Embed Size (px)
Citation preview
Mass spectrometric characterization of urinary
fibrinogen-derived peptides in prostate cancer and renal
cell carcinoma
Samuel Mesihää
Pro Gradu
Mass spectrometric characterization of urinary
fibrinogen-derived peptides in prostate cancer and
renal cell carcinoma
Samuel Mesihää
University of Oulu
Faculty of Biochemistry and Molecular Medicine
2015
This work was carried out at the Department of Clinical Chemistry, University of Helsinki
Helsinki, Finland
Supervisor:
Emeritus professor Ulf-Håkan Stenman, MD, PhD
Acknowledgements
I’m grateful for having an opportunity to work not only in one, but three laboratories under
the Medical Faculty of University of Helsinki. First of all, I thank Ulf-Håkan Stenman for
sharing his interest in the field of clinical chemistry and mass spectrometry and for
supervising me with this project. Suvi Ravela worked with the clinical validation related to
this project and her efforts encouraged this project. Maarit Leinimaa was invaluable help in
practical laboratory procedures, data interpretation and planning of the project.
Mass spectrometric analyses of this project were done in Professor Risto Renkonen’s
laboratory at Haartman Institute and Professor Marc Baumann’s laboratory at Meilahti
Clinical Proteomics Core Facility. I thank Risto and Marc for their co-operation. I’m grateful
to Rabah Soliymani and Suvi Saarnio from Baumann’s laboratory and Sakari Joenväärä from
Renkonen’s laboratory for introducing me to practical mass spectrometry.
Abbreviations
ACN acetonitrile
CHCA α-cyano-4-hydroxy cinnamic acid
CID collision-induced dissociation
DDA data dependent acquisition
DIA data independent acquisition
DTT N1–(p-isothiocyanatobenzyl)-diethylenetriamine-N1,N2,N3,N3-
tetraacetic acid
ESI electrospray ionization
FA formic acid
FDPs fibrinogen degradation products
HLB hydrophilic-lipophilic balance
HPLC high-performance liquid chromatography
IEX Ion exchange chromatography
IS internal standard
LC liquid chromatography
LOD limit of detection
mAbs monoclonal antibodies
MALDI matrix-assisted laser desorption/ionization
MCX mixed-mode cation exchange
MS mass spectrometry
MSE data independent mass spectrometry acquisition method
MS/MS tandem mass spectrometry
NMR nuclear magnetic resonance
m/z mass-to-charge ratio
PC prostate cancer
pAbs polyclonal antibodies
QTOF quadrupole time-of-flight
RCC renal cell carcinoma
RPC reversed phase chromatography
SEC size-exclusion chromatography
SPE solid phase extraction
TBS tris-buffered saline
TFA trifluoroacetic acid
TOF time-of-flight
TOF/TOF tandem time-of-flight
TR-IFMA time-resolved immunofluorometric assay
UHPLC ultra-high performance liquid chromatography
TABLE OF CONTENTS
Acknowledgements
Abbreviations
Table of contents
I LITERATURE SECTION
1 Introduction ................................................................................................................... 1
2 Review of the Literature ............................................................................................ 2
2.1 Biology of fibrinogen ....................................................................................................... 2
2.1.1 Fibrin(ogen) in cancer ......................................................................................... 4
2.2 Tumor markers ................................................................................................................. 5
2.2.1 Quantitative tumor marker discovery .................................................................. 7
2.3 Renal cell carcinoma and prostate cancer ........................................................................ 8
2.4 Proteases and peptides in cancer ...................................................................................... 8
2.5 Mass spectrometry ......................................................................................................... 10
2.5.1 Mass spectrometry in clinical laboratories ........................................................ 10
2.5.2 Electrospray ionization and matrix-assisted laser desorption/ionization in
peptide analysis ................................................................................................. 11
2.5.3 Mass analyzers ................................................................................................... 13
2.5.4 Data collection in peptide tandem mass spectrometry: Data-dependent and
data-independent acquisition ............................................................................. 14
2.5.5 Discovery and analysis of low-abundant peptides ............................................ 15
II EXPERIMENTAL PART
3 Aim of the project ...................................................................................................... 17
4 Materials and Methods ............................................................................................ 18
4.1 Materials ........................................................................................................................ 18
4.1.1 Urine samples .................................................................................................... 18
4.1.2 Chemicals and reagents ..................................................................................... 18
4.2 Biotinylation and Eu- Labelling .................................................................................... 19
4.3 Competitive time-resolved immunofluorometric assay ................................................. 19
4.4 Solid phase extraction .................................................................................................... 20
4.5 Mass spectrometry ......................................................................................................... 22
4.5.1 Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry . 22
4.5.2 Ultra-performance liquid chromatography quadrupole time-of-flight mass
spectrometry ...................................................................................................... 22
4.5.3 Data in table V ................................................................................................... 23
4.5.4 Data analysis ...................................................................................................... 24
4.5.5 Overview of the method development strategy ................................................. 24
5 Results ............................................................................................................................ 27
5.1 Evaluation of extraction methods in enrichment of urinary fibrinogen β peptides ....... 27
5.2 Peptide sequencing ......................................................................................................... 28
5.3 Immunological measurement of N-terminal fibrinogen β peptides ............................... 35
6 Discussion ...................................................................................................................... 36
7 Conclusions .................................................................................................................. 39
8 References ..................................................................................................................... 40
1
I LITERATURE SECTION
1 Introduction
Most cancer deaths are caused by metastases. Metastatic dissemination of cancer cell is
associated with increased proteolytic activity (López-Otin & Matrisian 2007), leading to
degradation of proteins, i.e., appearance of peptides, which may be specific for cancer. The
peptides are often excreted in urine, from where they are much easier to detect than in plasma.
Plasma protein concentration may be over 1000-fold more than in urine, while the peptide
levels are often similar due to glomerular filtration (Hortin & Sviridov 2007). This makes
urine attractive non-invasive candidate for tumor marker discovery.
Here we focused on fibrinogen-derived peptides, which we have previously found to be
potential tumor markers. These peptides occur at low concentrations in urine. Identification of
these peptides is therefore a challenging task and optimized sample treatment is required. We
investigated urine samples from patients with renal cell carcinoma (RCC) and prostate cancer
(PC) due to proximity of these tumors to urinary tract.
This thesis focuses on the development of efficient peptide enrichment strategies and
application of different mass spectrometric data acquisition modes to identify fibrinogen-
derived peptides with unknown mass and amino acid sequence in urine.
2
2 Review of the Literature
2.1 Biology of fibrinogen
Fibrinogen is a large soluble glycoprotein (340 kDa) in plasma that is synthesized by liver
hepatocytes. It consists of two sets of three different disulfide-linked subunits (α, β and γ).
The plasma fibrinogen levels are determined by the rate of fibrinogen synthesis and its
conversion to fibrin. Fibrinogen participates in two main physiological roles: in blood clotting
and ii) in acute phase reaction following the tissue inflammation (Jackson & Nemerson 1980,
Giavannini et al. 2004 & Weisel 2005).
In the clotting process (coagulation) soluble fibrinogen is converted by thrombin into an
insoluble fibrin polymer which forms a plug or “thrombus” to the coagulation site together
with red and white blood cells, platelets and other adhesive components. In the unplugging
process, termed as fibrinolysis, fibrinogen is degraded by plasmin and other proteolytic
enzymes. Regulation of coagulation and fibrinolysis is mediated by several different co-
factors, receptors, enzymes and inhibitors. Coagulation and fibrinolysis processes are
reviewed more extensively by Jackson & Nemerson 1980 and more recently by Chapin &
Hajjar 2015.
Briefly, in coagulation and fibrinolysis, liver-produced fibrinogen is converted into an
insoluble fibrin polymer from a soluble fibrinogen (coagulation) and fibrin in the clot is
degraded by several proteolytic enzymes (fibrinolysis). As a result of fibrinolysis, fibrinogen
degradation products (FDPs) derived from fibrinogen α, β and γ chains are formed. D-dimer
is also derived from fibrinogen, but it is only produced by a hard clot.
Some specific FDPs seem to have unique properties based on the cleavage pattern. For
instance fibrinogen beta (FIBβ) fragments seem to have protective and regenerative (FIBβ 45-
72, Krishnamoorthy et al. 2011), immunoregulatory (FIBβ 31-72, Skogen et al. 1998) and
anti-angiogenic properties (FIBβ 73-92, Krajewska et al. 2010) based on the site of the
cleavage and probably tissue involved.
3
The main proteolytic enzymes in these reactions are thrombin and plasmin, both of which
activation are discreetly regulated. Life-cycle of a fibrin(ogen) into a fibrinogen-derived
peptide is summarized in figure 1.
A
B
Figure 1. Fibrinogen in blood clotting and fibrinolysis A) at submolecular level and B) in fibrinogen beta
polypeptide). Coagulation: Fibrinopeptides A and B are released by the thrombin-catalyzed hydrolysis (blue
color) from α and β chains respectively. The release of fibrinopeptides expose the N-terminus of fibrinogen
alpha and beta chains, which subsequently leads to polymerization of the fibrin monomers into a soft clot.
Clotting factor XIIIA is a transglutaminase which is required to convert soluble soft clot into an insoluble hard
clot. Fibrinolysis: Activated plasmin degrades the insoluble fibrin hard clot and further breakdown of fibrinogen
is assisted by other proteases. Modified from and Adam et al. 2008 & UniProt Knowledgebase (2015).
Fibrinogen is measured in clinical laboratories. In United States plasma fibrinogen reference
values range from 2.0-4.3 g/L. Fibrinogen levels are decreased in fibrinolysis, severe liver
injuries, disseminated intravascular coagulation (DIC) or in hereditary hypofibrinogenemia.
Elevated levels of fibrinogen can be detected in inflammatory diseases and during pregnancy.
(Mayo Clinic 2015).
Widely used assay for fibrinogen-derived peptides in clinical laboratories is a D-dimer test
which is measured from plasma. Fibrinogen peptides containing D-dimer antigens are
markers to specifically indicate formation of a hard clot. Elevated D-dimer signifies enhanced
4
clot formation and/or fibrinolysis (Adam et al. 2009) and it is therefore a useful for exclusion
of acute pulmonary embolism (PE) and deep vein thrombosis (DVT) but abnormal D-dimer
levels may also indicate a liver disease, inflammation (Iba et al. 2007) or coagulation
abnormalities. Fibrinogen-derived peptides can be used for screening of bladder cancer
(Tsihlias & Grossman 2000 & Gromov et al. 2006). Other FDP-based assays are not yet in
wide use.
2.1.1 Fibrin(ogen) in cancer
Fibrinogen has been traditionally studied in context of thrombotic events. Fibrinogen is also
synthesized by many cancer cells (Sahni et al. 2008). Fibrin depositions can be found in some
malignant tumors, but not all (Hiramoto et al. 1960). Fibrin(ogen) seem to be especially
important in early stage metastasis of some cancer types (Thompson et al. 1985, Palumbo et
al. 2000, Ruiter et al. 2001).
Tumor angiogenic and the fibrinolytic activators also share common proteolytic enzymes as
such as urokinase plasminogen activator (uPa), plasmin, thrombin (Shibuya 2011) and the
matrix metalloproteinases (MMPs) (Lelongt et al. 2001). For instance, uPa is a serine protease
that modulates the fibrinolytic system by converting plasminogen to biologically active
plasmin. In addition to fibrinolysis, by plasmin activation, uPA system is associated with a
degradation of the extracellular matrix in many cancers and elevated levels of uPA are
associated with poor prognosis in some cancer types (Cantero et al. 1997, Duffy et al. 1998 &
Shiomi et al. 2000).
It has been discovered that FIBβ 73-92 has anti-tumor properties in vivo and in vitro by
inhibiting tumor vascularization with anti-endothelial cell differentiation mechanism
(Krajewska et al. 2010). So far no direct link between the origin of fibrinogen beta fragments
and cancer have been established. It is still not clear whether the elevation of these peptides
are derived from cancer-derived fibrinogen or whether these peptides are produced as
physiological anti-tumor response.
5
2.2 Tumor markers
Cancer is a common cause of death in developed countries. Cancer diagnosis by simple
measurement from human tissue or body fluids has been an important goal in medicine for
many years. Before late 20th century, precise cancer diagnosis was only done by exploratory
surgical operations until the development imaging technologies such as computed
tomography (CT), magnetic resonance imaging (MRI) and sonography. Biomedical imaging
technologies have some disadvantages depending on the technology used. Namely these
limitations can be low cost- or time-effectiveness, exposure to harmful irradiation or
requirement of skilled personnel.
