Developmental and Comparative Immunology 25 (2001) 725±743

www.elsevier.com/locate/devcompimm

The acute phase response and innate immunity of ®sh

Christopher J. Bayne*, Lena Gerwick
Department of Zoology, Center for Fish Disease Research and Environmental Health Sciences Center, Oregon State University, Corvallis,
OR 97331, USA
Received 25 October 2000; received in revised form 15 January 2001; accepted 17 January 2001

Abstract
Tissue trauma or invasion by pathogens or parasites induce changes in the quantities of several macromolecules in animal
body ¯uids. These changes comprise one aspect of the acute phase response (APR), which in toto involves metabolic changes in
several organ systems. One clear indication of the response is the increase in synthesis and secretion by the liver of several
plasma proteins, with simultaneous decreases in others. These acute phase proteins (APP) function in a variety of defense-
related activities such as limiting the dispersal of infectious agents, repair of tissue damage, inactivation of proteases, killing of
microbes and other potential pathogens, and restoration of the healthy state. Some APP are directly harmful to microbes, while
others modify targets thus marking them for cell responses. Some work alone while others contribute to cascades. Proteins that
are APP in mammals, and that have been identi®ed in both teleosts and elasmobranchs include C-reactive protein, serum
amyloid P, and several components of the Complement system. Others reported in teleosts include transferrin and thrombin. Of
these, only CRP has been reported to increase in acute phase plasma. In trout, a precerebellin-like protein is an APP with
unknown functions. A cDNA library enriched in fragments of transcripts that were more abundant in livers from ®sh under-
going an APR recently yielded sequences resembling 12 additional known APP, and as many others either not known to be
APP, or not similar to others yet in public databases. It appears that, as in mammals, hepatocytes are the prime source of APP in
®sh, and that pro-in¯ammatory cytokines induce transcription of their genes. q 2001 Elsevier Science Ltd. All rights reserved.
Keywords: Acute phase protein; Fish; Acute phase response; CRP; SAA; TCBP1; Transferrin; Pre-cerebellin-like protein; Fibrinogen; In¯am-
mation; Innate immunity

1. Introduction and mobile cells, and a large number of molecules

dissolved in body ¯uids. In comparison with the adap-
Innate immunity is the product of a varied and tive immune system in which all cells are lympho-
complex group of phenomena. It involves both ®xed cytes and all effector molecules are immunoglobulin
superfamily members, both the cells and the effector
molecules of the innate system are highly diverse
* Corresponding author. Tel.: 11-541-737-5352; fax: 11-541-
737-0501. (Table 1). Viewed in evolutionary context, this is
E-mail address: baynec@bcc.orst.edu (C.J. Bayne). not surprising; the lymphoid system is a relatively
Abbreviations: APR, acute phase response; APP, acute phase recent evolutionary development [1]. Prior to the
protein; cDNA, complementary DNA; CRP, C-reactive protein; origin of jawed vertebrates (gnathostomes), all
FIA, Freund's incomplete adjuvant; IL, interleukin; mRNA,
animals relied exclusively on innate, non-lymphoid
messenger RNA; NF, nuclear factor; RAG, recombination activat-
ing gene; SAA, serum amyloid A; TCBP, trout C-polysaccharide mechanisms to defend themselves against potential
binding-protein; TNF, tumor necrosis factor. colonization by viruses, bacteria, fungi, protists and
0145-305X/01/$ - see front matter q 2001 Elsevier Science Ltd. All rights reserved.
PII: S 0145-305 X(01)00 033-7
726 C.J. Bayne, L. Gerwick / Developmental and Comparative Immunology 25 (2001) 725±743

Table 1
Cells and effector molecules of the adaptive and innate arms of vertebrate immune systems. A predominant element of the acute phase response
is the increased concentrations of several humoral components of the innate arm (in bold)

Adaptive Innate

Cells T-, B-cells NK cells, monocytes/macrophages, granulocytes (predominantly neutrophils)

Tissues lymphoid liver, spleen, pronephros
Regulators cytokines cytokines
Humoral components Igs Complement system, clotting system, anti-proteases, metal-binding proteins, lectins, lysozymes,
antimicrobial peptides, opsonins
Kinetics slow fast

metazoan parasites, and to repair tissue damage. injury or infection' [7]. In its broadest context, the
Accordingly, a varied set of defenses and repair APR involves changes in at least the hepatic, neuroen-
processes evolved to meet these challenges [2]. docrine, hematopoietic, musculo-skeletal and immune
From the perspective of an evolutionary immunolo- systems. Additionally, involvement of the central
gist, the adaptive immune system arose as a means to nervous system is apparent through somnolence,
help the innate system function more effectively. fever, anorexia and other physiological parameters
The terminology of `adaptive' and `innate' came [8,9]. In the view of Kushner and Mackiewicz [8]
into use before the versatility of the innate immune the APR `represents the substitution of new `set
system was appreciated, but when antibodies and points' and a setting aside of many of the homeostatic
acquired cell mediated immunity constituted the processes that normally maintain stability in the face
primary focus of immunologists. The terms imply of an inconstant external environment'. Thus it is
that whereas T- and B-cells respond (`adapt') to anti- appropriate to use the term homeostasis to describe
gens, components of innate immunity are constitutive, the condition of not being in an acute phase response.
innate or `®xed'. This is far from reality [3,4]; both The APR is induced by plasma-borne signals.
cells and secreted molecules of innate immunity are These signals are the so-called pro-in¯ammatory
dynamic in regard to their readiness status. Other cytokines such as IL-1, IL-6 and TNFa. Their synth-
papers in this issue [5,6] address cell activationÐ esis and release are provoked by stimuli such as those
the process that constitutes the adaptive alteration of that result from wounding (e.g. kinins, tissue factors,
cell physiologyÐin the innate arm of the immune and metabolites of arachadonic acid) and as a conse-
system. Here we address adaptive alterations of the quence of microbial and parasitic colonization (e.g.
humoral component of the innate arm. We begin by lipopolysaccharides). The pro-in¯ammatory cyto-
identifying acute phase proteins (APP) and brie¯y kines induce altered rates of plasma protein synthesis.
describing the acute phase response (APR) in the The great majority of plasma proteins are synthesized
animals in which these are best knownÐthe in hepatocytes, though sites of extra-hepatic synthesis,
mammals. We then review what is known of APP such as the brain and leukocytes, are known for some.
and the APR in ®sh, and end by discussing the likely For many proteins, levels rise in the plasma even
value of further investigations into this complex set of before a septic state can be otherwise recognized
phenomena. [10]. The proteins whose rates of synthesis increase
in response to such stimuli include several that are
involved in repair of tissue damage, in ®ghting infec-
2. What is the acute phase response (APR)? tion, and in restoring the healthy (homeostatic) state
that prevailed before the stimulus [11]. On account of
The APR is a pervasive physiological response of this, the most responsive of these so-called acute
the body to injury, trauma or infection (Fig. 1). It has phase proteins (APP) have been used as indicators
been de®ned as `the entire array of metabolic and of the extent of tissue damage, or of incipient septi-
physiologic changes which occur in response to tissue cemia, in, for example, accident victims and patients
 C.J. Bayne, L. Gerwick / Developmental and Comparative Immunology 25 (2001) 725±743 727

Fig. 1. Diagram illustrating phenomena that induce an acute phase response, their immediate effects, and the mediators that are produced and
that induce systemic reactions (modi®ed from Koj and Gordon [12]).

after major surgery [12] and those suspected of septic quently, to avoid increasing the osmotic pressure and
infections of the central nervous system [13]. More viscosity of the blood, increases in levels of some
recently, the plasma level of one major acute phase plasma molecules are accompanied by decreases in
protein (C-reactive protein) has been found to have others. The energetic cost of protein synthesis [15]
clinical diagnostic value for predisposition to coron- may be another reason for reducing synthesis of
ary disease [14]. proteins that may not need to be maintained at existing
As physiological demands change, the ideal levels during the APR. We do not yet know if
composition of blood plasma changes. For example, decreased plasma levels result from reduced synthesis
when potential pathogens or other foreign agents need and secretion, or increased clearance, or both. In the
to be neutralized, or when damaged tissue needs to be APR, plasma proteins that increase by 25% or more
cleared away, the optimum levels of some plasma are called positive APP, and those that decrease by
proteins will be different from their optimum levels 25% or more are called negative APP.
when the body is in a healthy, homeostatic state. The
appropriate (adaptive) response to such circumstances
would seem to be to increase the concentrations of 3. Plasma components of the APR
defense and `clean-up' molecules in the plasma.
While this is not wrong, it is incomplete: the osmo- The acute phase proteins have been grouped
larity and viscosity of blood are determined by its total according to the extent to which their concentrations
solute concentration, and determine, in part, the work change (minor, intermediate and major), and the
that the heart must do to circulate the blood. Conse- direction of change (positive and negative), during
728 C.J. Bayne, L. Gerwick / Developmental and Comparative Immunology 25 (2001) 725±743

Table 2
Acute phase proteins known in mammals (based on Gabay and Kushner [9] and Steel and Whitehead [10], which should be consulted for
citation of the original papers). Note: Alpha-2-macroglobulin is an additional APP in rats [108] and ground squirrels [76]

