BOR Papers in Press. Published on June 24, 2015 as DOI:10.1095/biolreprod.115.128231

Applying SWATH Mass Spectrometry to Investigate Human Cervicovaginal Fluid During the Menstrual Cycle1 Kanchan Vaswani,3,4 Keith Ashman,3,4 Sarah Reed,5 Carlos Salomon Gallo,4 Suchismita Sarker,4 Jose A. Arraztoa,6 Alejandra Pérez-Sepúlveda,6 Sebastian E. Illanes,4,6 David Kavaskoff,4 Murray D. Mitchell,4 and Gregory E. Rice2,4 4 University of Queensland Centre for Clinical Research, Herston, Queensland, Australia 5 AB SCIEX, Mt Waverley, Victoria, Australia 6 Laboratorio Biología de la Reproducción , Facultad de Medicina, Universidad de los Andes, Chile 1

Financial support for this project provided by Therapeutics Innovations Australia (TIA), the Australian Government Smart Futures Scheme and the Queensland State Government Smart Futures Fund. G.E.R. was in receipt of an NHMRC Principal Research Fellowship. The worked detailed herein was funded, in part, by a University of Queensland Commercial Industry Engagement Fund award (G.E.R.) and by RoCan (http://rocan.com.au). 2 Correspondence: Greg Rice, Building 71/918, Royal Brisbane Hospital, Faculty of Medicine and Biomedical Sciences, University of Queensland, Herston QLD 4029, Australia 3 These authors contributed equally to this work. ABSTRACT Inherent inter- and intra-individual variation in the length of the menstrual cycle limits the accuracy of predicting days of peak fertility. To improve detection of days of peak fertility, a more detailed understanding of longitudinal changes in cervicovaginal fluid (CVF) biomarkers during the normal menstrual cycle is needed. The aim of this study, therefore, was to characterize longitudinal changes in CVF proteins during the menstrual cycle using a quantitative, data-independent acquisition mass spectrometry approach. Six serial samples were collected from women (n=10) during the menstrual cycle. Samples were obtained at two time points for each phase of the cycle: early and late pre-ovulatory, ovulatory and post-ovulatory. Information-Dependent Acquisition (IDA) of mass spectra from all individual CVF samples was initially performed and identified 278 total proteins. Samples were then pooled by time of collection (n=6 pools) and analyzed using IDA and information-independent acquisition (SWATH). The IDA library generated contained 176 statistically significant protein identifications (p < 0.000158). The variation in the relative abundance of CVF proteins across the menstrual cycle was established by comparison with the SWATH profile against the IDA library. Using time-series, pooled samples obtained from 10 women, quantitative data were obtained by SWATH analysis for 43 CVF proteins. Of these proteins, 28 displayed significant variation in relative abundance during the menstrual cycle (assessed by ANOVA). Statistical significant changes in the relative expression of CVF proteins during pre-ovulatory, ovulatory and post-ovulatory phases of menstrual cycle were identified. The data obtained may be of utility not only in elucidating underlying physiological mechanisms but also as clinically useful biomarkers of fertility status. INTRODUCTION Accurate prediction of the length and phases of the menstrual cycle is fundamental to the success of fertility awareness-based methods for the management of infertility and natural family

1 Copyright 2015 by The Society for the Study of Reproduction.

planning [1]. The efficacy of fertility awareness-based methods that rely on counting the days of the menstrual cycle, however, is affected by variability in cycle length, the timing of ovulation and the window of peak fertility [2, 3]. For example, menstrual cycle length in normal healthy women varies from 21-35 days and within individuals, maximum cycle length may vary by 6 to 11 days [4]. Cycle length is also affected by recent pregnancies or childbirth, breastfeeding, menarche or menopause, inherent cycle variation, or illness. The incidence of irregular cycles varies from 9 - 43% and is age related [3]. The window of fertility during the menstrual cycle spans from five days prior to ovulation to 1 day after ovulation [5]. Given these multiple sources of variation, the ability of women to successfully identify ovulation within the fertile period of the menstrual cycle is only ~55% and less than 30% on days of peak fertility. In addition, FAMBs, such as the Rhythm Method, require an accurate history of the menstrual cycles over the previous 8 to 12 months to be considered reliable [6]. Even with such records they are associated with unplanned pregnancy rates of up to 18% [7]. Cervicovaginal fluid has been a valuable source of clinical information about the physiological and pathophysiological status of the female reproductive tract in both non-pregnant and pregnant women. Changes in the physical properties of cervical secretions are predictive of ovulation [8, 9] and monitoring such changes have been used successfully in fertility awareness-based methods, including the Billings Ovulation Method [10] and Creighton Model [11]. As changes in the physiochemical composition of cervical secretions reflect underlying physiological processes, these methods are associated with lower rates of unplanned pregnancies. To improve the accuracy of such approaches to identify days of peak fertility requires a more detailed understanding of longitudinal (i.e serial) changes in cervicovaginal fluid biomarkers during the normal menstrual cycle. Currently, no such data are available. To date, menstrual phasespecific changes in cervicovaginal fluid proteins and inter-subject variability during the menstrual cycle remain to be established. In the absence of such baseline data, the physiological and pathophysiological significance of documenting cervicovaginal fluid proteomes is unknown. Previous studies that utilized mass spectrometry-based approaches have only identified peptides and proteins and not applied quantitative methods. To date, most studies that have attempted to characterize cervicovaginal fluid proteins have used samples obtained from pregnant and or laboring women [12-14] or in association with cervical cancer [15]. Such samples would be expected to display very different proteomes that those obtained from non-pregnant women. Indeed, there is a paucity of data to establish any concordance between the quantitative expression of proteins in cervicovaginal fluid obtained from non-pregnant and pregnant women. The aim of this study, therefore, was to establish the quantitative and phase-specific expression of proteins present in cervicovaginal fluid during the normal menstrual cycle (i.e. non-pregnant women). Two unique aspects of this study are: the collection and analysis of serial (i.e. longitudinal) samples of cervicovaginal fluid from individual women; and the application of a quantitative mass spectrometry profiling method (SWATH MS- sequential windowed acquisition of all theoretical mass spectra) to identify changes in protein abundance during the different phases (pre-ovulatory, ovulatory and post-ovulatory) of the menstrual cycle. SWATH-MS is a data-independent acquisition method that systematically performs MS/MS in the mass range (350-1,000 Da), repeatedly cycling in 25 Da increments over the entire chromatographic elution profile [16]. This generates a complete list of fragment ions and provides quantitative data of peptides peaks, identified in a preliminary IDA run, to determine

