Module – 3 Assignment: Analytical Chemistry

and Molecular Modelling – November 2009

Importance of Mass Spectrometry in

Biomedicine Illustrated by Cancer
Screening Applications

Jyotika Govil, MSc Molecular Medicine (044322)

Abstract: Mass Spectrometry is one of most widely applicable analytical

technique used for identification and characterization of molecules of a
large range of molecular weights. There are different components of a
mass spectrometer, which can be altered according to experimental
needs, and are briefly described in this paper. To visualize impact of MS
in biomedicine, applications are mentioned in drug discovery. However,
for the purpose of recognizing a clinical impact on general population,
use of protein profiling by MALDI-TOF MS in detection of cancer is the
main focus of this review.
1. Introduction
The technique used and advanced for the characterization and identification of
unknown and known compounds in a charged molecular state, since 1912, when
J.J. Thomson first devised it, is Mass Spectrometry. The technology started as
spectroscopy using fluorescent graphs was later modified into magnetic field
detection and hence termed as spectrometry .

The basic principle of mass spectrometry is to convert the introduced sample into
gaseous state charged ions which are accelerated and separated under a
magnetic field on the basis of their mass/charge (m/z) ratio, analyzed and
detected to form a mass spectrum. The schematic diagram in Figure-1 shows the
basic components of a mass spectrometer.

Figure-1: General schematic of parts of a Mass Spectrometer

Moving across the various components, as shown above, the processes that occur
are:

1. Introduction of sample through an inlet, which can be direct or through a

chromatograph.

2. Vaporization and ionization of sample to impart a charge (often positive) to the

molecule..
3. The ions are propelled to a mass analyzer, which consists of a magnetic field
and due to difference in m/z ratios, the molecules follow different trajectories and
speeds which are then scanned by the analyzers.

4. To produce an interpretable signal, detectors are employed which convert the

molecular analysis into electronic current signals, producing a visual spectrum,
recorded and displayed by computers.

1.1 Ionization Techniques

Electron Ionisation (EI): this is the oldest method, where a beam of electrons
from a cathode is bombarded upon the gaseous sample, leading to knocking out of
electrons. The positively charged sample ions are then accelerated and analyzed.
Being a strong ionization technique, often leads to fragmentation of sample.

Fast Atom Bombardment (FAB): The sample solubilised in a non-volatile matrix

(eg. Glycerol) is bombarded with a high current beam of atoms/ions (Argon or
Xenon) which does not fragment the sample and only eject surface ions dissolved
in the matrix. Has advantage of producing ions from biological compounds of high
molecular weight, polar samples and also generating ions for longer duration .

Electrospray Ionisation (ESI): Commonly used for large biomolecules, it ionizes

at atmospheric pressure, from solution. The sample is converted into fine spray
and is ionized in presence of strong electric field. The nozzle at high electric
potential forms the droplets, which are evaporated by action of heat/gas at
atmospheric pressure. Concentrating charge upon smaller droplets increases the
charge density. The ion formation and conversion into smaller droplets is
illustrated in Figure-2.
 Figure–2: Mass Spec by ESI

Matrix Assisted Laser Desorption Ionisation (MALDI): This is an advanced

soft ionisation method for analysis if biomolecules of up to 500,000 Daltons with
high accuracy and without causing fragmentation. The principle is to co-crystallize
the sample onto a matrix of UV-absorbing material which is irradiated with a laser.
The photons excite and vaporize matrix molecules, along with the analyte and
ionisation occurs. The ions are accelerated by the presence of electric field in
between the MALDI plate and analyser. It enables simultaneous analysis of many
samples of complex molecules. A similar technique is Surface Enhanced Laser
Desorption/Ionisation (SELDI) also used for biomolecules .

1.2 Analyzers

Magnetic Field Analyzers: A magnetic field is applied by an electromagnet

which is perpendicular to the path followed by the accelerated ions. The path is
curved and ions have the same kinetic energy when coming from the ionizers.
They separate depending upon their m/z ratio and only the ions with
corresponding trajectory are able to pass through. The strength of the magnetic
field is changed and hence due to varied m/z ratios, the components of the sample
are analyzed.

Time of Flight (TOF) Analyzers: The ions after being expelled from the
sample/matrix are accelerated by an electric potential and move a certain
distance before hitting the detector. By measuring this time-of-flight of the ions,
their m/z ratios are determined. The ions with lower mass have high speed and are
detected first. This detection method has no mass limit hence can be used for high
mass biomolecules and is most often combined with MALDI for detection.