Tumor markers provide complementary information in clinical decision-making. They can be
any biomolecules that are present in tissue, blood or in other body fluids during at least some
stages of cancer progression. Tumor markers can be used in diagnosis/confirmation of
diagnosis, prognosis, disease monitoring or in determination of a suitable drug treatment.
Some of examples of the commonly used tumor markers are listed in table I.
Table I. List of tumor markers in common use. Modified from National Cancer Institute web page (2011).
Tumor marker Sample type Cancer type Use Alpha-fetoprotein (AFP) blood liver cancer and germ cell
tumors
diagnosis,
prognosis and
monitoring
Beta-2-microglobulin (B2M) blood, urine and
CSF
multiple myeloma, chronic
lymphocytic leukemia and
some lymphomas
prognosis and
monitoring
Beta-human chorionic
gonadotropin
blood and urine choriocarcinoma and
testicular cancer
prognosis and
monitoring
BCR-ABL fusion gene blood and bone
marrow
chronic myeloid leukemia diagnosis and
monitoring
BRAF mutation V600E tumor cutaneous melanoma and
colorectal cancer
determination of
target therapy
CA15-3/CA27.29 blood breast cancer monitoring
CA19-9 blood pancreatic cancer,
gallbladder cancer, bile duct
cancer and gastric cancer
monitoring
CA-125 blood ovarian cancer monitoring
Calcitonin blood medullary thyroid cancer monitoring
Carcinoembryonic antigen (CEA) blood colorectal cancer and breast
cancer
monitoring
CD20 blood non-Hodkin lymphoma determination of
target therapy
Chromogranin A (CgA) blood neuroendocrine tumors diagnosis and
monitoring
Chromosomes 3, 7, 17 and 9p21 urine bladder cancer monitoring
Fibrin/fibrinogen urine bladder cancer monitoring
6
Field of tumor marker discovery offers many tempting promises and inspiration to medicine
and health economics. Early detection of cancer significantly improves therapeutic success
rate (McPhail et al. 2015). This has been exploited by clinical screening procedures such
opportunistic prostate-specific antigen based (PSA) screening of healthy men, which have
ultimately led to detection of prostate cancer 5-10 years before arising of any clinical
symptoms (Stenman et al. 2005).
An ideal tumor marker should have following characteristics: i) high specificity and
sensitivity, ii) appearance at early stage of cancer, iii) concentration of the marker should
reflect the severity of a tumor, iv) practical assay for the detection of the marker can be
developed. In clinical practice such markers do not exist and currently used tumor markers
contain only some elements of the ideal tumor marker (Sharma 2009). Ultimately the purpose
for tumor markers is to improve therapeutic treatment and reduce healthcare costs.
The performance of a diagnostic tumor markers in practice can be evaluated with receiver
operating characteristic (ROC) using the true positive and false positive rates from research
datasets. These values are computed using non-parametric or parametric statistical methods.
In ROC analysis sensitivity versus 1-specificity plot is created and the area under the curve
(AUC) is calculated. ROC is useful for discriminating the diseased from healthy subjects.
Desired advantage with ROC analysis is that it is helpful in determining the cut-off values in
medical decision-making because it can show the trade-off between true positive fraction
(TPF) and false positive fraction (FPF) when sensitivity and specificity criteria are changed
(Hajian-Tilaki 2013).
Human epididymis protein 4 (HE4) blood ovarian cancer monitoring
HER2/neu tumor breast cancer, gastric cancer,
gastroesophageal junction
adenocarcinoma
determination of
target therapy
Immunoglobulins blood and urine Waldenström
macroglobulinemia and
multiple myeloma
diagnosis and
monitoring
Lactate dehydrogenase blood germ cell tumors prognosis and
monitoring
Nuclear matrix protein 22 urine bladder cancer monitoring
Prostate-specific antigen (PSA) blood prostate cancer diagnosis and
monitoring
Thyroglobulin blood thyroid cancer monitoring
7
As high-throughput techniques are being developed, increasing amount of metabolic and
proteomic tumor marker patterns for cancer diagnostics (i.e. detection of multiple compounds
simultaneously) are being reported (Petricoin et al. 2002, Villanueva et al. 2004, Asiago et al.
2010, Frantzi et al. 2014)
Although a many new promising tumor markers are being reported, only few of them are
practically applied. The most common bottlenecks are analytical and clinical validation or
weak practicability of the candidate marker (Drucker & Krapfenbauer 2013, de Gramont et al
2015).
2.2.1 Quantitative peptide tumor marker discovery
A tumor marker is not only defined by its presence but also by the change of its levels. A
major obstacle in tumor marker discovery is reliable quantitation. Accurate mass
spectrometric quantitation is usually based on comparison of ion signal intensities between
the deuterated reference standard and the analyte. In tumor marker discovery the
overwhelming amount of analytes of unknown identities makes the use of reference standards
undesirable. Finding appropriate reference standard is often costly and ordering can be time-
consuming especially if one has to be synthesized.
Isobaric tags for relative and absolute quantitation (iTRAQ) is a labelling-based quantification
method which can be applied in tumor marker discovery. In this method only groups e.g.
diseased versus healthy, since the all peptides in one sample are subjected to the labelling
agents. Despite this limitation Chen et al. (2010) discovered with this approach potential
bladder cancer markers from pooled urine samples. Tonack et al. (2013) used iTRAQ to
compare serum and pancreatic juice profiles between healthy subjects and patients with
pancreatic cancer. Their results correlated with the previous markers but also new potential
tumor markers were identified.
Measuring relative ion abundances is an alternative method to semi-quantify peptides without
using labels. The main advantage over the labelling method is that samples can be analyzed
natively without external manipulation since this approach is based on data processing.
Although this method is not suitable for absolute quantification, information about relative
changes can be often sufficient in finding a tumor marker. Several statistical software
8
specialized in biomarker discovery from metabolic or proteomic datasets have been developed
that are helpful in identification and visualization of candidate marker level changes (For
example, Progenesis QI for LC-MS marketed by Waters Corp. and ClinProTools for MALDI
by Bruker Daltonics Corp.). Due to previously mentioned limitations, tumor marker discovery
with both label-free counting of relative ion abundances and labelling based quantification
methods should be validated using either immunoassays or quantification using deuterated
internal standard (IS).
2.3 Renal cell carcinoma and prostate cancer
Renal cell carcinoma (RCC) and prostate cancer (PC) are lethal urological cancers that are
common cause of cancer-deaths (Siegel et al. 2014).
Transrectal ultrasonography (TRUS) guided biopsy, digital rectal examination (DRE) and
measuring PSA levels are the gold standard in diagnosis of prostate cancer (Arcangeli et al.
1997). Opportunistic PSA measurements are helpful in early detection of prostate cancer and
measurement of protein-bound fractions, optimization of cut-off values (Agnihotri et al. 2014)
and several calculation strategies have reduced the overtreatment and the false positive rate
(Stenman et al. 1991, Stenman 2005). PSA however is still not specific marker for prostate
cancer and despite high cut-off values, the false-positive rate is alarmingly high (Kilpeläinen
et al. 2010). Even a low false-positive rate will predispose many patients to psychological
stress and to expensive unnecessary diagnostic procedures.
Although some tumor markers for RCC have being reported (Frantzi et al. 2014, Chinello et
al. 2015), routine prognostic risk assessment in RCC currently relies on histopathological
cancer staging (Ficarra et al. 2010). Furthermore, the median 2-year survival rate for patients
diagnosed with metastatic RCC is 10-20% with median survival time of 6-12 months
(Flanigan et al. 2003).
2.4 Proteases and peptides in cancer
Proteases have a significant influence on tumor growth, invasion and formation of metastases
(Mason & Joyce 2011). Protease activities are strictly regulated and certain proteases can be
only found in certain tissues or cell types (Turk et al. 2012). Tumor cells produce proteases
9
and they also regulate the microenvironment to their favour by inducing neighbouring cells to
express proteases (Zucker et al. 2000). Proteolytic activity also stimulate intracellular
cascades of the neighbouring cells via protease-activated receptors (PARs) (Ramachandran et
al. 2012). PAR- signalling can mediate physiological events such as haemostasis via
thrombin-induced activation of PAR4 in platelets (Sambrano et al. 2001). Interestingly,
fibrinogen specific thrombin is known to activate PARs of endothelium cells and platelets to
induce production of vascular endothelial growth factor (VEGF) which is involved in vascular
growth, promotion of tumor angiogenesis (Shibuya 2011), decrease in leukocyte adhesion
(Troump et al. 2000) and increased endothelial permeability (Lal et al. 2001).
One approach to study protease activities is to analyse the end-products, peptides. Specific
peptide cleavage products may relate to cancer-specific activity. It is already known that PSA
in prostate cancer is cleaved by endopeptidases to produce a nicked form of PSA. By
determining the ratio of nicked-to-total PSA and total and free PSA, Steuber et al. 2002 were
able to discriminate more accurately malignant prostate tumor from benign.
In 2006 Villanueva and co-workers reported that specific exoprotease products or patterns
may reflect presence of a cancer or a specific cancer type. Li et al. 2014 incubated C3f
peptide in interstitial fluids of breast cancer and demonstrated resulting specific cleavage
products correlated with plasma samples of the patients with early stage breast cancer but not
healthy women. Low molecular weight peptides are therefore attractive for tumor marker
discovery. Glomerulus is known to filtrate small molecules to urine and therefore potential
tumor markers are naturally concentrated in urine (figure 2). We assume that such tumor
markers can be discovered more easily from cancers that reside in proximity to the urinary
tract.
10
Figure 2. Tumor protease-generated peptides in urine. Each urine sample contains different peptides due to
differential expression of proteases. Consequently there may be cancer-specific patterns ion intensity abundances
and in peptide identifications. Many peptides form “ladders” in which the derivative peptide gives rise to new
shorter peptides as a result of exopeptidase activity. Some of these peptides are found in urine during cancer as
indicated by a red boxes and lines in the hypothetical peptide presented in the figure
2.5 Mass spectrometry
2.5.1 Mass spectrometry in clinical laboratories
Immunological assays, such as enzyme-linked immunosorbent assays (ELISA) and
commercially available dipsticks and lateral flow devices are widely used in clinical
laboratories and in point-of-care applications. Immunological assays are rapid and
inexpensive. The main disadvantage of these techniques is that they rely on their antibody-
binding properties. Cross-reactivity with unknown epitopes in the matrix results in false-
positive findings. Furthermore, cross-reactivity makes analysis of similar compound groups
difficult and the level of cross-reactivity can vary between different manufacturers. Cross-
reactivity may for example result to false-positives in amphetamine or methamphetamine
measurements in people who have taken over-the-counter cold medicines containing
ephedrine (D’Nicuola et al. 1992).
11
Mass spectrometry is more reproducible, sensitive and specific alternative to immunological
methods and mass spectrometer can detect and quantify multiple unknown components in a
single analysis. Nowadays GC-MS and LC-MS/MS instruments can found in many clinical
laboratories. Modern clinical applications of mass spectrometry is applied in detection of
inborn errors of metabolism (Wilcken & Wiley 2008), toxicological screenings (Sundström et
al. 2013) and in steroid (Koal et al. 2012) and peptide analysis (Inoue et al. 2012).
2.5.2 Electrospray ionization and matrix-assisted laser desorption/ionization in peptide
analysis
Ionization of the analyzed compounds is required in mass spectrometric analysis.
Atmospheric pressure chemical ionization (APCI), electrospray ionization (ESI) and matrix-
assisted laser desorption/ionization (MALDI) are all “soft ionization” techniques that yield
protonated intact precursor ion since only small amount of fragmentation has occurred within
the peptide bond. On the contrary, “hard ionization” methods such as EI (electron ionization)
produce highly fragmented spectra without a precursor ion that are not useful in structural
elucidation of molecules with molecular weight > 400 Da.
In MALDI, the samples are co-crystallized with a sample matrix often by drying the sample-
matrix mixture on a MALDI target plate. Laser with at specific UV-wavelength is used to
induce desorption and ionization of the matrix components. Ionization of the analyte is
thought to occur in a gas-phase proton transfer reaction between the matrix and the analyte. A
common characteristics of a conventional MALDI-spectra is the predominant presence of
singly charged molecules. The exact model of the ion formation is still under a debate
(Knochenmuss 2006).