2

peptide charge and retention time information. Thus, this provides an unbiased quantitative tool to compare the relative abundance of targeted peptides and identified proteins in a sample. SWATH establishes a comprehensive and permanent digital record of the fragment ion spectra of all the analytes in a biological sample for which the precursor ions are within a predetermined m/z versus retention time window [17] and allows targeted extraction of quantitative data from the spectral libraries [17-23]. The application of both SWATH MS and SRM (Selected Reaction Monitoring) requires the a priori generation of reference spectral maps by IDA that provide spectral coordinates for quantification. Herein, we demonstrate that the application of the mass spectrometric reference maps and the acquisition of SWATH maps to the analysis of human cervicovaginal fluid holds promise for accelerating the process of biomarker discovery. In particular, the clinical significance of not only defining the cervicovaginal fluid proteome but also quantifying phase-specific changes in protein abundance during the menstrual cycle resides in more accurate identification of female fertility status. Such data may be of utility in developing more effective multimarker in vitro diagnostics and point-of-care devices for the accurate and timely identification of periods of maximum fertility. Furthermore, the methods identified phase-specific CVF proteins not previously associated with the menstrual cycle. MATERIALS AND METHODS Reagents Trypsin Gold Mass Spectrometry Grade (Cat no. V5280) was purchased from Promega Madison, WI, USA. Ammonium bicarbonate salt (Cat no. A6141-500g >99%), PNGase (Cat no. P3620), MS reaction buffer (Cat no. R0154), formic acid (1ml Ampules ~98%; Cat no. 56302-10X1MLF Fluka), iodoacetamide (Cat no. I1149), ultrapure dithiothreitol (DTT, Cat no.15508-013) bovine serum albumin (BSA, Cat no A2153-10G), Empore™ C18 (Octadecyl) SPE Extraction Disk (Sigma Cat no 66883-U), acetonitrile (Cat no. 34998-2.5L), trifluoroacetic acid reagent plus (99%; Cat no T6508), HPLC water (Cat no 95304-2.5L); 2-propanol (Cat no 19516- 25ml), methanol (Cat no. 34860-2.5L-R) Chromasolv for HPLC, bicinchoninic assay kit (Cat no. BCA1-1KT) were obtained from Sigma-Aldrich St Louis, Mo, USA Poros Oligo R3 Bulk Media Reverse Phase Packing (Cat no. 1133226) was purchased from Applied Biosystems Foster City CA 94404. Protein Extraction Reagent M-PER (Cat no. 78501) was purchased from Thermo Fisher (Scoresby Vic, Australia) Study Design The study design was a prospective cohort study (Figure 1) and was performed under ISO17025 requirements in a NATA accredited research facility. Serial samples of cervicovaginal fluid were collected from the posterior fornix by the application of a sterile swab at speculum examination from 10 women at 6 time points during their menstrual cycles (with informed consent, and Human Research Ethics approval obtained from University Los Andes, Chile). Samples were collected from each woman at two time points during: the pre-ovulatory (sample code a and b, n=18, two missing samples); ovulatory (sample code c and d, n=20) and postovulatory (sample code e and f, n=20) phase of the menstrual cycle (Table 1). Individual cervicovaginal fluid samples (n=58) were reduced, alkylated, deglycosylated, trypsin digested and subjected to standard liquid chromatography, tandem mass spectrometry (information dependent acquisition, IDA) analysis (Experiment 1). Peptide ions were identified by database

3

search strategies to define the cervicovaginal fluid proteome obtained using the sample collection protocol. To identify phase-specific changes in the abundance of cervicovaginal fluid proteins during the menstrual cycle, samples were pooled by time of collection (n=6 pools) and subjected to liquid chromatography, tandem mass spectrometry and information independent acquisition (SWATH Experiment 2). Six pooled samples (a,b,c,d,e and f) of cervicovaginal fluid were extracted, processed as three independent technical replicates (assay replicates). Each technical replicate was processed separately for IDA and SWATH data acquisition. The data obtained from each pool was combined to establish a sample- and study-specific peptide ion database (IDA library). The variation in the abundance of cervicovaginal fluid peptides across the menstrual cycle was established by comparing SWATH peptide ion profiles against IDA library using Peakview™ software. Cervicovaginal Fluid Collection Healthy, non-pregnant, normally cycling women who were not taking a contraceptive pill were recruited to the study, with informed consent, from a natural family planning clinic at Hospital Parroquial de San Bernardo, San Bernardo, Chile (see Table 2). Serial samples of cervicovaginal fluid were collected from individual healthy, infection-free, non-pregnant women (n=10) of reproductive age during the pre-ovulatory, ovulatory and post-ovulatory phases of the menstrual cycle. All cervicovaginal fluid samples were collected in the absence of recent sexual activity (> 24 h) per vagina at speculum examination by placing a sterile swab in the posterior fornix for 10 s. The swab was removed and placed directly into a polypropylene tube and, within 10 min of collection, transported on ice and stored at -80oC until extracted. Cervicovaginal Fluid Swab Extraction Frozen swabs (swab tip downwards) were extracted in M-PER Mammalian Protein Extraction Reagent (500μl) with no added protease inhibitors. Tubes were vortexed for 10 s and then placed on a shaking platform (Thermo Fisher) at 200 rpm for 30 min at room temperature. Samples were incubated for 30 min at 4oC and centrifuged for 15 min at 15,000 g at 4°C (Heraeus Fresco21, Thermo Fisher). The swab was inverted in the tube (swab tip upwards) and then centrifuged at 15,000 g for 15 min at 4°C. Supernatant fluid (~500µl) was aspirated and used for further mass spectrometric processing and analysis. The remaining sample was stored at -80°C. Cervicovaginal Fluid Protein Analysis Total protein recovered from swab samples was determined using Bicinchoninic acid assay (BCA assay) and Bovine Serum Albumin (BSA) as standard. Absorbance was measured at 562nm using a microplate reader (BMG Labtech, Ortenberg/Germany). Reduction and Alkylation Each sample was extracted in protein extraction reagent and prepared for mass spectrometry analysis in triplicate (i.e. three independent technical replicates). Aliquots (100 μl, 1 μg protein/μl) of extracted swab supernatant fluid were mixed with ammonium bicarbonate (100 μl of 50 mM) for 10 s on a vortex mixer. DTT/bicarbonate (10μl of 20mM) was added to each sample and then incubated for 1h at 60oC. Iodoacetamide (10 μl of 1M in 100mM bicarbonate)