Group 1 (Minor and intermediate APP; concentrations may increase by up to three times the initial concentration; n ˆ 14).
Coagulation factors: ®brinogen, prothrombin, Factor VIII, plasminogen.
Proteinase inhibitors: heparin cofactor II, plasminogen activator inhibitor I, a1-antitrypsin, a1-antiplasmin, a1-antichymotrypsin.
Transport proteins: a1-acid glycoprotein, haptoglobin, hemopexin, ceruloplasmin, ferritin.
Group 2 (Intermediate APP; concentrations may increase three- to ®ve-fold; n ˆ 10).
Complement components: Complement(C)1s, C2, factor B, C3, C4, C5, C9, C4b-binding protein, C1 inhibitor, mannose-binding protein (also
known as mannose-binding lectin).
Group 3 (Major APP; concentrations may increase up to 1000-fold; n ˆ 4).
Serum amyloid A1, Serum amyloid A2, C-reactive protein (in man) a, Serum amyloid P (in mice) a.
Group 4 (Negative APP; concentrations decline; n ˆ 9).
Albumin, prealbumin, transferrin, apolipoprotein A 1 and A 2, properdin, a2-HS glycoprotein, inter-a-trypsin inhibitor, histidine-rich
glycoprotein.
a
CRP and SAP are both pentraxins, and are considered to have evolved from a common ancestor [10,55]. This is discussed in the text.

an acute phase response [10], or according to impressive capacity to speci®cally recognize and
function [9]. Table 2 combines these. respond to a great variety of non-self determinants
may have reduced some of the selection pressures on
the innate system. In any case, it appears reasonable to
4. Reasons to expect robust acute phase responses expect that, compared to more derived taxa (birds,
(APRs) in ®sh mammals), the ®rst taxon to have the bene®t of RAG
genes and lymphocytes would have retained an APR
Despite their lack of a recombination activating more closely resembling that which typi®ed ancestral
gene (RAG) system for the generation of diversity in groups that relied exclusively on mechanisms of innate
B- and T-cell receptors [1], the agnathans (jawless immunity. At least three sets of empirical data support
®shÐhag®sh and lampreys) and all invertebrates this contention: ®rst, ®sh injected with bacteria in FCA
cope successfully with potential pathogens. As the acquire the ability to survive a challenge before an
RAG and lymphoid systems are ®rst evident in adaptive humoral immune response (serum antibody)
gnathostomes [16], their immediate ancestors are can be detected [20]. Second, while not as effective as
presumed to have relied entirely on innate means of adjuvant and antigen, adjuvant alone will induce a state
internal defense. We have reason to believe that these of heightened resistance to pathogen challenge,
animals had robust APRs. First, the ability to mount an presumably due to stimulation of the innate immune
APR was present long before the origin of the chor- system [20,21]. Third, salmon in which the ability to
dates: the essential elements of an APR are evident in produce serum antibody is severely compromised (by
arthropods. In insects, defense molecules (notably neonatal exposure to steroids) retain the normal level
antimicrobial and antifungal peptides) increase in the of resistance to Aeromonas salmonicida challenge
body ¯uids in response to in¯ammatory stimuli [17,18] [22], an observation that has been interpreted to
and both the receptors (Toll-like) and the components imply a pre-eminent role for the complement system
of signal transduction pathways involved in mounting in resistance to this pathogen.
the response (NFkB and others) are conserved [19]. As ®sh are ectotherms, their body temperatures
Second, with the acquisition of the adaptive arm of generally ¯uctuate with their environments. At least
immunity, the innate arm may have been relieved of in carp and cat®sh, effective lymphoid responses have
the pressure to be as comprehensive and redundant in been found to require temperatures above the mini-
regard to pathogen recognition as it had formerly been. mums that these species may naturally experience
As the adaptive immune system evolved and [23,24]. It is probable that, under such non-permissive
approached its condition in modern endotherms, its circumstances, the innate arm of the immune system
 C.J. Bayne, L. Gerwick / Developmental and Comparative Immunology 25 (2001) 725±743 729

meets the animal's internal defense needs [25]. thatÐdue to low temperature inhibition of adaptive
However, it remains to be discovered if the APR in immunity and the low environmental temperatures
®sh is less constrained by temperature than is the experienced by this ®sh speciesÐ`it would be advan-
adaptive response. tageous for primitive vertebrates to have serum
proteins with broad speci®city for determinants
commonly found in the cell walls or surface structures
5. Acute phase proteins and the acute phase of invading organisms', and that `non-antibody preci-
response in ®sh pitins (may) form part of the ®shes' humoral defenses
against invasion by microorganisms and parasites'.
5.1. Targeted studies of APP and the APR in ®sh These were prophetic words indeed, published
16 years before Janeway's more widely known, land-
Historically, the abilities of ®sh to mount APRs mark proposal of pattern recognition molecules for
were ®rst approached by targeted proteomics: inves- pathogen associated molecular patterns [38].
tigators selected in advance the proteins they would The pentraxins are a family of ancient and multi-
study, and measured the amounts of such proteins in functional pentameric proteins that bind their ligands
blood or mucus samples taken from control and in a calcium-dependent manner. For CRP, the ligands
experimental ®sh. This approach was used with rain- are Pneumococcal C-polysaccharide, phosphorylcho-
bow trout (Oncorhynchus mykiss) [21,26,27,28], line and related structures that are expressed on many
Atlantic salmon (Salmo salar) [29], Japanese eel microbes and on the surfaces of disrupted eukaryotic
(Anguilla japonica) [30], carp (Catla catla) [31], cells; SAP binds to phosphoryl-ethanolamine and can
channel cat®sh (I. punctatus) [32,33], plaice (Pleuro- be af®nity puri®ed as a result of its af®nity for agarose.
nectes platessa) [25,34], murrel (Channa punctatus) CRP is capable of activating complement, opsonizing
[35] and tilapia (Oreochromis mossambicus) [36]. bacteria, fungi and parasites, binding free chromatin,
These studies have been fruitful, and have clearly inducing cytokine release [39] and agglutinating parti-
established that homologs of at least some known cles. Additional functions are likely [10]. CRP is typi-
APP are present in both elasmobranchs and teleosts, cally prepared by passing a body ¯uid or extract in a
and that in some teleosts some of these plasma calcium-containing buffer over an af®nity column of
proteins increase in concentration in response to solid-phase (Sephadex) C-polysaccharide or phos-
in¯ammatory stimuli. phorylcholine, and, after washing the column, eluting
the bound molecules with either a solution of one of
5.1.1. The pentraxinsÐC-reactive protein (CRP) and the ligands, or a divalent cation-free, EDTA-contain-
serum amyloid P (SAP) ing buffer. Proteins resembling CRP have been
Interestingly, the protein that ®rst revealed (in 1930 isolated from several invertebrates: a horseshoe crab
[37]) the existence of an APR in mammals (CRP, Limulus polyphemus [40], a land snail Achatina fulica
named for its af®nity for the C-polysaccharide of [41] and a marine bivalve mollusc [42]. In ®sh, CRP-
Pneumococcus) was the same protein that, in 1973, like proteins have been reported from the eggs of the
marked the ®rst discovery of an acute phase protein in lumpsucker (Cyclopterus lumpus [43]), and from the
®sh [25]. That year, Baldo and Fletcher reported that sera of tilapia (Oreochromis mossambicus) [36],
precipitates formed when sera from plaice (P. cat®sh (I. punctatus) [32], rainbow trout (O. mykiss)
platessa, a marine teleost) was mixed with extracts [21,26±28,44], Atlantic salmon (S. salar) [29], plaice
of various microorganisms, some fungi or the nema- (P. platessa) [45,46], Japanese eel (Anguilla japonica)
tode Ascaris lumbricoides. The precipitates dissolved [30], and a carp Catla catla [31]. In the only study of
in 1 mM phosphorylcholine solution, and in 0.1 M relevant serum proteins in an elasmobranchÐthe
EDTA. These data suggested a CRP-like nature of dog®sh Mustelus canisÐa CRP-like molecule was
the precipitin. While CRP was known to be a major also detected [47]. Of these molecules in lower
APP in mammals, Baldo and Fletcher [25] considered animals, only those from trout (X99385) and salmon
it probable that the CRP-like plasma molecule was (the pentraxin X99386) have been sequenced fully.
constitutive in plaice, and went on to speculate Partial amino acid sequences, binding speci®city,
730 C.J. Bayne, L. Gerwick / Developmental and Comparative Immunology 25 (2001) 725±743

