386
Cytomegalovirus, 222 Diethylnitrosamine, 74
Cytotoxicity, 115 3,5-Diethyloxycarbonyl-1,4-dihydro-2,4,6-
trirnethylpyridine, 68
Data, Di-isopropylphosphorofluoridate, 71
acquisition, 488 N,N’-Dimethyl-4,4’-bipyridilium chloride. 62
continuous - see Continuous data Dimethylnitrosamine, 61, 62, 68
dissemination, 489 Diphenyl ether, 205
interpretation, 489 Discordance, 272
mu1t iva r ia t e, 420 Discriminant analysis, 438
paired, 438 Disease,
presenceiabsence - see PresencejAbsence background, 27
Data cardiovascular, 125
ranked - see Ranked data congenital, 360
raw, 531 Tyzzer’s, 124
recording, 4 19 zoonotic, 85
558 Index
Disfunction, immune, 204 Energy, heat, 91

Disorders, metabolic, 21 5 Engineering requirements, 520
Distribution, 134 Enterocytes, 196
binomial, 408 Environment, 87
chi-squared, 410 Environmental influences, 23
normal, 408 Environmental Protection Agency, 523
Student’s t-. 410 Enzyme interaction, 209
DNA, 223, 316 Enzymes,
cDNA, 315 oxidative, 121
DNA, hydrolytic, 121
intercalators, 388 Enzymic defects, 85
ligase, 315 Eosin, 121
precursor pools, unbalanced, 245 Epidemiology, 464
repair, 245 Epidermis, 155
synthesis, 225 Equations, Michaelis-Menten, 189
Dog, 18, 95 Equipment, 534
Dose, 15 Erythrocytes, polychromatic, 256
high, 42 Esterases, I51
human exposure/rodent potency, 456 Ethanol, 387
integrated, 77 Ethics, 542
intermediate, 42 Ethylene glycol monomethylether, 388
levels, 41 Evaluation.
low, 42 of risk, 459
maintenance, 187 statistical, 237
maximum tolerated ( M T D ) , 39, 447 Events, coronary, 469
multiple, I86 Exact trend test, 425
repetitive, 189 Examination, electron microscopic, 120
selection of, 343, 412 Exon, 318
threshold, 74, 448 Exotic species, 96
virtually safe (VSD), 449, 450 Experiment, duration of, 319, 415
Dose-response curve, 465 ‘Experimental unit,’ 420
diphasic, 266 Experiments,
linear, 266 animal, 546
Dose-response relationship. 74, 444 human, 544
Dosing, 100 Exposure chambers, 109
Down’s syndrome, 244 Exposure, 465
Drosophila, 224, 262, 362 inhalation, 316
Drugs, 108 nasal, 110
anti-cancer. 22 1 ‘nose only’, 13, 14
anti-viral, 221 ‘whole body’, 13, 14
prolactin releasing, 30 oral, 316
Duodenum, 157 routes of, 6
Duration, of experiment, 319, 415 Extraction ratio, 185
Extrapolation,
Echinodcrrns, 361 interspecies, 267
Ectodcrm, 122 linear, 450
ED,,,. 446 low-dose, 449
EEC Directive 86/609/EE, 505 Extrapolation method, linear, 450
Electron microscopy, 120 Eyes, 102
Elution, alkaline, 259
Elutriation, 196 Facilities, 535
Embryo, 219 Factor analysis, 438
whole, 361 Factors,
Embryogenesis. 361 confounding, 478
‘Embryonic rests,’ 226 safety, 15
Enclosure, 5 13 False negative rate, 414
Endocrine system, 384 False positive rate, 413
Endocrine tumours, 28 FDA, 523, 528
Endoplasmic reticulum, 151 ‘Feathering’, 187
Endoredu plica t ion, 253 Fecundity index, 397
Endrin, 224 Female fertility index, 397
Index 559
Ferret, 20 Glutaraldehyde, 121