Quadrupolar Analyzers: The apparatus has four parallel metal electrodes which
produce a Direct Current (DC) and Alternating Current (AC) field in between them.
At a particular time only a specific m/z ratio resonant ion can pass through the
field path and read by the detector as shown in Figure-3. By modifying the fields
produced by the electrodes, the whole spectra of different m/z ratios can be
established.

Figure-3: Quadrupole analysis after electron beam ionisation (Wittmann, 2007)

1.3 Detectors

Faraday-Cup Detector: This is one of the older methods of charge detection, the
ions passing through the analyzers, release their charge at the electrodes and the
current is amplified and measured. These are not highly sensitive detectors.

Electron Multipliers: The basic principle is use of dynodes, which release

secondary particles after receiving the analyte ions. There is cascade of
ions/electron release, amplifying the signal many folds, which can be through a
continuous dynode electron multiplier giving a measurable current.
Photon Multipliers: These detectors involve electron dynodes as well as
phosphorescent plates which convert electronic signal into photons. A visible
signal is recorded by the photomultiplier.

1.4 Tandem Mass Spectrometry

This process involves better analysis of components of a large molecule,

composition and structure elucidation with higher sensitivity. Usually abbreviated
as MS/MS it applies two stages of mass spectrometry. Firstly, the molecular ions
are selected by the m/z ratio using a particular MS technique, then the molecule is
fragmented by use of inert gas to be analysed again by a MS technique.

1.5 MS Coupling

MS is coupled with various separation techniques to analyse complex mixtures

and separate natural products and to identify unknown compounds. Coupling with
various chromatographic techniques like Gas Chromatography (GC-MS), Liquid
(LC-MS), High Performance Liquid Chromatography (HPLC-MS) or Capillary
Electrophoresis (EC-MS) is done to co-ordinate separation, purification of analytes
before characterisation with MS .

2. Literature Review

Mass spectrometry in its various forms is widely used in pharmaceutical and

biomedical field. Apart from drug analysis, MS is applicable for characterization
and elucidation of biomolecules like nucleic acid, lipids, biopolymers and proteins.
MS is used to initiate serum and urine diagnosis tools for detecting marker
proteins, drugs, metabolites etc .

MS is coupled with proteomics and bioinformatic tools to build systems that can
screen protein libraries and generate new ones. With the availability of high-
throughput screening by MS, it is applied for detecting protein biomarkers of
diseases in plasma/serum/urine which is a highly sought after field .
Hence, the focus of this review is particularly on the use of MALDI/SELDI – TOF for
screening of cancer by biomarker characterization in serum/saliva/cells.

2.1 Article: Use of proteomic patterns in serum to identify ovarian

cancer – Petricoin III et al, 2002; The Lancet.

Background: With the availability of treatment for ovarian cancer with only
surgery, and high survival rate, detection in stage-I can have a major impact on
women at high risk due to familial history or genetic predisposition with BRCA1
and BRCA2 genes. Biomarkers like CA125 have been used for detection, but with
low positive predictive value. Hence, developing a bioinformatic technique for
high-risk population is important and feasible with the aid of high-throughput
SELDI/MALDI-TOF screening to measure low molecular weight serum proteins that
can profile a diseased condition.

Aim: To establish a population based screening technique with MS and proteomic

patterns by bioinformatic algorithms, for early and consistent detection of ovarian
cancer.

Methods: A training set of data was accumulated by acquiring serum samples

from 50 healthy females and 50 with confirmed ovarian cancer, and generating
the proteomic mass spectra to segregate cancer from non-cancer spectra by use
of genetic algorithms and cluster analysis. To test the method, 116 masked
samples (50 with cancer including stage-I cancer, 66 healthy) were analysed by
Protein Biology System 2 SELDI-TOF mass spectrometer on a protein chip, with
specificity for molecular mass range 0-20,000Da. Positives and controls were
analysed simultaneously on same as well as multiple chips.

The pattern established with training set was compared with the test set data to
distinguish patterns of proteins different in cancer patients by repetitive algorithm
matching and classified into clusters of – unaffected, diseased and new cluster
(completely new pattern).
Results: The reproducibility of the experiment was successfully tested. The
results show that the algorithm correctly segregates the cancer and non-cancer
spectra. In the masked set, there was 100% sensitivity, 95% specificity and also
indentifying all 50 cancer samples with the 18 stage-I cancer. It shows high
positive predictive value (94%) in contrast to low results with CA125 biomarker.