4

was added to each tube and incubated for 1h at 37o C in the dark. Parallel incubations were performed using BSA (0.66 μg) as a quality control for reduction/alkylation and digestion reactions. PNGase and Trypsin Treatment Reduced and alkylated samples were deglycosylated by PNGase (4μl, 200 units 10μl of 5 X MS reaction buffer, for 1 h at 37 oC. Samples were then digested with Trypsin Gold (1μg) for 18h at 37 oC. Post-digestion, formic acid (100 μl, 0.1%) was added and samples were centrifuged (15,000g for 15min) through 10K Da size exclusion membrane (PALL, Nanosep Cheltenham Vic, Australia ) [24]. The flow through fraction was retained for analysis. Desalting Peptides were desalted using a modified version of the stage tip protocol. Briefly, a 3mm piece of an Empore™ C18 (Octadecyl) SPE Extraction Disk was excised and placed in a gel loader tip and POROS slurry (5 μl) was added to form a micro-column[25]. Triflouracetic acid (1 volume, 0.1% in water) was added to the sample and loaded onto the micro-column. The micro-column was washed with trifluoroacetic acid (20µl, 0.1% in water). Peptides were eluted from the micro-column by three washes of acetonitrile (20 μl X 3, 0.1% formic acid). Eluates were pooled and samples were dried at room temperature in vacuum evaporator for 45 min. Samples were reconstituted in formic acid (50 μl, 0.1%), vortexed for 10 s and centrifuged for 2 min at 10,000 g to remove particulates. Mass Spectrometry Proteins present in cervicovaginal fluid samples were identified by liquid chromatography, tandem mass spectrometry (10 μl of reconstituted sample per injection, Eksigent 1D plus nanoflow, ultra performance liquid chromatography system equipped with a Nanoflex cHiPLC Eksigent and a 5600 Triple TOF, ABSciex). The cHPLC was configured as a trap/column system. The trap dimensions were 200µm (internal diameter) by 6mm long and was packed with 3µm 120Å C18-CL ChromXP reversed phase support. The column dimensions were 200µm (internal diameter) by 15cm long and was filled with the same material. Peptides were separated using a linear gradient (60 min from 5 to 80% B at 500 nl/min) of 0.1% formic acid in water (Solvent A) and 0.1% formic acid in acetonitrile (Solvent B) and were delivered by a Nanospray III electrospray interface (105 mm stainless steel emitter, Thermo Fischer THIES528). For standard information-dependent analysis, a 250-ms survey scan (Time off flight (TOF)-MS) was collected, from which the top 25 ions were selected for automated MS/MS analysis in subsequent experiments where each MS/MS event consisted of a 100-ms scan. The selection criteria for parent ions included intensity (> 150 counts/s) and a charge state > 2. Once an ion was fragmented by MS/MS, its mass and isotopes were excluded for a period of 6s. Ions were isolated using a quadrupole resolution of 0.7 Da and fragmented in the collision cell using collision energy ramped from 15 to 45 eV within the 100-ms accumulation time. In the instances where there were less than 25 parent ions that met the selection criteria, ions were subjected to longer accumulation times to maintain a constant total cycle time of 1.25 s. For SWATH MS-based experiments (Experiment 2) the instrument was configured as described by Gillet et al. [26]. Briefly, the mass spectrometer was operated in a looped product ion mode.