Fertilisation, 381 Glutathione, 147
Fibrosis, myocardial, 125 conjugation with, 67
Filters, air, 99 Glutathione transferase, 146
Filtration, glomerular, 184 Glycine, 148
Fingerprint, 203 conjugation with, 67
Fisher’s exact test, 423 ‘GnRH test,’ 392
Flow, laminar, 51 I , 513 Gnotobiotic animals, 97
Fluid, seminal, 392 Gnotobiotic stock, 98
Fly ash, 205 Gonadotrophin releasing hormonc. 384
Foci, 321 Gonads, 159
r-Foetoprotein, 228 Gradient technique, 259
Foetus, 219 Groups,
Follicle, 385 control. 476
Food additives, 221, 229 number of, 421
Food Additives Legislation, 495 of people, 464
Food and Drug Administration (FDA), 523 Growth, arrested. 216
Food and Drug Association, 505 Guidelines, 469
Food deprivation, 90 Guinea pig, 18
Food Laws, 491
enforcement, 494 Haematology, 49, 50
Foodstuffs, adulteration of, 492 Haematoxylin, I2 I
Form, physical, 4 Haemoglobin, 70
Formaldehyde, 121, 467, 5 18 reaction with electrophilcs. 77
2-Formyl-1 -methyl pyridinium salt, 62 Haemorrhage, antepartum, 214
Fraction, microsomal, 137 Hair, 102
France, 527 Halothane, 209, 220
Freedom, degrees of, 410 Hamster, 28
Fume, 108 Chinese, 20
Fume cupboards, 51 1 Syrian, 20
Function, reproductive, 390 Haptens, 207
‘Functional development,’ 217 antigen formation with, 208
Hatchability, 262
Gap-junction intercellular tests, 282 Hatching, 363
Gametogenesis, 376 Hazard, 443
Ganglia, dorsal root, 197 Health and Safety at Work Act, 506
Gases, 108, 111 Heating, 87
anaesthetic, 220 Heliotrine, 68
Gastro-intestinal tract, 134 Hen, 20
Gavage, 8 Hepatitis. mouse, 124
Gene, Hepatitis virus, 100
amplification, 322 Hepatocytes, 196
cloning, 313 Heptachlor, 224
conversion, 259 Herbicides.
dominant, 243 diphenyl ether, 205
mutation of, 245 phenoxy, 205
recessive, 243 HERP, 456
Generations, future, 243 Herpes, 228
Genetic influences, 23 Herpesvirus simiae (B-virus), 100
Genetic polymorphism, 149 Heterogeneity, 42 1
Genetic purity, 85 Hexachlorophane, 62
Genetic risk, assessment of, 268 Hexachlorophene, 205
Genetic toxicology, 243 Hexane, 68, 386
‘Genomic library,’ 3 15 2,5-Hexanedione, 68. 386
Genotoxic carcinogens, 277 Hippocampus, 64
Germ cells, 377, 387 Histochemical methods. 121
Glands, mammary, 126 Histopathology, 44
Glomerulonephrosis, chronic, 125 Hormone,
Glucose-6-phosphatase, I22 follicle stimulating, 382
Glucuronic acid, conjugation with, 67 gonadotrophin releasing, 384
Glutamine, 149 1u t ci n isi n g . 3 82
560 I rtde?c
Hormone (cont.) Influences, genetic P . environmental, 23