divalent cation requirements, subunit size and penta- increased 10- to 20-fold in just 48 h when ®sh were
meric structure are the criteria on which identi®ca- injected with 3-methyl-cholanthrene, 2,3,7,8-tetra-
tions have typically been assigned. In view of its chlorodibenzo-p-dioxin or Arochlor 1254, but not to
presence in both invertebrates and gnathostomes, we some other agents [27]; interestingly, simultaneous
predict the expression of a CRP-like pentraxin in injection of Blue Tongue virus (which was itself not
agnathans. stimulatory) signi®cantly inhibited production. The
The ®rst study to report the de novo appearance or implication that the CRP-like protein may provide
increase in concentration of a plasma protein in ®sh some protection from the toxic effects of, or assist
following an in¯ammatory stimulus was that of in the clearance of, certain toxins is consistent with
Ramos and Smith [36]: tilapia (O. mossambicus) other studies. Under more natural conditions, tropical
experiencing tissue necrosis were found to contain a freshwater murrel exposed to various xenobiotics
CRP-like component in both serum and body mucus, increased their plasma levels of a CRP-like protein
as early as 24 h after injury. Subsequently, Winkel- as early as 12 h after exposure [35], demonstrating
hake and Chang [26] injected turpentine (a classical the possibility of using the plasma level as a biomar-
inducer of the APR) into trout (O. mykiss) muscles, ker for pollutant exposure. In light of increases in CRP
and subjected other trout to thermal stress, then in response to microbial preparations, its in vivo func-
measured bacterial agglutinating activity in the tions have generally been thought to relate more to
serum 24±48 h later. Levels were found to increase anti-microbial defenses, and this was addressed by
15-fold, from 0.07 mg/ml in untreated ®sh to 1.06 mg/ Szalai [33]; however, if the cat®sh pentraxin aids
ml 48 h after turpentine injection, and 18-fold, to defense against the fungus Saprolegnia or any other
1.27 mg/ml 48 h after a temperature shock. Levels parasite or pathogen in this or any other ®sh, this
declined after 48 h. Both the scale and the timing of remains to be demonstrated. In trout, sera with high
these changes are classical indicators of an APR. levels of a C-polysaccharide-binding protein (TCBP),
While alternative interpretations are admissible, the induced by injection of FCA or FCA with V. anguil-
activity was ascribed to a CRP-like pentameric mole- larum [21]), were found to have more potent opsoniz-
cule with monomers of ca. 20,000 Da. ing capacity than sera with lower levels. Though the
In several of the more recent studies, macromole- puri®ed TCBP itself was opsonic [49], this ability to
cules have been af®nity puri®ed from blood collected enhance phagocytosis was further improved in the
1±2 days after injections of in¯ammatory agents. In presence of complement [50]. The propensity to
trout and cat®sh, for example, molecules that bound in bind to foreign particles is re¯ected in the erythro-
a calcium-dependent manner to C-polysaccharide or agglutinating activity reported for eel CRP-like
phosphorylcholine appeared to be pentameric, which protein [30].
is typical of both CRP and serum amyloid P (SAP) The versatility of CRP is illustrated by its apparent
(the pentraxins), two of the major APP in mammals ability to serve as a surface receptor on cytotoxic
(Table 2). The subunit mass of the trout protein is cells. Trout lymphocytes were recognized by a rabbit
26 kDa [44]. The amino acid composition of the ca. antiserum to trout CRP-like protein [51], and when
100 kDa cat®sh pentraxin [48] indicated similarity to these cells were treated with the antiserum plus
human, dog®sh and trout CRPs. However, in the complement (to lyse reactive cells), a reduction in
presence of EDTA the molecule showed anomalous cytotoxicity against a murine myeloma cell line was
behavior, and the authors argued that it is held observed. A similar property has been reported for
together by protected disul®de bonds. The subunit human NK cells [52]. Yet another property of the
size of the Atlantic salmon pentraxin was 37 kDa trout CRP-like protein is its ability to stimulate migra-
[29], and, following an injection of Escherichia coli tion (chemokinesis) within a macrophage-enriched
LPS, plasma levels increased marginally and transi- population of homologous head kidney leukocytes
ently from 50 to 300 mg/ml, peaking at 2 days. The [49].
highest value reported was 450 mg/ml. The two pentraxins that occur in mammals are the
Few studies have focussed on the functions of ®sh products of homologous genes. It appears that only
pentraxins. In trout, a CRP-like molecule in plasma one of them is an APP in any one species. It has
 C.J. Bayne, L. Gerwick / Developmental and Comparative Immunology 25 (2001) 725±743 731

Fig. 2. Alignment of a rainbow trout SAP-like pentraxin characterized by Jensen and colleagues (GenBank X99385) with the deduced amino
acid sequence from an unpublished nucleotide sequence obtained from a rainbow trout liver cDNA library enriched for fragments of acute
phase protein genes (unpublished observations). Identical residues are marked with asterisks. The full length of the SAP-like trout protein is
shown below (N-terminal to C-terminal), where the region equivalent to the new fragment is shown in bold.

been of interest to discover at what stage of phylogeny authors concluded that this was a negative acute phase
the ancestral pentraxin gene duplicated. The presence protein; however, the ®sh (Salvelinus alpinus) in this
of two distinct pentraxin genes in ®sh has been ques- study had been injected with a live pathogen (A.
tioned [53±55]. Lund and Olafsen [56] found just one salmonicida). They were seriously sick, with livers
pentraxin in salmon, wolf-®sh, cod and halibut. On the that, at the later time points in the study, were `yellow
other hand, two pentraxins have been reported in and rigid'. Thus, despite the presence in these animals
dog®sh [47] and plaice [57]. In the plaice, just one of elevated mRNA for another APP (SAA), it seems
of these (`CRP') increased following an injection of plausible that the decline in pentraxin message may
LPS, cortisol or adrenaline. In a 1995 paper [58], have been pathological, and not re¯ective of the
Jensen and colleagues interpreted their ®ndings and response potential under more favorable circum-
those of Murai et al. [28] to indicate that `there are two stances. In light of (i) the absence of knowledge of
different pentraxin-like proteins present in rainbow the ligand speci®city (phosphorylcholine or phosphor-
trout serum', a view they later reversed [55]. Evolu- ylethanolamine) of the products in that study, and (ii)
tionary analyses by Whitehead and colleagues [53,54] the frequency of other reports of two distinct pentrax-
suggested to them the presence of a single pentraxin ins in both teleosts and elasmobranchs, it seems more
gene in ®sh. In a more recent study [55], Whitehead reasonable to consider that the duplication of the
and colleagues concluded that salmonids have a single ancestral pentraxin gene may have occurred
pentraxin gene. That conclusion appears unwarranted, beforeÐor at the time ofÐthe appearance of
since (i) to screen two cDNA libraries, the authors gnathostomes. This is consistent with our recent
relied on only an SAP-based probe, and (ii) they discovery of a trout liver pentraxin sequence that
sequenced only a single clone from their salmon shares more identity with the murine CRP (44%)
cDNA library and just two clones from their trout than with the other, `SAP-like' pentraxin of trout
library. The salmonid sequences appeared most (37%) (Fig. 2). An SAP-like protein has been
closely related to CRP of Xenopus (Fig. 2 in Ref. described in the urochordate Botrylloides leachii[59],
[55]). While considering that `it is not possible to and other studies [60,61] provide evidence for the
determine whether the salmonid pentraxin is a homo- existence of at least two pentraxins in the horseshoe
logue of mammalian CRP or SAP', the authors stated crab Limulus polyphemus, implying that their origin
that it is `mostly SAP-like'. Messenger RNAs for this was over 500 million years ago.
pentraxin declined in most ®sh over 120 h, but were CRP-like proteins were the main focus of early
`strikingly high' in two individuals. Nevertheless, the work on ®sh acute phase proteins. However, the
732 C.J. Bayne, L. Gerwick / Developmental and Comparative Immunology 25 (2001) 725±743

other pentraxinÐserum amyloid P (SAP)Ðhas been in¯ammation [65] and transporting lipids [66]. It is
reported not only in salmonids [55,56,58] but also in possible that they lack any role in combating
European plaice [45,46] and in dog®sh [47]. Two microbes, and may serve to help repair tissue damage.
pentraxins were separately prepared from plaice sera SAA is strongly up-regulated in Arctic char (S. alpi-
on the basis of their calcium-dependent binding to nus) liver during a septicemia [55]. In that study, ®sh
Sepharose-linked pneumococcal C-polysaccharide were injected with live A. salmonicida, the causative
(CRP) and to native Sepharose (SAP) respectively. agent of furunculosis, then examined for altered levels
The authors [57] presented evidence for the existence of transcripts for serum amyloid A, serum amyloid P
of two distinct molecular species, and were of the and albumin. While SAA mRNA was undetectable in
view that `plaice C-reactive protein and serum control samples, it was detected 48 h after injections,
amyloid P component are homologous with each and at 120 h it reached levels approximately 40 times
other and with their human counterparts and indicate those seen at 48 h. As the ®sh sampled at 120 h `had
that there has been stable conservation of this protein entered the ®nal stage of furunculosis and showed
family throughout vertebrate evolution'. clinical signs and gross pathology consistent with a
Uncertainty also prevails regarding a related issue: severe systemic bacterial infection' [55], it is possible
rainbow trout had earlier been reported [28] to have that these ®sh failed to mount the most vigorous APR
two plasma proteins that bind the C polysaccharide of of which the species is capable. Albumin mRNA
Streptococcus pneumoniae in a calcium-dependent levels declined late in the course of the infection.
manner. However, it appears that just one of theseÐ SAA was up-regulated also in Cyprinus carpio in
so-called trout C-polysaccharide-binding protein 2 the course of an in¯ammatory response to turpentine
(TCBP2)Ðis a pentraxin [44]. Its major subunits are oil. This was observed in both in¯ammatory leuko-
25 kDa, its native mass is 135 kDa, and its N-terminal cytes [67] and hepatopancreas/liver (unpublished
20 amino acid sequence shares .40% homology with personal observations, CJB, 1999).
CRPs of other species. It also appears as pentameric Recently it has become clear that transcription of
disks in the electron microscope. The other trout mole- SAA genes in S. salar [68] is regulated by factors
cule (TCBP1) [44] is a structurally distinct, 81.4 kDa similar to those that regulate their expression in
heterodimeric protein with subunits of 26.6 and mammals. In primary cultures of trout hepatocytes,
43.7 kDa, and was tentatively considered to not be a SAA messenger RNA levels were measured following
CRP homolog. Recent studies (Bayne et al., below) exposure to recombinant human cytokines, bacterial
con®rm this. For the present, the situation remains LPS, or media conditioned by macrophages. These
unclear: some teleosts appear to have two functionally macrophages had been stimulated by either LPS, or
`C-reactive' proteins, only one of which is a pentraxin. It a yeast-derived glucan, or by media recovered from
is not known if both are APP. The resolution of this stimulated homologous leukocytes. Levels of SAA
enigma, and the determination of the number of mRNA increased modestly in hepatocytes exposed
pentraxin genes in ®sh will both require further work. to media conditioned by LPS-stimulated macro-
Trout plasma contains several proteins that bind to phages, or to human recombinant IL-6, TNFa or IL-
A. salmonicida LPS [62]. Among these, both the trout 1b. Macrophage-conditioned media were ineffective
CRP (TCBP2) and SAP [58,63] occur along with two unless prepared with LPS or glucan in addition to the
other unidenti®ed molecules, and it will be of interest putative leukocyte-derived macrophage-activating
to discover if either of these is an APP. factors; such media induced up to a 5.2-fold increase
in SAA mRNA levels. The greatest increases (.10-
5.1.2. Serum amyloid A fold) occurred in hepatocytes exposed directly to LPS,
Acute phase SAAs (which are not pentraxins) are though the effect could have been indirect via other
major APPs in mammals (Table 1). They `have been cell types in these primary cultures.
identi®ed in all vertebrates investigated to date and
are highly conserved' [64]. While their physiological 5.1.3. Transferrin
functions are incompletely known, they are likely to Transferrin is one of several plasma proteins
be multifunctional, involved in at least modulating capable of binding iron. Though in rabbits and rats
 C.J. Bayne, L. Gerwick / Developmental and Comparative Immunology 25 (2001) 725±743 733