5

In this mode, the instrument was specifically tuned to allow a quadrupole resolution of 25 Da/mass selection. The stability of the mass selection was maintained by the operation of the Radio Frequency (RF) and Direct Current (DC) voltages on the isolation quadrupole in an independent manner. Using an isolation width of 26 Da (25 Da of optimal ion transmission efficiency 1 Da for the window overlap), a set of 32 overlapping windows was constructed covering the mass range 400–1200 Da. Consecutive swaths need to be acquired with some precursor isolation window overlap to ensure the transfer of the complete isotopic pattern of any given precursor ion in at least one isolation window and thereby to maintain optimal correlation between parent and fragment isotopes peaks at any LC time point. Data Analysis To identify proteins present in individual cervicovaginal fluid samples (Experiment 1), mass spectra were analyzed using MASCOT™ and Protein Pilot™ search engines against the Swissprot database with the species set as human, specifying trypsin as the enzyme, 1 missed cleavage variable modifications were cysteines as carbamidomethyl and oxidized methionine. Significance of protein identification by IDA mass spectrometry was ascribed using a Bonferroni correction for multiple hypothesis testing (i.e. 0.05/number of comparisons). Protein identifications that achieved significance and were identified in all three technical replicates were subjected to further statistical analyses (Mascot score cut off was 38 or p < 0.000158) . The changes in the relative abundance of proteins present in cervicovaginal fluid pools across the menstrual cycle were established by comparing the extracted-ion peak intensities of the three technical replicates for each sample (Experiment 2- SWATH MS). Variation in the relative expression of proteins was assessed by two-way ANOVA with variance partitioned between sample collection time and protein. Where significance was established, group means were compared using post-hoc test (corrected for multiple hypothesis testing, Bonferroni). Data are presented as mean ± SE (n=3). Ingenuity Pathway Analysis™ software was used to categorize cervicovaginal fluid proteins identified into functional groups. Variation in the recovery of protein from swab samples was assessed by ANOVA (STAT 11, StataCorp LP, College Station TX, USA), with the variance partitioned between menstrual cycle phase and subject. Data are presented as mean ± S.E. RESULTS Cervicovaginal Fluid Swab Protein Recovery: The overall recovery of protein from swabs averaged 1129 μg protein (mean ± SE, n=58, 2 missing values). Data were subjected to ANOVA with the variance partitioned between patients and phase of the cycle. No statistically significant effect was identified for either patient or phase of cycle (p > 0.38, n=58). The total protein recovered from pre-ovulatory (a and b), ovulatory (c and d) and post-ovulatory (e and f) phase swabs that average 1209 ± 124, 1090 ± 183 and 1090 ± 90 μg protein, respectively. A swab control was used to demonstrate that no protein signatures were detected from an extracted blank swab (data not shown). Cervicovaginal Fluid Proteome: Experiment 1: Standard LC MS/MS (IDA) of Individual Samples. A total of 278 cervicovaginal fluid proteins were identified (Supplemental Table S1; Supplemental Data are available online at www.biolreprod.org). Of these 176 cervicovaginal fluid proteins were ascribed statistical

6

significance (i.e. p < 0.000158, using a Bonferroni correction for multiple hypothesis testing) (Supplemental Table S1). Experiment 2: IDA SWATH Method for Pooled Samples. IDA and SWATH profiles were generated for each pooled sample (n=6) in three independent technical replicates (i.e. independent sample processing). A combined IDA library was generated from the six individual IDA profiles (a,b,c,d,e and f) using ProteinPilot ™. The IDA library was used to identify peptide ions that were present in SWATH ion profiles. Proteins were identified and quantified by comparing SWATH generated peptide ion profiles for each individual pool against the IDA library (PeakView™). For example, the extracted-ion chromatograms for NEALIALLR (a peptide from plastin-2) at each phase of the three phases (pre-ovulatory, ovulatory and post ovulatory) of the menstrual cycle are presented in Figure 2. This figure shows how the intensity of the peptide ion changes across the 3 phases and the corresponding high-resolution product ion MS/MS spectrum. Here data from only 3 out of the 6 pools are represented in Figure 2, namely A and D: pool a (early pre-ovulatory phase); B and E: pool c (early ovulatory phase) and C and F: pool f (late post-ovulatory phase), where the figure clearly shows how intensity increases from a to c and c to f for the Plastin-2 peptide. Subsequently, only proteins that were identified in all three technical replicates of Experiment 2, were included in subsequent quantitative analyses. All uncommon proteins between the three replicates were therefore eliminated. Pathway analyses All the proteins identified in the study were categorised into pathways. A canonical pathway analysis of all the CVF proteins was generated using Ingenuity™ (Figure 3). This figure displays the top 25 pathways. Figure 4 presents the menstrual cycle variation in the relative abundance (mean ± SE, n=3) of proteins (n=43) that were identified in all 3 SWATH technical replicates. 28 proteins displayed significant changes in the relative in proteins across the menstrual cycle (as assessed by 2-way ANOVA and Bonferroni post hoc tests). These proteins included, proteins previously associated with changes in the menstrual cycle, such as MMP9, defensins, and cystatin, haemoglobin and keratin isoforms were also identified. Other proteins are displayed in Supplemental Figure S1. DISCUSSION Cervicovaginal fluid is composed of fluids that originate from the oviducts, endometrium, cervix and vagina. The most important of these fluids is cervical mucus that regulates the overall viscosity of cervicovaginal fluid. Ovarian steroids regulate changes in the secretion of cervical mucus throughout the menstrual cycle. During the ovulatory phase of the menstrual cycle the volume of cervicovaginal fluid increases and is characterized by changes in water content, mucins, electrolytes, enzymes and proteins [27]. Daily tracking of changes in the vulvar perception of cervicovaginal fluid allows a woman to estimate her current phase of the menstrual cycle. Vulvar perception of cervicovaginal fluid changes function as a fertility biomarker and is extensively used to establish the daily fecundity [9, 11, 28]. Inherent inter- and intra-individual variation in the length of the menstrual cycle, however, limits the accuracy of predicting days of peak fertility. In addition, some medical situations impair the assessment of cervicovaginal fluid condition, including cervical ectropion or infection, ovarian cyst, vulvar dermatitis and