luteinising-hormone releasing, 382 Influenza, 222, 228
luteotropic, 395 virus, 219
Hormones, ovarian, 395 Inhalation, 10, 52
Hormones, steroidal, 155 chambers, 109
Host susceptibility, altered, 206 exposure, 341
HSE, 525 toxicology, 107, 109
Human Exposure/Rodent Potency dose Inhibin, 383
(HERP), 456 Inhibition, differential zones of, 258
Human experiments, 544 Inhibitors, 142
Human population, 466 Inject ion,
Humidity, 87 intravenous, 15
Hurler’s syndrome, 225 p a r e n t e d , 342
Husbandry, of animals, 535 subcutaneous, 100
Hybrid vigour, 33 Insects, 361
Hybridisation, 317 Inspection, 530
Hydra, 362 Intake, daily, acceptable, 447
Hydrolysis, 135 Interaction, 410
Hydrophilic properties, 61 covalent, 57, 59
Hydroxyurea, 257 of enzyme, 209
Hygiene, occupational, 107 reversible, 57
Hygienist, 465 Interpretation, 480
Hyperplasia/neoplasia, adrenal medullary, statistical, 405
29 Intestine, small, 194, 195
Hyperplasia, parathyroid, 29 Intravenous injection, 15
Hypersensitivity, as immunotoxic response, Intron, 318
207 Inversion, 253
Hypertension, 214 In vitro, 10
Hyperthermia, 218, 220, 228 Ionising irradiation, 244
Hypervitaminosis A, 223, 226, 227 Ionising radiation, 225
Hypervitaminosis D, 92 Ions,
Hypochlorite, 210 cadmium, 234
Hypothalamus, 388 lead, 234
Hyperthermia, 95 metal, 221, 223
Hypoxia, 218, 220, 228 Irradiation, 223
ionising, 244
IARC, 268, 506 Irritation, 9, 10
IgE, 210 Isolator technology, 97
Imipramine. 232 Isolators, 99, 5 13
Immune function, in vitro, 206 Isotopes, stable, 193
Immunisation, typhus, 545 Italy, 527
Immunological markers, 86
Immunomodulation, inadvertant, 202 Janus green B, 226
Immunosuppression, 203 Japanese Pharmaceutical Affairs Bureau, 525
Immuno surveillance, 203 Judaism, 543
Immunotoxicant, 202, 203, 204 Junction, gap, 368
Imperative, categorical, 542
‘Inborn errors of metabolism’, 223 Karryorhexsis, 123
Inbred strains, 24 Kepone, 224
Incubation, 363 Kidney, 158
Index, cadmium in, 72
fecundity, 397 cells, 196
female fertility, 397 perfused, 195
live birth, 397 Kinetics, first order, 182
male fertility, 397 linear, 182
mating, 396 Michaelis-Menten, 182
Induction process, 142 Klinefelter’s syndrome, 244
Infection, 89, 100, 102 Kruskal- Wallis one-way analysis of
viral. 221 variance, 429
zoonotic, 86, 96
Inflammation, 123 Laboratory animals, 16
1ndc.x 56 1
Laboratory, Mantel-Bryan log-probit method, 451