Fig. 3. Tandem crossed immuno-electrophoretic analysis of sera from rainbow trout. Samples collected before and 10 days after injection with
an in¯ammatory stimulus were loaded at 0 and 10 respectively, and electrophoresed ®rst from LEFT to RIGHT. Then agarose gel containing
rabbit antiserum to trout plasma proteins was poured above the ®rst dimension separated trout plasma components, and electrophoresis was
carried out at 908 to the ®rst dimension. During this stage, antigen-antibody precipitates formed, and those visible here as bimodal peaks were
stained with Coomassie blue. `1' and `2' indicate positive APP, identi®ed by the higher levels of the regions of their precipitates to the RIGHT.

it is a positive APP [69], in some mammals it is a may be a positive APP in rainbow trout. This is
negative APP. This is at ®rst surprising, since redu- consistent with a report [71] that levels of free iron
cing the availability of iron to bacteria is one means in plasma were lower in trout 24±48 h after injections
used by vertebrates to control pathogens. In our of LPS. Further indication that transferrin may be a
comparisons of control and acute phase trout plasma positive APP in trout was seen in a recent molecular
using tandem crossed immuno-electrophoresis genetic study (below [72]).
(Fig. 3), one protein that was increased in acute
phase sera was subjected to trypsin digestion followed 5.1.4. Alpha-2-macroglobulin
by electro-spray ionization mass spectrometry, and Alpha-2-macroglobulin (a2-M) is the most versatile
this revealed fragment masses suggestive of transfer- anti-protease known, capable of trapping and func-
rin [70]. Encouraged by this, we blotted 2-D gels on to tionally silencing all classes of proteases [73]. It
PVDF membranes and probed them with an antiserum may also serve to bind and transport cytokines
to human transferrin, which reacted with the ®sh anti- (reviewed in Ref. [74]), thus potentially modulating
gen. When compared with samples taken from ®ve immune responses. In several mammals a2-M is an
®sh before they were injected to elicit an APR, acute phase protein [75±77]. It occurs in plaice [78],
those taken 10 days later were found to have either and is also thought to play a role in the different
(1) no increase, (2) small increases or (3) a strong susceptibilities of distinct trout species to bacterial
increase in transferrin. Thus it appears that transferrin pathogens [79]. Accordingly, we raised rabbit
734 C.J. Bayne, L. Gerwick / Developmental and Comparative Immunology 25 (2001) 725±743

Fig. 4. Rocket immuno-electrophoretic quanti®cation of a2-macroglobulin in serum samples from rainbow trout. Plasma samples from rainbow
trout were loaded at 5 ml/well in plain 1% agarose. Then 1% agarose containing rabbit antiserum to trout alpha-2-macroglobulin (5% in
agarose) was poured on the neighboring area of the glass, and the samples were electrophoresed into the antibody-containing gel. As the
quantity of antibody was the same throughout the gel, the height of each peak re¯ects the quantity of alpha-2-macroglobulin in the sample. Fish
no. 1, 2 and 3 were injected with Freund's complete adjuvant; Fish no. 4 was injected with turpentine.

antiserum against puri®ed trout a2-macroglobulin, plasma level, and the putative pre-existing pool is
and determined relative levels of this antigen in located.
trout acute phase plasmas (unpublished). Small differ-
ences were detected following an injection of Vibrio
bacterin-in-oil [72] or of Freund's incomplete adju-
5.1.6. Lysozyme
vant or turpentine (Fig. 4). However, these were not
Lysozymes occur not only in plasma, but also in
consistent. It seems that a2-macroglobulin is not an
mucus secretions. They cleave peptidoglycans in
APP in trout, a view arrived at independently by B.
bacterial cell walls [81], and have been reported to
Robertsen (personal communication, 2000).
increase in concentration in Atlantic salmon [82],
rainbow trout [83], yellow tail [84] and turbot [85]
in response to injections of various glucans (reviewed
5.1.5. Complement component C3
in Ref. [86]). When trout were given an acute stress
Levels of C3 rise three- to ®ve-fold in the course of
(see C3 above) several were found to experience a
mammalian APRs. In rainbow trout, within just
pre-acute phase rise in plasma lysozyme activity
10 min of the initiation of acute stressors, C3 was
[87]. Lysozymes are synthesized in both the liver
found to increase in plasma [80]. In view of the cool
and extra-hepatic sites, however the kinetics of synth-
water temperature in these experiments (12C) and the
esis and secretion are not yet clear.
need for multiple steps in the processes of gene acti-
vation and protein synthesis, the short response time
seen here was interpreted to mean that the response
was probably not due to the classical APR (involving 5.1.7. Lectins
de-novo synthesis of product). It appears likely that a Mannan binding lectin (MBL) is one of the
pool of presynthesized C3 is held in store and released mammalian serum lectins that is an APP and involved
to the plasma as part of a pre-acute phase response. in activation of the Complement cascade. While there
The immune system, we suggest, is a participant in the is no evidence that it is increases in the APR in ®sh, it
classical ®ght-or-¯ight response. This remains spec- has been isolated from an eel (Anguilla anguilla [88]
ulative until de-novo synthesis is proven not to occur and from Atlantic salmon [89] in which it serves a
within the short period required for the change in defensive role.
 C.J. Bayne, L. Gerwick / Developmental and Comparative Immunology 25 (2001) 725±743 735

Fig. 5. One-dimensional 12% SDS±PAGE gel (without reducing agent) showing plasma samples taken from one individual trout on Day 0,
before treatment, and on Days 3, 5, 7, 10,14 and 21 after treatment. H indicates a total protein load of 173 mg (heavy load), L indicates a total
protein load of 39.6 mg (light load). The arrows indicate some of the proteins that increased in concentration following injection of Vibrio
bacterin into the ®sh. Molecular weight markers were run in the center lane; masses (kDa) are shown to the left.

5.2. Shotgun approaches to study of the ®sh APR potent APR (Fig. 5), we examined plasma samples by
means of two-dimensional electrophoresis (iso-elec-
Proteomic and genomic approaches have recently tric focussing following by SDS±PAGE). Several
been applied to studies of the APR in some teleosts. In protein spots were found to change consistently in
general, these studies have not targeted pre-selected samples taken after ®sh were injected with a Vibrio
proteins for study. Exploiting the potentials of bioin- bacterin emulsi®ed in Freund's incomplete adjuvant
formatics to identify sequences of interest, these (FIA). One strongly up-regulated protein (Fig. 6) was
studies have surveyed the plasma proteins and the excised from gels and a partial amino acid sequence
mRNAs in livers for those that increase following was obtained. This information was used to design
in¯ammatory stimuli. This approach has proven to degenerate oligonucleotide primers that were used to
be very productive in terms of identifying candidate amplify a fragment of the speci®c cDNA; this was
APPs. then sequenced. Using new primers, the sequence of
In the 1990s, the tools of molecular biology opened the complete cDNA was then obtained (GenBank
the doors to more speci®c and sensitive means to Accession number AF192969). The APP is a precer-
measure increases and decreases in speci®c gene ebellin-like protein, known previously only from the
expression, and these tools were ®rst used to study mammalian brain [91]. Northern analysis (unpub-
the APR in ®sh by Jensen et al. [55], as discussed lished observations) shows that mRNA is increased
above. More recently, combined proteomic and geno- in the liver at least as early as 3 days post-injection.
mic approaches have been used in studies of the APR The molecule resembles the complement component
in rainbow trout, O. mykiss [90]. After ®rst determin- C1q B chain, but lacks the collagen portion that is
ing which of several protocols would induce the most required for the formation of the C1q complex of A,
736 C.J. Bayne, L. Gerwick / Developmental and Comparative Immunology 25 (2001) 725±743

Fig. 6. Two-dimensional gel electrophoretic analysis of plasma samples from a rainbow trout bled before (LEFT) and 10 days after (RIGHT)
intra-peritoneal injection with a Vibrio bacterin emulsi®ed in mineral oil. Each spot represents a polypeptide component of a plasma protein. To
the left of the star in the Day 10 sample but not the Day 0 sample there is a spot which, when analyzed (see text), was found to be a
precerebellin-like protein that increased from undetectable to 320 ^ 20 mg/ml in plasma.

