Identification of Novel Putative Urinary Markers of Endometriosis by High-Resolution Quantitative Proteomics

sooner medical or surgical treatment and follow-up. The current gold standard for the diagnosis of the disease requests laparoscopic evaluation and biopsy of the visualized lesions [3]. Although laparoscopy is a minimally invasive procedure, it requires general anesthesia, developed surgical skills and it has a high procedural cost. In addition, laparoscopy is associated with a, although rare, risk of potential intra-operative or post-operative complications [4,5]. The AbsTRACT D, Ricciardi D, Mancini F, et al., Identifica tion of Novel Putative Urinary Markers of Endometriosis by High-Resolution Quantitative Proteomics. Biomed J Sci & Tech this study was to compare urine proteomic profiles in patients with endometriosis versus controls, trying to identify a panel of specific proteins representing novel non-invasive biomarkers of endometriosis. Eight 30-40 years old women with laparoscopically verified ovarian endometriotic cysts were enrolled. Five fertile women with non-endometriotic ovarian cystsand no visual signs of peritoneal endometriosis served as control group. Morning urine were collected in the day of the scheduled surgery and processed to be analyzed by proteomic. Eleven proteins were increased in endometriosis, while prothrombin was reduced. Seven proteins were found to fall into one large protein-protein interaction network, mainly involved with inflammation. This is the first report on the application of quantitative proteomics in the search of an array of urinary proteins, which may guide further studies for the validation of novel non-invasive biomarkers of endometriosis.

development of high-throughput technologies, such as proteomics, has led to an early diagnosis of medical diseases through the identification of proteins linked to different pathological conditions [6]. In previous studies we have demonstrated that cervical mucus is a source of putative protein biomarkers [7]. We have also characterized a pattern of cervical mucus proteins involved in the pathophysiology of endometriosis [8], which might represent putative non-invasive markers of endometriosis. Urine proteomic analysis is a valuable source to identify specific proteins and peptides. Compared to cervical mucus collection, urine proteomic analysis has the advantage to obtain proteins in a non-invasive, simple method and in a large quantity.
It has been defined as a fluid biopsy of the urogenital tract, which provides considerable information about these organs [9].
Consequently, many changes in urogenital tract function may be detected by urinary proteomic analysis. Currently, urinary proteomics is mainly used as diagnostic tool to identify biomarkers of kidney diseases, such as IgA nephropathy, renal cell carcinoma, and to detect biomarkers for chronic allograft dysfunction after kidney transplantation [10]. In addition, urinary proteomics have been used to study a variety of diseases, such as obstructive sleep apnea and obesity [11] and brain [12] and cardiovascular diseases [13]. Protein metabolites excreted in urine may reflect the state of the body and serve as informative biomarkers for some diseases. Through urinary proteomics, the molecule metabolites in urine can be comprehensively analyzed (metabolomics).
Urinary metabolomics has been widely used in the study of various diseases, including diabetic nephropathy [14], chronic heart failure [15] and offers potential early diagnosis of ovarian and breast cancer [16]. Previous studies have been published attempted to identify the proteomic profiling of endometriosis in urine samples. Using two-dimension gel separation coupled with mass spectrometry, Tokushige et al. [17] demonstrated the unique presence of a highly up-regulated protein, the CK-19, in the urine of women with endometriosis compared to a control group of women without endometriosis. Recently, urine peptide biomarkers have been searched using magnetic beads-based matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS). Wang et al. [18] identified in urine samples of patients with endometriosis a specific cluster of peptides. However, an exhaustive identification of the list of putative urine markers of endometriosis has not been still provided. The aim of this study was to identify, in urine of patients affected by endometriosis, a panel of candidate proteins, which may be further validated as novel putative biomarkers of this pathological condition.

Ethical Approval
The Ethical board of "Fondazione Policlinico "A. Gemelli" approved the protocol (registration number 0019148/16) on May, 5 th 2016.

Patients
Eight 30 to 40 years old infertile women, who underwent laparoscopy for ovarian endometriotic cysts, confirmed at histological exam, were enrolled. They were compared with 5 fertile women, with non-endometriotic ovarian cysts, confirmed at histological exam. Patients and controls were all operated upon by a surgeon with a long previous experience in the laparoscopic diagnostic and treatment of endometriosis (R.M.) [1,19,20]

Urine Collection
Urine culture was performed in the week preceding admission for surgery, to exclude urinary tract infection.
Morning midstream urine samples were collected by patients and controls after hospitalization in the morning of surgery.
Immediately after urine collection, a tablet of cocktail of protease inhibitor (Roche-Basel, Switzerland) was added to 10 ml of samples.

Sample Preparation
Ten milliliters of each urine samples were centrifuged at 1.500×g, 10 minutes at 4°C. The supernatant has been collected and stored at -80°C. One milliliter of each urine sample has been added with acetonitrile (4 volumes), vortex mixed and incubated at -20°C overnight. After centrifugation at 10.000×g, 20 minutes at 4°C, the supernatant was discharged and the precipitated pellet were resuspended in 100µL of Urea 6M/Tris 100 mM pH 7.8 buffer.
Total protein content was quantified by Bradford assay. 10μg of each sample have been subjected to in-solution digestion protocol as described previously [21]. Briefly, each sample corresponding to  This quantification method has been used to define the relative quantities of all peptides in a sample. The Proteome Discoverer application calculates peptide areas during processing, using them to automatically calculate protein areas for the proteins in the report. It calculates the area of any given protein as the average of the three most abundant distinct peptides identified in the protein.