animal, 508 Maps, 468
bench, 51 1 Markers, immunological, 86
design, 505 Marmosets, 19
general, 51 1 Mass, critical, 225
Laboratory practice, good, 84, 89, 102 Matched-pair f-test, 439
Laboratory Practice, Regulations ( G L P ) , Material, inhaled, 114, 117
36 Mating index, 396
Laboratory worker, protection of, 506 Maturation, 380
Lactate, 383 Maximum Tolerated Dose (MTD), 39
Lambda bacteriophage shuttle vector, 250 Mean, 409
LC,,, 14 Mean square, 434
LD,,, 439 Measles, 100
Lead acetate. 25 Mechanisms, studies of, 26
Lead, Medulla, 159
ions, 219, 234 Melanocytes, 368
compounds, 91 Membrane, I32
Legal aspects, 83 receptors, 383
Legislation, 548 plasma, 154
Legislation, Food Additives, 495 Mental retardation, 92
Lesions, of foot, 89 Mercapturic acids, 145
Level. ‘no effect,’ 286 Mercury,
Leydig cell, 377, 382, 387 poisoning, 220
tumours, 28 salts, methyl, 63
Life-force, 543 salts, phenyl, 63
Light, ultraviolet, 223 Mesencephalon, 367
Lighting, 87 Mesoderm, 122
Limb bud, 226, 361 Metabolic rate, 23
Lindanc, 224 Metabolism, 130
Linkages, covalent, 208 comparative, 25
Lipids, 168 detoxification by, 67
Lipophilic properties, 61 inborn errors of, 223
Live birth index, 397 Phase 1, 67
Liver, 158 Phase 2, 67, 135
cadmium in, 72 Metal ions, 221, 223
perfused, 194 Metallothionein-cadmium complex, 73
tumours, 28 Metals, heavy, 121
Longevity, 90 Metasystox, 63
Lung, 107. 134, 156 Methods,
perfused, 195 histochemical, 121
tumours, 28 pre-incubation, 252
Luteinisation, 386 Methylation, 149
Lymphocytes, ‘neurotoxic esterase’ in. 75 Met h y I -n - bu t y 1ketone, 68
Lymphoid tissue. 126 3-Methylcholanthrene, 146
Lymphoma, malignant, 28 2,2’-Methylenebis-3,4,6-trichlorophenol,
Lysozomes, 154, 156, 219 62
M H W , 525
McNemar’s test, 439 Michigan, PBBs in, 204
Macrophage activity, I 15 Microflora, 157
Magnesium oxide, 204 Micromass culture, 367
Malaoxon. 66 Micronutrients, 168
Malathion, 66, 224 Microsomes, 151
Male fertility index, 397 Milk, 230
Malformation, 360 Minamata. 21 5
Malnutrition, 90 Mirex, 224
Mammary tumours, 28 Miscarriage, 2 15
Man-years, 477 Mitochondria, 142, I50
Management, Mitogens, 256
of study, 533 Mitotic combination, 259
of risk, 460 Mixtures, complex, 340
Mann-Whitney U-test, 429 Model systems, host resistant, 206
562 Index
Model, dose threshold, 75

animal, 190, 191 Neuropathy target esterase, 71
‘flip-flop’, 185 Neurotoxic esterase, 7 1, 75
log-probit, 450 NIH3T3,321
mat hematical, 181 Nitro-reductase activity, 144
multi-hit, 453 Nitrobenzene, 393
multi-hit, gamma, 453 Nitroso-carbaryl, 224
multistage, 453 Nitrous oxide, 220
one-compartment, 182 Non-genotoxic carcinogens, 277
one-hit, 451 ‘No Observable Effect’ level, 447
pharmacokinetic, 453 Number of groups, 421
Weibull. 453 Nutrition, 89
Modelling, mathematical, 449 Nutritional status, 167
Models, immunotoxological, in animals, 206
Monkey, cynomolgus, 19 ‘Observational unit,’ 420
Monoamine oxidase, 143 Observations, multiple, on same animal, 438
Morphology, sperm, 393 Occupational hygiene, 107
Motility, spermatozoal, 392, 393 Oculomucutaneous. 209
Mouse, 17, 93 OECD, 524, 525
ascitic. 361 Oestradiol, 395
scoliotic, 85 Oestrogen, 385, 395
Mouse lymphoma t k . 327 Oncogene, 32 1
Mouth, 102 uhl, 322
MTD, 39, 43,447 f o s , 321
Mucus, cervical, 394 mvc, 321
Multigeneration. 397 N - m y c , 322
Muscle cells, striated, 60 ras, 321
Mutation, Oocytes, 381, 389
backward, 247 Oogenesis, 376
base-pair, 245 Oral, 8
deleterious, 244 Oral exposure, 341
dominant skeletal, 458 Organ, perfused, 192
forward, 247 Organisation for Economic Co-operation
frameshift, 245 and Development (OECD), 524
point, 322 Organisms, 246
Myocardial fibrosis, 125 Organochlorine pesticides, 224
Myocarditis, chronic fibrosing, 29 Organophosphorus
Myxoid degeneration, 125 compounds, 71, 72
pesticides, 224
NAD(P)H-cytochrome c reductase, 154 O,S,S-Trimethyl phosphorodithioate, 61, 62
Nails, 95 Outbred strains, 24
Nasal cavity, 117 Outliers, 43 1
Nasal exposure, 110 Ovarian function, 385
Necropsy, 348 Ovary, 126
Necrosis, 123 Ovotoxins, 389
of liver, 63 Oxidase activity, mixed function, 136, 193
Negative control, 41 Oxidase, monoamine, 143
Negative, false, 461 Oxygenase system, mixed function, 209
Nematodes, 361
Neonatal absorption, 134 p53, 323
Neoplasia, adrenal medullary, 29 P450, 318
Neoplasms, Pachytene, 378
pancreas-endocrine, 29 Pain, 83, 551
pancreas-exocrine, 29 Pair-feeding, 93, 101
thyroid-C-cell, 29 Palate, 361
Nephrocalcinosis, 125 cleft, 226, 233, 234
Nephropathy, chronic, 24 Pancreas-endocrine, neoplasms of, 29
Nephrosis, 125 Pancreas-exocrine, neoplasms of, 29
Nerve cells, 196 Paraffin wax, 121
Neurons, 60, 367 Parallelogram approach, 286
Neuropathy, delayed, 71 Paraquat, 62, 64
Index 563
Parathion, 224 Pre-incubation method, 252