7

cervical surgery among them. Such cervical problems are also the cause of 5-10% of cases of infertility [29, 30]. Changes in cervical fluids may also influence sperm penetrability, nutrition, and survival. Given these contributing factors, the ability to successfully identify ovulation within the fertile period of the menstrual cycle is only ~55% and less that 30% on days of peak fertility. The identification of menstrual cycle phase-specific protein signatures, however, would represent an important clinical advance in assisting women to identify the presence of the fertility phase of the cycle and as a diagnostic aid to understand the etiology in some infertility cases or other gynecological diseases. To improve the detection of days of peak fertility, a more detailed understanding of longitudinal changes in cervicovaginal fluid biomarkers during the normal menstrual cycle is needed. Recent advances in mass spectrometry data acquisition (such as SWATH MS) provide the opportunity to establish comprehensive digital reference libraries of peptides present in biospecimens that can be used to identify and quantify the presence of constituent peptides in test samples.[17, 19-21] The aim of this study, therefore, was to utilize a SWATH MS method to profile proteins present in cervicovaginal fluid collected longitudinally across the menstrual cycle. Swab samples of cervicovaginal fluid were collected from normal, non-pregnant cycling women and analyzed by a traditional liquid chromatography mass spectrometry approach and by SWATH. The reproducibility of sample collection was assessed by measuring total protein recovery and elution from swabs. The total cervicovaginal fluid protein collected by swab application did not vary significantly across the menstrual cycle or between subjects. Initially, using a liquid chromatography mass spectrometry approach on individual cervicovaginal fluid samples, 176 proteins were identified (only proteins that achieved statistical significance using a Bonferroni correction for multiple hypothesis testing are reported). These proteins were categorized into canonical pathways. The principal pathways represented included: the LXR/RXR activation; FXR/RXR activation, Acute Phase Response Signaling; Clathrin-mediated Endocytosis Signalling; and Production of Nitric Oxide and Reactive Oxygen Species pathways. Previous MS-based studies have identified a greater number of proteins present in cervicovaginal fluid, for example, Shaw et. al., [31] reported the identification of more than 600 cervicovaginal fluid proteins. The difference in number of proteins identified most likely reflects different sampling techniques, site of collection, processing methods and the attribution of statistically. In Shaw’s study, cervicovaginal fluid was collected using a piece of gauze (5cm x 5 cm) inserted into the vagina for 1h. In this study, only posterior fornix fluid was collected. Where similar, more specific cervicovaginal fluid collection methods have been used, a similar number of proteins were identified (e.g. Di Quinzio et al., [32] 157 proteins; Dasari et al., [33] 105 proteins and Pereira et al., [34] 205 proteins). While liquid chromatography mass spectrometry is of utility in identifying cervicovaginal fluid proteins it is characterized by poor quantitation, reproducibility and biases to higher abundance precursor-ion signals, thus, the fidelity for physiological time series comparisons may be compromised and low abundance species may not be identified. In contrast, the alternative SWATH MS approach used in this study and applied to cervicovaginal fluid samples collected at 6 time points across the menstrual cycle, is based upon information independent acquisition.

8

Using this approach, precursor-ions are fragmented independently of their MS1 signal and a complete MS2 map is generated. The MS2 data are available for semantic analysis by comparison with sample-specific spectra library (i.e. a list of peptide fragment masses generated at the same time). Quantitation is achieved by targeted data extraction using a list of fragment masses and, thus, is similar to multiple Selected Reaction Monitoring approach (SRM). Collins et al., report the advantages of the SWATH as: high-throughput; accurate quantitation; and a wide dynamic range [20]. In this study, pooled samples were used for assessing changes in cervicovaginal fluid protein relative abundance across the menstrual cycle. Similarly, previous studies have used pooled samples to build spectral libraries [21] [26]. As SWATH profiles exist for all proteins within the spectral library, variation in the abundance of the all proteins identified in the technical replicates across the menstrual cycle may be potentially evaluated (n=43, only proteins that were identified in all three independent replicate analyses are reported). Menstrual phase-specific changes (i.e. pre-ovulatory (a and b); ovulatory (c and d) and post-ovulatory (e and f) in the relative abundance of identified proteins were established for 28 of the 43 proteins identified. Proteins that display phase specific changes identified using the SWATH method included proteins previously implicated in menstrual cycle changes and pathologies (e.g. MMP9, defensins, cystatin, haemoglobin isoforms). In addition, proteins, not previously associated with menstrual physiology or pathophysiology were identified, including: cornulin; plastin-2 and CD69; BPI fold-containing family B member 1, and MELT (a plasma membrane protein, thought to be involved in phospholipid binding). Cornulin is an S100 fused-type protein (SFTP), that is differentially expressed in cornifying keratinocytes of the epidermal layer of skin [35]. Its presence has been observed in the epithelia of the esophagus, within the inner root sheath of the hair follicles and in the upper epidermis. Cornulin mRNA transcripts have been identified by quantitative PCR in other skin sites such as, scalp skin, foreskin, and in cultured primary keratinocytes [36] while the fetal brain, adult lung, uterus and kidney, skeletal muscle and heart have also shown expression of this protein at a low level. Calcium-induced differentiation of keratinocytes has been linked to an increased expression of cornulin [36]. Other studies of the protein, have further suggested that cornulin may be expressed in response to cellular homeostatic challenges and function as a survival factor [37]. It has been similarly proposed that cornulin may be a marker of late stage epithelial cell differentiation. Interestingly it has been noted that vaginal epithelial cells undergo hormonedependent differentiation during the menstrual cycle that is associated with phase dependent expression of specific proteins, such as keratins [38]. Cornulin has been previously identified in vaginal secretions by mass spectrometry, however, its precise role remains to be elucidated. Members of the S100 family are small, calcium binding proteins present in many cells. There are more than 20 different S100 proteins identified to date. They display differential tissue and celltype expression profiles [39]. Apart from being involved in many intracellular processes in reaction to increases in intracellular calcium, there is now more focus on the extracellular functions of S100 family proteins. Of relevance to this study, S100s were shown to have antimicrobial functions and were capable of blocking bacterial propagation in mucosal epithelium. It has been reported that S100 proteins secreted from different cells during inflammation serve as

