Journal of Proteomics xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Journal of Proteomics
journal homepage: www.elsevier.com/locate/jprot

Proteomic analysis of first trimester maternal serum to identify candidate

biomarkers potentially predictive of spontaneous preterm birth
⁎
Arlene M. D'Silvaa, Jon A. Hyettb, Jens R. Coorssenc,d,
a
Department of Molecular Physiology, The Molecular Medicine Research Group, School of Medicine, Western Sydney University, Campbelltown, NSW 2150, Australia
b
Department of High Risk Obstetrics, RPA Women and Babies, Royal Prince Alfred Hospital, University of Sydney, Sydney, SW 2050, Australia
c
Department of Health Sciences, Faculty of Applied Health Sciences, Brock University, St. Catharines, ON L2S 3A1, Canada
d
Faculty of Mathematics and Science Brock University, St. Catharines, ON L2S 3A1, Canada

A R T I C L E I N F O A B S T R A C T

Keywords: Spontaneous preterm birth (sPTB) remains a major clinical dilemma; current diagnostics and interventions have
Premature labour not reduced the rate of this serious healthcare burden. This study characterizes differential protein profiles and
Two-dimensional gel electrophoresis post-translational modifications (PTMs) in first trimester maternal serum using a refined top-down approach
Post-translational modifications coupling two-dimensional gel electrophoresis (2DE) and mass spectrometry (MS) to directly compare subsequent
LCMS/MS
term and preterm labour events and identify marked protein differences. 30 proteoforms were found to be
Protein species
significantly increased or decreased in the sPTB group including 9 phosphoproteins and 11 glycoproteins.
Proteoforms
Changes occurred in proteins associated with immune and defence responses. We identified protein species that
are associated with several clinically relevant biological processes, including interrelated biological networks
linked to regulation of the complement cascade and coagulation pathways, immune modulation, metabolic
processes and cell signalling. The finding of altered proteoforms in maternal serum from pregnancies that de-
livered preterm suggests these as potential early biomarkers of sPTB and also possible mediators of the disorder.
Biological significance: Identifying changes in protein profiles is critical in the study of cell biology, and disease
treatment and prevention. Identifying consistent changes in the maternal serum proteome during early preg-
nancy, including specific protein PTMs (e.g. phosphorylation, glycosylation), is likely to provide better oppor-
tunities for prediction, intervention and prevention of preterm birth. This is the first study to examine first
trimester maternal serum using a highly refined top-down proteomic analytical approach based on high re-
solution 2DE coupled with mass spectrometry to directly compare preterm (< 37 weeks) and preterm
(≥37 weeks) events and identify select protein differences between these conditions. As such, the data present a
promising avenue for translation of biomarker discovery to a clinical setting as well as for future investigation of
underlying aetiological processes.

1. Introduction Proteoforms are of fundamental importance to all biological pro-

Preterm birth (PTB), defined as delivery < 37 weeks' gestation, re-
mains the leading cause of perinatal mortality and morbidity worldwide
[1]. An estimated 15 million babies are born preterm each year and
28% of deaths in infancy are related to prematurity [1,2]. In addition to
the risk of immediate neonatal complications, preterm infants have
increased risks of neurodevelopment complications (such as cerebral
palsy and autism) and of cognitive, cardiovascular and metabolic dis-
orders [3–5]. Understanding and predicting PTB has been a challenge
as the risk factors and pathophysiological pathways associated with
PTB appear to be complex and poorly understood [6,7]. Consequently,
there has been little success in predicting or preventing PTB.
cesses [8,9]. Exploring and defining the serum proteome in pregnancies
that are later affected by PTB should improve our insight into the
evolution and progression of this disease. Quantification of disease-as-
sociated protein alterations can be achieved by using two-dimensional
gel electrophoresis (2DE) coupled with mass spectrometry (MS)
[10–12]. 2DE was specifically chosen because it is the only available
proteomic technique that can simultaneously resolve hundreds-to-
thousands of intact protein species in a single analytical run, while also
enabling multiple parallel analyses. As the only such routine top down
analytical protocol, it is thus the only available approach that enables
quantitative profiling of large sets of complex mixtures of protein spe-
cies; that is, as part of the routine analytical protocol, this approach

⁎
Corresponding author at: Department of Health Sciences, Faculty of Applied Health Sciences, Brock University, St. Catharines, ON L2S 3A1, Canada.
E-mail address: jcoorssen@brocku.ca (J.R. Coorssen).

https://doi.org/10.1016/j.jprot.2018.02.002
Received 15 August 2017; Received in revised form 29 January 2018; Accepted 2 February 2018
1874-3919/ © 2018 Elsevier B.V. All rights reserved.