Parathyroid hyperplasia, 29 Precision, 47
Parenteral injection, 342 Preparations, hormonal, 229
Particle size, 5, 12 Presenceiabsence data,
distribution, 1 12, 1 13 stratified analysis, 426, 427
Particles, ‘respirable’, 13 unstratified analysis, 422-426
Partition coefficients, 132 Pressure sensor, I I 5
Pathology, 49, 50 Primates,
Pathway, oxidative, 210 cynomolgus monkey, 19
Penetration, of chemicals, 61 marmoset, 19
Penicillenic acid, 208 Principles, 21 7, 2 18
Penicillin, 208 embryological, 21 6
Penicilloyl conjugate, 208 Probit analysis, 439
Pens, 88 Procedure,
Peptides, regulatory, 27 autoradiographic, I93
Periarteritis, 33 multistage, linearised. 453
Period, latent, 472 non-invasive. 192
Peroxisome proliferator activated receptor standard operating, 529
(PPAR), 328 Process, induction, 142
Peroxisomes, 154 Profile,
Personnel, 534 immunological, 203
Pesticides. 22 1, 229 immunopharmacological, 204
organ oc h I o rine , 224 Prolactin, 384
organophosphorus, 224 Prolactin-releasing drug, 30
Pharmacokinetics, 164, 181 Properties,
non-linear, 189 hydrophilic, 61
Phase, lipophilic, 61
diploid, 262 Prophage, 248
haploid, 262 Prostaglandins, 385
Phenobarbitone, 146 Prostate, 380, 389
Phenyl saligenin phosphate. 68 Protection, of workers. 5 13
Phenylketonuria, 225 Protein, 167
Phosphate conjugates, 149 androgen binding, 380
Physical form, 4 binding, 220
Phytohaemagglutinin, 202 reactivity. direct, 208
Pisces, 361 Protocol, 87, 530, 537
Pituitary tumours, 28 experimental, 93
Placenta, 160 Purine, 225
chorioallantoic, 219 Purity, genetic, 85, 86
yolk sac, 219 p-Value. 406
Placentation, 365 PYB6 sarcoma cells, 207
Plasma membrane. I54 Pyknosis, 123
Plasma, seminal, 392 Pyrimidine, 225
Plasmids, 3 15 Pyruvate. 383
Plug, copulatory, 365
Point mutation, 322 QSAR, 445
Pollution, air, 108 Quality assurance. 98
Polybrominated biphenyls (PBBs), 204 Quality Assurance Unit, 530
Polychlorinated biphenyls, 252 Quarantine, 86, 98, 99
Polycyclic aromatic hydrocarbons, 389 Questionnaire, 479
Polymerase chain reation (PCR), 317
Polymerisation, 208 Rabbit, 18,994
Polymorphism, Radiation, 389
genetic, 149, 167, 192 ionising, 220, 225
restriction fragrncnt length, 3 15 Radicals, 244
Population, human, 466 Radionuclides, 194
Positive control, 41 Randomisation, 49, 51, 52, 417
Positive. false, 461 Ranked data,
Power, statistical, 446 stratified analysis, 430, 431
Practolol, 209 unstratified analysis, 428, 429
Pralidoxime, 62 Rat, 16, 93
564 Index
Rate constant, Risk, 443, 465