In ESI, a liquid-phase sample (often separated by HPLC) is infused into a thin capillary and
high voltage is applied to the electrospray cone. Charged aerosol droplets are sprayed from
the ionization source to the mass spectrometer. Heat-assisted evaporation of the solvent result
in a charge accumulation in a droplet containing the analytes until the Rayleigh limit is
reached. At Rayleigh limit the droplet dissociates as result of electrostatic repulsion created
by increased charge density in a droplet. After multiple Coulomb fissions the stream of ions
will be sent to a mass analyzer. Different models have been proposed as a mechanism of ion
formation from a charged droplet in ESI such as ion evaporation model, charged residue
12
model and chain ejection model. Konermann et al. 2013 suggested the real mechanism may
be a combination of these models depending on the molecular weight and structure of the
analytes.
Stapels & Barovsky (2004) compared LC-ESI-QTOF and off-line LC-MALDI TOF/TOF in
analysis of digested DNA-binding proteins. In that study they found that majority of the
proteins were identified by both techniques (131/253) with high confidence, but 37 were
specific to LC-ESI-QTOF while 85 were only discovered using LC-MALDI TOF/TOF.
Average peptide mass errors were higher with MALDI. They also concluded that comparing
the superiority of each technique over one another is difficult since nature and the
optimization of variables such as MALDI matrix, ESI conditions and differences in HPLC
conditions in these techniques is very different.
MALDI and ESI should be viewed rather as complementary techniques. Both ionization
techniques hold certain biases. ESI tends to favor aliphatic amino acids or amino acids with
hydroxyl groups, while MALDI favors peptides with basic or aromatic groups (Stepels &
Barovsky 2004). Highly acidic peptides produce a very poor ion yield with MALDI (Juhasz
& Biemann K 1994).
Matrix effect can affect the limit of detection (LOD) and quantitative performance in mass
spectrometric analysis. The effect is caused by the presence of matrix or other components in
solution such as endogenous compounds, metabolites, salts, detergents and buffers. Matrix
effect can either enhance or suppress the signal of the ion (Annesley 2003). All peptides also
have unequal ionization efficiency. Polar molecules seem to be more susceptible to ion
suppression (Bonfiglio et al. 1999) and high mass signals can suppress smaller mass signals
(Sterner et al. 2000). Different ionization methods are more susceptible to matrix effect than
the others and the mechanisms of ion suppression or enhancement can be different. In ESI,
change in droplet properties may result to ion suppression (Annesley 2003). MALDI is more
tolerant to salt (Xu et al. 2006) than ESI, but is less fit for comparative quantitative analysis
due to low and variable intraexperiment reproducibility of its individual peaks (Albrethsen
2007) , which may be a result of a competitive ionization (Cohen & Chait 1996).
13
2.5.3 Mass analyzers
Ions that are generated in the ion source are separated in the mass analyser based on the mass
to charge ration (m/z). Separation is done based on different principles in different mass
analyzers, but the separation is commonly achieved by manipulation of RF voltages or
electric and magnetic fields. Most commonly used mass analyzers in proteomic analyses are
fourier transform ion cyclotron resonance (FTICR), ion trap (IT), orbitrap, quadrupole (Q)
and time-of-flight (TOF) mass analyzers. Hybrid mass analyzers consist of two or more
different mass analyzer units. A more comprehensive overview about different analyzers and
soft ionization methods in context of proteomic analysis is written by Mann et al. 2001. More
recently developed orbitrap is reviewed by Scigelova & Makarov 2006.
Choice of a mass spectrometer is often determined by the instrumentation costs and by the
performance that is required. Mass resolution, analysis time, sensitivity, hyphenation
compatibility, quantitation and MS/MS performance, reproducibility and dynamic range are
all factors to be considered.
Common hybrid mass analyzers include linear trap quadrupole (LTQ), quadrupole ion trap
(QIT) and quadrupole time-of-flight (QTOF). Hybrid mass analyzers are especially well
suited for MS/MS analysis mainly because they are combining the advantages and
overcoming the disadvantages in both analyzers. In LC-QTOF high mass accuracy and
sensitivity can be reached with relatively fast spectral acquisition by selecting ions first with
sensitive quadrupole and further separating ions with rapid and accurate mass resolution TOF
analyzer with high dynamic range (Stolker et al. 2004). These characteristics make LC-QTOF
useful in high throughput clinical screening and in identification of unknowns based on either
mass accuracy (Wolff et al. 2001) and/or retention time or spectral library search (Lee et al.
2015).
In conventional ESI-QTOF MS/MS, ions from ESI source are guided in the first quadrupole
(Q1) and filtered by alternating the magnitude of the RF amplitude and DC potentials in the
electrodes at fixed ratio. Specific ratios of RF to DC facilitate the trajectory of ions with m/z
to reach the collision cell (Q2). Ion with specific m/z can be targeted or the whole range of
predetermined m/z window can be scanned (scan mode) if the potentials are swept from
minimum value to maximum value or vice versa. In Q2 peptides are fragmented by collision
14
induced dissociation (CID) in which the accelerated ions are collided with inert gas molecules
such as helium. Collision may result to but is not limited in fragmentation of peptide bonds
(these fragments are called b and y- ions depending on which terminus of the peptide had the
fragmentation occurred and therefore where the charge has retained). Fragmented ions (also
known as product or daughter ions) are further separated in a flight tube of a TOF analyzer, in
which the products ions in electric field are directed to detector. Ions with low m/z reach the
detector first. The time of an ion reaching the detector is measured and the each time point
correlates with specific m/z value.
2.5.4 Data collection in peptide tandem mass spectrometry: Data-dependent and data-
independent acquisition
Conventionally MS/MS mode is operated using data dependent acquisition (DDA). In DDA
fixed number of precursor ions are selected using specific rules that are set by the algorithms
in mass spectrometer. With this approach, peptide signals with high signal-to-noise ratio are
matched to the spectra in a database. Peptides that are sampled by a mass spectrometer are
more prone to pick up stronger signals and therefore DDA method is not as reproducible in
analysis of low-abundant peptides (Geromanos et al. 2009, Doerr 2015).
MSE is a data independent acquisition (DIA) method used typically with LC. It is also known
as “all ions” or “broadband CID” (bbCID) depending of the mass spectrometer vendor.
Unlike DDA, DIA does not apply ion transmission time in Q1 to maximize duty cycle. In
DIA the collision energies are rapidly alternated between high and low collision energies in a
single analysis to obtain simultaneously full precursor and product ion spectra for the
matching retention time point of the LC separation (figure 3). Advantage in such parallel
fragmentation is that all peptides are simultaneously analyzed regardless of the ion intensities
(Geromanos et al. 2009).
15
Figure 3. A Principle of data-independent (MSE) acquisition in QTOF. The first quadrupole (Q1) is
operating in scanning mode sending all ions to the second quadrupole (Q2), which acts as a collision cell. In Q2,
ions are scanned in at subsequent points in low energy and high energy fragmentation. In low energy scans
eluting precursor ions are detected after TOF separation, while in high energy scans corresponding product ions
are detected. Post-processing software will then link precursor ions with their corresponding product ions based
on the same retention time.
DIA is less laborious in non-target analysis than DDA since the selection and identification of
an unknown precursor ion from MS data set is automatically done in a single LC-MSE run.
Many articles of improved protein and peptide coverages have been reported with DIA was
compared to DDA. Blackburn et al. 2010 compared DDA and DIA (LC-MSE) in tomato leaf
proteome. With DDA 234 peptide matches and 162 protein identifications were, whereas
using DIA 1492 and 576 matches respectively were made with lower false positive rate.
Similar tendencies have been reported by Distler et al. 2013 and Tsou et al. 2015.
Recent developments in DIA involve use multiple isolation windows together with DIA
(SWATH-MS) to further increase the method sensitivity and consistency (Gillet et al. 2012).
Ion mobility separation (IMS)-enhanced MSE (HDMSE) and collision energy-optimized
UDMSE) can improve the alignment of precursor ion to product ion and thus increase the
method sensitivity. IMS can provide another dimension of separation as drift time and
retention time are both used for the alignment (Distler et al. 2013).
2.5.5 Discovery and analysis of low-abundant peptides
Analysis of low-abundant peptides near limit of detection (LOD) is challenging. Two of the
main techniques that are applied in tumor marker discovery are nuclear magnetic resonance
16
(NMR) spectroscopy and mass spectrometry (MS). Although NMR spectroscopy is superior
to MS in terms of quantitation (peak area in NMR is directly related to the nuclei
concentration of specific nuclei such as 1H or 13C), it is not sensitive enough for discovery of
low-abundant molecules. With mass spectrometry, even attomole levels of peptides have been
reported (Moyer et al. 2003). While the high sensitivity and resolution of the mass
spectrometric instruments are required, sample preparation is also critical in detection of low
abundant peptides.
Circulating carrier proteins such as albumin are not filtrated in urine and thusly spend more
time in plasma than the filtrated peptides (Hortin & Sviridov 2007). Non-covalent binding of
the low-molecular weight molecules to circulating carrier proteins have been long known and
such well characterized examples of these binding proteins include human serum albumin
(Vallner 1979), sex hormone-binding globulin (Siiteri et al. 1982), thyroxine-binding globulin
(Hoffenberg & Ramsden 1983) and previously mentioned alpha 1-antichymotrypsin
(Stenman et al. 1991). Circulating carrier proteins may therefore be a potential source of
tumor markers (Mehta et al. 2003-2004) and therefore the sample preparation could be also
focused on high molecular weight fraction.
Simultaneous reduction of matrix effect (i.e. removal of undesired components) and
concentration of target analytes are the two main objectives in the sample preparation. In
peptide analysis, commonly employed strategies include centrifugation, high-pressure liquid
chromatography (HPLC), reversed phase chromatography (RPC), hydrophilic interaction
chromatography (HILIC), ion exchange chromatography (IEX) size-exclusion
chromatography (SEC) and affinity chromatography (AC). Peptide extraction and mass
spectrometric analysis can be directly coupled to provide faster analysis and more
reproducible results. In LC-MS, MS is coupled with liquid-liquid extraction (LLE) and in
surface-enhanced laser desorption/ionization (SELDI) solid phase extraction (SPE) is coupled
to TOF-MS. Sample treatment and preanalytical variables must be handled with care in order
to avoid loss of target analytes, production of artefacts and unreproducible results.
Developments in sample preparation, resin composition, automation and mass spectrometry
will further facilitate the discovery of tumor markers. Emerging chromatographic products
will focus more on reducing the matrix effect producers such as phospholipids (Ahmad et al.
2012).
17
II EXPERIMENTAL PART
3 Aim of the project
The main objective of this project was to extend the amount of data related to biology of
cancer specific proteases. Due to glomerular filtration of peptides <10kDa, we believe that
differential proteolysis products in cancer are enriched in urine and that urine may serve as
promising medium for early prognostic marker of RCC or PC. The potential of fibrinogen
beta as a biomarker for prostate cancer was earlier linked in unpublished study from our
group. Mass spectrometry is the method of choice in this study.
The specific aims of the project were:
To develop a suitable workflow for enrichment of different unknown products of the
fibrinogen beta peptides in urine specimen.
To overcome challenges associated in identification of low-abundant new peptide
tumor markers using different mass spectrometric techniques.
To find suitable marker or peptide ladder profile for detection of early RCC or PC.
18
4 Materials and Methods
4.1 Materials
4.1.1 Urine samples
Urine samples from controls and patients with RCC and PC were collected at the Department
Urology, Helsinki University Central Hospital (Helsinki, Finland). Controls had comparable
age and gender to patients. Samples were collected during the period of 2004-2010 and mass
spectrometric analyses were performed in 2013-2014.
First morning urine samples were collected from patients prior to any treatment in plastic
containers and 50 mL was transferred to 50 mL Falcon tubes. Urine samples were
centrifugated and cells were removed prior storing at -80 ºC. After thawing, urine was
centrifugated at 1500 g for 10 min and pellet was removed. Specific gravity (S.G.) was
measured with a digital urine refractometer (UG-1, Atago, Honcho, Japan). Samples with
abnormal S.G. values (<1.002) or > 1.030) were not used in experiments. Aliquoted samples
were stored at -20 ºC without protease inhibitors. The study was approved by the institutional
review board (16/E6/2004) and informed consent was obtained from all patients. Sample
handling after collection included 2-3 freeze-thaw cycles.
The donor of the prostate cancer urine PC8 sample used in the most identifications in this
study was classed as T4 Nx M0 in TNM tumor staging system.
4.1.2 Chemicals and reagents
Methanol, ethanol, formic acid (FA), trifluoroacetic acid (TFA) and acetonitrile (ACN) were
HPLC grade (Sigma Aldrich). Ultrapure water was obtained with a Milli-Q purification
system (Millipore, Bedford, MA, USA).