Please cite this article as: Coorssen, J., Journal of Proteomics (2018), https://doi.org/10.1016/j.jprot.2018.02.002
A.M. D'Silva et al.
resolves protein isoforms, splice variants, and the vast range of post-
translationally modified protein species that define biological func-
tionality. The ability to resolve protein species, including those invol-
ving specific post-translational modifications (PTMs) is likely to prove
quite important; for example, changes in glycosylation have been re-
ported in trophoblast tissue in pregnancies affected by both pre-
eclampsia and PTB [13]. Identification of consistent alterations in the
abundance of particular protein species may thus provide new oppor-
tunities for prediction of PTB.
A number of aetiological pathways have been implicated in PTB
including infection (chorioamnionitis), uteroplacental ischaemia, hae-
morrhage and mechanical over-distension of the uterus [14]. Risk fac-
tors for PTB have been identified but screening on the basis of maternal
history fails to predict the majority of women who present in sponta-
neous preterm labour or with preterm prelabour rupture of membranes
[15,16]. Current investigational tools, such as ultrasound assessment of
cervical length and measurement of vaginal fetal fibronectin appear to
focus on a final common pathway toward PTB [17–19]. In contrast,
obstetricians screen for a variety of pregnancy complications at
11–13+6 weeks' gestation and have demonstrated that early prediction
enables effective therapeutic intervention and disease prevention
[20,21]. In this study, we identify changes in the proteome of first
trimester maternal serum collected from pregnancies that subsequently
delivered preterm.
2. Materials and methods
2.1. Study design
Proteomic analyses were carried out on a cohort of maternal serum
samples prospectively collected and stored during first trimester
screening at 11–13+6 weeks' gestation. The samples were collected
between 2011 and 2014 and serum was separated within 4 h of col-
lection. Aliquots of residual serum, available after measurement of free
βhCG and PaPP-A, were stored immediately at −80 °C. Details of
pregnancy outcome were collated at the end of pregnancy allowing
identification of a cohort of women who laboured spontaneously before
37 weeks' gestation (n = 10). These were matched (one to one) to a
cohort of women who laboured spontaneously and delivered at term
(≥37 weeks' gestation) (n = 10). Control samples were matched by
maternal age, gestational age, BMI, parity, smoking status, sex of the
fetus and sample storage time. Samples were excluded if preterm de-
livery was elective, had occurred after spontaneous rupture of mem-
branes or involved a multiple pregnancy or a pregnancy affected by
chromosomal or structural abnormality. The study was approved by the
local hospital Human Ethics Committee (Protocol No: X11-0305).
2.2. 2-Dimensional gel electrophoresis
A ‘top down’ proteomic approach was used to resolve and identify
maternal serum proteins and PTMs via 2DE and liquid chromatography
mass spectrometry/mass spectrometry (LC/MS/MS). As described pre-
viously, 500 μg total serum protein was resolved on a 17 cm 3–10 non-
linear immobilised pH gradient strip in the first dimension followed by
second dimension resolution on 7–20% gradient gels with a combina-
tion of lithium dodecyl sulfate (LDS) and sodium dodecyl sulfate (SDS)
detergent [22]. All analyses were carried out using three technical re-
plicates to ensure reproducibility (Fig. 1).
Resolved proteins were subsequently detected in-gel using the cur-
rent gold standard protocol with Colloidal Coomassie Brilliant Blue
(cCCB) as a near-IR dye [23]. To enable resolution of co-migrating
proteins that appear as hyper-abundant spots following the second di-
mension of resolution, and to facilitate detection of low abundance
proteins, a post-fractionation approach was used, combining 2DE with
third dimension electrophoresis (3DE) and a well-established ‘deep
imaging’ protocol, as previously described [22,24–27]. 3DE is used to
2
Journal of Proteomics xxx (xxxx) xxx–xxx
further resolve co-migrating proteins that appear as hyper-abundant
spots after initial resolution by 2DE, using a gradient gel customized to
provide optimal resolution within the target molecular weight range
[24,25]. Deep imaging involves excising saturating spots/regions prior
to imaging the gel again, thus enabling detection of lower abundance
protein species [25,26]. The specific sub-proteomes associated with
protein phosphorylation and glycosylation were identified using the
fluorescent stains Pro-Q Diamond and Pro-Q Emerald, respectively,
according to the manufacturer's protocols (Life technologies, Carlsbad,
CA). After detection of phosphoproteins and glycoproteins, gels were
finally stained with cCCB for total protein as previously described
[23,28,29]. Stained gels were imaged using the FLA-9000 (FujiFilm/GE
Health Sciences) for phosphoprotein (555/580 nm excitation/emission
(ex/em)) and glycoprotein detection (510/520 nm ex/em; 500 V PMT
and 100 μm pixel size). Optimal near-IR imaging of cCBB-stained gels
using the FLA-9000 was carried out at 685/ > 750 ex/em with a PMT
setting of 600 V and pixel resolution set to 100 μm [23,26].
2.3. Image analysis and mass spectrometry
Image analysis was carried out using the software Delta2D V4.0
(DECODON Gmbh, Greifswald, Germany) in accordance with the
manufacturer's specifications. Protein patterns revealed by each stain
were used to create three unique maps for each patient sample. The
‘phosphorylation’ and ‘glycosylation’ images were directly super-
imposed over the total-protein stain image to enable unambiguous
matching of protein spots. This enabled quantitation of changes in both
the PTM and amount of each resolved protein species. Images were
divided into sPTB and control groups. For the purpose of this analysis,
we used union setting by aligning each gel image to form a consensus
pattern. This pattern was then transferred to all raw images to enable
spot comparisons between the sPTB and control groups. The union
setting enabled selection of only those spots that were 100% re-
producible within all three replicate gels. The total number of resolved
protein species for each proteome (i.e. total proteome, phosphopro-
teome and glycoproteome) were calculated using the three individual
raw gel images for every patient (i.e. n = 3 per patient) and are re-
ported as mean ± SEM (Table 2). To enable calculation of the ob-
served molecular weight (MW) and pI of a given protein spot, a set of
2DE standards (Bio-Rad) was routinely resolved over the course of the
study. A standard curve was created based on the average migration of
the protein standards.
For inclusion in the analysis, changes in spot volume (i.e. the
abundance of a resolved protein species) had to (i) differ significantly
between samples from sPTB and matched control patients (t-test;
p < 0.05, n = 3) and (ii) have a fold change of > 1.5 [30]. To account
for patient-to-patient variability and ensure that only consistently
changing protein species would be included for identification by MS,
greys values for individual spot volumes were calculated across the
sPTB and control sample set for all technical replicates. This was further
confirmed by visual inspection of both the gels and imaging data.
Protein species meeting the criteria were manually excised from the
gels and identified using LC/MS/MS; peptides were isolated for LC/MS/
MS analysis and data analysed as described previously [23,26]. The
peptide sequences from the MS/MS spectra were identified by corre-
lation with the peptide sequences of proteins in the SwissProt database
(version 2011_06) using the Mascot Daemon search algorithm (V2.2.2)
(Matrix Science, Boston, MA, USA). The PANTHER (Protein Analysis
Through Evolutionary Relationships) classification system and UniProt
database were employed to assess the identified proteins for biological
context, involvement in various physiological pathways and association
with disease pathophysiology.
3. Results
The demographic and clinical characteristics of the sample set are
A.M. D'Silva et al. Journal of Proteomics xxx (xxxx) xxx–xxx

sPTB (<37 wks) Controls (>37 wks)

n = 10 n = 10

Protein assay

Isoelectric focussing on 17 cm, 3-10 NL

(first dimension)

Phosphoproteins: Glycoproteins:
Total protein:
ProQ Diamond, ProQ Emerald,
CCB, n=3
n=3 n=3

Image analysis
(Delta 2D)

Mass spectrometry
(LCMS/MS)

PANTHER analysis

Protein class/ molecular funcon/ biological process/ pathway

Fig. 1. Schematic representation of the experimental strategy for proteomic analysis of alterations in the human serum proteome in sPTB and matched control cohorts.