absorption, 186 assessment, 442, 443
bimolecular, 59 estimates, quantitative, 449
elimination, 188 evaluation, 459
first-order elimination, 183 genetic, assessment of, 268
Rate, management, 460
false negative, 414 Risk-benefit analysis, 460
false positive, 413 RNA, 316
metabolic, 23 Rubella, 216, 220, 221, 225, 228
Ratio, extraction, 185
Rats, ovariectomised, 390 Safety evaluation, 26
Reactions, Safety factor approach, 447
autoallergic, 202 Safety factors, 15
drug, adverse, 446 Salicylates, 219, 220, 224, 225, 227
hydrolytic, 67 Samples, blood, 44
Phase 1, 67, 135 Sampling, blood. 102
pseudo-allergic, 2 10 Scaling ,
reversible, 57 based on body weight, 455
Reagents, 535 based on surface area, 456
Receptors, membrane, 383 Scandinavia, 527
Recombination, mitotic, 259 Schedule, master, 530
Recording, of data, 419 Sendai virus, 124
Records, 102 Sensitivity, 270, 41 1
Reduction, 144 Sensitivity analysis, 458
Region, Sensor, pressure, 115
centrilobular, 158 Sertoli cells, 377, 382, 386
periportal, 158 Serum protein, assay, 86
Regression, 4 1 1 Serum, production, 97
analysis, linear, 435 Seveso, 205
analysis, multiple, 438 Sex, 166
Regulated procedure, 84 SF, 447
Regulations, Sialodacroadenitis, 124
Good Laboratory Practice (GLP), 36 Sign test, 439
quarantine, 86 Significance, statistical, 406
Relationship, Single-strand conformation polymorphism,
conceptus, 369 325
dose - response, 444 Skin, 102, 133
maternal, 369 grafting, 86
Replication, 49, 51, 52 Smoke, 108
Report, 533, 538 Smoking, 467
final, 530 Society, industrialised, 552
Reproduction, 91 Sodium aurothiomalate, 234
Reproduction studies, multigeneration, 52, Sodium chloride. 233
396 Solutions, 535
Reproductive Assessment by Continuous Solvents, 468
Breeding, 397 Somites, 364
Reproductive system, female, 380 Species, 338, 549
Reproductive system, male, 377 choice of, 16, 391, 412
Reproductive toxicology, 37 differences, 164, 165, 168
Requirements, specific, 529 exotic, 96
Resin, epoxy, 121 primate, 96
Resistance, 249 similarities, I92
Response, ‘Speciesism,’ 546
variable, 420 Specificity, 270, 41 1
probit, 450 Specified pathogen free (SPF), 99
Restriction Fragment Length Sperm,
Polymorphism, 3 15 morphology, 393
Restriction enzymes, 3 15 numbers, 390
Retardation, mental, 92 production rate, 390
Retroviruses, 321 Spermatids, 388
RFLP, 315 Spermatocytes, 197, 378
Index 565
Spermatogenesis. 236, 376, 378 steroid, 379