Conclusion: Screening for ovarian cancer in high-risk groups is important to

tackle the disease, hence such individuals were source of samples. The
combination of proteomic mass spectra and bioinformatic genetic algorithms
enables identification and segregation of low molecular weight protein patterns
that discriminate cancer from non-cancer. The technique shows higher promise
than other screening methods and with requiring small samples for MS and being
cost-effective, this high-throughput screening, if combined with ultrasonography
can be applied for clinical diagnostics .

2.2 Article: Mass spectrometry-based serum proteome pattern analysis

in molecular diagnostics of early stage breast cancer – Pietrowska et al,
2009; Journal of Translational Medicine

Background: With the establishment of proteomic pattern analysis for developing

diagnostics for cancer, this experiment was carried out for early breast cancer
detection as it is the most common malignancy in females. The analysis of
complete protein profiles and not individual peptides is enabled by MALDI-TOF
spectrometry which characterizes proteins of low molecular mass with high
accuracy and low time from complex samples (like serum). These spectral pattern
analyses can be compared as more efficient than individual biomarker analysis by
immunoassays.
Aim: To develop a serum based proteome blueprint analysis, by statistical
classifiers, of spectral patterns generated by MALDI-TOF and comparing it with
protein biomarker immunoassays.

Methods: Analysis was done on samples obtained (by informed consent) from 92
breast cancer (stage I and II) patients and from 104 healthy females (same age
group) as controls. The serum was removed of albumin and other high – molecular
weight proteins and analysed with Autoflex MALDI-TOF spectrometer to obtain
spectra between 2,000-10,000Da mass ranges, recording four spectra for every
sample. For comparison, immunoassays (ELISA, CMIA, TRACE & MEIA) were
conducted for selected markers. The data analysis was carried out by identifying
the [M+H]+ ions recorded on spectra, into Gaussian-Bell curves, computing
algorithms and forming classifiers of patterns to distinguish between cancer and
non-cancer samples.

Results: Using the MALDI spectrum of proteins, a classifier was established to

identify protein patterns and pick similar patterns. With this classifier, spectral
patterns that discriminate between healthy and cancer samples were configured.
Three positions on the spectra: m/z= 2865.54, 3578.73 & 2303.48Da were most
important for segregating the diseased samples, two of which were exclusively
found in this study due to sample pre-treatment and albumin removal. These
patterns were independent of age differences in the source subjects. The classifier
was specific even with other pathologic conditions. In comparison to other
biomarker immunoassays, the biomarker from these three spectral positions, gave
a higher sensitivity (88%) and specificity (78%). Due to observation of osteopontin
in only cancer samples, it was also tested as a classifier biomarker in combination,
but gave only 28% specificity.

Conclusion: Due to high-throughput and accurate proteome analysis with MALDI-

TOF, its combination with proteomics and bioinformatics for the analysis of serum
protein fingerprints and patterns promises as a better diagnostic for early breast
cancer detection. This procedure overcomes the limitations of existing
immunological testing for protein breast cancer biomarkers .
2.3 Article: Detection of colorectal cancer using MALDI-TOF serum
protein profiling – de Noo et al, 2006; European Journal of Cancer

Background: Serum protein profiling for distinguishing cancer from non-cancer

has been used and established, as shown in many previous experiments. The
same concept is being applied for early diagnosis of colorectal cancer, which has
the highest cancer morbidity and mortality rates. There is a requirement for
biomarkers for this cancer as well for detection, monitoring and therapeutic
studies. Although the protein profiling by MS has been appreciatively used, there
have been criticisms about the statistical analyses of data and chances of bias or
over-interpretation. Hence, this study uses “double cross-validation” for authentic
results.

Aim: To use a stringent statistical analysis for non-biased development of a

diagnostic technique for early detection of colorectal cancer. This is based on Mass
Spectrometric analysis of serum protein profiles.

Methods: For conducting the study, samples from 66 stage-IV colorectal cancer
patients were taken, and as controls from 50 healthy individuals. To minimize
batch effects, all samples were equally randomized and the procedure repeated a
week after the first test. The peptides were isolated with use of Magnetic Bead-
Hydrophobic Interaction Chromatography. The samples were then subjected to
MALDI-TOF (Ultraflex instrument) operating with a linear mode SCOUT ion source.
Peaks were specifically measured in the range of 960 – 11,169Da. The resultant
data was normalized for statistical analysis and also distributed into inter-quartile
ranges of intensity.

For the interpretation, the “double cross-validation” method was employed, to

exclude every possible bias.

Results: The method displayed a 92.6% total recognition rate, 95.2% sensitivity
and 90.0% specificity with the first spectra, accurately classifying 60/63 cancer
samples (including all stage III,IV and 9/10 stage I) and 45/50 controls. To remove
bias entire process was repeated after a week and a permutation calculation of
the cross-validation was done with no significant bias found.