3
A.M. D'Silva et al. Journal of Proteomics xxx (xxxx) xxx–xxx

Table 1 proteome, phosphoproteome and glycoproteome are summarised in

Baseline characteristics of sPTB vs term control pregnancies. Table 2. Not surprisingly, several abundant serum proteins were de-
tected, including albumin, immunoglobulins (Ig) (gamma-1 chain C
Demographic Early sPTB Controls Significance⁎ (p
feature (24–33 weeks) (≥ 37 weeks) value) region, alpha-1 chain C region, mu chain C region), alpha-2-macro-
(n = 10) (n = 10) globulin, haptoglobin and serotransferrin.
30 protein species and/or PTM of statistically significant differential
Maternal characteristics
abundance (p < 0.05) were characterized, including 9 phospho- and
Age (years) 33.3 ± 1.42 33.4 ± 1.09 0.9561
Gestational age at 12.3 12.4 0.7278 11 glyco-proteoforms. There were no significant increases or decreases
serum in the total number of protein species resolved from the serum samples
sampling of sPTB and term control cohorts, including for the phospho- and gly-
Gestational age at 30.0 ± 0.39 40.0 ± 0.30 0.0001⁎ coproteomes (Table 2). This was also true for the deep imaging
delivery
(Table 2). Changes in protein species that were not 100% reproducible
(weeks)
Parity: =0 60% 70% 1.0000# across all technical replicates were not included in our final analyses.
>0 40% 30% This initial analysis focussed explicitly on changes detected most con-
Ethnicity: East 30% 30% 1.0000# sistently among ≥8/10 patients and thus represented the most sub-
Asian
stantial alterations in relative protein abundance between sPTB and
White 70% 70%
BMI (kg/m2) 25.25 24.23 0.6612 matched control samples with 100% reproducibility among technical
Smoking status: – – 1.0000# replicates (Tables 3, 4 and 5).
Yes LC/MS/MS data coupled with Mascot Daemon searches of SwissProt
No 10 10 and LudwigNR databases yielded high-quality identifications of all 30
Outcome characteristics of these select protein species from the standard 2DE analysis (Tables 3,
Birth weight (g) 2025 ± 354.9 3768 ± 130.2 0.0001⁎ 4 and 5). Several were related species of the same ‘parent’ protein,
Sex: Male 80% 70% NA
differing in molecular weight (MW) and pI. It is interesting to note that
Female 20% 30%
among these protein species, for three (i.e. alpha-1-antitrypsin, vitamin
Values have been represented as mean ± SEM or %, statistical analysis involved an D binding protein, apolipoprotein A-1) differences in abundance were
unpaired t-test. detected in spots that, according to specific stains, appear to be both
⁎
Denotes statistical significance. phosphorylated and glycosylated; all these proteins are resolved into
#
Denotes Fisher 2 × 2 contingency test. different spots on the 2D gels, corresponding to multiple, separate
species.
presented in Table 1. Maternal age distribution, parity, ethnicity, BMI, Deep imaging of the resolved 2DE gels resulted in enhanced vi-
smoking during pregnancy and concurrent medical history were not sualisation of low abundance protein species for individual serum
significantly different in the two groups. The sex of the fetus was proteomes (Fig. 5). Post-fractionation by third dimension resolution of
comparable in the two groups and the mean birth weights were the hyper-abundant protein regions removed from serum proteomes
2025 ± 354.9 g in early sPTB and 3768 ± 130.2 g in the control initially resolved by 2DE (i.e. particularly those known to correspond to
group. albumin, immunoglobulin heavy and light chains, and serotransferrin)
Representative gel images of 10 samples from each sPTB and term enabled resolution of additional protein species from the more abun-
control cohort for total protein differences and PTM differences dant co-migrating proteins (Fig. 6). This was consistent with our pre-
(phosphorylation and glycosylation) are shown in Figs. 2–4. All samples vious results [22,24–27]. However, comparative analyses of sPTB and
yielded well-resolved proteomes encompassing the full MW/pI range of term labour serum samples indicated no significant differences in the
the gels (Figs. 2–4). The number of resolved protein species for the total amounts of the lower abundance protein species resolved.

3 pI 10 3 pI 10

200 200
150 8A 150
6A 29 A
120 19 A 120
100 6A 100
85 85
MW (KDa)

70 70
MW (KDa)

60 60
50 144 A 50
40 73 A 40
151 A
30 30
25 25
20 105 A 20

15 106 A 15
10 10

Term Preterm
Fig. 2. Comparative 2DE and quantitative differential analysis of the maternal serum proteome.
Proteins (500 μg) were resolved in the first dimension by isoelectric focussing (IEF) on a 17 cm, 3–10 non-linear (NL) immobilised pH gradient (IPG), and the second dimension using
7–20% gradient gel with 0.1% sodium dodecyl sulfate (SDS) + 0.1% lithium dodecyl sulfate (LDS) stained by cCBB. Numbers listed represent protein spots of interest identified across
three replicate 2DE gels; these are listed in Table 3. Quantitative image analysis indicates protein spots that differ significantly in abundance (increased abundance in SPTB labelled red;
reduced abundance labelled green) t-test, p < 0.05, n = 10 each cohort.

4
A.M. D'Silva et al. Journal of Proteomics xxx (xxxx) xxx–xxx

3 pI 10 3 pI 10

200 200
150 150
120 120
100 100
85 8B 85

70 44 B 70
MW (KDa)

MW (KDa)
60 60
46 B
50 22 B 50 34 B
40 40
24 B
30 31 B 30
25 48 B 25
20 39 B 20

15 15
10 10

Term Preterm
Fig. 3. Representative 2DE gel images of sPTB and matched controls stained with ProQ.
Diamond to identify phosphoproteins. Analysis was as described in Fig. 2. Numbers represent protein spots of interest identified across three replicate 2DE gels; these are itemized in
Table 4. Quantitative image analysis indicates protein spots that differ significantly in abundance (increased abundance in SPTB labelled red; reduced abundance labelled green) t-test,
p < 0.05, n = 10 each cohort.

According to the molecular function analysis, most of the differen- Table 2

tially expressed proteins were related to binding and catalytic activities. Total protein species resolved by 2DE in maternal serum from sPTB and term cohorts.
Some were found to be involved in receptor, transport and signal
Stain Protein species detected
transducer regulated activities as well (Fig. 7A). Protein class analysis
using PANTHER indicated differentially expressed candidates among sPTB Control
the sPTB and term labour groups were mostly involved in enzyme
modulation, defence/immunity, and signalling (Fig. 7B). PANTHER Total proteins (cCBB) 620 ± 14 628 ± 15
Deep imaging 728 ± 16 734 ± 11
analysis also revealed the association of proteins in different biological Phosphoproteins (ProQ Diamond) 315 ± 18 303 ± 33
processes, including immunological, developmental, and metabolic Glycoproteins (ProQ Emerald) 198 ± 24 192 ± 18
(Fig. 7C). The identified differentially expressed proteins were linked to
4 pathways: blood coagulation, plasminogen activation, vitamin D Figures given are mean ± SEM.
metabolism, and angiogenesis (Fig. 7D).
determine therapeutic interventions. There is a substantial gap in our
understanding of the molecular mechanisms underlying PTB. To that
4. Discussion
end, we have utilized a top-down proteomic approach to analyse first
trimester maternal sera from pregnant women, including a targeted
The identification of first trimester biomarkers holds promise in
examination of PTMs, for detection of early markers of premature de-
both facilitating the early identification of those patients at greatest risk
livery. The data identify numerous protein species that are associated
and enhancing our understanding of the disease process(es) to

3 pI 10 3 pI 10

200 200
150 150
120 1C 120
100 100
85 85
9C
MW (KDa)

70 19 C 70 12 C
MW (KDa)

60 42 C 13 C 60 16 C
50 45 C
50
40 40 51 C
18 C
30 59 C 30
25 25
20 20

15 15
10 10

Term Preterm
Fig. 4. Representative 2DE gel images of sPTB and matched controls stained with ProQ Emerald to identify glycoproteins. Analysis was as described in Fig. 2. Numbers represent protein
spots of interest identified across three replicate 2DE gels; these are itemized in Table 5. Quantitative image analysis indicates protein spots that differ significantly in abundance
(increased abundance in SPTB labelled red; reduced abundance labelled green) t-test, p < 0.05, n = 10 each cohort.