Synthetic peptide mimicking fibrinogen β (50-71) sequence, 2264.19 Da (monoisotopic
mass), with a deletion of glycine residue at position 70 was purchased from Davids
19
Biotechnologie, Regensburg, Germany (DKKREEAPSLRPAPPPISGGY) and stored at -20
ºC.
Polyclonal DKKREEAPSLRPAPPPISGGGY- specific rabbit anti-human IgG antibody was
purchased from Agilent Technologies. Eu-labelling kit, streptavidin-coated 96-well plates,
assay buffer (50 mM Tris-HCL, pH 7.8 containing per liter 9 g NaCl, 5g bovine serum
albumin (BSA), 0.5 g bovine gamma globulin, 0.1 g tween 40 and
diethylenetriaminepentaacetic acid (DTPA) and 0.5g NaN3), wash solution, and enhancement
solution used in time-resolved immunofluorometric assays (TR-IFMAs) were from
PerkinElmer Life Sciences, Wallac Oy, Turku, Finland.
4.2 Biotinylation and Eu- Labelling
DKKREEAPSLRPAPPPISGGGY-specific anti-peptide pAb was biotinylated using Pierce
Biotechnology EZ-Link Sulfo-NHS-LC-Biotin (sulfosuccinimidyl-6-[biotinamido]hexonate)
reagent (Thermo Scientific, Rockford, IL, USA) according to manufacturer’s protocol. 20-
fold molar excess of biotin was used to label the anti-fibrinogen peptide.
DKKREEAPSLRPAPPPISGGGY peptide was labelled with N1–(p-isothiocyanatobenzyl)-
diethylenetriamine-N1,N2,N3,N3-tetraacetate (DTTA) chelated to Eu(III) using Delfia Eu-
labelling kit (reference number 1244-302, PerkinElmer). Labelling procedures were done
according to manufacturer’s description in 40-fold molar excess of europium chelate.
4.3 Competitive time-resolved immunofluorometric assay
The competitive TR-IFMA for quantification of DKKREEAPSLRPAPPPISGGGY peptide
(method principle illustrated in figure 4) was done using following procedure: (a) Addition of
biotinylated DKKREEAPSLRPAPPPISGGGY- peptide (stock solution 1.5 µg/mL) diluted
with assay buffer to 0.007 µg/mL (100 µl/well) in streptavidin-coated 96-well plates,
incubation in a shaker for 60 minutes, wash two times. (b) Addition of 100 µl of urine sample
or peptide standard (c) immediately followed by addition of 100 µl of Eu-labelled anti-
DKKREEAPSLRPAPPPISGGGY polyclonal antibody (1 µg/mL), incubation in a shaker for
60 minutes, wash four times. 200 µl of Delfia® Enhancement solution was added and the
20
fluorescence of Eu3+ chelates (613 nm) was quantified using Victor2 V 1420 multilabel HTS
counter (PerkinElmer Life Sciences). The assay range was 0.03-7.3 µg/mL.
Figure 4. Principle of competitive time-resolved immunofluorometric assay. A) Biotin-conjugated peptide
binds to streptavidin-coated 96-well plate via formation of biotin-streptavidin complex. B) Clinical sample
provides competitive target peptides. C) Europium (III)-labelled pAbs bind to the target peptides that are present.
Peptides in the sample compete for the pAb binding sites with the biotin-streptavidin-linked peptides. After
washing step, europium fluorescence emission is measured only from a well-bound biotin-streptavidin-peptide-
pAb complexes. Thus the europium fluorescence is inversely correlated to the peptide concentration.
4.4 Solid phase extraction
Resource S Cation exchange column, Superdex 200 10/300 GL column and Superdex Peptide
HR 10/30 column (GE Healthcare Biosciences, Uppsala, Sweden) were connected with
ÄKTApurifier FPLC system (GE Healthcare Biosciences). All columns were equilibrated
according to manufacturer protocol with 20% ethanol or start buffer with a flow rate 0.7
mL/min. 0.5 mL fractions were collected and absorbance at 214 nm, 218 nm and 280 nm were
simultaneously monitored. System was controlled by UNICORN software v.4.00 (GE
Healthcare Biosciences).
1.5×5 cm Sephadex® G-25M PD-10 gel filtration columns were purchased from GE
Healthcare, Buckimhamshire, UK. Samples were handled according to manufacturer’s
21
instructions applying gravity protocol using 2.5 mL as sample volume. 100 µl aliquots of
fractions 2 and 3 were combined prior ZipTip C18 extraction.
All procedures from following products were done according to manufacturer’s instructions
with subsequently described exceptions. ZipTip C18 desalting and purifications were done
using ZipTip C18 reversed phase resins (Millipore, Billerica, MA, USA). C18 purified samples
were eluted with 3-10 µl of 40% ACN in 0.1% trifluoroacetic acid (TFA) prior to MALDI-
TOF and in 3 µl of 40% ACN in 0.1% formic acid (FA) prior to ESI-QTOF experiments. C18
Spin Columns were purchased from Thermo Scientific (Rockford, IL, USA) and the samples
were eluted to 40 µl of 40% ACN in 0.1% FA. Oasis HLB (hydrophilic-lipophilic balance)
and Oasis MCX (mixed-mode cation exchange) SPE cartridges and were obtained from
Waters Corporation (Milford, MA, USA). All Oasis® solid phase extracted samples were
diluted with formic acid (FA) in 2:1 sample/0.2% FA ratio before sample extraction. Peptide
extractions using Oasis® SPE cartridges were performed using Alltech vacuum manifold
(Alltech). Modified sample volumes and elution conditions for all sample preparation
experiments are listed in table II.
Table II. Summary of sample preparation conditions tested.
SPE method
SPE typea
Sample
volume (µl)b
Elution conditions
Optimal SEC
fractionation
range (Da)
Oasis HLB
Pierce C18 Spin Column
ZipTip C18
Oasis MCX
Resource S
Superdex Peptide HR
10/30
Superdex 200 10/300 GL
PD-10 Desalting Column
RP
RP
RP
RP+CEX
CEX
SEC
SEC
SEC
660
100-300b
200
660
1000
500-1000
500
2500
70%ACN, 0.1% FA
40%ACN, 0.1% FA
40%ACN, 0.1% (T)FAc
5% NH4OH in MeOH
1M NaCl
5% ACN, 0.1% TFA or
TBS, pH 7.5
TBS, pH 7.5
Same as sample
conditions
100-7000
10 000-600 000
>5000
a RP: reversed phase chromatography, CEX: cation exchange chromatography, SEC: size-exclusion
chromatography b Depending on the available sample volume. Sample binding step was done twice in volumes over 150 µl. c Trifluoroacetic acid (TFA) was used in MALDI-TOF experiments and formic acid (FA) was used in LC-QTOF
experiments.
22
4.5 Mass spectrometry
4.5.1 Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry
MALDI-TOF MS and MALDI-TOF/TOF MS/MS measurements were performed with
Bruker UltrafleXtreme 2000 Hz instrument (Bruker Daltonics Corporation, Bremen,
Germany) equipped with a SmartBeam II laser (355nm). 0.5 µl of extracted urine sample was
spotted on an AnchorChipTM target plate (Bruker Daltonics), dried and overlaid with 0.5 µl of
CHCA matrix solution (Sigma-Aldrich, USA) and allowed to dry prior to insertion into mass
spectrometer. All analyses were operated in positive and reflector modes in a mass range of
700-4000 Da. Spectra were acquired by accumulating spectra of 10000 laser shots for MS and
30000 for MS/MS analysis. External calibration was performed using Bruker Peptide
Calibration Standard (#206195 Peptide Calibration Standard, Bruker Daltonics). Smoothing
and baseline corrections of the MALDI-TOF peptide mass spectra were processed using
FlexAnalysis v3.4 (Bruker Daltonics). Protein database search was done with Biotools v3.2
software (Bruker Daltonics) using University of Helsinki Mascot server (Matrix Science Ltd.,
UK) as search engine. Peptide identifications searches were done against UniProt Human
reference database (dated 16.4.2014 to 7.5.2014) with the following settings: no enzyme,
precursor ion mass tolerance 0.1 Da, fragment ion mass tolerance 0.5 Da, no modifications
was selected as fixed modification box and oxidation of methionine as variable modification.
Identification threshold was set as p<0.05.
4.5.2 Ultra-performance liquid chromatography quadrupole time-of-flight mass
spectrometry
Ultra-performance liquid chromatography (UPLC)-MSE analysis was conducted on a Waters
nanoACQUITY UPLC connected to a Waters Synapt G2-S HDMS quadrupole time-of-flight
(QTOF) mass spectrometer (Waters, Milford, MA, USA). All analyses were performed in
positive- mode ESI with nano flow, flow rate 300 nL/min and a 0.5 s scan time. Calibration
was performed using sodium formate. UPLC system was equipped with a nanoAQUITY
UPLC Symmetry C18, 5µm, 180 µm x 20 mm trap column (Waters) and BEH130 C18, 1.7
µm, 75 µm x 150 mm analytical reversed-phase column (Waters). 1-4 µl of each sample was
injected depending on the ion intensities in prior screening in LC-MS analysis of equal
23
injections. The UPLC was operated under the following conditions: Solvent A: H2O + 0.1%
formic acid, Solvent B: ACN + 0.1% formic acid. Gradient: 0-1 min 1% B, 1-2 min 8% B, 2-
30 min 30% B, 30-31 min 50% B, 31-32 min 85% B, 32-36 min 85%B, 36-37 min 1% B.
Total run time was 45 min. Mass spectrometer was operated in MSE mode in sensitivity mode
with mass range 100-2500 m/z for precursors and 100-2500 m/z for fragments. Fragmentation
data was collected for one second scan time at low energy using 4 V and for high energy
scans collision energy was ramped between 14 and 42 V. 1 ng/µl leucine encephalin peptide
was infused for lock mass calibration.
Raw data was imported to Waters PLGS 3.0 software (Waters) and processed using low
energy ion intensity threshold of 100 counts and high energy ion intensity threshold of 25
counts. Peptide identifications were searched against UniProt Human reference database
(dated 21.2.2014 to 17.6.2014) with following settings: Precursor ion mass tolerance 1 Da,
fragment ion mass tolerance 5 ppm, digest reagent was set to trypsin or no enzyme with 0-2
allowed missed cleavages. Carbamidomethylated cysteine or no fixed modifications was set
as fixed modification depending from a search and oxidation of methionine as variable
modification. Identification threshold was set as p<0.05. LC-MS data was converted into 2D
profile data using Progenesis LC-MS software (Nonlinear Dynamics, version 2.4). The runs
were time aligned automatically using manually chosen reference. Samples with different
extraction treatments were aligned separately. The aligned peaks were picked from stacked
data files. Peaks with charge state from + 1 to + 5 were chosen for analysis and were
normalized by Progenesis LC-MS.
4.5.3 Data in table V
Targeted LC-MS analysis peptide identification information received from Suvi Ravela
(University of Helsinki) was adapted to table V. This is indicated as DDA QTOF in the table.
LC-MS/MS was done using QTOF Ultima Global (Waters, Manchester, UK) and LTQ
Orbitrap XL (Thermo Electron, Bremen, Germany). Conditions in this analysis were briefly
following:
QTOF Ultima Global. 10µl of extracted peptide sample was injected onto a RP-HPLC
(CapLC, Waters, Manchester, UK) with a C18 trap column (Atlantis dC18, NanoEase Trap
24
Column, 5µm, Waters) and eluted with a linear gradient of CAN (from 5 to 50%% in 30 min)
in 0.1% FA at a flow rate of 0.3 µl /min. The MS was calibrated using 2 pmol/µl
glufibrinogenic peptide B fragments and the LC-MS repeatability was validated using a
commercial standard BSA-digest (Waters).
LTQ Orbitrap XL. 5µl of the sample was injected onto a nano RP-HPLC (nanoLC Ultimate
3000, Dionex, Sunnyvale, CA, USA) with a C18 trap column (PepMap C18, 5 µm, 100Å,
Dionex) and a 0.075 x 150 mm C18 analytical column (PepMap C18, 3 µm, 100 Å, Dionex)
at a flow rate of 0.3 µl/min. The peptides were eluted with a binary gradient of CAN; from 0
to 20% in 120 min and from 20 to 40% in 60 min. MS/MS spectra were obtained by CID and
detected in the linear IT. Dynamic exclusion was used, with a repeat count of one; exclusion
duration was set at 3 min and exclusion width at ±5 ppm.