9

useful markers of disease activity for a variety of indications including chronic obstructive pulmonary disease, asthma, rheumatoid arthritis, colitis, Alzheimer’s disease and cancer among others and may regulate inflammatory responses [39]. A subset of S100s, the calgranulins, specifically S100A8 (Calgranulin A, MRP8), S100A9 (Calgranulin B, MRP14) and S100A12 (Calgranulin C, EN-RAGE) are present in neutrophils and monocytes and can be induced in endothelial and epithelial cells. S100A12 may induce cytokine expression and/or release. S100A8 and S100A9, are proinflammatory factors that induce chemotaxis of neutrophils and monocytes [39]. Infection-induced pathological inflammation causes chemotactic S100 proteins to be expressed by vaginal epithelial cells and present in vaginal secretions [40, 41]. Consistent with the data obtained in this study, Birse et al., also [42] identified an increased luteal phase expression on S100-A12 in a cross sectional analysis of cerviovaginal lavage samples. Plastin proteins are actin bundling proteins [43]. L-Plastin has been identified in hematopoietic linage cells [43] and implicated in the activation of T-cells and the modulation of cell surface expression of IL2RA/CD25. L-Plastin plays a role in innate and adaptive immunity [44] and is involved in NKG2D recruitment into lipid rafts and NKG2D-mediated NK cell migration [45]. The changes in L-Plastin across the menstrual cycle may reflect changes in the number of leukocytes present. Interestingly, a study by Larsson and Platz-Christensen [46] on vaginal cells during normal menstrual cycle, reported a mid-cycle increase in the ratio of WBC/epithelial cells in posterior fornix. In this study, we identified unique L-Plastin peptides in CVF, (Supplemental Table S1, Supplemental Figure S2) and an increase in their relative abundance during late phases of the menstrual cycle. Another protein expressed in CVF proteome of normal menstrual cycle was CD69. This protein is an leukocytic early activation marker [47]. Interestingly this protein like L-Plastin has been found in circulating leukocytes and is presence in intestinal mucosa where by it functions in an inflammatory role. It has been suggested to play a role in preventing bacterial pathogens from crossing the external layers of mucosa into the deeper inner tissues [47]. CD69 is expressed in memory T cells and regulatory T cells, in intestinal tissue. Both L-Plastin and CD-69 are important regulators of T-cell activation. Similar to L-Plastin, CD69 may reflect the change in T-cells members during menstrual cycle. More research will need to be undertaken to determine if the expression of these two proteins correlates in the cervicovaginal environment. Bactericidal permeability increasing protein (BPI) is a 55–60kDa cationic anti-microbial protein that is stored in primary azurophilic granula of neutrophil granulocytes and synthesised by mucosal epithelial cells. It is a member of a large family of lipid transfer proteins. Like CD69 and L-Plastin it may have an inflammatory role in the cervicovaginal environment. BPI is thought to neutralise the activity LPS from Gram-negative bacteria by attaching to the LPS lipid A molecule on bacterial membrane and hence destroying inner cell membrane resulting in cell lysis [51]. Previously, antibacterial proteins have been identified in CVF, including defensins and cathelicidin [48, 49], which we also identified in (Supplemental Table S1). BPI has not previously identified in CVF. Elucidating the role of these proteins and their clinical utility is beyond the scope of this Phase I biomarker study. In conclusion, to date, definition of the cervicovaginal fluid proteome has not adequately controlled for clinical factors such as biological variation (e.g. during the menstrual cycle) and

10

the use of contraceptives [50]. In this study, cervicovaginal fluid was obtained from healthy, ovulatory, non-contraceptive women to establish a baseline proteomic profile. The identification of such profiles is the first step to create a gold standard to aid in the identification of abnormal and pathological conditions. The development of a sample processing method, including a small sample volume and application of SWATH MS to the longitudinal profiling of proteins present in cervicovaginal fluid collected from the posterior fornix affords opportunity for high throughput, quantitative assessment of menstrual phase-specific changes. Such data may be of utility not only in elucidating underlying physiological mechanisms but also as clinically useful biomarkers of fertility status. The latter being of relevance in the context of family planning and the management of infertility. ACKNOWLEDGEMENTS The assistance of Hassendrini N. Peiris and Nileshkumar R. Patel for figure formatting is greatly appreciated. The authors’ gratefully acknowledge the long-standing support of the members Rotary Club of Williamstown, Victoria, Australia. REFERENCES 1. Pallone SR, Bergus GR. Fertility awareness-based methods: another option for family planning. J Am Board Fam Med 2009; 22:147-157. 2. Bortot P, Masarotto G, Scarpa B. Sequential predictions of menstrual cycle lengths. Biostatistics 2010; 11:741-755. 3. Widholm O, Kantero RL. Statistical Analysis of Menstrual Patterns of 8,000 Finnish Girls and Their Mothers. Acta Obstetricia Et Gynecologica Scandinavica 1971:3-&. 4. Munster K, Schmidt L, Helm P. Length and Variation in the Menstrual-Cycle - a CrossSectional Study from a Danish County. British Journal of Obstetrics and Gynaecology 1992; 99:422-429. 5. Wilcox AJ, Dunson D, Baird DD. The timing of the "fertile window" in the menstrual cycle: day specific estimates from a prospective study. British Medical Journal 2000; 321:1259-1262. 6. Fehring RJ. New low- and high-tech calendar methods of family planning. J Midwifery Womens Health 2005; 50:31-38. 7. Kambic RT, Lamprecht V. Calendar rhythm efficacy: a review. Adv Contracept 1996; 12:123-128. 8. Scarpa B, Dunson DB, Colombo B. Cervical mucus secretions on the day of intercourse: An accurate marker of highly fertile days. European Journal of Obstetrics Gynecology and Reproductive Biology 2006; 125:72-78. 9. Bigelow JL, Dunson DB, Stanford JB, Ecochard R, Gnoth C, Colombo B. Mucus observations in the fertile window: a better predictor of conception than timing of intercourse. Human Reproduction 2004; 19:889-892. 10. Dorairaj K. The modified mucus method in India. Am J Obstet Gynecol 1991; 165:20662067. 11. Hilgers TW, Stanford JB. Creighton model NaProEducation Technology for avoiding pregnancy - Use effectiveness. Journal of Reproductive Medicine 1998; 43:495-502.

11

12.

13.

14. 15.

16.

17.

18.

19. 20.

21.

22.

23.

24. 25.