5
A.M. D'Silva et al. Journal of Proteomics xxx (xxxx) xxx–xxx

with several clinically relevant biological processes, including inter-

No. of patients related biological networks linked to regulation of complement cas-
cade, metabolic processes, blood coagulation, signalling and immune
8/10
8/10
9/10
8/10
8/10
9/10
8/10
8/10
8/10
8/10
modulation (Fig. 7).
The most significant changes identified from the total protein profile
were related to complement activation (Table 3); there were increased
0.04/0.01
0.04/0.01
0.02/0.05
0.02/0.01
0.02/0.04
0.01/0.03
0.06/0.04
0.02/0.01
0.02/0.01
0.03/0.01
SD sPTB/

levels of complement factors (C4-A, factor B and H) and a decreased

level of complement C3 in the first trimester serum of women who
cont

delivered prematurely (Table 3). The current data are thus consistent
with previous investigations suggesting that inflammatory and immune
Fold change

related events in early pregnancy are part of the pathogenic mechan-

isms of sPTB [31,32]. Notably, patients selected in this study were
4.7
3.5
2.7
1.9
1.9
1.8
1.7
1.5
1.5
1.5

spontaneous idiopathic cases with no signs of intra-amniotic infection.

In contrast to previous studies, we observed a 1.9 fold decrease of C3 in
No. of unique peptides

sPTB compared to controls (Table 3). One possible reason for this could
be that our study was carried out on first trimester maternal serum in
comparison to studies that were carried out on late trimester biological
fluids. It would thus be of interest to assess larger populations at dif-
ferent time-points in pregnancy to understand the role of these com-
32
11
64
32

19
32

24
17
3

plement associated proteins in sPTB in particular as to whether or not

there is a ‘rebound’ effect due to low early levels of these mediators.
Seq. coverage %

Despite the literature suggesting a link between infection/in-

flammation and PTB and the central role complement plays in the pa-
thogenesis of infectious and inflammatory disease there has been lim-
30

73
17

22
36

17
17
6

ited research examining the relationship between complement

Total proteins stained by cCBB varying 1.5-fold or greater in abundance in sPTB and term control groups identified using 2DE coupled with LCMS/MS.

activation and PTB [33,34]. A systematic review indicated that certain

inflammatory cytokines in cervico-vaginal fluid (CVF) and in amniotic
Mascot protein

fluid, but not in plasma, are strongly associated with spontaneous PTB,
suggesting that inflammation at the maternal-fetal interface, rather
than systemic inflammation, may play a major role in the aetiology of
score

1947
1287
399
388

784
957
345
746
530
87

PTB [35]. Notably, here, we identified several proteins associated with

inflammation; indicating that inflammation/immune activation is in-
Calculated MW (kDa)/pI

itiated as early as 11 weeks of pregnancy in women subsequently ex-

periencing a spontaneous preterm delivery.
To our knowledge, this is the first study to investigate PTMs as early
as the first trimester of pregnancy. Here, several proteoforms with
100.3/4.4
133.6/3.9

133.6/4.1
47.8/9.7
57.9/9.1
30.3/3.3

92.0/4.4
48.4/3.5

77.6/5.3
28.4/3.9

significant differences in expression in sPTB vs. control group were

identified. Notably, a key finding here was the identification of vitamin
D binding protein (VDBP) and apolipoprotein A1 (ApoA1) across three
Theoretical MW (kDa)/pI

Three technical replicates were assessed for each of the 10 control and 10 preterm birth samples.

different assessments: namely total protein profiles, as well as phos-

phoprotein and glycoprotein detection; these occur as different species.
VDBP is a multifunctional serum protein with several physiological
roles [36]. Aside from its classical function of transporting and meta-
192.6/6.6

187.0/6.0
163.1/6.0

139.0/6.2
36.0/8.4

30.7/5.5

69.3/5.9
52.9/5.4

85.4/6.6
45.1/6.1

bolising vitamin D, VDBP regulates global placental function [37],

promotes actin clearance during tissue remodelling following the aug-
mentation of pro-inflammatory response [38], engages with immune
sPTB/cont

cells [39,40] and is a precursor macrophage activating factor [41].

Thus, VDBP is capable of modifying inflammation and protecting
↑sPTB
↑sPTB

↑sPTB
↑sPTB
↑sPTB
↑sPTB
↑sPTB
↑sPTB
↑cont
↑cont

against vascular dysfunction.

Bodnar et al. established that vitamin D is associated with PTB and
Ig gamma-1 chain C region

Vitamin D-binding protein

that this relationship is similar in spontaneous and medically indicated

Alpha-2-macroglobulin

PTB [42]. Previous research has documented increased levels of VDBP

Complement factor H
Complement factor B
Apolipoprotein A-I
Complement C4-A

in the CVF from mid to late pregnancy as an indicator of spontaneous

Protein identified

Complement C3

Serum albumin

labour [43,44]; it is suggested that the increase in VDBP may be in-

Haptoglobin

dicative of increased inflammation and tissue damage integral to cer-

vical remodelling and fetal membrane weakening with approaching
labour onset. The data here indicated decreased levels of phosphory-
lated VDBP but elevated levels of total and glycosylated VDBP species
Gene names

in the sPTB cohort (Tables 4 and 5). Post-translationally modified forms

APOA1
IGHG1

of VDBP have not been previously reported in the literature. Interest-

A2M

CFH
C4A

ALB

CFB

ingly, a dual biomarker model of albumin/VDBP in CVF was found to

HP
GC
C3

be more effective than fetal fibronectin (currently used clinically as a

Spot number

diagnostic marker) in predicting sPTB in the second trimester although

the critical proteoform of either protein was not established [45].
151 A
144 A
105 A

106 A
Table 3

19 A

13 A
73 A

29 A

ApoA1 is the predominant high density lipoprotein and is a primary

6A

8A

acceptor for cholesterol [46,47]. Although ApoA1 is dysregulated in

6
A.M. D'Silva et al. Journal of Proteomics xxx (xxxx) xxx–xxx

Table 4
Alterations in the phosphoproteome varying 1.5-fold or greater in abundance in sPTB and term control groups.