Conclusion: The experiment followed a previously explored technique of

discriminating cancer from non-cancer patients on basis of MS spectra protein
profiling, in this case for colorectal cancer. The method was highly reproducible
and statistical and experimental bias was endeavoured to be highly reduced. A
classifier was successfully constructed and achieved high sensitivity and
specificity (95.2% & 90% respectively). Due to possibility of discrepancies in
previously conducted similar experiments, here special focus was on proper
sample collection, randomized testing and “double cross-validation” of
experimental data. The classifier was chosen as “Fisher linear discrimination”
which a widely used discrimination method has given a high separation with
97.3% of data under the curve. The method yet to be completely validated is
highly promising for colorectal diagnosis .

2.4 Article: Analysis of the saliva from patients with oral cancer by
matrix-assisted laser desorption/ionization time-of-flight mass
spectrometry – Chen et al, 2002; Rapid Communications in Mass Spectrometry

Background: There are changes in the biomolecular constitution of saliva with

physiological changes in the body. The starch digesting enzyme alpha-amylase
also changes in concentration in patients of oral cancer with altered hydrolytic
activity of the enzyme. To proceed with this as a basis of distinguishing normal
people from oral cancer patients, a combination of MALDI-TOF with SDS-PAGE is
done. As electrophoresis was the only characterization method for proteins before,
due to high sensitivity and rapid result generation, MALDI is used.

Aim: To develop a swift method of oral cancer diagnosis by use of MALDI-TOF

mass spectrometry.
Methods: Untreated saliva samples were collected from 20 healthy individuals
(for control) and from 20 oral cancer patients. Also, the control samples were
divided into groups pertaining to blood groups. Samples were prepared for
spectrometry by mixing them with sinapinic acid matrix in acetonitrile in varying
ratios. For enhanced characterization, the samples were electrophoresed with
SDS-PAGE. The bands were cut and digested with trypsin and then used on
sinapinic acid matrix for MALDI-TOF linear spectrometry.

Results: The peak of alpha-amylase was particularly observed and studied at

57kDa in the spectra. It was seen in 13/20 control samples (in blood group-O, no
peaks seen which could be due to some unknown blood group component, but
could not be explained here). Another peak at 66kDa was observed, but only in 3
controls and more pronounced and present in the samples from cancer patients
(15/20) with the 57kDa peak disappearing. The 66kDa peak corresponds to
albumin which was confirmed with in-gel trypsin digestion fingerprinting.
Electrophoresis also shows higher concentration of 66kDa band (albumin) in
cancer patients.

Conclusion: These results show a possibility of changed protein composition of

saliva (especially in terms of albumin concentration increase and alpha-amylase
suppression) in cancer patients. However, no confirmation of this direct relation
could be made with this experiment. The amylase concentration is not suppressed
in SDS-PAGE, and in MALDI the reduction could be attributed to a general
suppression of alpha-amylase concentration due to increase in albumin
concentration (as seen by comparing spectra of different volume ratios). The
combination method of SDS-PAGE with MALDI-TOF gives a rapid diagnostic
approach which is minimally invasive and has easier sample acquisition, being a
saliva test. The saliva profile changes give discrimination between cancer and
non-cancer, but the relation still needs to be validated .
2.5 Article: High-Throughput Genotyping of Oncogenic Human
Papilloma Viruses with MALDI-TOF Mass Spectrometry - Soderlund-Strand
et al, 2008;

Background: Cervical cancer is principally caused by the Human Papilloma Virus

(HPV), which exists in various forms. Apart from the need to develop
immunization, the genotyping and screening of HPV is important for detection of
cervical cancer as well as monitoring of HPV load in vaccinated individuals. The
different genotypes confer different risks for populations for development of
cervical cancer and cervical intraepithelial neoplasia (CIN).

Aim: To develop a high-throughput, screening methodology for genotyping 14

HPV types, known of causing cervical cancer by MALDI-TOF analysis. This is done
in comparison with the standard HPV genotyping procedure of PCR amplimers
undergoing Reverse Dot Blot Hybridisation (RDBH).

Methods: The analysis was done with 532 cervical cell samples, 271 f which
obtained from cancer patients undergoing treatment and follow-up. The samples
were prepared by freeze-thawing followed by centrifugation and with 10-fold
dilution for comparison of methods. The HPV DNA was amplified with PCR followed
by target identification with immunoassay. The mass spectrometry was carried out
using multiples primary PCR products on a Sequenom MassARRAY. Various
primers were designed for PCR to get optimum annealing, followed up with
homogenous mass extend (hME) reaction to specify each genotype. There were
discrepancies in the results obtained from the two methods and reanalysis with
MS and RDBH was conducted with excluding exceptions to have a total of 502
samples. Statistical analysis was done by ĸ statistics for concordance between the
methods.