B and C chains. The relevance of this gene product to of which is shown in Table 3. Some of this work has
internal defense, implied by its elevated plasma levels been published [72].
following a bacterin injection, is unclear. As false positives are common in SSH libraries, the
Most recently, suppression subtractive hybridiza- presence of a sequence in this library is not proof of its
tion (SSH) has been used in a non-targeted (shotgun) up-regulation in the course of an APR; for each gene,
way to identify genes whose transcription is altered in it will be necessary to independently verify transcript
rainbow trout responding to an in¯ammatory stimulus levels in control and experimental ®sh. Nevertheless,
[72]. This approach does not rely on pre-selection of the contents of this library imply that rainbow trout
speci®c proteins or genes for study. Instead, messen- mount an impressively broad APR, with many of the
ger RNAs are prepared from appropriate tissuesÐin classical APP represented.
this case livers of control (non-injected) ®sh and from Clones in this SSH library were picked at random
®sh injected with the Vibrio bacterin in FIA. Reverse for sequencing. Under such circumstances, the
transcription to obtain complementary DNAs is frequency with which particular cDNAs are found
followed by mixing and annealing steps designed to will re¯ect the abundance on the corresponding
remove the cDNAs that are common to both prepara- mRNA in the original tissue extract. Among the
tions, and to enrich for less abundant transcripts. In clones in Table 3, the most frequently encountered
our work, fragments of cDNA of length 300±700 base was the so-called differentially regulated trout
pairs were cloned into E. coli, and 440 clones expres- protein-1 (DRTP1, GenBank accession no.
sing trout sequences were randomly selected for AF281355 [92]). This is a homolog of CD59, whose
sequencing. The outcome was a remarkable library function is to prevent target cell lysis by the terminal
of potential acute phase gene fragments, a selection membrane attack complex of complement [93]. We
 C.J. Bayne, L. Gerwick / Developmental and Comparative Immunology 25 (2001) 725±743 737

Table 3
Names of nucleotide sequences selected from a cDNA library enriched for fragments of putative acute phase proteins expressed in livers of
rainbow trout. The data in this table summarize work by the authors, most of which has been published [72]

Group A. Genes for proteins known to be APP in mammals (n ˆ 11).

Complement components C3-1, C7, Bf-2, B/C2-B and C1 inhibitor.
Clotting components Fibrinogen and Thrombin
Haptoglobin
Hemopexin
Serum amyloid A
Lysozyme

Group B. Genes for proteins that may be secreted into the plasma and have defensive functions (n ˆ 6).
Cysteine proteinase
Chemotaxin
C-type lectins (2)
Putative antibacterial peptide, hepcidin.
Trout C-polysaccharide-binding protein 1.

Group C. Genes for cell proteins that may be involved in defense-related functions (n ˆ 4).
Glutathione peroxidase
Prostaglandin D synthase
Toll-like receptor
Cathepsin

Group D. Genes for novel proteins.

Names remain unknown.

speculate that this re¯ects an increase in the danger of (AF281345, an extension of a previously known
complement-mediated damage to self during the protein [44]), a cysteine proteinase (AF281331), a
APR, due to increased complement activity. chemotaxin (AF271114), two C-type lectins
Genes which appear in the trout SSH library and are (AF281349 and AF281351), and a putative antibacter-
known mammalian APP include ®brinogen ial peptide (AF281354). The ®rst of these (TCBP1)
(AF281352 new for ®sh), thrombin (AF281359, veri®ed the suggestion [44] that TCBP1, while it binds
previously known for this species [94]), haptoglobin to C-polysaccharide, is not a pentraxin. Its 20 N-term-
(three fragments AF279136, AF281329, AF281330 inal amino acid sequence is identical to the only avail-
new for ®sh), hemopexin (AF281339, previously able sequenced region of TCBP1, and the peptide
known [95]), a pentraxin resembling serum amyloid continues as a homolog of the precerebellin-like
P (previously known [55]), and lysozyme (previously protein[90]. Second, there are several genes whose
known [96]). In addition, ®ve proteins involved in the functions are cellular yet related to oxidative stress
activation and control of complement are present: C3- [glutathione peroxidase (AF281338)], production of
1 (previously known [97]), C7 (AF281336), Factor mediators [prostaglandin D synthase (AF281353)],
Bf-2 (previously known [98]), Factor B/C2-B [99], recognition of pathogen-associated molecular patterns
C1 inhibitor (AF281335, new for ®sh), and the or of cytokines [a member of the Toll-like family of
previously mentioned DRTP1 (AF281355). non-self receptors (AF281346)], bacteriolysis (cathe-
What are perhaps the most valuable components of psin, AF281353), and regulation of acute phase gene
the SSH library fall into three additional groups, in transcription (IL-6 NF, unpublished). The third group,
which most members have not been described in tele- comprising .40% of the library, includes sequences
osts. First, the genes that, on account of being plasma that do not resemble any sequences in the GenBank
proteins with probable defense-related functions, can database. Within this group, some have multiple
reasonably be considered prospective additional APP: `stop' codons in all six reading frames, and these are
the trout C-polysaccharide binding protein 1 likely to be 3 0 untranslated regions (UTRs). Some of
738 C.J. Bayne, L. Gerwick / Developmental and Comparative Immunology 25 (2001) 725±743

these are probably the 3 0 UTRs of identi®ed sequences personal communication) that it is possible to raise the
(just listed). However, some unidenti®ed sequences resistance of a ®sh to a particular parasite or pathogen
appear to be open reading frames for novel genes. In by treating the ®sh with an unrelated agent. This was
view of the prevalence of immune-related genes in the clearly evident when fewer `control' coho salmon,
identi®ed members of the library, it is anticipated that injected with a modi®ed Freund's complete adjuvant,
additional defense-related genes will be found as but without immunogen, died than naõÈve ®sh when
these unidenti®ed sequences are functionally charac- subsequently challenged with A. salmonicida, A.
terized. hydrophila or Vibrio ordalii [20]. Enhanced resistance
It is also of interest that the SSH library contains was evident as early as 4 days post-injection. This
cDNA fragments for more than 10 genes that are not phenomenonÐrapidly acquired, enhanced resistance
suspected of involvement in the classical APR. These to a pathogen due to an injection of an unrelated
may re¯ect metabolic changes in hepatocytes as they agentÐis not limited to ®sh. Such enhancement is
adapt to the circumstance signaled by the in¯amma- likely due, in part, to leukocyte activation [21].
tory stimulus. For example, ribosomal proteins However, in the ®sh examined by Kodama and collea-
(AF281334 and three other unpublished sequences) gues, levels of CRP increased even earlier than cells
may increase to facilitate synthesis of additional became activated. It is most probable that both cellu-
acute phase or other proteins; a TAT-binding protein lar and humoral components of the innate immune
(AF281342) and the IL-6 NF gene (see above, unpub- system act co-operatively to provide enhanced
lished) may increase due to their roles in regulation of resistance to pathogens following stimulation by
the APR; biotinidase (AF281332 and AF281333) may appropriate danger signals, such as microbial lipopo-
play a role in denying microbes access to biotin, and lysaccharides or glucans.
so on. These ideas encourage consideration of the It is commonly stated that innate immunity is `non-
hepatic APR as involving a lot more than the increase speci®c'. That is, of course, one of its strengths, since
and decrease of synthesis and export of plasma a relatively small number of molecules and mechan-
proteins. These ideas remain to be tested. isms can protect against any of a number of microbes
or parasites expressing one or more common molecu-
5.2.1. The nature of the APR in ®sh in comparison to lar signatures [4]. But is the non-speci®city of the
what might have been predicted/anticipated effector molecules mirrored in the nature of the
In light of suggestions that the anti-protease a2- APR? Is the APR stereotyped? The evidence suggests
macroglobulin may play a signi®cant role in minimiz- that ®sh, like humans [102], mount qualitatively
ing microbial pathogenesis [79] and of its up-regula- distinctive APR to viral, fungal and microbial agents
tion in the APR of some mammals, it was expected to [103]. A full dissection of this must await a wider
increase in trout. This appears not to be the case (see knowledge of cytokines and of the regulation of
earlier). However, this versatile molecule may be held gene transcription in these animals. In mammals,
at relatively high constitutive levels in trout plasma, acute phase gene expression is regulated by the spec-
suf®cient to meet needs. Alternatively, stimuli other trum of identities and quantities of cytokines released
than Vibrio bacterin may induce its up-regulation. It is in response to distinct in¯ammatory stimuli. Further-
also possible that different anti-proteases in plasma more, each gene appears to respond in its own way
[100] may preclude the need for more a2-macroglo- (scale, direction, timing of initiation, kinetics),
bulin. depending on the cytokine mix and (probably) the
Transferrin binds iron, removing it from the pool metabolic status of the cell at the time of its stimula-
that is otherwise accessible to microbes. Its identity as tion. There is no reason to expect a less complex
an APP in trout was implied in our proteomic work situation to prevail in ®sh.
[70]. Consequently, it was not surprising to ®nd trans-
ferrin as one of the genes that turned up repeatedly 5.2.2. The potential of the ®sh APR as a biomarker for
(®ve times) in the SSH library, even though it is a environmental insults and as an index of ®sh health
positive APP in only some mammals. Alteration of the levels of speci®c mRNAs after
It has long been known ([20,86,101]; Richard Holt, appropriate stimuli requires a multi-step activation
 C.J. Bayne, L. Gerwick / Developmental and Comparative Immunology 25 (2001) 725±743 739