Calculated Mascot Seq. No. of

Spot UniProt Protein Theorecal Fold SD No. of
sPTB/cont MW protein coverage unique
ID ID idenfied MW (kDa)/pI change sPTB/cont paents
(kDa)/pI score % pepdes
Alpha-1-
31 B SERPINA1 ↑sPTB 46.7/5.37 45.6/3.10 2201 59 73 3.2 0.21/0.10 9/10
antrypsin

46 B VTN Vitronecn ↑sPTB 54.3/5.55 61.7/2.94 83 6 5 3.0 0.12/0.18 8/10

Ig alpha-1
24 B IGHA1 ↑sPTB 37.6/6.08 51.6/3.13 78 7 3 3.0 0.16/0.22 8/10
chain C region
22 B ALB Serum albumin ↑sPTB 69.3/5.92 52.3/3.37 570 34 37 2.8 0.10/0.20 8/10

39 B HP Haptoglobin ↑sPTB 45.2/6.13 39.8/3.49 668 35 42 2.4 0.02/0.04 9/10

8B PL Plasminogen ↑sPTB 90.5/7.04 77.6/6.74 110 3 5 2.1 0.02/0.04 8/10

Vitamin D-
34 B GC ↑cont 53.0/5.4 48.4/3.76 937 35 52 2.0 0.02/0.04 8/10
binding protein

Apolipoprotein
44 B APOA1 ↑sPTB 31.0/5.56 61.7/3.13 134 36 7 1.7 0.05/0.06 8/10
A-I

Complement
48 B C4A ↑sPTB 192.7/6.65 33.0/6.37 411 3 14 1.7 0.02/0.04 9/10
C4-A

Rows highlighted in blue indicate proteins identified in the total protein analyses (Table 3), green indicates proteins also identified in the glycoprotein analyses (Table 5) and orange
indicates proteins identified in both total protein and glycoprotein analyses.

Table 5
Alterations in the glycoproteome varying 1.5-fold or greater in abundance in sPTB and term control groups identified using 2DE coupled with LCMS/MS.

Calculated Mascot Seq. No. of

Spot UniProt Protein Theorecal Fold SD No. of
sPTB/cont MW protein coverage unique
ID ID idenfied MW (kDa)/pI change sPTB/cont paents
(kDa)/pI score % pepdes

19 C ALB Serum albumin ↑sPTB 69.3/5.92 61.7/4.78 1499 36 107 9.3 0.17/0.03 8/10

Vitamin D-binding
51C GC ↑sPTB 53.0/5.4 47.8/3.79 270 26 25 5.9 0.06/0.02 8/10
protein

59 C APOA1 Apolipoprotein A-I ↑sPTB 30.8/5.56 30.0/3.70 2670 73 99 3.3 0.04/0.03 8/10

Ig gamma-1 chain
45 C IGHG1 ↑cont 36.1/8.46 51.6/3.70 333 32 22 3.3 0.02/0.06 8/10
C region

Ig mu chain C
9C IGHM ↑cont 49.4/6.35 70.0/5.57 421 23 14 2.8 0.01/0.06 8/10
region

Alpha-1-
1C SERPINA1 ↑sPTB 46.7/5.37 103.5/3.37 876 38 22 2.5 0.01/0.09 9/10
antrypsin

13 C TF Serotransferrin ↑cont 77.0/6.81 65.0/5.62 1400 25 41 2.4 0.06/0.10 9/10

Alpha-1B-
18 C A1BG ↑cont 54.2/5.56 41.4/3.58 362 21 25 2.3 0.05/0.09 8/10
glycoprotein
Complement
16 C C3 ↑cont 187.0/6.02 61.7/6.74 475 13 4 2.0 0.02/0.07 9/10
C3
Anthrombin-
42 C SERPINC1 ↑cont 52.6/6.32 53.1/3.64 632 26 5 2.0 0.05/0.07 8/10
III
Inter-alpha-
trypsin
12 C ITIH4 ↑sPTB 103.8/6.43 65.0/3.35 394 19 10 1.8 0.05/0.04 8/10
inhibitor heavy
chain H4

Rows highlighted in blue represents proteins identified in the total protein analyses (Table 3), green represent proteins also identified in the phosphoprotein analyses (Table 4) and orange
represent proteins identified in both total protein and phosphoprotein analyses.

7
A.M. D'Silva et al.
A B
3 pI
200
150
120
100
85
MW (KDa)
70
60
50
40
30
25
20
15
10
Fig. 5. Representative images of an initial 2DE resolved proteome used for excision of high abundance regions (A), and a deep imaged serum proteome after excision of highly saturating
regions (B).
diverse tissues and body fluids in a variety of diseases [48,49], there is
little information concerning its relation to PTB. Previous studies have
shown a potential role of ApoA1 in pregnancy complications such as
preeclampsia and intrauterine growth restriction [50,51]. A very recent
study has shown lower levels of ApoA1 in the cord blood profile of
preterm infants [52]. In the same study, the authors report a sig-
nificantly higher ApoB/ApoA1 index in preterm neonates although,
again, critical proteoforms were not identified. Our data confirm dis-
tinctly increased serum levels of phosphorylated and glycosylated
ApoA1 in the sPTB cohort. Longitudinal studies on both prenatal (i.e.
maternal serum, urine, CVF) and post-natal samples (cord blood) would
help us to understand the significance of this protein in PTB and the
consequences of these differences across these biological fluids.
Additional proteins of interest identified in our study include alpha-
1-antitrypsin and alpha-1β-glycoprotein. Alpha-1-antitrypsin was
identified as clearly different phosphorylated and glycosylated species
whereas alpha-1β-glycoprotein was identified only as glycosylated
proteoform. Levels of alpha-1-antitrypsin and alpha-1β-glycoprotein
increase in response to systemic tissue injury, inflammation or infection
[53,54] and changes in protein concentrations have been correlated
with PTB in mid-to-late gestation. For both proteins, our results are
consistent with previous studies involving proteomic profiling in am-
niotic fluid and CVF in the second trimester to understand mechanisms
leading to PTB [55–58].
A further notable finding of this study was the identification of
single-chain glycoprotein antithrombin III (AT) (Table 5). Thrombin
plays a central role in the coagulation cascade and participates in pla-
telet activation and transforming fibrinogen into fibrin [59]. The re-
action of thrombin with its major inhibitor (AT), results in the forma-
tion of an inactive stable complex, the thrombin-antithrombin III (TAT)
complex. The TAT complex is widely accepted to assess thrombin ac-
tivation; AT complexes with vitronectin (VTN) causing a conforma-
tional change facilitating heparin binding [60].
Elevated concentrations of TAT in the maternal circulation have
been reported in patients with PTB relative to women with a normal
pregnancy [61,62]. The most likely reason for decreased levels of AT in
our study could be that the samples analysed were collected at a dif-
ferent time point (i.e. in the first trimester of pregnancy) in contrast to
the second trimester of pregnancy. An important implication of our
findings is the possibility that the TAT complex concentration only
starts to increase as the pregnancy advances in response to the activa-
tion of the fibrinolytic cascade, suggesting perhaps a complicating ‘re-
bound’ effect of these proteins after the first trimester of pregnancy.
This is the first study to analyse VTN expression in first trimester
Journal of Proteomics xxx (xxxx) xxx–xxx
10 3 pI 10
200
150
B C
120
100
85 A D
MW (KDa)
70
60
50 F
40
30
25 E
20
15
10
maternal circulation and its association with PTB (Table 4). VTN is not
only involved in adhesion, but plays a significant role in tissue repair,
regulation of coagulation and fibrinolysis, and immune defence [63].
VTN is essential for maintaining the haemostasis and deficiency in this
protein is associated with increased risk for complications such as
thromboembolic diseases [64]. Increased level of this glycoprotein may
stabilize cell membrane damage through its adhesive properties thus
maintaining blood vessel integrity. As such, these processes may play a
preventive role in the occurrence and development of PTB. In light of
the association between proteins of the coagulation system (AT and
VTN) and sPTB, the role of these as potential predictive markers PTB is
quite relevant.
Previously, our group identified candidate placental proteins asso-
ciated with PTB [30]. These proteins were differentially characterized
at the time of preterm (25–32 weeks) or term (≥37 weeks) delivery.
Despite different sample types and time points, comparison of the
findings with the current study is illustrative of the cascade of events
leading to PTB. Both studies identified changes in proteins involved in
the coagulation cascade; annexin A4 was not detected in the preterm
placenta [30], whereas the current study identified decreased levels of
AT in the maternal serum of the PTB cohort. Furthermore, here we
identified an increase in VDBP. Butt et al. reported actin changes in
placental tissue in both PTB (membrane fraction) and the term (soluble
fraction) conditions [30]. VDBP has high affinity binding sites for actin
[65]. The dynamic VDBP-actin complex is found in the serum of hu-
mans and animals that have sustained injuries and/or inflammation
(e.g., trophoblastic emboli, severe acute hepatitis, acute lung injury)
[66]. However, the role of this complex in PTB has yet to be confirmed.
Several protein species identified in our study have been reported in
amniotic fluids and CVF, however, these are somewhat invasive tests,
and relatively late in the gestational process, when PTB may be in-
evitable. It would be worthwhile to investigate PTMs in different bio-
logical fluids to improve our understanding of how these are associated
with PTB and other gestational disorders. PTMs identified in our study
potentially indicate alterations to different regulatory pathways in sPTB
such as RAF/MAP kinase cascade, MAP kinase activation in TLR cas-
cade, reelin signalling pathway, and the adenylate cyclase inhibitor
pathway. It is thus likely that no single biomarker will ever achieve the
desired predictive efficacy due to the multifaceted aetiology of PTB.
Therefore, combined use of markers has been shown to have better
predictive accuracy than individual markers alone. As such, identifying
key proteins and their PTMs help improve our understanding of PTB
and may also prove to be attractive targets for therapeutic interven-
tions.
8
A.M. D'Silva et al. Journal of Proteomics xxx (xxxx) xxx–xxx