Results: The MS method when compared with histopathology gave 91.4%

sensitivity, only 46% specificity and 58.9% positive predictive value. It identified
all samples shown positive by RDBH and also 5 more samples which could not be
identified with RDBH. Overall concordance between MS analysis and RDBH was
0.945. There were discrepancies in some data when compared for the two
methods, however, these were mostly in the samples taken from stage I of CIN.

Conclusion: MS-based screening and HPV genotyping proved to be consistent,

more sensitive and also economical. 10 x 384 samples could be analysed in 2 days
with an individual cost of US$2 per sample. Due to the study being conducted on a
large sample, the results can be considered as generally valid. MS enabled
detection of HPV types which are missed by other tests (type 16 and 18) and
carrying out detection of 1-100 HPV DNA copies of 14 types with high efficacy and
speed. These plus points could enable a clinical application of MS high-throughput
screening for cervical cancer.

3. Conclusion

Before the development of Mass Spectrometry, the characterization of absolute

molecular weight was dependent on conformations and calculations based on data
obtained from chromatography, electrophoresis and centrifugation. The error rates
with these are high and other molecular properties also interfere. Large molecules
like biomolecules have been successfully characterized with MS purely on basis of
experimental proof and not theoretical optimizations. With the ability to convert
samples into molecular ions and highly specific separation on basis of
mass/charge, MS has been developed as a prime analytical tool. With advances
like MALDI, ESI, SELDI and TOF identification of blood and plasma components has
been possible as they are soft and more specific techniques and do not cause
fragmentation of biomolecules.

In the process of drug development MS is applied for characterization of

compounds at each step of discovery – screening of purity, physical and chemical
characteristics, lead optimization, in vitro screening of pharamacokinetic
properties (ADME- Absorption, Distribution, Metabolism & Excretion) of drugs,
metabolite identification both in vitro and in vivo, therapeutic assessments,
toxicology studies, drug-protein interactions etc .

Mass Spectrometry is used in combination with separation techniques like Liquid

Chromatography, HPLC, Electrophoresis and the analytes are directly sent from
the purification columns into the ionization chambers. The purification of
compounds is very important as MS being highly sensitive obtains an m/z peak for
all components. Occurrence of impurities and interfering particles causes
suppression of peaks or distorted spectra. As it has been observed from the
articles, in the complex constitution of serum various proteins alter observations
of the desired proteins. Albumin has been shown to suppress peaks of small
molecular weight proteins needed for screening as biomarkers. Hence, it is
important to pre-treat the samples before analysing with MS . The right choice of
ionization technique is very important as a strong technique like EI or CI can lead
to fragmentation, which could be important for certain molecules, for example, in
metabolite analysis in ADME studies of drugs where Desorption ESI and MALDI are
used to build high-throughput screening libraries for fast development process .
The benefit of high-throughput screening has increased usage in drug screening,
however, the complexity of MS restricts usage for analysing millions of lead .

In the reviewed articles, MS has been extensively applied to combine proteomic

libraries with spectral data to identify biomarkers. MS was used since small
proteins could only be detected with this technique. However, individual
biomarkers are not identified, hence, profiling of complete spectra of components
is done. There have been successful discriminations built with bioinformatic and
statistical tools, yet there have been high chances of optimistic bias and
reproducibility issues which have prevented such techniques to be clinically
applied. On one hand, where the high sensitivity and specificity of MS enables
such characterizations, on the other hand, the same qualities limit its ability to be
consistently used in a clinical environment. Some of the studies have been
conducted on small sample sizes and in a specific population which limits the
potential of being applied for different races, as they may differ in their protein
profiles. For every study a training set is to be made for comparisons and if
clinically applied it would be difficult to have a standard profile with which
comparisons could be made.

To summarize, Mass Spectrometry due to adaptability and applicability with fast

and accurate analysis, is utilized and has a high impact in drug discovery,
screening, lead optimization, drug monitoring, biomedicine, disease diagnosis,
molecular modelling, imaging, isotopic identification, biomarker identification,
molecular weight determination, fingerprinting and structural analysis of peptides
in proteomics, lipids, nucleotides and other biopolymers , .
 BIBLIOGRAPHY

Wittmann, (2007), “Schematic view of the ion source based on electron impact
ionization and the quadrupole mass filter typically found in a GC-MS
instrument”, Microbial Cell Factories, vol.6.