of gene transcription [104]. Consequently it is not and their susceptibility to ¯ooding (with consequent
subject to change as rapidly as, for example, blood uncontrolled dispersal of the ®sh), as well as the prob-
levels of stress hormones during the ®ght-or-¯ight ableÐhighly undesirableÐselection of resistance
response [105]. Indeed, the kinetics of the APR are genes in pathogens, indicate the need for proceeding
such that the quantities of speci®c transcripts present cautiously if at all in this area.
in the cell starting an hour or less (but not less than
10 min) after the initiating stimulus and lasting for 5.2.3. The future
several days thereafter re¯ect the APR `status' of While it appears that a Vibrio bacterin emulsi®ed in
the organism. This, of course, is the basis for the mineral oil will elicit a strong APR in trout, it would
high value now placed on `transcriptional pro®ling' be premature to conclude that this is among the most
made possible by gene arrays. Equipped with the potent of stimulators. Among the tasks that lie ahead
knowledge of what ranges of transcript levels are indi- is the pro®ling of APRs to compounds that have been
cative of a healthy physiological status (homeostasis), found to enhance resistance to viral and bacterial
we can sample transcript levels in, for example, biop- pathogens (reviewed in Ref. [86]), with the aims of
sies of liver tissue taken from feral or other ®sh (or discovering the plasticity of the responses and their
other) species, and draw inferences about their recent other attributes.
experiences such as exposure to viral, bacterial or Gene array technology that allows us to gain a `snap
fungal pathogens, thermal stress, tissue trauma, etc. shot' of the levels of mRNAs for many genes at any
And it should be possible to detect incipient septice- chosen time is sure to be applied fruitfully to extend
mia. By such means, ®sh can be used as sensitive our knowledge of innate immunity in ®sh. For example,
sentinels capable of indicating the presence of toxins, microarrays manufactured to measure levels of tran-
other pollutants, pathogens or parasites, or of report- scripts for genes coding for APP should have utility
ing other environmental perturbations. not only in basic immunology research, but also in the
A bene®cial outcome of molecular studies of the APR aquaculture industry as an additional tool for monitoring
in any species is the increased ease with which variant the health of the ®sh, and in the hands of those respon-
and novel defense molecules should be identi®ed, while sible for the management of wild ®sh stocks.
at the same time gaining the sequence data required for As in other areas of contemporary cellular and
further characterization of the gene, and for transfection molecular biology, the challenges of the immediate
and expression in heterologous systems. For both future center on de®ning the functional signi®cance
complement component C3 and the related a2-macro- of proteins, in this case acute phase reactants. The
globulin, already it is clear that more diverse forms means by which such knowledge will be gained
occur in ®sh than in mammals, and this has been inter- must include two hybrid systems and appropriate
preted to imply a more diverse recognition repertoire of biosensors to evaluate protein±protein interactions,
these components of the innate arm of the defense in vivo transfection and over-expression, and both
system [106,107]. The possible existence of multiple knock-down and knock-outs of gene expression.
genes for other molecules of the innate immune system Transfection for in vitro expression and puri®cation
therefore carries implications for the diversity of foreign of products will also have their place in both facilitat-
determinants within the entire recognition repertoire. ing in vitro assays for functions, and boosting in vivo
As the main selective advantage of the APR is levels for whole animal evaluations of functions.
clearly the increase in disease resistance and healing The breadth of knowledge we now have of the APR
capacity that it provides, there is obvious appeal for in ®sh encourages us to do several things. We must
the concept of transgenic ®sh engineered to over- determine the extent of any temperature constraints on
express selected APP. However, it should be obvious the APR and other aspects of immune potentiation
that the competitive abilities of such genetically modi- working through the innate immune systems of both
®ed organisms would be different from their wild-type cold water and warm water species. We need to iden-
con-speci®cs, perhaps giving selective advantages to tify and characterize not only the cytokines, but also
the transgenic individuals. Several considerations, the second messenger pathways, and mechanisms of
including the open nature of typical ®sh habitats, transcriptional and translational regulation of acute
740 C.J. Bayne, L. Gerwick / Developmental and Comparative Immunology 25 (2001) 725±743

phase genes. And the details of post-translational and their role in host defense. Dev Comp Immunol
processing, secretion and ®nal clearance of APP 2001;25:807±825.
[6] Ellis AE. Innate host defense mechanisms of ®sh against
must be mapped. Functions of all APP need to be
viruses and bacteria. Dev Comp Immunol 2001;25:827±839.
determinedÐa challenge that is magni®ed by the [7] Kushner I. The phenomenon of the acute phase response.
possible multifunctionality of some of these proteins. Ann NY Acad Sci 1982;389:39±48.
Uniquely related to the APR in ®sh, we must ask if, in [8] Kushner I, Mackiewicz A. The acute phase response: an
addition to C3 and a2-macroglobulin, other APPs overview. In: Mackiewicz A, Kushner I, Baumann H,
editors. Acute phase proteins. Molecular biology, biochem-
occur in multiple functionally distinct forms. And
istry and clinical applications, CRC Press, 1993. p. 4±19.
our picture of the evolution of the innate immune [9] Gabay C, Kushner I. Acute phase proteins and other
system will remain unclear until the innate systems, systemic responses to in¯ammation. New England J
including the APRs, are adequately described in the Med 1999;340:448±53.
cyclostomes. [10] Steel DM, Whitehead AS. The major acute phase reactants:
As advances are made in the study of the APRs in C-reactive protein, serum amyloid P component and serum
amyloid A protein. Immunol Today 1994;15:81±88.
®sh, we predict that formerly unknown defense mole- [11] Hatherill M, Tibby SM, Murdoch IA. Diagnostic markers of
cules will be identi®ed, and their mechanisms of infection: comparison of procalcitonin with C-reactive
action described. This is certain to yield bene®ts protein and leukocyte count. Arch Dis Child 1999;81:417±
with value beyond the arena of immuno-evolution, 21.
to applications in ®sh husbandry, and perhaps [12] Koj A, Gordon AH, editors. Introduction. In: Dingle JT,
Gordon JL, general editors. Research monographs in cell
human health.
and tissue physiology. The acute-phase response to injury
and infection. Elsevier, 1985.
[13] Noris-Garcia E, Dorta-Contreras AJ, Escobar-Perez X,
Acknowledgements Gonzalez-Hernandez M. Haptoglobin in cerebrospinal ¯uid
as a marker of infectious process in central nervous system
(in Spanish). Rev Neurol 1999;29:117±20.
We thank Borre Robertsen for suggesting improve- [14] Albert MA, Ridker PM. The role of C-reactive protein in
ments to an earlier version of this paper. Particular cardiovascular disease risk. Curr Cardiol Rep 1999;1:99±
thanks go to Kazuhiro Fujiki, Miki Nakao, Tomoki 104.
Yano and Wendy Reynolds whose careful teaching [15] Pannevis MC, Houlihan DF. The energetic cost of protein
greatly facilitated our own molecular research, and synthesis in isolated hepatocytes of rainbow trout (Oncor-
hynchus mykiss). J Comp Physiol [B] 1992;162:393±400.
to Virginia Weis for providing access to her lab.
[16] Flajnik MF. Churchill and the immune system of ectothermic
Our work has been supported by a fellowship to vertebrates. Immunol Rev 1998;166:5±14.
CJB from the Japan Society for the Promotion of [17] Boman HG, Nilsson I, Rasmuson B. Inducible antibacterial
Science, and by the National Institutes of Health defence system in Drosophila. Nature 1972;237(5352):232±
(ES-7060 to LG and F06 TW-02305 to CJB). 5.
[18] Medzhitov R, Janeway Jr C. The toll receptor family and
microbial recognition. Trends Microbiol 2000;8:452±6.
[19] Imler JL, Hoffmann JA. Signaling mechanisms in the anti-
References microbial host defense of Drosophila. Curr Opin Microbiol
2000;3:16±22.
[1] Agrawal A, Eastman QM, Schatz DG. Transposition [20] Olivier G, Evelyn TPT, Lallier R. Immunity to Aeromonas
mediated by RAG 1 and RAG2 and its implications for the salmonicida in coho salmon (Oncorhynchus kisutch) induced
evolution of the immune system. Nature 1998;394:744±51. by modi®ed Freund's complete adjuvant: Its non-speci®c
[2] Beck G, Habicht GS, Cooper EL, Marchalonis JJ. Primordial nature and the probable role of macrophages in the phenom-
immunity: foundations of the vertebrate immune system. enon. Dev Comp Immunol 1985;9:419±32.
Ann New York Acad Sci 1994;712:376. [21] Kodama H, Yamada F, Murai T, Nakanishi Y, Mikami T,
[3] Bayne CJ. Adaptive thinking in immuno-nomenclature and Izawa H. Activation of trout macrophages and production of
immuno-evolution. Immunol Today (Letters) 1995;15:598± CRP after immunization with Vibrio anguillarum. Dev
9. Comp Immunol 1989;13:123±32.
[4] Medzhitov R., Janeway Jr C. Innate immune recognition: [22] Sakai DK. Defence mechanisms in salmonid ®sh. In: Kimura
mechanisms and pathways. Immunol Rev 2000;173:89±97. T, editor. Proceedings of the OJI Symposium on Salmonid
[5] Neumann NF, Stafford JL, Barreda D, Ainsworth AJ, Diseases. Hokkaido University Press, 1992. p. 233±40.
Belosevic M. Antimicrobial mechanisms of ®sh phagocytes [23] Serero M, Avtalion RR. Regulatory effect of temperature and
 C.J. Bayne, L. Gerwick / Developmental and Comparative Immunology 25 (2001) 725±743 741