3 pI 10 3 pI 10

A1-A5 B1-B3

3 pI 10 3 pI 10

C1-C3 D1-D5
pI 10 3 pI 10
3

D6-D10 E1-E3
pI 10 pI 10
3 3

F1-F5 F6-F8
(caption on next page)

9
A.M. D'Silva et al. Journal of Proteomics xxx (xxxx) xxx–xxx

Fig. 6. Representative images of third-dimension separations of high-density (i.e., hyper-abundant) protein regions excised from serum proteomes resolved by 2DE; red circles indicate
protein species resolved from co-migrating hyperabundant proteins. Designations A-F refer to excised gel regions (see Fig. 5), and in each case the associated numbers refer to specific
subsections of those excised gel pieces (i.e., A1–A5 indicates that excised region A was subdivided into five approximately equal sized gel pieces that were then resolved in parallel on
third gels (see Materials and methods)).

(A) Molecular funcon (B) Protein class

transfer/carrier protein cell adhesion molecule

9.5 % 4.8 %

catalyc acvity signalling molecule

38.1% 14.3 % defence/immunity protein
binding 23.8 %
42.9%

receptor
9.5 %

receptor acvity hydrolase

9.5 % 9.5 %
28.6 %
transporter acvity signal transducer acvity
4.8 % 4.8 %

(C) Biological process (D) Pathway

biological adhesion
4.1%
response to smulus blood coagulaon
biological regulaon 62.5 %
18.4 %
12.2%

mulcellular
or biogenesis
organismal process
6.1%
14.3 %
angiogenesis
12.5 %

metabolic process cellular process

16.3% 22.4 %

vitamin D metabolism and pathway

12.5 % plasminogen acvang cascade
12.5 %
10.2 %
developmental process
2.0 %
immune system process
2.0 %

Fig. 7. Functional clustering and physiological pathways associated with the differentially expressed proteins identified in sPTB. Pie charts showing the protein classes (A) molecular
function, (B) protein class, (C) biological process, and (D) pathway.

5. Conclusions
This study has shown that several protein species associated with
clinically relevant biological processes are altered in first trimester
serum from pregnancies that are subsequently complicated by sPTB.
The data identify a series of biomarker candidates potentially predictive
of sPTB and also highlight potential mechanisms that contribute to the
pathophysiology of PTB. Notably, the identified PTMs appear as early
as 11 weeks' gestation and are thus likely strongly predictive of sPTB.
Despite the sample size (i.e. 10 sPTB cases and 10 matched con-
trols), the uniformity and reproducibility of the results serve to high-
light the likely importance of the identified protein species and specific
PTMs as biomarkers for sPTB. The methods used also provide detection
sensitivity in the low-to-sub femtomole range, further emphasizing the
quantitative rigor of the approach [67]. The identification of key
protein species at the 11–13-week time point suggests that early serial
sampling of patients may provide even more in-depth knowledge of the
progression of the disease and prove to be useful in determining when
select interventions (e.g. antenatal steroids) should be administered.
Thus, the strengths of this study are: (i) evaluation of early gestational
serum samples in a well-characterized and relatively homogenous po-
pulation of women using a refined, sensitive, and well-established top-
down proteomic approach; and (ii) only patients with spontaneous
preterm delivery were included. The next phase of this work will be to
validate the data with independent cohorts and further integrate the
proteomic data with demographic information to develop additional
reliable predictive tools for PTB. The predictive usefulness of the cur-
rent potential biomarkers to detect subtypes of sPTB as well as other
gestational disorders (e.g. preeclampsia) also warrants detailed study.