Splenic atrophy, 126 System,
Square, mean. 434 endocrine, 384
Squares, sum of, 434 immunological, 203
Stain, metabolising, 365
Van Gieson, 121
periodic Schiff, 121 Table, contingency, 410
Standard Operating Procedures. 529 Taurine, 148
Standard deviation. 409 TCDD, 146, 205
Standard error, 409 Teeth, 95
Stathmokinetic agents, 225 Temperature, 87
Statistical evaluation, 237 Teratogenesis, 222
Statistics, 405 principles of, 217
Status. nutritional, 167 Teratogens, non-, 362
Steady state, 186 Test,
Sterilisation, 89, 94, 98 Ames, 247
Stock, exact trend, 425
breeding, 102 Fisher’s exact, 423
gnotobiotic, 98 ‘GnRH,’ 392
Strains, 338 ‘Testes barrier,’ 264
differences, I67 Testicular atrophy, 126
inbred, 24 Testis, size, 391
outbred, 24 Testosterone, 377, 378
Stratification, Tests, mutagenicity, 246
in experimental design, 416 cells used in, 246
in statistical analysis, 421 organisms used in, 246
Structural activity relationships, Tetracycline, 227
quantitative, 445 Tetraethyl-lead, 68
Study, 530 Thalidomide, 215, 219, 220, 225, 229, 232,
case-control, 470, 472 233, 235, 238
case-referent, 472 Threshold Limit Value (TLV), 444
cohort, 470 Thrombosis, atrial, 28
cohort, prospective, 470 Thyroid, 125
cohort. retrospective. 470 Thyroid-C-cell, neoplasms of, 29
cross-sectional, 475 Tissue, lymphoid, 126
descriptive, 476 TLV, 444
design of, 44, 45, 46 Toluidine blue, 121
duration of. 47 2-Tolyldiphenylphosphate.mono-, 68
duration of, acute, 47 Topical application, 342
duration of, chronic, 48 Toxicity,
duration of, sub-acute, 47 acute, 59
duration of, sub-chronic, 48 chronic, 26, 60
ED,,,446 developing, phases of, 65
epidemiological, 444 reproductive, principles of, 217
intervention. 476 selective, 63
management, 533 Toxicology,
metabolic. 190 genetic, 243
mortality, 474 inhalation, 107, 109
Substrates, reproductive, 37
Type I, 137 Tract, gastro-intestinal, 134
Type 11, 137 ‘Traits,’ 243
Sulphotransferases, 148 Transfection, 321
Sum of squares, 434 Transformation, of cell, 249
Synacthen, 210 Transformations, metabolic,
Syndrome, phase I, 196
Down’s, 244 phase 11, 196
Hurler’s, 225 Transgenic, 250
Klinefelter’s, 244 Translocation, 255
Turner’s, 244 Transplantation, 545
Synthesis, Transport.
dr not‘o, 258 passive, 131
566
Transport (cont.) Vehicle, 5, 8

specialised, 133 Vehicle control, 40
Trend, 421 Ventilation, 88, 99, 518
Triethyl-lead, 68 Vesicles, seminal, 389
Triethyltin chloride, 62 Vinyl chloride, 69, 284
Triethyltin hydroxide. 62 Vinyl chloride monomer, 471
Trimethyltin. 64 Virus,
Trypan blue. 219, 234 influenza. 219
Tubules, scminiferous, 377 Sendai, 124
Tumours, 102, 205. 334 Viruses, 353
chemically induced, 324 Vitamin A, 92, 225
endocrine, 28 Vitamin C, 92
fatal. 428 Vitamin D,, 93
genetic, 351 Vitamin E. 92
hormonal, 352 Vitamin K, 92
incidental, 428 Volume,
Leydig cell, 28 apparent, 189
liver. 28 distribution, 183
lung. 28 VSD, 449,450
mammary, 28
pituitary, 28 Water, 89
spontaneous, 90, 325 Wax, paraffin, 121
visible, 428 Weibull model, 453
Tumour suppressor genes, 322 W H O , 525
Turner's syndrome, 244 Wilcoxon matched-pair signed-ranks test,
Typhus, immunisation, 545 439
Tyzzer's disease, 124 Wilcoxon rank sum test, 429
Wild state, 92
UDP-glucuronyltransferase, I45 Woodworkers, 467
U K Health and Safety Executive (USE), 525 World Health Organisation (WHO), 525
Ultraviolet light. 223
Units, immunogenic, 207
X chromosome, 249
Urinalysis. 50
Xenopus, 363
Uterus. 389
Vaccination. 86. 103 Yolk sac, 363