Bioinformatics and Statistical Analysis
The relative protein level ratios between the group of controls (n=5) and the group of patients (n=8) have been determined from the respective averages of protein abundances expressed for each protein in the two groups. All the proteins detected with a ratio>1.5 STRING is a database for predicted protein-protein interactions at EMBL clusters the extracted results from many protein-protein interactions databases, like Mint, BioGrid, etc. It also uses the information from KEGG pathways and Reactome to provide the best annotations for the interactions of one protein [22]. Analysis has been performed considering a high confidence of interactions (0.7), evaluated for experiments, databases, co-occurrence and coexpression. MCL clustering has been performed defining 3 as MCL inflation parameter.

Results
Protein identification led to the characterization in the endometriosis group of 65-92 different urinary proteins per sample in the group of patients with endometriosis and of and 32-77 proteins per sample in the control group. Eleven proteins were found to be increased in patients with endometriosis (Table   1), while one protein (prothrombin) was found to be decreased in patients with endometriosis versus the control group. Seven out of the 12 differentially expressed proteins fall into a large protein-protein interaction network based on the prediction results of STRING system (Figure 1). Two clusters have been built,

Discussion
In the present study, the human urinary proteomic pattern was investigated for the first time to search for altered molecules in patients with endometriosis. We identified a panel of 12 differentially expressed proteins, which might represent novel putative markers for non-invasive diagnosis of endometriosis. immune system cells were moreover found to be dysfunctional in endometriosis [24]. As a consequence after the implantation of endometriotic cells within the abdominal cavity, there is an endometriosis-induced secretion of cytokines, chemokines, nitric oxide, immunoglobulins. Supporting this physio-pathological hypothesis we previously demonstrated that IL-10, although at low doses, exerts its action on endometrial cells in modulating antiinflammatory and anti-apoptotic effects in vitro [25]. In the present study we identified a cluster of proteins linked with inflammatory reactions, constituted by 5 proteins (KNG1, ITIH4, CLU, EGF and F2).
Kininogen 1 is a protein that in humans is encoded by the KNG1 gene. The KNG1 gene uses alternative splicing to generate two different proteins: high-molecular-weight kininogen (HMWK) and low-molecular-weight kininogen (LMWK). HMWK in turn is cleared by the enzyme kallikrein to produce bradykinin. Kallikrein in the presence of HMWK stimulates mononuclear chemotaxis [26] and induces the release of neutrophil elastase. Bradykinin stimulates inflammation, releases prostaglandins and nitric oxide, and enhances microvascular flow and permeability [27]. KNG-1 plays moreover an important role in regulating, at gene level, urokinase plasminogen activator receptor, which is involved in cell migration and proliferation. The increased urinary excretion in endometriosis of KNG1 might so reflect an increase of KNG1 production and release, as part of the inflammatory reaction which is observed in endometriosis. Inter-alpha-trypsin inhibitor heavy chain H4 (ITIH4) is an acute-phase inflammatory response protein that belongs to a super family of protease inhibitors [28].
Winden et al. [29] reported ITIH4 detection in blood up to 3 years prior to the diagnosis of some diseases, such as breast cancer and colorectal cancer, which supports its potential use as a diagnostic biomarker. Since both KNG1 and ITIH4 have been reported in urines as reduced in ovarian cancer at early stage [30], while we with the C5b-9 complex within immune deposits in human membranous glomerulonephritis [32] where it is supposed to act as a regulator of the C5b-9 cell injury. Depletion of clusterin enhances immune glomerular injury in the isolated perfused kidney [33] and clusterin -/-knockout mice are more prone to injury due to immunocomplexes [34]. In 1995 Megalin was identified as the target of the C5b-9 injury in experimental membranous glomerulonephritis [35]. Megalin is the cell receptor for clusterin in many tissues [36], so that the presence of clusterin in glomeruli represents a host limitation to complement injury, as it prevents C5b-9 insertion into cell membrane, where the complement complex competes with clusterin for the access to the same receptor, which is Megalin. Several previous studies reported that complement pathway is frequently altered in endometriosis, with an increase in the levels of C5b-9 component [37]. As a consequence we may speculate that the increase in the levels of clusterin in the Urinary EGF (epidermal growth factor) has been proposed as a marker for different non-urological diseases [38]. Intriguingly, the EGF system is strictly involved in the pathogenesis of endometriosis [39]. It is known that the expression of EGF system in eutopic endometrium from women with endometriosis varies from the endometrium of healthy women, with significant quantitative and qualitative differences [40]. In addition, a dysregulation of EGF system has been demonstrated in the setting of severe versus mild endometriosis, suggesting functional and biochemical dissimilarities between these two types of endometriosis [41].
As a consequence, the higher urinary EGF levels observed in patients with endometriosis might reflect the dysfunctional EGF system in this disease. Finally, in the present study we reported for the first time the lower urinary prothrombin levels in patients with endometriosis. It might reflect the changes in coagulation status that has been described in patients with endometriosis [42]. Recent evidences in fact have reported higher prevalence of trombophilic disorders in endometriosis patients [43]. The lower levels of prothrombin in urines of endometriosis patients might reflect this specific coagulation status. This is the first reported application of quantitative high-resolution mass spectrometrybased proteomics that detected a specific array of proteins inthe urines of patients with endometriosis. The findings of the present study confirm that endometriosis could be viewed as a disease with systemic inflammatory involvement, as reflected by metabolites excreted in urines. These data call for further confirming studies for the validation of these non-invasive biomarkers of endometriosis.