4.5.4 Data analysis
Paws open-source software (ProteoMetrics, LCC) was used to analyze peak lists from
MALDI-TOF-MS spectra. N-terminal FIBβ sequences were manually searched among
collected peak lists that were exported as Microsoft Excel files.
Theoretical precursor ion masses were calculated using ExPASy PeptideMass tool
(http://web.expasy.org/peptide_mass/) and theoretical singly charged fragment ions were
calculated using Fragment ion calculator from Institute for systems biology
(http://db.systemsbiology.net:8080/proteomicsToolkit/FragIonServlet.html)
Data analysis with all software were performed in 7.11.2013-7.5.2014.
4.5.5 Overview of the method development strategy
Intensity and signal-to-noise ratio values from MALDI-TOF spectra were used to measure
peptide extraction efficiencies. To obtain high FIBβ peptide yields, several SPE sorbents and
size-exclusion chromatography methods were tested. Each method was optimized using
synthetic N-terminal FIBβ standard (2264.19 Da). One example of the method optimization
was determination of proper elution conditions as demonstrated in figure 4.
25
Figure 4. Optimization of acetonitrile (ACN) conditions for C18 Spin columns. ACN was used as an eluent
in C18 Spin column protocol. 1µ𝑔 of internal FIBβ standard (2264.19 Da) was spiked in urine prior peptide
extraction. Average (n=2) post-extraction intensities from MALDI-TOF were compared at each ACN
concentration.
Sample preparation method or combination of two methods with a high-intensity yield of
standard peptide was chosen for the subsequent peptide sequencing analyses. Retention times
and accurate masses of all identified peptides were used to evaluate MS spectra of other
samples or experiments with different sample preparation protocol. In LC-analysis, the most
intense charge state of the peptide was used in comparative analyses. The workflow used is
summarized in figure 5.
0
50 000
100 000
150 000
200 000
250 000
300 000
350 000
400 000
450 000
500 000
40% ACN 50% ACN 60% ACN 70% ACN
Intensity
FIBβ standard intensity at different ACN concentrations
26
Figure 5. Outlines of the sample processing strategy carried out in this study. MS results from MALDI-TOF
are used as quality-check during the workflow. Identification is facilitated when sample with high marker
concentration is found and optimized pre-analytics are established. Subsequently, retention time and accurate
mass values of successfully identified peptides were used to find the same peptides in MS spectra.
27
5 Results
5.1 Evaluation of extraction methods in enrichment of urinary fibrinogen β peptides
We evaluated different sample preparation methods by comparing the intensities of previously
identified peptides. The most prominent peptide extraction methods are compared in tables III
(MALDI-TOF) and IV (LC-ESI-QTOF). Highest intensity yields in MALDI-TOF were
received with SP+SC (Superdex peptide and C18 spincolumn) workflow while the generic PD-
10 + C18 ZipTip approach was not suitable due poor (co)crystallization process.
Conclusively, using generic method, wider range of FIBβ peptides were detected with higher
intensity while the use of size-exclusion chromatography to exclusively beneficial to some
peptides. Among the chromatographic methods, Superdex peptide column was superior over
the other methods.
Table III. Comparison of different sample preparation method efficiencies of fibrinogen β peptides in
MALDI-TOF spectra.
Calculated FIBβ mass (Da)a
Peak intensityb / Signal-to-noise ratio (in gray color)
S200 + SC RS MCX SP + SC
1031.56
1107.53
1464.77
2078.09
2234.19
2321.21
2367.16
2421.19
2594.34
2894.47
55
-
78
17
26
-
16
-
11
-
5510
5
31
30
22
19
54
14
5
31
21
-
9
-
-
-
115
26
5
44
33
14
71
-
-
13
6
-
26
20
-
-
126
14
7
34
26
-
-
278
156
206
10
111
-
29
10
376
104
152
9
53
85
19 a Mass error window of 10 ppm was allowed. b Signal intensities are 1000- fold. Signal-to-noise ratio threshold was set as 6.
28
Table IV. Comparison of FIBβ intensity yields in LC-ESI-QTOF (Synapt G2S) using different enrichment
strategies. Size-exclusion intensities examined in relation to generic extraction with 10 kDa size exclusion filter
and C18 resin (PD-10 + C18 ZipTip).
Calculated Mass
(Da)
Retention
time (min)
PD-10 +
ZT
Fold change of intensity compared to PD10 + ZT
Superdex peptide + C18
spin column
Superdex 200 +
C18 spin column
1031.56
1107.53
1464.77
1620.87
1630.83
1722.79
1767.78
1793.90
2015.92
2022.94
2078.09
2272.07
2234.19
2321.21
2367.16
2421.19
2459.32
2594.34
2608.25
2835.42
2894.47
21.83
11.71
12.59
12.26
19.11
28.24
22.23
22.59
11.19
24.20
17.91
19.79
15.49
23.20
19.03
28.48
32.85
19.34
24.01
23.01
19.58
1811
180
48
3
7
101
16
26
14
631
167
35
568
11
1570
69
11
230
36
22
36
-
-
1.9
4.0
-
3.9
3.3
-1.4
-
2.7
6.0
7.3
8.0
3.1
-1.3
5.4
-5.5
1.7
-7.2
1.9
9.4
-5.0
-4.1
-
-
-
-
-
-
-
4.3
-1.3
-1.6
1.3
-
-26.2
-
1.6
-
-
3.1
-
Only samples with similar charge states, measured accurate mass difference of 0.02 Da and retention time
difference of 0-2.5 minutes were regarded as same peptide.
5.2 Peptide sequencing
Presentation of the tandem mass results was done using the guidelines suggested by Carr and
co-workers (2004) in the Molecular & Cellular Proteomics journal with following exceptions:
a) Peptide charge states are not presented. b) Specific search engine score values in
identification, are not presented. c) Some specific values of observed masses and mass errors
are not reported.
MS and MS/MS spectra of positive FIBβ identifications contained expected b and y ions as
presented in figure 6 using FIBβ 354-374 as example. MS/MS spectra from each method used
are included.
29
A
B C
D
30
E
Figure 6. De novo identification of AHYGGFTVQNEANKYQISVNK (FIBβ 354-374, M= 2367.16 Da)
peptide using different ionization and data acquisition methods.
A) MALDI-TOF spectrum. B) Sequence ions obtained from fragment ion calculator provided by Institute for
systems biology. C) LIFT-TOF/TOF tandem mass spectrum. D) Tandem mass spectrum using data dependent
acquisition (DDA) with QTOF Ultima Global. Original data of this figure could not been obtained. E) Fragment
ions detected in LC-MS/MS (QTOF) with MSE data acquisition. In MSE, charge states from 2-4 with average
precursor mass of 2368.178 (M+H) were observed and fragmented. Sample in MALDI/TOF/TOF was prepared
using Superdex peptide column extraction followed by C18 Spin column. In LC-MS/MS and MSE, PD10 gel
filtration followed by C18 ZipTip was used.
In total, we have successfully identified 28 different FIBβ peptides in urine (table V), of
which 26 were not mentioned in the literature. Positive identification results were obtained
using following sample preparation strategies: i) PD-10 followed by C18 ZipTip desalting, ii)
Superdex peptide followed by C18 Spin column or iii) Resource S column.
31
Table V. List of fibrinogen β chain products successfully identified from urine samples of RCC (n=3) and
PC (n=6) patients.
Peptide sequencea Positionb Mass(Da)c rt
(min)d
Sample
preparatione
Data acquisition
method
K.SQGVNDNEEGFFSARGHRPLD
K.K
S.QGVNDNEEGFFSARGHRP.L
L.DKKREEAPSLRPAPPPISGGGY
.R
D.KREEAPSLRPAPPPISGGGY.R
D.KREEAPSLRPAPPPISGGGYR.A
R.EEAPSLRPAPPPISGGG.Y
R.EEAPSLRPAPPPISGGGY.R
E.EAPSLRPAPPPISGGGY.R
A.PSLRPAPPPISGG.G
A.PSLRPAPPPISGGG.Y
A.PSLRPAPPPISGGGY.R
A.PSLRPAPPPISGGGYR.A
K.DLWQK.R
K.DLWQKR.Q
K.DNENVVNEYSSELEK.H
K.GGETSEMYLIQPDSSVKPYR.V
R.QDGSVDFGRK.W
K.ISQLTRMGPTELLIEMEDWK.G
I.SQLTRMGPTELLIEMEDWKGD
K.V
I.SQLTRMGPTELLIEMEDWKGD
KVK.A
R.MGPTELLIEMEDWK.G
R.MGPTELLIEMEDWKGDK.V
K.GDKVKAHYGGFTVQNEANKY
QISVNK.Y
K.VKAHYGGFTVQNEANKYQISV
NK.Y
K.VKAHYGGFTVQNEANKYQISV
NKYR.G
K.AHYGGFTVQNEANKYQISVNK
.Y
K.YQISVNK.Y
K.IRPFFPQQ-
30-51
31-48
50-71
52-71
52-72
54-70
54-71
55-71
57-69
57-70
57-71
57-72
153-157
153-158
164-178
248-267
286-295
329-348
330-351
330-353
335-348
335-351
349-374
352-374
352-376
354-374
368-374
484-491
2459.16
2015.95
2321.22
2078.09
2234.21
1630.83
1793.91
1664.85
1244.69
1301.71
1464.72
1620.84
688.34
844.44
1767.78
2272.07
1107.53
2421.19
2608.25
2835.42
1722.79
2022.94
2894.49
2594.34
2913.51
2367.16
850.45
1031.56
32.85
11.19
23.20
17.91
15.49
19.11
22.59
12.14
11.16
11.26
12.26
14.19
14.61
12.88
22.23
19.79
11.71
28.48
24.01
23.01
28.24
24.20
19.58
19.34
-
19.03
12.80
21.83
PD10+ZT
PD10+ZT
PD10+ZT
PD10+ZT,
SP+SC, RS
PD10+ZT,
SP+SC, RS
PD10+ZT
PD10+ZT
PD10+ZT
PD10+ZT
PD10+ZT
PD10+ZT
PD10+ZT
PD10+ZT
PD10+ZT
PD10+ZT
PD10+ZT
PD10+ZT
PD10+ZT
PD10+ZT,
SP+SC
SP+SC
PD10+ZT
PD10+ZT
SP+SC
SP+SC
SP+ZT
PD10+ZT,
SP+SC
PD10+ZT
PD10+ZT
DDA (QTOF)
DDA (QTOF)
DDA (QTOF),
DDA (TOF)
DDA (QTOF),
DDA (TOF)
DDA (QTOF),
DDA (TOF)
DDA (QTOF)
DDA (QTOF)
DDA (QTOF)
DDA (QTOF)
DDA (QTOF)
DDA (QTOF)
DDA (QTOF)
MSE
MSE
MSE
MSE
MSE
MSE
MSE
MSE
MSE
MSE
MSE
MSE
DDA (TOF)
MSE, DDA
(QTOF), DDA
(TOF)
MSE
MSE, DDA
(QTOF) a Underlined M indicates oxidized methionine.
b Amino acid position according to UniProt knowledge base numbering of fibrinogen beta chain
(entry number P02675).
c Measured masses shown. Mass error window of 0.03 Da was allowed between calculated exact mass and
measured accurate mass.
d Retention times are based on MSE based identifications or precursor ion observations in QTOF Synapt G2S.
e PD10= PD10 Desalting column, ZT= C18 ZipTip, SP= Superdex Peptide HR 10/30 column, SC= C18 Spin
column.
32
Although the focus of our study was fibrinogen-derived peptides, some other peptides were
identified as well due to non-target nature of MSE mode. All identifications made are listed in
table VI.
Table VI. List of non-FIBβ identifications acquired with MSE data acquisition method from urine
samples of RCC (n=3) and PC (n=6) patients.