10
A.M. D'Silva et al.
Transparency document
The Transparency document associated with this article can be
found, in online version.
Acknowledgments
The authors acknowledge the School of Medicine (Western Sydney
University) for support and RPA Women and Babies (Royal Prince
Alfred Hospital) for providing maternal serum samples. AMD ac-
knowledges a WSU School of Medicine Research Scholarship. The au-
thors thank the Cerebral Palsy Alliance for funding to assess PTMs.
Author contributions
A.M.D. and J.R.C. conceived and designed the experiments; J.A.H.
collected serum samples; A.M.D. performed the experiments; A.M.D.
and J.R.C. analysed the data; A.M.D., J.R.C., and J.A.H. wrote the
paper. All authors read and approved the final manuscript.
Conflicts of interest
The authors declare no conflict of interest.
Appendix A. Supplementary data
Supplementary data associated with this article can be found in the
online version, at http://galen.westernsydney.edu.au/mmrg/DSilva/
proj1_jpr/.
References
[1] H. Blencowe, et al., National, regional, and worldwide estimates of preterm birth
rates in the year 2010 with time trends since 1990 for selected countries: a sys-
tematic analysis and implications, Lancet 379 (9832) (2012) 2162–2172.
[2] J.E. Lawn, S. Cousens, J. Zupan, 4 million neonatal deaths: when? Where? Why?
Lancet 365 (9462) (2005) 891–900.
[3] F. Soleimani, F. Zaheri, F. Abdi, Long-term neurodevelopmental outcomes after
preterm birth, Iran Red Crescent Med J 16 (6) (2014) e17965.
[4] H. Carr, et al., Preterm birth and risk of heart failure up to early adulthood, J. Am.
Coll. Cardiol. 69 (21) (2017) 2634–2642.
[5] V. Mericq, et al., Long-term metabolic risk among children born premature or small
for gestational age, Nat. Rev. Endocrinol. 13 (1) (2017) 50–62.
[6] M.S. Esplin, Overview of spontaneous preterm birth: a complex and multifactorial
phenotype, Clin. Obstet. Gynecol. 57 (3) (2014) 518–530.
[7] E.J. Hodgson, C.J. Lockwood, Preterm birth: a complex disease, Preterm Birth,
Wiley-Blackwell, 2010, pp. 8–16.
[8] N.F.V. Ardala Breda, Osmar Norberto de Souza, Richard Charles Garratt, Protein
structure, modelling and applications, in: D.A. Gruber, C. Huynh, et al. (Eds.),
Bioinformatics in Tropical Disease Research: A Practical and Case-Study Approach,
2006.
[9] P.R. Jungblut, et al., Proteomics in human disease: cancer, heart and infectious
diseases, Electrophoresis 20 (10) (1999) 2100–2110.
[10] J.R. Coorssen, A.L. Yergey, Proteomics is analytical chemistry: fitness-for-purpose
in the application of top-down and bottom-up analyses, Proteome 3 (4) (2015)
440–453.
[11] B.M. Oliveira, J.R. Coorssen, D. Martins-de-Souza, 2DE: the phoenix of proteomics,
J. Proteome 104 (2014) 140–150.
[12] B. Thiede, et al., High resolution quantitative proteomics of HeLa cells protein
species using stable isotope labeling with amino acids in cell culture (SILAC), two-
dimensional gel electrophoresis (2DE) and nano-liquid chromatograpohy coupled
to an LTQ-OrbitrapMass spectrometer, Mol. Cell. Proteomics 12 (2) (2013)
529–538.
[13] G.F. Clark, Functional glycosylation in the human and mammalian uterus, Fertil.
Res. Pract. 1 (2015) 17.
[14] R. Romero, et al., The preterm parturition syndrome, BJOG 113 (Suppl. 3) (2006)
17–42.
[15] M.S. To, et al., Prediction of patient-specific risk of early preterm delivery using
maternal history and sonographic measurement of cervical length: a population-
based prospective study, Ultrasound Obstet. Gynecol. 27 (4) (2006) 362–367.
[16] B.M. Mercer, et al., The preterm prediction study: a clinical risk assessment system,
Am. J. Obstet. Gynecol. 174 (6) (1996) 1885–1895.
[17] J.D. Iams, et al., The length of the cervix and the risk of spontaneous premature
delivery, N. Engl. J. Med. 334 (9) (1996) 567–573.
[18] V.C. Heath, et al., Cervical length at 23 weeks of gestation: prediction of sponta-
neous preterm delivery, Ultrasound Obstet. Gynecol. 12 (5) (1998) 312–317.
11
Journal of Proteomics xxx (xxxx) xxx–xxx
[19] C.J. Lockwood, et al., Fetal fibronectin in cervical and vaginal secretions as a pre-
dictor of preterm delivery, N. Engl. J. Med. 325 (10) (1991) 669–674.
[20] F. Park, et al., Prediction and prevention of early-onset pre-eclampsia: impact of
aspirin after first-trimester screening, Ultrasound Obstet. Gynecol. 46 (4) (2015)
419–423.
[21] D.L. Rolnik, D. Wright, L.C. Poon, N. O'Gorman, A. Syngelaki, C. de Paco Matallana,
et al., Aspirin versus Placebo in Pregnancies at High Risk for Preterm Preeclampsia,
N. Engl. J. Med. 377 (7) (2017) 613–622.
[22] A.M. D'Silva, J.A. Hyett, J.R. Coorssen, A routine ‘top-down’ approach to analysis of
the human serum proteome, Proteome 5 (2) (2017).
[23] V.J. Gauci, M.P. Padula, J.R. Coorssen, Coomassie blue staining for high sensitivity
gel-based proteomics, J. Proteome 90 (2013) 96–106.
[24] R.H. Butt, J.R. Coorssen, Postfractionation for enhanced proteomic analyses: rou-
tine electrophoretic methods increase the resolution of standard 2D-PAGE, J.
Proteome Res. 4 (3) (2005) 982–991.
[25] E.P. Wright, et al., Deep imaging: how much of the proteome does current top-down
technology already resolve? PLoS One 9 (1) (2014) e86058.
[26] E.P. Wright, et al., Top-down proteomics: enhancing 2D gel electrophoresis from
tissue processing to high-sensitivity protein detection, Proteomics 14 (7–8) (2014)
872–889.
[27] M.A. Partridge, et al., An initial top-down proteomic analysis of the standard cu-
prizone mouse model of multiple sclerosis, J. Chem. Biol. 9 (1) (2016) 9–18.
[28] R.H. Butt, J.R. Coorssen, Pre-extraction sample handling by automated frozen dis-
ruption significantly improves subsequent proteomic analyses, J. Proteome Res. 5
(2) (2006) 437–448.
[29] R.H. Butt, et al., Enabling coupled quantitative genomics and proteomics analyses
from rat spinal cord samples, Mol. Cell. Proteomics 6 (9) (2007) 1574–1588.
[30] R.H. Butt, et al., An initial proteomic analysis of human preterm labor: placental
membranes, J. Proteome Res. 5 (11) (2006) 3161–3172.
[31] A.M. Lynch, et al., Complement activation fragment Bb in early pregnancy and