Vapours, 108, I l l YY bodies. 285
Variance. 409
homogeneity of, 431 Zinc, 389
Vasopressin deficient, 85 Zoonotic diseases, 85
Vector. 313 Zoonotic infections, 86, 96

Diana Anderson, D M Conning - Experimental Toxicology-Royal Society of Chemistry (1993) PDF

Uploaded by

Copyright:

Available Formats

Diana Anderson, D M Conning - Experimental Toxicology-Royal Society of Chemistry (1993) PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Diana Anderson, D M Conning - Experimental Toxicology-Royal Society of Chemistry (1993) PDF

Uploaded by

Copyright:

Available Formats

EXPERIMENTAL TOXICOLOGY

The Basic Issues

ISBN 0-85186-451-1 (Hardback)

First published in hardback 1988

Second Edition 0 The Royal Society of Chemistry 1993

All Rights Reserved

Published by The Royal Society of Chemistry,

Printed and bound in the UK by Hartnolls Ltd., Bodmin

W. N. Aldridge formerly The Robens Institute,

Chapter 1 Introduction to Experimental Toxicology

Chapter 2 Effects of Physical Form, Route, and

Chapter 3 Influence of Animal Species, Strain, Age,

Chapter 4 Experimental Design

1 Regulatory and Experimental Toxicology 35

Chapter 5 The Biochemical Principles of Toxicology

Chapter 6 Animal Husbandry

Chapter 7 Inhalation Toxicology

Chapter 8 Histopathology in Safety Evaluation

Chapter 9 The Metabolism and Disposition of

Chapter 10 Theory and Practice in Metabolic Studies

1 Theoretical Considerations 181

Chapter 11 Immunotoxicology-Conceptual Problems

Chapter 12 Perspectives-The Evaluation of Reproductive

Chapter 13 Genetic Toxicology

3 Usefulness of Some of the Systems for

Chapter 14 Molecular Toxicology

Chapter 15 Testing for Carcinogenicity

Chapter 16 In Vitro Methods for Teratology Testing

Chapter 17 Assessing Chemical Injury in the

2 Anatomy and Physiology of the Male

Chapter 19 Risk Assessment of Chemicals

Chapter 21 Information and Consultancy Services in

Chapter 22 Regulations and Advisory Requirements in

Chapter 23 The Influence of a Growing

Chapter 24 Good Laboratory Practice

Chapter 25 Ethics in Experiments on Animals

Toxicology is defined, classically, as the study of the adverse effects of

study of the adverse effects of chemicals on living systems in order to

human communities against such consequences. The diligent pursuit of

Eflects of Physical Form,

B Vehicle and Concentration

standard reference on inhalation toxicology. While it is relatively simple

Generation of Gases, Vapours, and Aerosols

been the first inhalation experiment in the literature to report uninten-

Particle Size as a Variable

penetrate to the alveoli. Below that diameter an increasing proportion

Design of Inhalation Experiments

no. of parts by volume of the compound

i.e. 10p.p.m. means 10 ml of the vapour in lo6 ml of air. It is a volume:

mg m-3 and mg 1-1 are weight:volume measurements suitable for

i.e. 10mgm-3 means 10mg of the compound (vapour, liquid, or

mgm-3 = Pep .m. x molecular weight

E Intravenous and Other Routes

Bailey, K. (1983). Physiological factors affecting drug toxicity. Regulat. Toxicol.

Znfluence of Animal Species,

2 Genetic Versus Environmental Influences

3 Versus Outbred Strains

5 Numbers of Animals: Safety Evaluation as

The experimental requirements for mechanism studies and safety