Peptide sequence Position Mass (Da) rt
(min)
Sample Sample
preparation
α-2-HS-glycoprotein (Fetuin-A)
(FETUA_HUMAN)
I.DYINQNLPWGYK.H
Apolipoprotein A-I
(APOA1_HUMAN)
R.ARAHVDALRTHLAPYSDELRQRLA
A.R
R.AHVDALRTHLAPYSDELRQRLAA.R
K.VSFLSALEEYTK.K
K.VSFLSALEEYTKKLNTQ-
C-reactive protein
(CRP_HUMAN)
K.AFVFPKESDTSYVSLK.A
R.ALKYEVQGEVFTKPQLWP-
Fibrinogen α chain (FIBA_HUMAN)
T.ADSGEGDFLAEGGGV.R
T.ADSGEGDFLAEGGGVR.G
T.ADSGEGDFLAEGGGVRGPR.V
A.DSGEGDFL.A
A.DSGEGDFLAEGGGV.R
A.DSGEGDFLAEGGGVR.G
D.SGEGDFL.A
A.SGEGDFLAEGGGV.R
A.SGEGDFLAEGGGVR.G
S.GEGDFLAEGGGV.R
S.GEGDFLAEGGGVR.G
G.EGDFLAEGGGV.R
G.EGDFLAEGGGVR.G
E.GDFLAEGGGVR.G
D.FLAEGGGVR.G
F.LAEGGGVR.G
F.SSANNRDNTYNRVSEDLRSRIEVLK
RKVIEKVQHIQL.L
K.MKPVPDLVPGNF.K
K.MKPVPDLVPGNFK.S
P.DLVPGNF.K
P.DLVPGNFK.S
P.DLVPGNFKSQ LQKVPPE.W
46-57
176-200
178-200
251-262
251-267
26-41
207-224
20-34
20-35
20-38
21-28
21-34
21-35
22-28
22-34
22-35
23-34
23-35
24-34
24-35
25-35
27-35
28-35
118-154
226-237
226-238
231-237
231-238
231-247
1509.73
2829.53
2602.39
1385.70
1970.04
1816.93
2132.13
1379.58
1535.69
1845.86
838.33
1308.55
1464.65
723.31
1193.52
1349.62
1106.49
1262.59
1049.47
1205.57
1076.53
904.48
757.41
4407.54
1328.68
1456.77
760.37
888.47
1895.02
28.73
23.06
24.32
36.42
38.27
25.10
31.22
25.00
18.74
16.86
23.76
26.82
19.83
21.08
24.39
18.29
24.60
19.84
24.06
17.32
16.62
13.50
18.23
24.66
26.11
20.47
26.61
18.81
17.59
PC8
PC8
PC8
PC8
PC8
PC8
PC8
PC38
PC38, PC8
PC38
PC38
PC38
PC38, PC8
PC38
PC38
PC38, PC8
PC38
PC38
PC38
PC38
PC38
PC38
PC38
PC38
PC38
PC38
PC38
PC38
PC38
SP+SC
SP+SC
SP+SC
S200+SC
S200+SC
PD10+ZT
PD10+ZT
S200+SC
S200+SC,
MCX
S200+SC
S200+SC
S200+SC
S200+SC,
MCX
S200+SC
S200+SC
S200+SC,
MCX
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
33
K.SQLQKVPPEWK.A
P.SSAGSWNSGSSGPGSTGNRNPGSSG
T.G
P.GSTGNRNPGSSGTGGTATWKPGSSG
P.G
P.GSTGSWNSGSSGTGSTGNQNPGSPR
PG.S
L.DGFRHRHPDEAAFFDTASTGKTFPG
FFSPMLGEFVSETESRGSESG.I
H.PDEAAFFDTASTGKTFPGFFSPMLG
EFVSETESRGSESGIFTNTKESSSHHPG
IAEFPS.R
P.DEAAFFDTASTGK.T
K.ESSSHHPGIAEFPSRGK.S
K.SSSYSKQFTSSTSYNRGDSTFES
.K
K.SSSYSKQFTSSTSYNRGDSTFESK.S
K.SSSYSKQFTSSTSYNRGDSTFESKS.
Y
K.SSSYSKQFTSSTSYNRGDSTFESKSY
.K
K.SSSYSKQFTSSTSYNRGDSTFESKSY
KM.A
S.SYSKQFTSSTSYNRGDSTFES.K
S.SYSKQFTSSTSYNRGDSTFESK.S
S.YSKQFTSSTSYNRGDSTFES.K
S.YSKQFTSSTSYNRGDSTFESK.S
S.KQFTSSTSYNRGDSTFES.K
A.DEAGSEADHEGTHSTKRGHAKSRP
V.R
A.GSEADHEGTHSTKRGHAKSRPV.R
E.GTAGDALIEGSVEEGAEYTSHNNM
QFSTFDRDADQWEENCAEVYGGGW
WYNNCQAANLNG.I
E.GAEYTSHNNMQFST.F
Fibrinogen β chain (FIBB_HUMAN)
See table V
Fibrinogen γ chain (FIBG_HUMAN)
R.TSTADYAMFK.V
K.AIQLTYNPDESSKPNMIDAATLK.S
K.AIQLTYNPDESSKPNMIDAATLK.S
K.RLDGSVDFK.K
R.LDGSVDFK.K
K.KNWIQYK.E
K.NWIQYK.E
K.EGFGHLSPTGTTEFWLGNEK.I
R.TSTADYAMFK.V
239-249
290-315
303-328
329-355
507-552
514-572
515-527
559-575
576-598
576-599
576-600
576-601
576-603
578-598
578-599
579-598
579-599
581-598
605-629
608-629
760-819
774-787
80-88
89-111
89-111
223-231
224-231
232-238
233-238
239-258
283-292
1338.73
2351.99
2374.09
2490.07
5040.28
6301.90
1358.60
1821.88
2552.09
2680.19
2767.22
2930.28
3205.41
2378.03
2506.13
2290.99
2419.10
2040.91
2658.25
2343.14
6590.73
1601.64
1044.60
2519.26
2535.26
1035.52
879.43
978.53
850.44
2206.03
1149.50
14.48
11.80
12.91
12.44
27.58
24.60
20.73
12.16
15.22
13.56
13.62
15.17
13.67
14.81
12.82
14.19
13.22
13.12
10.28
7.91
28.39
18.74
13.08
27.34
22.36
13.87
14.57
13.74
16.51
29.75
14.27
PC38
PC38
PC38
PC38
PC38
PC38
PC38
PC38
PC38
PC38
PC38
PC38
PC38
PC38
PC38
PC38
PC38
PC38
PC38
PC38
PC38
PC38
PC8
PC8
PC8
PC8
PC8
PC8
PC8
PC8
PC8
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
S200+SC
PD10+ZT,
S200+SC
SP+SC
PD10+ZT,
S200+SC
PD10+ZT,
S200+SC
PD10+ZT
PD10+ZT,
S200+SC
PD10+ZT,
S200+SC
PD10+ZT
S200+SC,
PD10+ZT
34
K.VGPEADKYR.L
R.LTYAYFAGGDAGDAFDGFDFGDDP
SDK.F
K.FFTSHNGMQFSTWDNDNDKFEGNC
AEQDGSGWWMNKCHAGHLNGVYY
QGGTYSK.A
K.ASTPNGYDNGIIWATWK.T
Haptoglobin
(HPT_HUMAN)
R.ILGGHLDAKGSFPWQAK.M
R.VGYVSGWGR.N
R.VGYVSGWGRNA.N
R.VGYVSGWGRNANFK.F
V.GYVSGWGRNA.N
V.GYVSGWGRNANFK.F
G.YVSGWGRNANFK.F
Hemoglobin subunit α
(HBA_HUMAN)
K.TYFPHFDLSHGSAQVK.A
Ig γ-1 chain C region
(GC1_HUMAN)
K.GPSVFPLAPSSK.S
K.ALPAPIEKTISK.A
Ig λ chain C regions
(LAC_HUMAN)
K.AAPSVTLFPPSSEELQANK.A
K.YAASSYLSLTPEQWK.S
K.YAASSYLSLTPEQWKSHR.S
Metallothionein-3
(MT3_HUMAN)
M.DPETCPCPSGGSCTCADSCK.C
Metallothionein-1B
(MT1B_HUMAN)
-MDPNCSCTTGGSCACAGSCK.C
Metallothionein-1H
(MT1H_HUMAN)
-MDPNCSCEAGGSCACAGSCK.C
-MDPNCSCEAGGSCACAGSCKCK
.K
M.DPNCSCEAGGSCACAGSCK.C
293-301
302-328
329-382
383-399
162-178
278-286
278-288
278-291
279-288
279-291
280-291
42-57
5-16
210-221
4-22
65-79
65-82
2-21
1-20
1-20
1-22
2-20
1033.52
2833.20
6165.56
1892.90
1823.97
979.49
1164.57
1553.78
1065.50
1454.71
1397.68
1832.89
1185.63
1266.76
1985.02
1742.85
2123.04
1959.67
1894.65
1892.62
2123.69
1761.45
11.29
38.08
26.28
31.54
24.00
18.69
18.77
20.02
16.08
17.41
19.29
22.70
22.39
16.87
26.74
29.36
23.30
34.53
30.43
38.88
33.48
21.85
PC8
PC8
PC8
PC8
PC8
PC8
PC8
PC8
PC8
PC8
PC8
PC8
PC8
PC8
PC8
PC8
PC8
PC8
PC38
PC38
PC38
PC38
PD10+ZT
PD10+ZT,
S200+SC
SP+SC
PD10+ZT,
S200+SC
PD10+ZT
PD10+ZT,
S200+SC
S200+SC
PD10+ZT,
S200+SC
S200+SC
S200+SC
S200+SC
PD10+ZT
S200+SC
PD10+ZT
PD10+ZT
PD10+ZT
PD10+ZT
SP+SC
SP+SC
SP+SC
SP+SC
SP+SC
35
Serotransferrin
(TRFE_HUMAN)
K.EGYYGYTGAFR.C
Transthyretin
(TTHY_HUMAN)
K.ALGISPFHEHAEVVFTANDSGPRRY
TI.A
L.GISPFHEHAEVVFTANDSGPRRYTIA
.A
I.AALLSPYSYSTTAVVTNPK
531-541
101-127
103-128
128-147
1282.57
2983.51
2870.42
2111.09
21.85
28.34
27.85
28.18
PC8
PC8
PC8
PC8
S200+SC
SP+SC
SP+SC
SP+SC
5.3 Immunological measurement of N-terminal fibrinogen β peptides
We chose PC8 for optimization of the peptide enrichment methods since this sample was
known from previous experiments to contain target fibrinogen β peptides as presented in table
V. TR-IFMA was used to rapidly scan which gel filtration fractions (figure 7) contained target
peptides sequences from FIBβ50-71.
Figure 7. Quantitative FGB peptide immunoreactivity measurements of Superdex peptide HR 10/30 size
exclusion chromatography fractions by TR-IFMA. Samples with positive identifications were chosen for the
analysis. Fractions 23 and 24 represent all of our positive 8 positive identifications listed in Table III.
It must be noted that we have not yet focused on peptides from high-mass fractions 18 and 19
that exhibit high response to FIBβ50-71 specific pAbs (0.1 µg/mL and 0.5 µg/mL
respectively). In this project, however, we focused in small sized peptides less than 4000
Daltons. These overlooked fractions still show great promise for future studies.
0
0,1
0,2
0,3
0,4
0,5
0,6
16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39
FG
B p
epti
de
IFM
A (
µg/m
L)
Fraction number
PC8
36
6 Discussion
Endogenic peptides are generated by two main pathways: synthesis of a new peptide or
degradation of a longer peptide or protein into new smaller peptides. Tumor
microenvironment can activate proteases that cause cancer-specific proteolysis in the target
cancerous tissue microenvironment (Lopez-Otin & Matrisian 2007 & Deng et al. 2015).
Differential protease activity may be a mode of function for this (Villanueva et al. 2006). Li
and co-workers (2014) have already demonstrated that proteolysis products of
carboxypeptidese N (CPN) in plasma can be found in early stages of a breast cancer. Changes
in the presence of these peptide markers may therefore reflect cancer-specific protease
activity.
Using MALDI-TOF as primary instrument for workflow optimization in biomarker discovery
has the advantage of avoiding long analysis time that is associated in most LC hyphenated
ESI techniques due to complexity of analysis by formation of multiply charged ions and
larger datasets produced in chromatographic separation. Disadvantages of using MALDI prior
screening are i) lower reproducibility (Albrethsen 2007), which obscures quantitative studies
and ii) insufficient sensitivity due to detect low abundant peptides due lack of liquid
chromatography separation. Additionally these two different ionization mechanisms favor
ions differently which results in unique identifications with one ionization mechanism than
the other. MALDI is highly tolerant to salt and impurities (Xu et al. 2006), which is can be
considered as an advantage when studying biological matrix, such as blood and urine. With
MALDI-TOF, we optimized protocols by choosing sample treatment that had the highest
intensity according to one N-terminal FIBβ standard peptide, which by no means represent all
FIBβ peptides.