spontaneous preterm birth, Am. J. Obstet. Gynecol. 199 (4) (2008) 354.e1–354.e8.
[32] A.M. Lynch, et al., The relationship of circulating proteins in early pregnancy with
preterm birth, Am. J. Obstet. Gynecol. 214 (4) (2016) 517.e1–517.e8.
[33] A. Elimian, et al., Amniotic fluid complement C3 as a marker of intra-amniotic
infection, Obstet. Gynecol. 92 (1) (1998) 72–76.
[34] E. Soto, et al., Evidence for complement activation in the amniotic fluid of women
with spontaneous preterm labor and intra-amniotic infection, J. Matern. Fetal
Neonatal Med. 22 (11) (2009) 983–992.
[35] S.Q. Wei, W. Fraser, Z.C. Luo, Inflammatory cytokines and spontaneous preterm
birth in asymptomatic women: a systematic review, Obstet. Gynecol. 116 (2 Pt 1)
(2010) 393–401.
[36] R.F. Chun, New perspectives on the vitamin D binding protein, Cell Biochem. Funct.
30 (6) (2012) 445–456.
[37] J.K. Cleal, et al., Placental amino acid transport may be regulated by maternal vi-
tamin D and vitamin D-binding protein: results from the Southampton Women's
Survey, Br. J. Nutr. 113 (12) (2015) 1903–1910.
[38] S.E. Lind, et al., Role of plasma gelsolin and the vitamin D-binding protein in
clearing actin from the circulation, J. Clin. Invest. 78 (3) (1986) 736–742.
[39] M. Petrini, et al., Gc (vitamin D binding protein) binds to cytoplasm of all human
lymphocytes and is expressed on B-cell membranes, Clin. Immunol. Immunopathol.
31 (2) (1984) 282–295.
[40] R.R. Kew, et al., Localization and quantitation of the vitamin D binding protein (Gc-
globulin) in human neutrophils, Blood 82 (1) (1993) 274–283.
[41] H. Nagasawa, et al., Gc protein (vitamin D-binding protein): Gc genotyping and
GcMAF precursor activity, Anticancer Res. 25 (6A) (2005) 3689–3695.
[42] L.M. Bodnar, R.W. Platt, H.N. Simhan, Early-pregnancy vitamin D deficiency and
risk of preterm birth subtypes, Obstet. Gynecol. 125 (2) (2015) 439–447.
[43] S. Liong, et al., Is vitamin D binding protein a novel predictor of labour? PLoS One 8
(10) (2013) e76490.
[44] H.-N. Yoo, et al., Non-invasive prediction of preterm birth in women with cervical
insufficiency or an asymptomatic short cervix (≤25 mm) by measurement of bio-
markers in the cervicovaginal fluid, PLoS One 12 (7) (2017) e0180878.
[45] S. Liong, et al., New biomarkers for the prediction of spontaneous preterm labour in
symptomatic pregnant women: a comparison with fetal fibronectin, BJOG Int. J.
Obstet. Gynaecol. 122 (3) (2015) 370–379.
[46] E. Taylor, V. Gomel, The uterus and fertility, Fertil. Steril. 89 (1) (2008) 1–16.
[47] R. Kiwi, Recurrent pregnancy loss: evaluation and discussion of the causes and their
management, Cleve. Clin. J. Med. 73 (10) (2006) 913–921.
[48] M.K. Bohlmann, et al., Hysteroscopic findings in women with two and with more
than two first-trimester miscarriages are not significantly different, Reprod.
BioMed. Online 21 (2) (2010) 230–236.
[49] S.E. Brown, et al., Evaluation of outpatient hysteroscopy, saline infusion hyster-
osonography, and hysterosalpingography in infertile women: a prospective, ran-
domized study, Fertil. Steril. 74 (5) (2000) 1029–1034.
[50] U. Rosing, et al., Serum levels of apolipoprotein A-I, A-II and HDL-cholesterol in
second half of normal pregnancy and in pregnancy complicated by pre-eclampsia,
Horm. Metab. Res. 21 (7) (1989) 376–382.
[51] N. Radunovic, et al., Plasma apolipoprotein A-I and B concentrations in growth-
retarded fetuses: a link between low birth weight and adult atherosclerosis, J. Clin.
Endocrinol. Metab. 85 (1) (2000) 85–88.
[52] T. Katragadda, et al., Comparison of cord blood lipid profile in preterm small for
gestational age and appropriate for gestational age newborns, J. Clin. Diagn. Res.
11 (1) (2017) SC05–SC07.
[53] J.M. Hunt, R. Tuder, Alpha 1 anti-trypsin: one protein, many functions, Curr. Mol.
Med. 12 (7) (2012) 827–835.
[54] T. Fournier, N.N. Medjoubi, D. Porquet, Alpha-1-acid glycoprotein, Biochim.
A.M. D'Silva et al.
Biophys. Acta 1482 (1–2) (2000) 157–171.
[55] N. Izumi-Yoneda, et al., Alpha 1 antitrypsin activity is decreased in human amnion
in premature rupture of the fetal membranes, Mol. Hum. Reprod. 15 (1) (2009)
49–57.
[56] L. Pereira, et al., Identification of novel protein biomarkers of preterm birth in
human cervical−vaginal fluid, J. Proteome Res. 6 (4) (2007) 1269–1276.
[57] I.A. Buhimschi, et al., Multidimensional proteomics analysis of amniotic fluid to
provide insight into the mechanisms of idiopathic preterm birth, PLoS One 3 (4)
(2008) e2049.
[58] J. Hitti, et al., Noninvasive diagnosis of intraamniotic infection: proteomic bio-
markers in vaginal fluid, Am. J. Obstet. Gynecol. 203 (1) (2010) 32.e1–32.e8.
[59] J.T. Crawley, et al., The central role of thrombin in hemostasis, J. Thromb.
Haemost. 5 (Suppl. 1) (2007) 95–101.
[60] R.D. Rosenberg, P.S. Damus, The purification and mechanism of action of human
antithrombin-heparin cofactor, J. Biol. Chem. 248 (18) (1973) 6490–6505.
[61] M.A. Elovitz, J. Baron, M. Phillippe, The role of thrombin in preterm parturition,
12
Journal of Proteomics xxx (xxxx) xxx–xxx
Am. J. Obstet. Gynecol. 185 (5) (2001) 1059–1063.
[62] T. Rosen, et al., Plasma levels of thrombin-antithrombin complexes predict preterm
premature rupture of the fetal membranes, J. Matern. Fetal Med. 10 (5) (2001)
297–300.
[63] B. Singh, Y.-C. Su, K. Riesbeck, Vitronectin in bacterial pathogenesis: a host protein
used in complement escape and cellular invasion, Mol. Microbiol. 78 (3) (2010)
545–560.
[64] K.T. Preissner, D. Seiffert, Role of vitronectin and its receptors in haemostasis and
vascular remodeling, Thromb. Res. 89 (1) (1998) 1–21.
[65] R. Bouillon, et al., The isolation and characterization of the 25-hydroxyvitamin D-
binding protein from chick serum, J. Biol. Chem. 255 (22) (1980) 10925–10930.
[66] W.M. Lee, R.M. Galbraith, The extracellular actin-scavenger system and actin
toxicity, N. Engl. J. Med. 326 (20) (1992) 1335–1341.
[67] N. Noaman, et al., Coomassie staining provides routine (sub)femtomole in-gel de-
tection of intact proteoforms: expanding opportunities for genuine Top-down
Proteomics, Electrophoresis 38 (2017) 3086–3099.