Liong S, Di Quinzio MKW, Heng YJJ, Fleming G, Permezel M, Rice GE, Georgiou HM. Proteomic analysis of human cervicovaginal fluid collected before preterm premature rupture of the fetal membranes. Reproduction 2013; 145:137-147. Lo JO, Reddy AP, Wilmarth PA, Roberts VH, Kinhnarath A, Snyder J, Rincon MP, Gravett MG, Nagalla SR, Pereira LM. Proteomic analysis of cervical vaginal fluid proteins among women in recurrent preterm labor. J Matern Fetal Neonatal Med 2014; 27:1183-1188. Heng YJ, Liong S, Permezel M, Rice GE, Di Quinzio MK, Georgiou HM. The interplay of the interleukin 1 system in pregnancy and labor. Reprod Sci 2014; 21:122-130. Van Raemdonck GA, Tjalma WA, Coen EP, Depuydt CE, Van Ostade XW. Identification of protein biomarkers for cervical cancer using human cervicovaginal fluid. PLoS One 2014; 9:e106488. Arnhard K, Gottschall A, Pitterl F, Oberacher H. Applying 'Sequential Windowed Acquisition of All Theoretical Fragment Ion Mass Spectra' (SWATH) for systematic toxicological analysis with liquid chromatography-high-resolution tandem mass spectrometry. Anal Bioanal Chem 2015; 407:405-414. Liu Y, Huttenhain R, Collins B, Aebersold R. Mass spectrometric protein maps for biomarker discovery and clinical research. Expert Review of Molecular Diagnostics 2013; 13:811-825. Liu Y, Chen J, Sethi A, Li QK, Chen L, Collins B, Gillet LC, Wollscheid B, Zhang H, Aebersold R. Glycoproteomic analysis of prostate cancer tissues by SWATH mass spectrometry discovers N-acylethanolamine acid amidase and protein tyrosine kinase 7 as signatures for tumor aggressiveness. Mol Cell Proteomics 2014; 13:1753-1768. Law KP, Lim YP. Recent advances in mass spectrometry: data independent analysis and hyper reaction monitoring. Expert Review of Proteomics 2013; 10:551-566. Collins BC, Gillet LC, Rosenberger G, Rost HL, Vichalkovski A, Gstaiger M, Aebersold R. Quantifying protein interaction dynamics by SWATH mass spectrometry: application to the 14-3-3 system. Nature Methods 2013; 10:1246-+. Lambert JP, Ivosev G, Couzens AL, Larsen B, Taipale M, Lin ZY, Zhong Q, Lindquist S, Vidal M, Aebersold R, Pawson T, Bonner R, et al. Mapping differential interactomes by affinity purification coupled with data-independent mass spectrometry acquisition. Nat Methods 2013; 10:1239-1245. Scheidweiler KB, Jarvis MJ, Huestis MA. Nontargeted SWATH acquisition for identifying 47 synthetic cannabinoid metabolites in human urine by liquid chromatography-highresolution tandem mass spectrometry. Anal Bioanal Chem 2015; 407:883-897. Zhu X, Chen Y, Subramanian R. Comparison of information-dependent acquisition, SWATH, and MS(All) techniques in metabolite identification study employing ultrahighperformance liquid chromatography-quadrupole time-of-flight mass spectrometry. Anal Chem 2014; 86:1202-1209. Wisniewski JR, Zougman A, Nagaraj N, Mann M. Universal sample preparation method for proteome analysis. Nat Methods 2009; 6:359-362. Rappsilber J, Ishihama Y, Mann M. Stop and go extraction tips for matrix-assisted laser desorption/ionization, nanoelectrospray, and LC/MS sample pretreatment in proteomics. Anal Chem 2003; 75:663-670.

12

26.

Gillet LC, Navarro P, Tate S, Rost H, Selevsek N, Reiter L, Bonner R, Aebersold R. Targeted data extraction of the MS/MS spectra generated by data-independent acquisition: a new concept for consistent and accurate proteome analysis. Mol Cell Proteomics 2012; 11:O111 016717. 27. Gorodeski GI. The Cervical Cycle. In: Adashi EY, Rock JA, Rosenwaks Z (eds.), Reproductive Endocrinology, Surgery and Technology. Philadelphia: Lippincott-Raven 1996: 301-324. 28. Stanford JB, Smith KR, Dunson DB. Vulvar mucus observations and the probability of pregnancy. Obstet Gynecol 2003; 101:1285-1293. 29. Moriyama A, Shimoya K, Ogata I, Kimura T, Nakamura T, Wada H, Ohashi K, Azuma C, Saji F, Murata Y. Secretory leukocyte protease inhibitor (SLPI) concentrations in cervical mucus of women with normal menstrual cycle. Mol Hum Reprod 1999; 5:656-661. 30. Fritz MA, Speroff L. Clinical Gynecologic Endocrinology and Infertility Philadelphia: Lippincott Williams & Wilkins; 2011: 1157. 31. Shaw JLV, Smith CR, Diamandis EP. Proteomic analysis of human cervico-vaginal fluid. Journal of Proteome Research 2007; 6:2859-2865. 32. Di Quinzio M, Oliva K, HOLDSWORTH S, Ayhan M, Walker S, Rice G, Georgiou H, Permezel M. Proteomic analysis and characterisation of human cervico‐vaginal fluid proteins. Australian and New Zealand Journal of Obstetrics and Gynaecology 2007; 47:915. 33. Dasari S, Pereira L, Reddy AP, Michaels JEA, Lu XF, Jacob T, Thomas A, Rodland M, Roberts CT, Gravett MG, Nagalla SR. Comprehensive proteomic analysis of human cervical-vaginal fluid. Journal of Proteome Research 2007; 6:1258-1268. 34. Pereira L, Hitti J, Lapidus J, Eschenbach D, Gravett M, Nagalla S. Identification of cervicalvaginal biomarkers of recurrent preterm birth by proteomic analysis. American Journal of Obstetrics and Gynecology 2007; 197:S48-S48. 35. Mlitz V, Strasser B, Jaeger K, Hermann M, Ghannadan M, Buchberger M, Alibardi L, Tschachler E, Eckhart L. Trichohyalin-like proteins have evolutionarily conserved roles in the morphogenesis of skin appendages. J Invest Dermatol 2014; 134:2685-2692. 36. Contzler R, Favre B, Huber M, Hohl D. Cornulin, a new member of the "fused gene" family, is expressed during epidermal differentiation. J Invest Dermatol 2005; 124:990997. 37. Yagui-Beltran A, Craig AL, Lawrie L, Thompson D, Pospisilova S, Johnston D, Kernohan N, Hopwood D, Dillon JF, Hupp TR. The human oesophageal squamous epithelium exhibits a novel type of heat shock protein response. Eur J Biochem 2001; 268:5343-5355. 38. Schaller G, Genz T. Immunohistochemical detection of keratins 1 and 13 as differentiation markers in the hormone-dependent human vaginal epithelium. Gynecol Obstet Invest 1990; 30:94-96. 39. Saieg A, Brown KJ, Pena MT, Rose MC, Preciado D. Proteomic analysis of pediatric sinonasal secretions shows increased MUC5B mucin in CRS. Pediatr Res 2015; 77:356362. 40. Yano J, Kolls JK, Happel KI, Wormley F, Wozniak KL, Fidel PL, Jr. The acute neutrophil response mediated by S100 alarmins during vaginal Candida infections is independent of the Th17-pathway. PLoS One 2012; 7:e46311. 13