Data-independent MSE is shows great promise for routine clinical and toxicological screening
laboratories to replace current immunological methods due capabilities to identify target
compounds in single run in low concentrations due to its high mass accuracy and resolving
power. Moreover, MSE reduces the total analysis time as it does not require optimization with
collision energy parameters. Furthermore, data post-processing and interpretation can done in
minutes (Sundström et al. 2013, Chindarkar et al. 2014). In this project, we exploited the non-
targeting nature of this method (since no information of the precursor ion is required), which
37
helped us to discover novel fibrinogen derived peptides along with other peptides that were
discovered in the process. This approach can provide other relevant information. In our model
sample PC8 for example, we discovered a fragment from fetuin-A. Urinary fetuin-A has been
previously related as potential marker for acute kidney injury (Zhou et al. 2006), which may
occur when presence of enlarged prostate gland obstructs the urethra or during androgen
deprivation therapy (Lapi et al. 2013). In clinical environment, non-target analysis could be
implemented in opportunistic diagnosis or in metabolic profiling where the studied disease
could be more accurately classified.
While most peptides were found and identified using PD-10 followed by C18 ZipTip, some
peptides were identified only when using Superdex peptide size-exclusion column. This
enrichment was vital in identifying some of the peptides that had only a weak signal in former
method.
Immunological methods can only give a rough direction in biomarker discovery as it is
limited to specific targets. It is difficult to estimate whether a strong response from TR-IFMA
is a result of a low amounts of the high coverage peptides or high amounts of the low
coverage peptides. Mass spectrometry is required to study different ladder-peptides.
One interesting concept we found in this study was using gel filtration column optimized for
larger peptides (Superdex 200 column with fractionation range of 10 000 – 600 000) to
harvest aggregated peptides in urine. This phenomenon was not studied thoroughly but we
could identify large amount of different peptides, especially those derived from fibrinogen α
(see table VI for the complete list).
Limitations of our study were following: i) Use of polyclonal antibody that was specific to a
sequence with extra glycine. Therefore the TR-IFMA data cannot be surely trusted. ii) The
Mascot MS/MS database algorithms are originally optimized for tryptic peptides and
therefore detection of tryptic peptides are biased over non-tryptic peptides. Furthermore
tryptic peptides are more easily protonated as they automatically contain two basic sidechains
at each terminus. iii) Limited amount of samples and duplicates used in a study. iv) pH was
not normalized or optimized in methods used, which can considerably affect peptide yield in
extraction or the ionization process. iv) Lack of sufficient urinary peptide database. Some
attempts have been made but these databases are either incomprehensive, not up to date or
38
they are specific to one method, such as the human urinary peptide database (Siwy et al.
2011), which is only specific to CE-MS. Sufficient database containing peptidome of healthy
controls and individuals with specific diseases is difficult to create and maintain since there
are multiple sample preparation techniques, mass spectrometric techniques and products by
different manufacturers. Moreover, it is difficult to have quality control of such database.
39
7 Conclusions
In this work, we have defined a systematic workflow for the analysis of unknown non-tryptic
proteolytic peptides. We used MALDI-TOF in screening of suitable sample preparation
methods as MALDI-TOF has a rapid analysis time and is more tolerant to salt and
contamination. MSE requires LC separation making the analysis time longer, but as an
advantage we found it more suitable in identifying the unknown non-tryptic peptides. 28
FIBβ-derived peptides were identified, of which 26 have not been described in urine to best of
our knowledge. In one cancer sample alone we identified 15 peptides with MSE, 6 with DDA
MS/MS in ESI-QTOF and 4 with MALDI-TOF/TOF summing up to 22 different FIBβ
peptides altogether. We have also defined the urgency of a proper database for human urinary
peptides. This workflow is equally applicable to different peptides. In this context it is
important to realize the implications of improved sensitivity in mass spectrometry.
Compounds that were previously undetectable are being identified in many different
biological systems. These findings are likely inspire researchers to new discoveries.
In conclusion, different MS/MS techniques provided us valuable complementary information.
We proved usability of MSE in non-target analysis and application in biomarker discovery.
Presence or amount of FIBβ peptides were not linked to disease in this study. In the future we
aim to study whether these peptides could serve as prognostic markers for metastatic RCC
and PC.
40
8 References
Adam SS, Key NS & Greenberg CS (2009) D-dimer antigen: current concepts and future
prospects. Blood 113(13): 2878-2887.
Agnihotri S, Mittal RD, Kapool R & Mandhani A (2014) Raising cut-off value of prostate
specific antigen (PSA) for biopsy in symptomatic men in India to reduce unnecessary biopsy.
Indian J Med Res 139(6): 851-856.
Ahmad S, Kalra H, Gupta A, Raut B, Hussain A & Rahman MA (2012) HybridSPE: a novel
technique to reduce phospholipid-based matrix effect in LC-MS bioanalysis. J Pharm Bioalled
Sci 4: 267-275.
Albrethsen J (2007) Reproducibility in protein profiling by MALDI-TOF mass spectrometry.
Clin Chem 53(5): 852-858.
Annesley TM (2003) Ion suppression in mass spectrometry. Clin Chem 49(7): 1041-1044.
Arcangeli CG, Ornstein DK, Keetch DW & Andriole GL (1997) Prostate-specific antigen as a
screening test for prostate cancer. The United States experience. Urol Clin North Am 24(2):
299-306.
Asiago VM, Alvarado LZ, Shanaiah N, Gowda GA, Owusu-Sarfo K, Ballas RA & Raftery D
(2010) Early detection of recurrent breast cancer using metabolite profiling. Cancer Res
70(21): 8309-8318.
Blackburn K, Mbeunkui F, Mitra SK, Mentzel T & Goshe MB (2010) Improving protein and
proteome coverage through data-independent multiplexed peptide fragmentation. J Proteome
Res 9(7): 3621-3637.
Cantero D, Friess H, Deflorin J, Zimmermann A, Bründler MA, Riesle E, Korc M & Büchler
MW (1997) Enhanced expression of urokinase plasminogen activator and its receptor in
pancreatic carcinoma. Br J Cancer 75(3): 388-395.
41
Carr S, Aebersold R, Baldwin M, Burlingame A, Clauser K & Nesvizhskii A (2004) The need
for guidelines in publication of peptide and protein identification data. Mol Cell Proteomics
3:531-533.
Chapin JC & Hajjar KA (2015) Fibrinolysis and the control of blood coagulation. Blood Rev
(1): 17-24.
Chen YT, Chen CL, Chen HW, Chung T, Wu CC, Chen CD, Hsu CW, Chen MC, Tsui KH,
Chang PL, Chang YS & Yu JS (2010) Discovery of novel bladder cancer biomarkers by
comparative urine proteomics using iTRAQ technology. J Proteome Res 9(11): 5803-5815.
Chindarkar NS, Wakefield MR, Stone JA & Fitzgerald RL (2014) Liquid chromatography
high-resolution TOF analysis: Investigation of MSE for broad-spectrum drug screening. Clin
Chem 60(8): 1115-1125.
Chinello C, Cazzaniga M, De Sio G, Smith AJ, Grasso A, Rocco B, Signorini S, Grasso M,
Bosari S, Zoppis I, Mauri G & Magni F (2015) Tumor size, stage and grade alterations of
urinary peptidome in RCC. J Transl Med 13(1): 332 10.1186/s12967-015-0693-8.
Cohen SL & Chait BT (1996) Influence of matrix solution conditions on the MALDI-MS
analysis of peptides and proteins. Anal Chem 68: 31-37.
Distler U, Kuharev J, Navarro P, Levin Y, Schild H & Tenzer S (2013) Drift time-specific
collision energies enable deep coverage data-independent acquisition proteomics. Nat
Methods 11(2): 167-170.
D’Nicuola J, Jones R, Levine B & Smith ML (1992) Evaluation of six commercial
amphetamine and methamphetamine immunoassays for cross-reactivity to
phenylpropanolamine and ephedrine in urine. J Anal Toxicol 16(4): 211-213.
Doerr A (2015) DIA mass spectrometry. Nat Methods 35 10.1038/nmeth.3234.
42
Drucker E & Krapfenbauer K (2013) Pitfalls and limitations in translation from biomarker
discovery to clinical utility in predictive and personalized medicine. EPMA J 4(1): 7.
Duffy MJ, Duggan C, Mulcahy HE, McDermott EW & O’Higgins NJ (1998) Urokinase
plasminogen activator: a prognostic marker in breast cancer including patients with axillary
node-negative disease. Clin Chem 44(6): 1177-1183.
Ficarra V, Brunelli M, Cheng L, Kirkali Z, Lopez-Beltran A, Martignoni G, Montironi R,
Novara G & Van Poppel H (2010) Prognostic and therapeutic impact of the histopathologic
definition of parenchymal epithelial renal tumors. Eur Urol 58(5): 655-568.
Flanigan RC, Campbell SC, Clark JI & Picken MM (2003) Metastatic renal cell carcinoma.
Curr Treat Options Oncol 4(5): 385-390.
Frantzi M, Metzger J, Banks RE, Husi H, Klein J, Dakna M, Mullen W, Cartledge JJ,
Schanstra JP, Brand K, Kuczyk MA, Michak H, Vlahou A, Theodorescu D & Merseburger
AS (2014) Discovery and validation of urinary biomarkers for detection of renal cell
carcinoma. J Proteomics 98: 44-58.
Geromanos SJ, Vissers JP, Silva JC, Dorschel CA, Li GZ, Gorenstein MZ, Bateman RH &
Langridge JI (2009) The detection, correlation, and comparison of peptide precursor and
product ions from data independent LC-MS with data independent LC-MS/MS. Proteomics
9(6): 1683-1695.
Giavannini I, Chiarla C, Giuliante F, Vellone M & Nuzzo G (2004) Modulation of plasma
fibrinogen levels in acute-phase response after hepatectomy. Clin Chem Lab Med 42(3): 261-
265.
Gillet LC, Navarro P, Tate S, Röst H, Sevelsek N, Reiter L, Bonner R & Aebersold R (2012)
Target data extraction of the MS/MS spectra generated by data-independent acquisition: new
concept for consistent and accurate proteome analysis. Mol Cell Proteomics 11(6)
10.1074/mcp.O111.016717.
43
de Gramont A, Watson S, Ellis LM, Rodón J, Tabernero J, de Gramont A & Hamilton SR
(2015) Pragmatic issues in biomarker evaluation for targeted therapies in cancer. Nat Rev
Clin Oncol 12(4): 197-121.
Gromov P, Moreira JMA, Gromova I & Cells JE (2006) Proteomic analysis of urinary
fibrinogen degradation products in patients with urothelial carcinomas. Clin Proteomics 2(1):
45-65.
Hajian-Tilaki K (2013) Receiver operating characteristic (ROC) curve analysis for medical
diagnostic test evaluation. Caspian J Intern Med 4(2): 627-635.
Hiramoto R, Bernecky J, Jurandowski J & Pressman D (1960) Fibrin in human tumors.
Cancer Res 20: 592-593.
Hoffenberg R & Ramsden DB (1983) The transport of thyroid hormones. Clin Sci 65(4): 337-
342.
Hortin GL & Sviridov D (2007) Diagnostic potential for urinary proteomics.
Pharmacogenomics 8(3): 237-255.
Iba T, Gando S, Murata A, Kushimoto S, Saitoh D, Equchi Y, Ohmoto Y, Okamoto K,
Koseki K, Mayumi T, Ikeda T, Ishhikura H, Uyema M, Ogura Y, Endo S & Shimazaki S
(2007) Predicting the severity of systemic inflammatory response syndrome (SIRS)-
associated coagulopathy with hemostatic molecular markers and vascular endothelial injury
markers. J Trauma 63(5): 1093-1098.
Inoue K, Ikemura A, Tsuruta Y, Tsutsumiuchi K, Hino T & Oka H (2012) On-line solid-
phase extraction LC-MS/MS for the determination of Ac-SDKP peptide in human plasma
from hemodialysis patients. Biomed Chromatogr 26(2): 137-141.
Jackson C & Nemerson Y (1980) Blood coagulation. Annu Rev Biochem 49: 765-811.
44
Juhasz P & Biemann K (1994) Mass spectrometric molecular-weight determination of highly
acidic compounds of biological significance via their complexes with basic polypeptides. Proc
Natl Acad Sci U S A 91(10): 4333-4337.
Kilpeläinen TP, Tammela TLJ, Määttänen, L, Kujala P, Stenman UH, Ala-Opas M, Murtola
TJ & Auvinen A (2010) False-positive screening results in the Finnish prostate cancer
screening trial. Br J Cancer 102(3): 469-474.