antigen upon immunity in ectothermic vertebrates. III. Estab- [41] Agrawal A, Mitra S, Ghosh N, Bhattacharya S. C-reactive
lishment of immunological suppression in ®sh. Dev Comp protein (CRP) in haemolymph of a mollusc, Achatina fulica
Immunol 1978;2:87±94. Bowdich. Indian J Exp Biol 1990;28:788±9.
[24] Bly JE, Quiniou SM, Clem LW. Environmental effects on [42] Olafsen JA. Role of lectins (C-reactive protein) in defense of
®sh immune mechanisms. Dev Biol Stand 1997;90:33±43. marine bivalves against bacteria. Dev Immunol 1992;3:67±
[25] Baldo EA, Fletcher TC. C-reactive protein-like precipitins in 84.
plaice. Nature 1973;16(246(5429)):145±6. [43] White A, Fletcher TC, Towler CM, Baldo BA. Isolation of a
[26] Winkelhake JL, Chang RJ. Acute phase (C-reactive) protein- C-reactive protein-like precipitin from the eggs of the lump-
like macromolecules from rainbow trout (Salmo gairdneri). sucker (Cyclopterus lumpus L.). Comp Biochem Physiol C
Dev Comp Immunol 1982;6:481±9. 1978;61:331±6.
[27] Winkelhake JL, Vodicnik MJ, Taylor JL. Induction in rain- [44] Murata M, Kodama H, Onuma M. Characterization of rain-
bow trout of an acute-phase (C-reactive) protein by chemi- bow trout C-polysaccharide binding proteins. J Vet Med Sci
cals of environmental concern. Comp Biochem Physiol 1995;57:419±25.
1983;74C:55±58. [45] White A, Fletcher TC, Pepys MB, Baldo BA. The effect of
[28] Murai T, Kodama H, Naiki M, Mikami T, Izawa H. Isolation in¯ammatory agents on C-reactive protein and serum
and characterization of rainbow trout C-reactive protein. Dev amyloid P-component levels in plaice (Pleuronectes platessa
Comp Immunol 1990;14:49±58. L.) serum. Comp Biochem Physiol C 1981;69(2):325±9.
[29] Lund V, Olafsen JA. Changes in serum concentration of a [46] White A, Fletcher TC, Pepys MB. Serum concentrations of
serum amyloid P-like pentraxin in Atlantic salmon, Salmo C-reactive protein and serum P component in plaice (Pleur-
salar L., during infection and in¯ammation. Dev Comp onectes platessa L.) in relation to season and injected lipo-
Immunol 1999;23:61±70. polysaccharide. Comp Biochem Physiol B 1983;74:453±8.
[30] Nunomura W. C-reactive protein in eel: puri®cation and agglu- [47] Robey FA, Tanaka T, Liu T-Y. Isolation and characterization
tinating activity. Biochim Biophys Acta 1991;1076:191±6. of two major serum proteins from the dog®sh, Mustelus
[31] Paul I, Mandal C, Mandal C. Effect of environmental pollu- canis, C-reactive protein and amyloid P component. J Biol
tants on the C-reactive protein of a freshwater major carp, Chem 1983;258:3889±94.
Catla catla. Dev Comp Immunol 1998;22:519±32. [48] Szalai AJ, Bly JE, Clem LW. Chelation affects the confor-
[32] Szalai AJ, Norcum MT, Bly JE, Clem LW. Isolation of an mation, lability and aggregation of channel cat®sh (Ictalurus
acute-phase phosphorylcholine-reactive pentraxin from punctatus) phosphorylcholine-reactive protein. Comp
channel cat®sh (Ictalurus punctatus). Comp Biochem Biochem Physiol B 1992;102:545±50.
Physiol B 1992;102:535±43. [49] Kodama H, Arimitsu H, Mukamoto M, Sugimoto C.
[33] Szalai AJ, Bly JE, Clem LW. Changes in serum concentra- Enhancement of phagocytic and chemokinetic activities of
tions of channel cat®sh (Ictalurus punctatus Ra®nesque) rainbow trout head kidney cells by C-reactive protein. Am J
phosphorylcholine-reactive protein (PRP) in response to Vet Res 1999;60:240±4.
in¯ammatory agents, low temperature-shock and infection [50] Nakanishi Y, Kodama H, Murai T, Mikami T, Izawa H.
by the fungus Saprolegnia sp. Fish Shell®sh Immunol Activation of rainbow trout complement by C-reactive
1994;4:323±36. protein. Am J Vet Res 1991;52:397±401.
[34] White A, Fletcher TC, Pepys MB, Baldo BA. The effect of [51] Edagawa T, Murata M, Hattori M, Onuma M, Kodama H.
in¯ammatory agents on C-reactive protein and serum Cell surface C-reactive protein of rainbow trout lympho-
amyloid P-component levels in plaice (Pleuronectes platessa cytes. Dev Comp Immunol 1993;17:119±27.
L.) serum. Comp Biochem Physiol 1981;69C:325±9. [52] Baum LL, Johnson B, Berman S, Graham D, Mold C. C-
[35] Ghosh S, Bhattacharya S. Elevation of C-reactive protein in reactive protein is involved in natural killer cell-mediated
serum of Channa punctatus as an indicator of water pollu- lysis but does not mediate effector-target cell recognition.
tion. Indian J Exp Biol 1992;30:736±7. Immunology 1987;61:93±99.
[36] Ramos F, Smith AC. The C-reactive protein (CRP) test for [53] Seery LT, Schoenenberg DR, Barbaux S, Sharp PM, White-
the detection of early disease in ®shes. Aquaculture head AS. Identi®cation of a novel member of the pentraxin
1978;14:261±6. family in Xenopus laevis. Proc R Soc Lond 1993;253:263±8.
[37] Tillett WS, Francis T. Serologic reactions in pneumonia with [54] Rubio N, Sharp PM, Rits M, Zahedi K, Whitehead AS. Struc-
non-protein somatic fraction of Pneumococcus. J Exp Med ture, expression, and evolution of Guinea pig serum amyloid
1930;52:561±71. P component and C reactive protein. J Biochem
[38] Janeway CA. Approaching the asymptote? Evolution and 1993;113:277.
revolution in immunology. Cold Spring Harbor Symposia [55] Jensen LE, Hiney MP, Shields DC, Uhlar CM, Lindsay AJ,
in Quantitative Biology 1989;54:1±13. Whitehead AS. Acute phase proteins in salmonids. Evolu-
[39] Ballou S, Lozanski G. Induction of in¯ammatory cytokine tionary analyses and acute phase response. J Immunol
release from cultured human monocytes by C-reactive 1997;158:384±92.
protein. Cytokine 1992;4:361±8. [56] Lund V, Olafsen JA. A comparative study of pentraxin-like
[40] Robey FA, Liu TY. Limulin: a C-reactive protein from Limu- proteins in different ®sh species. Dev Comp Immunol
lus polyphemus. J Biol Chem 1981;256:969±75. 1998;22:185±94.
742 C.J. Bayne, L. Gerwick / Developmental and Comparative Immunology 25 (2001) 725±743