41.

42.

43.

44. 45.

46. 47. 48.

49.

50.

Yano J, Noverr MC, Fidel PL, Jr. Cytokines in the host response to Candida vaginitis: Identifying a role for non-classical immune mediators, S100 alarmins. Cytokine 2012; 58:118-128. Birse K, Novak RM, Westmacoot TB, Burgener A. Proteomic analysis of cervicovaginal fluid uncovers immune pathway variation between the follicular and luteal phases of the menstrual cycle - implications for HIV susceptibility. In: AIMS 2014, 20th International AIDS Conference , . Melbourne; 2014. Lin CS, Park T, Chen ZP, Leavitt J. Human plastin genes. Comparative gene structure, chromosome location, and differential expression in normal and neoplastic cells. J Biol Chem 1993; 268:2781-2792. Morley SC. The actin-bundling protein L-plastin supports T-cell motility and activation. Immunol Rev 2013; 256:48-62. Serrano-Pertierra E, Cernuda-Morollon E, Brdicka T, Hooejsi V, Lopez-Larrea C. L-plastin is involved in NKG2D recruitment into lipid rafts and NKG2D-mediated NK cell migration. J Leukoc Biol 2014; 96:437-445. Larsson PG, Platz-Christensen JJ. The vaginal pH and leucocyte/epithelial cell ratio vary during normal menstrual cycles. Eur J Obstet Gynecol Reprod Biol 1991; 38:39-41. Radulovic K, Niess JH. CD69 Is the Crucial Regulator of Intestinal Inflammation: A New Target Molecule for IBD Treatment? J Immunol Res 2015; 2015:497056. Zegels G, Van Raemdonck GAA, Coen EP, Tjalma WAA, Van Ostade XWM. Comprehensive proteomic analysis of human cervical-vaginal fluid using colposcopy samples. Proteome Science 2009; 7. Young IR, Young IR, Rice GE, Rice GE, Palliser HK, Palliser HK, Ayhan M, Ayhan M, Dellios NL, Dellios NL, Hirst JJ, Hirst JJ. Identification of bactenecin-1 in cervicovaginal fluid by two-dimensional electrophoresis in an ovine model of preterm labour. Proteomics 2007; 7:281-288. Zegels G, Van Raemdonck GAA, Tjalma WAA, Van Ostade XWM. Use of cervicovaginal fluid for the identification of biomarkers for pathologies of the female genital tract. Proteome Science 2010; 8.

FIGURE LEGENDS Figure 1. Flowchart showing the workflow of both Experiment 1 and Experiment 2. Figure 2. SWATH analysis of cervicovaginal fluid proteins. Extracted ion chromatogram of NEALIALLR peptide of plastin-2 ; SWATH intensities (Y-axes) across the menstrual cycle (A, B, and C). these SWATH profiles were generated on PeakView. D, E, F) High-resolution product ion MS/MS spectrum of NEALIALLR peptide (plastin-2). Each chromatogram shows SWATH accumulated MS/MS spectra (top, positive peaks) compared to the reference peptide ion library (bottom, negative peaks). The figure displays 3 out of the 6 phases, i.e Pre- ovulatory, pool a (D); Ovulatory, pool c (E); and Post-ovulatory, pool f (F). Intensities are displayed on Yaxes and mass/charge; m/z on X-axes. Graphs from pool b, pool d and pool e are not shown. Figure 3. Ingenuity Pathway Analysis of proteins identified in cervicovaginal fluid during the menstrual cycle. Swab samples were collected form 10 women at six sampling times during 14

gestation. Samples were processed and analysed by routine IDA LC M/MS (n=58, two missing samples). A total of 278 proteins were identified (Supplemental Table S1). Protein identifications that achieved statistical significance (using a Bonferroni correction for multiple hypothesis testing, n=176) were categorized by canonical pathway analysis. 25 canonical pathway were identified with –log(p value) less than 2.8 and containing 2 or more proteins. Figure 4. Variation in the relative abundance of 5 proteins across the menstrual cycle. Variation in proteins Plastin-2; PLSL (A), Cornulin; CRNN (B), Bactericidal permeability increasing protein Protein; BPI (C) and MELT Protein (D). Complement C3; CO3 in E) displays no significant changes across the menstrual cycle. SWATH intensity data across the menstrual cycle was assessed by 2-way ANOVA. Group means (phase of the menstrual cycle) were compared using Bonferroni multiple comparison tests. Statistical differences presented represent comparison of group means with collection time “a”. whereby, (*=p