Knochenmuss R (2006) Ion formation mechanisms in UV-MALDI. Analyst 131(9): 966-986.
Koal T, Schmiederer D, Pham-Tuan H, Röhring C & Rauh M (2012) Standardized LC-
MS/MS based steroid hormone profile analysis. J Steroid Biochem Mol Biol 129(3-5): 129-
138.
Konermann L, Ahadi E, Rodriguez AD & Vahidi S (2013) Unraveling the mechanism of
electrospray ionization. Anal Chem 85(1): 2-9.
Krajewska E, Lewis CE, Chen YY, Welford A, Tazzyman S & Staton CA (2010) A novel
fragment derived from the beta chain of human fibrinogen, beta 43-63, is a potent inhibitor of
activated endothelial cells in vitro and in vivo. Br J Cancer 102(3): 594-601.
Krishnamoorthy A, Ajay A, Hoffmann D, Kim TM, Ramirez V, Campanholle G, Bobadilla
MA, Waikar SS & Vaidya VS (2011) Fibrinogen β-derived Bβ(15-42) peptide protects
against kidney ischemia/reperfusion injury. Blood 118(7): 1934-1942.
Lapi F, Azoulay L, Niazi MT, Yin H, Benayoun S & Suissa S (2013) Androgen deprivation
therapy and risk of acute kidney injury in patients with prostate cancer. J Am Med Assoc
310(3):289-296.
Lal BK, Varma S, Pappas PJ, Hobson RW 2nd & Durán WN (2001) VEGF increases
permeability of the endothelial cell monolayer by activation of PKB/akt, endothelial nitric-
oxide synthase, and MAP kinase pathways 62(3): 252-262.
45
Lee S, Ji D, Park M & Chung KH (2015) Development of a comprehensive spectral library of
sildenafil and related active analogues using LC-QTOF-MS and its application for screening
counterfeit pharmaceuticals. Forensic Sci Int 257: 182-188.
Lelongt B, Bengatta S, Delauche M, Lund LR, Werb Z, Ronco PM (2001) Matrix
metalloproteinase 9 protects mice from anti-glomerular basement membrane nephritis through
its fibrinolytic activity. J Exp Med 193(7): 793-802.
Li Y, Li Y, Chen T, Kuklina AS, Bernard P, Esteva FJ, Shen H, Ferrari M, Hu Y (2014)
Circulating proteolytic products of carboxypeptidase N for early detection of breast cancer.
Clin Chem 60(1): 233-242.
López-Otin C & Lynn MM (2007) Emerging roles of proteases in tumour suppression. Nat
Rev Cancer 7(10): 800-808.
Moyer SC, Budnik BA, Pittman JL, Costello CE & O’Connor PB (2003) Attomole peptide
analysis by high-pressure matrix-assisted laser desorption/ionization Fourier transform mass
spectrometry. Anal Chem 75(23): 6449-6454.
Mann M, Hendrickson RC & Pandey A (2001) Analysis of proteins and proteomes by mass
spectrometry. Annu Rev Biochem 70: 437-473.
Mason SD & Joyce JA (2011) Proteolytic networks in cancer. Trends cell boil 21(4): 228-
237.
Mayo Clinic (2015) Plasma Fibrinogen. Presented as part of Mayo Medical Laboratories
information sheet about plasma fibrinogen measurements. URL:
http://www.mayomedicallaboratories.com/test-catalog/Clinical+and+Interpretive/8484. Cited
2015/9/10.
McPhail S, Johnson S, Greenberg D, Peake M & Rous B (2015) Stage at diagnosis and early
mortality from cancer in England. Br J Cancer 112: 108-115.
46
Mehta AI, Ross S, Lowenthal MS, Fusaro V, Fishman DA, Petricoin EF & Liotta LA (2003-
2004) Biomarker amplification by serum carrier protein binding. Dis Markers 19(1): 1-10.
National Cancer Institute (2011) Tumor markers. Presented as part of National Cancer
Institute’s tumor marker information sheet. URL: http://www.cancer.gov/about-
cancer/diagnosis-staging/diagnosis/tumor-markers-fact-sheet. Cited 2015/8/8.
Palumbo JS, Kombrink KW, Drew AF, Grimes TS, Kiser JH, Degen JL & Bugge TH (2000)
Fibrinogen is an important determinant of the metastatic potential of circulating tumor cells.
Blood 96(10): 3302-3309.
Petricoin EF, Ardekani AM, Hitt BA, Levine PJ, Fusaro VA, Steinberg SM, Mills GB,
Simone C, Fishman DA, Kohn EC & Liotta LA (2002) Use of proteomic patterns in serum to
identify ovarian cancer. Lancet 359(9306): 572-575.
Ramachandran R, Noorbakhsh F, Defea K & Hollenberg MD (2012) Targeting proteinase-
activated receptors: therapeutic potential and challenges. Nat Rev Drug Discov 11(1): 69-86.
Ruiter DJ, van Krieken JH, van Muijen GN & de Waal RM (2001) Tumour metastasis: is
tissue an issue? Lancet Oncol 2(2): 109-112.
Sambrano GR, Weiss EJ, Zheng YW, Huang W & Coughling SR (2001) Role of thrombin
signalling in platelets in haemostasis and thrombosis. Nature 413(6851): 74-78.
Sahni A, Simpson-Haidaris PJ, Sahni SK, Vaday GG & Francis CW (2008) Fibrinogen
synthesized by cancer cells augments the proliferative effect of fibroblast growth factor-2
(FGF-2). J Thromb Haemost 6(1): 176-183.
Scigelova M & Makarov A (2006) Orbitrap mass analyzer—overview and application in
proteomics. Proteomics 2: 16-21.
Sharma S (2009) Tumor markers in clinical practice: General principles and guidelines.
Indian J Med Paediatr Oncol 30(1): 1-8.
47
Shibuya M (2011) Vascular Endothelial Growth Factor (VEGF) and Its Receptor (VEGFR)
Signaling in Angiogenesis. A Crucial Target for Anti- and Pro-Angiogenic Therapies. Genes
Cancer 2(12): 1097-1105.
Shiomi Y, Eguchi Y, Tani T, Kodama M & Hattori T (2000) Cellular distribution and clinical
value of urokinase-type plasminogen activator, its receptor, and plasminogen activator
inhibitor-2 in esophageal squamous cell carcinoma. Am J Pathol 152(2): 567-575.
Siegel R, Ma J, Zou Z & Jemal A (2014) Cancer statistics, 2014. CA Cancer J Clin 64(1): 9-
29.
Siiteri PK, Murai JT, Hammond GL, Nisker JA, Raymoure WJ & Kuhn RW (1982) The
serum transport of steroid hormones. Recent Prog Horm Res 38: 457-510.
Siwy J, Mullen W, Golovko I, Franke J & Zürbig P (2011) Human urinary peptide database
for multiple disease biomarker discovery. Proteomics Clin Appl 5(5-6): 367-374.
Skogen WF, Senior RM, Griffin GL & Wilner GD (1988) Fibrinogen-derived peptide B beta
1-42 is a multidomained neutrophil chemoattractant. Blood 71(5): 1475-1479.
Stapels MD & Barofsky DF (2004) Complementary use of MALDI and ESI for the HPLC-
MS/MS analysis of DNA-binding proteins. Anal Chem 76: 5423-5430.
Stenman UH, Leinonen J, Alfthan H, Rannikko S, Tuhkanen K, Alfthan O (1991) A complex
between prostate-specific antigen and alpha 1-antichymotrypsin is the major form of prostate
–specific antigen in serum of patients with prostate cancer: assay of the complex improves
clinical sensitivity for cancer. Cancer Res 51(1): 222-226.
Stenman UH (2005) The value of free PSA in addition to total PSA in early detection of
prostate cancer. Eur Oncol Rev 60: 1-4 http://doi.org/10.17925/EOH.2005.0.0.1f.
Stenman UH, Abrahamsson PA, Aus G, Lilja H, Bangma C, Hamdy FC, Boccon-Gibod L &
Ekman P (2005) Prognostic value of serum markers for prostate cancer. Scand J Urol Nephrol
Suppl 216: 64-81.
48
Steuber T, Nurmikko P, Haese A, Pettersson K, Graefen M, Hammerer P, Huland H & Lilja
H (2002) Discrimination of benign from malign prostatic disease by selective measurements
of single chain, intact free prostate specific antigen. J Urol 168(5): 1917-1922.
Stolker AL, Niesing W, Fuchs R, Vreeken RJ, Niessen WM & Brinkman UA (2004) Liquid
chromatography with triple-quadrupole and quadrupole-time-of-flight mass spectrometry for
the determination of micro-constituents - a comparison. Anal Bioanal Chem 378(7): 1754-
1761.
Strathmann FG & Hoofnagle NF (2011) Current and future applications of mass spectrometry
to the clinical laboratory. Am J Clin Pathol 136(4): 609-616.
Sundström M, Pelander A, Angerer V, Hutter M, Kneisel S & Ojanpera I (2013) A high-
sensitivity ultra-high performance liquid chromatography/high-resolution time-of-flight mass
spectrometry (UHPLC-HR-TOFMS) method for screening synthetic cannabinoids and other
drugs of abuse in urine. Anal Bioanal Chem 405(26): 8463-8474.
Sundström M, Pelander A & Ojanperä I (2015) Comparison between drug screening by
immunoassay and ultra-high performance liquid chromatography/high-resolution time-of-
flight mass spectrometry in post-mortem urine. Drug Test Anal 7(5): 420-427.
Thompson WD, Campbell R & Evans T (1985) Fibrin degradation and angiogenesis:
quantitative analysis of the angiogenic response in the chick chorioallantoic membrane. J
Pathol 145(1): 27-37.
Tromp SC, oude Ebrink MG, Dings RP, van Velzen S, Slaaf DW, Hillen HF, Tangelder GJ,
Reneman RS & Griffioen AW (2000) Tumor angiogenesis factors reduce leukocyte adhesion
in vivo. Int Immunol 5: 671-676.
Tsihlias J & Grossman HB (2000) The utility of fibrin/fibrinogen degradation products in
superficial bladder cancer. Urol Clin North Am 27(1): 39-46.
49
Tsou CC, Avtonomov D, Larsen B, Tucholska M, Choi H, Gingras AC & Nesvizhskii AI
(2015) DIA-Umpire: comprehensive computational framework for data-independent
acquisition proteomics. Nat Methods 12(3): 258-264.
Turk B, Turk D & Turk V (2012) Protease signalling: the cutting edge. EMBO J 31(7): 1630-
1643.
UniProt Knowledgebase (2015) P02675 FIBB_HUMAN. URL:
http://www.uniprot.org/uniprot/P02675. Cited 2015/10/20, last entry update 2015/10/14.
Vallner JJ (1979) Binding of drugs by albumin and plasma protein. J Pharm Sci 66(4): 447-
465.
Villanueva J, Philip J, Entenberg D, Chaparro CA, Tanwar MK, Holland EC & Tempst P
(2004) Serum peptide profiling by magnetic particle-assisted, automated sample processing
and MALDI-TOF mass spectrometry. Anal Chem 76(6): 1560-1570.
Villanueva J, Shaffer DR, Philip J, Chaparro CA, Erdjument-Bromage H, Olshen AB,
Fleisher M, Lilja H, Brogi E, Boyd J, Sanchez-Carbayo M, Holland EC, Cordon-Cardo C,
Scher HI & Tempst P (2006) Differential exoprotease activities confer tumor-specific serum
peptidome patterns. J Clin Invest 116(1): 271-284.
Weisel JW (2005) Fibrinogen and fibrin. Adv Protein Chem 70: 247-299.
Wilcken B & Wiley V (2008) Newborn screening. Pathology 40(2): 104-115.
Wolff JC, Eckers C, Sage AB, Giles K & Bateman R (2001) Accurate mass liquid
chromatography/mass spectrometry on quadrupole orthogonal acceleration time-of-flight
mass analyzers using switching between separate sample and reference sprays. 2.
Applications using the dual-electrospray ion source. Anal Chem 73(11): 2605-2612.
Zhou H, Pisitkun T, Aponte A, Yuen PS, Hoffert JD, Yasuda H, Hu X, Chawla L, Shen RF,
Knepper MA, Star RA (2006) Exosomal Fetuin-A identified by proteomics : A novel urinary
biomarker for detecting acute kidney injury. Kidney Int 70:1847-1857.
50
Zucker S, Cao J, & Zhen WT (2000) Critical appraisal of the use of matrix metalloproteinase
inhibitors in cancer treatment. Oncogene 19(56): 6642-6650.