[57] Pepys MB, De Beer FC, Milstein CP, March JF, Feinstein A, response to lipopolysaccharide in rainbow trout (Oncor-
Butress N, Clamp JR, Taylor J, Bruton C, Fletcher TC. hynchus mykiss). Comp Biochem Physiol 1991;98A:195±
C-reactive protein and serum amyloid P component in 200.
the plaice (Pleuronectes platessa L.), a marine teleost, [72] Bayne CJ, Gerwick L, Kazuhiro F, Nakao M, Yano T.
are homologous with their human counterparts. Biochim Immune relevant (including acute phase) genes identi®ed
Biophys Acta 1982;704:123±33. in the livers of rainbow trout, Oncorhynchus mykiss, by
[58] Jensen LE, Petersen TE, Thiel S, Jensenius JC. Isolation of a means of suppression subtractive hybridization. Dev Comp
pentraxin-like protein from rainbow trout serum. Dev Comp Immunol 2001;25:205±17.
Immunol 1995;19:305±14. [73] Sottrup-Jensen L. a-Macroglobulins: structure, shape and
[59] Sco®eld VL, Puntambekar L, Schluter SF, Coombe DR. mechanism of proteinase complex formation. J Biol Chem
Serum amyloid P component (SAP)-like protein from botryl- 1989;264:11539±42.
lid ascidians provides a clue to amyloid function. Dev Immu- [74] James K. Interactions between cytokines and a2-macroglo-
nol 1992;3(1):67±84. bulin. Immunol Today 1990;11:163±6.
[60] Nguyen NY, Suzuki A, Boykins RA, Liu T-Y. The amino [75] Sarcione EJ, Bogden AE. Hepatic synthesis of alpha 2 (acute
acid sequence of Limulus C-reactive protein. J Biol Chem phase)-globulin of rat plasma. Science 1966;153:547±8.
1986;261:10456±65. [76] Srere HK, Belke D, Wang LC, Martin SL. Alpha 2-Macro-
[61] Shrive AK, Metcalfe AM, Cartwright JR, Greenhough TJ. globulin gene expression during hibernation in ground squir-
C-reactive protein and SAP-like pentraxin are both present rels is independent of acute phase response. Am J Physiol
in Limulus polyphemus haemolymph: crystal structure and 1995;268:R1507±12.
Limulus SAP. J Mol Biol 1999;290:997±1008. [77] Coutinho CM, van Leuven F, Araujo-Jorge TC. Detection of
[62] Hoover GL, el-Mowa® A, Simko E, Kocal TE, Ferguson alpha-macroglobulin in the heart of mice infected with
HW, Hayes MA. Plasma proteins of rainbow trout (Oncor- Trypanosoma cruzi. Parasitol Res 1999;85:249±55.
hynchus mykiss) isolated by binding to lipopolysaccharide [78] Starkey PM, Fletcher TC, Barrett AJ. Evolution of alpha 2-
from Aeromonas salmonicida. Comp Biochem Physiol B macroglobulin. The puri®cation and characterization of a
1998;120:559±69. protein homologous with human alpha 2-macroglobulin
[63] Murata M, Onuma M, Kodama H. Isolation and character- from plaice (Pleuronectes platessa L.) plasma. Biochem J
ization of rainbow trout (Oncorhynchus mykiss) serum 1982;205:97±104.
amyloid P component (SAP). J Vet Med Sci 1994;56:661±5. [79] Freedman SJ. The role of alpha-2-macroglobulin in furuncu-
[64] Uhlar CM, Whitehead AS. Serum amyloid A, the major losis: a comparison of rainbow trout and brook trout. Comp
vertebrate acute-phase reactant. Eur J Biochem Biochem Physiol 1991;98B:549±53.
1999;265:501±23. [80] Demers NE. Immediate effects of acute stress in innate
[65] Berliner JA, Navab M, Fogelman AM, Frank JS, Demer LL, immunity in rainbow trout (Oncorhynchus mykiss). PhD
Edwards PA, Watson AD, Lusis AJ. Atherosclerosis: basic thesis. Oregon State University, 1996.
mechanism-oxidation, in¯ammation, and genetics. Circula- [81] Ellison RT, Giehl TJ. Killing of Gram-negative bacteria by
tion 1995;91:2488±96. lactoferrin and lysozyme. J Clin Investig 1991;88:1080±91.
[66] Banka CL, Yuan T, de Beer MC, Kindy M, Curtiss LK, de [82] Engstad RE, Robertsen B, Frivold E. Yeast glucan induces
Beer FC. Serum amyloid A (SAA): in¯uence on HDL- increase in lysozyme and complement-mediated haemolytic
mediated cellular cholesterol ef¯ux. J Lipid Res activity in Atlantic salmon blood. Fish Shell®sh Immunol
1995;36:1058±65. 1992;2:287±97.
[67] Fujiki K, Shin D-H, Nakao M, Yano T. Molecular cloning [83] Jorgensen JB, Sharp GJE, Secombes CJ, Robertsen B. Effect
and expression analysis of carp (Cyprinus carpio) interleu- of a yeast cell wall glucan on the bactericidal activity of
kin-1b, high af®nity immunoglobulin E Fc receptor ( subunit rainbow trout macrophages. Fish Shell®sh Immunol
and serum amyloid A. Fish Shell®sh Immunol 2000;10:229± 1993;3:267±77.
42. [84] Matsuyama H, Mangindaan REP, Yano T. Protective effect
[68] Jorgensen JB, Lunde H, Jensen L, Whitehead AS, Robertsen of schizophyllan and scleroglucan against Streptococcus sp.
B. Serum amyloid A transcription in Atlantic salmon (Salmo infection in yellowtail (Seriloa quinqueradiata). Aquacul-
salar L.) hepatocytes is enhanced by stimulation with macro- ture 1992;101:197±203.
phage factors, recombinant human IL-1 beta, IL-6 and TNF [85] Santarem M, Novoa B, Figueras A. Effects of b-glucans on
alpha or bacterial lipopolysaccharide. Dev Comp Immunol the non-speci®c immune response of turbot (Scophthalmus
2000;24:553±63. maximus L.). Fish Shell®sh Immunol 1997;7:429±37.
[69] Schreiber G, Tsykin A, Aldred AR, Thomas T, Fung W-P, [86] Robertsen B. Modulation of the non-speci®c defence of ®sh
Dickson PW, Cole T, Birch H, de Jong FA, Milland J. The by structurally conserved microbial polymers. Fish Shell®sh
acute phase response in the rodent. Ann NY Acad Sci Immunol 1999;9:269±90.
1989;557:61±85. [87] Demers NE, Bayne CJ. The immediate effects of stress on
[70] Gerwick LG. Acute phase proteins in rainbow trout (Oncor- hormones and plasma lysozyme in rainbow trout. Dev Comp
hynchus mykiss). PhD thesis. Oregon State University, 2000. Immunol 1997;21:363±73.
[71] Congleton JL, Wagner EJ. Acute-phase hypoferremic [88] Gercken J, Renwrantz L. A new mannan-binding lectin from
 C.J. Bayne, L. Gerwick / Developmental and Comparative Immunology 25 (2001) 725±743 743

the serum of the eel (Anguilla anguilla L.): isolation, char- Cloning, structure, and function of two rainbow trout Bf
acterization and comparison with the fucose-speci®c serum molecules. J Immunol 1998;161:4106±14.
lectin. Comp Biochem Physiol Biochem Mol Biol [99] Nakao M, Fushitani Y, Fujiki K, Nonaka M, Yano T. Two
1994;108:449±61. diverged complement factor B/C2-like cDNA sequences
[89] Ottinger CA, Johnson SC, Ewart KV, Brown LL, Ross NW. from a teleost, the common carp (Cyprinus carpio). J Immu-
Enhancement of anti-Aeromonas salmonicida activity in nol 1998;161:4811±8.
Atlantic salmon (Salmo salar) macrophages by a mannose- [100] Ellis AE. Immunity to bacteria in ®sh. Fish Shell®sh Immu-
binding lectin. Comp Biochem Physiol C Pharmacol Toxicol nol 1999;9:291±308.
Endocrinol 1999;123:53±59. [101] Yano T, Mangindaan REP, Matsuyama H. Enhancement of
[90] Gerwick L, Reynolds WS, Bayne CJ. A precerebellin-like the resistance of carp Cyprinus carpio to experimental
protein is part of the acute phase response in rainbow trout. Edswardiella tarda infection, by some b-1,3-glucans. Bull
Dev Comp Immunol 2000;24:597±607. Japan Soc Sci Fish 1989;55:1815±9.
[91] Urade Y, Oberdick J, Molinar-Rode R, Morgan JI. Precere- [102] Bini L, Magi B, Marzocchi B, Cellesi C, Berti B, Raggiaschi
bellin is a cerebellum-speci®c protein with similarity to the R, Rossolini A, Pallini V. Two-dimensional electrophoretic
globular domain of complement C1q B chain. Proc Natl patterns of acute-phase human serum proteins in the course
Acad Sci USA 1990;88:1069±73. of bacterial and viral diseases. Electrophoresis 1996;17:612±
[92] Lee PH, Goetz FW. Characterization of a novel cDNA 6.
obtained through differential display PCR of phorbol-ester- [103] Gerwick L, Steinhauer R, LaPatra S, Sandell T, Ortuno J,
stimulated ovarian tissue from the brook trout (Salvelinus Hajiseyedjavadi N, Bayne, CJ. The acute phase response of
fontinalis). Mol Reprod Dev 1998;49:112±8. rainbow trout (Oncorhynchus mykiss) plasma proteins to
[93] Vakeva A, Lehto T, Takala A, Meri S. Detection of a soluble viral, bacterial and fungal in¯ammatory agents. Fish Shell-
form of the complement membrane attack complex inhibitor ®sh Immunol 2001 (in press).
CD59 in plasma after acute myocardial infarction. Scand J [104] Wegenka UM, Buschmann J, Lutticken C, Heinrich PC,
Immunol 2000;52:411±4. Horn F. Acute-phase response factor, a nuclear factor bind-
[94] Ban®eld DK, MacGillivray RTA. Partial characterization of ing to acute-phase response elements, is rapidly activated by
vertebrate prothrombin cDNAs: Ampli®cation and sequence interleukin-6 at the posttranslational level. Mol Cell Biol
analysis of the B chain of thrombin from nine different 1993;13:276±88.
species. Proc Natl Acad Sci USA 1992;89:2779±83. [105] Gerwick L, Demers NE, Bayne CJ. Modulation of stress
[95] Miot S, Duval J, Le Goff P. Molecular cloning of a hemo- hormones in rainbow trout by means of anesthesia, sensory
pexin-like cDNA from rainbow trout liver. DNA Seq deprivation and receptor blockade. Comp Biochem Physiol
1996;6:311±8. A 1999;124:329±34.
[96] Dautigny A, Prager EM, Pham-Dinh D, Jolles J, Pakdel F, [106] Sunyer JO, Zarkadis IK, Sahu A, Lambris JD. Multiple forms
Grinde B, Jolles P. cDNA and amino acid sequences of rain- of complement C3 in trout that differ in binding to comple-
bow trout (Oncorhynchus mykiss) lysozymes and their impli- ment activators. Proc Natl Acad Sci USA 1996;93:8546±51.
cations for the evolution of lysozyme and lactalbumin. J Mol [107] Mutsuro J, Nakao M, Fujiki K, Yano T. Multiple forms of
Evol 1991;32:187±98. alpha2-macroglobulin from a bony ®sh, the common carp
[97] Zarkadis IK, Sarrias MR, Sfyroera G, Sunyer JO, Lambris (Cyprinus carpio): striking sequence diversity in functional
JD. Cloning and structure of three rainbow trout C3 mole- sites. Immunogenetics 2000;51:847±55.
cules: a plausible explanation for their functional diversity. [108] van Vugt H, van Gool J, de Ridder L. Alpha-2-macroglobulin
Dev Comp Immunol 2001;25:11±24. of the rat, an acute phase protein, mitigates the early course
[98] Sunyer JO, Zarkadis I, Sarrias MR, Hansen JD, Lambris JD. of endotoxin shock. Brit J Exp Pathol 1986;67:313±9.