Proteomic Profiling of Human Uterine Fibroids Reveals Upregulation of the Extracellular Matrix Protein Periostin

Proteomic Profiling of Human Uterine Fibroids Reveals Upregulation of the Extracellular Matrix... Abstract The central characteristic of uterine fibroids is excessive deposition of extracellular matrix (ECM), which contributes to fibroid growth and bulk-type symptoms. Despite this, very little is known about patterns of ECM protein expression in fibroids and whether these are influenced by the most common genetic anomalies, which relate to MED12. We performed extensive genetic and proteomic analyses of clinically annotated fibroids and adjacent normal myometrium to identify the composition and expression patterns of ECM proteins in MED12 mutation–positive and mutation–negative uterine fibroids. Genetic sequencing of tissue samples revealed MED12 alterations in 39 of 65 fibroids (60%) from 14 patients. Using isobaric tagged–based quantitative mass spectrometry on three selected patients (n = 9 fibroids), we observed a common set of upregulated (>1.5-fold) and downregulated (<0.66-fold) proteins in small, medium, and large fibroid samples of annotated MED12 status. These two sets of upregulated and downregulated proteins were the same in all patients, regardless of variations in fibroid size and MED12 status. We then focused on one of the significant upregulated ECM proteins and confirmed the differential expression of periostin using western blotting and immunohistochemical analysis. Our study defined the proteome of uterine fibroids and identified that increased ECM protein expression, in particular periostin, is a hallmark of uterine fibroids regardless of MED12 mutation status. This study sets the foundation for further investigations to analyze the mechanisms regulating ECM overexpression and the functional role of upregulated ECM proteins in leiomyogenesis. Uterine leiomyomas, also known as fibroids, are the most prevalent tumors in women, occurring in >70% of premenopausal women (1–3). Although fibroids are benign, depending on their size, number, and location, they can cause substantial morbidity and burden on the health care system (4). The clinical signs and symptoms of fibroids include heavy or prolonged menstrual bleeding, discomfort, abdominal pain, pregnancy complications, and infertility (4, 5). Fibroids are also a major reason for hysterectomy (4). Treatment options for patients with fibroids are predominantly limited to surgery and hormonal agents because of an incomplete understanding of the etiology and pathogenesis of these tumors. Whole genome sequencing has provided insight into the genetic basis of fibroids (6). Fibroids harbor four main types of driver mutations: high mobility group AT-hook 2 (HMGA2) reorganizations; biallelic inactivation of fumarate hydratase; deletions in collagen, type IV α 5 (COL4A5) and collagen, type IV α 6 (COL4A6); and mutations in the mediator complex subunit 12 (MED12) (7). However, mutations in exon 1 and exon 2 of MED12 represent the most common genetic aberrations in fibroids (8, 9). MED12 is a subunit of the mediator complex that participates in genome-wide regulation as well as gene-specific transcription (10). The mediator complex connects transcription factors to RNA polymerase (10). MED12 mutations or deletions are mutually exclusive to certain other genetic alterations, such as fumarate hydratase inactivation, suggesting that distinct mechanisms operating downstream contribute to the pathogenesis of this disease (11). Conditional expression of a Med12 missense variant in a mouse model, a liaison commonly present in human patients, led to the development of uterine fibroids, providing further evidence that Med12 genetic aberrations are drivers of uterine leiomyogenesis (12). In contrast, fibroids did not form in another mouse model in which the Med12 was deleted, indicating that fibroids associated with Med12 defects arise from a gain of function mechanism (12). However, the downstream signaling pathways involved in the development of fibroids remain unknown. Fibroids are thought to arise from a single myocyte that undergoes hyperplastic transformation (13). Unlike malignant tumors, fibroids are characterized by modest rates of cell proliferation (5, 14). However, fibroids can expand to the size of a grapefruit because of their abundant extracellular matrix (ECM), which is a defining feature of uterine leiomyomas (4, 5, 13, 15). The ECM is a component of all organs and plays an integral role in tissue development, wound healing, and maintenance of normal tissue homeostasis (16). This three-dimensional matrix structure is composed of a complex network of proteins belonging to the collagen, proteoglycan, and glycoprotein families (17, 18). The ECM not only supports cell organization but also regulates cell response and behavior through ligand-integrin interactions and the release of bound soluble factors, including growth factors and Wnt signaling molecules (19). It is well established that aberrant remodeling of the ECM contributes to the development and progression of a range of diseases, including cancer and fibrosis (17). Despite disorganized and copious ECM being a hallmark feature of fibroids, no studies to date have investigated whether there is an association between underlying genetic anomalies of fibroids and ECM expression patterns. In this study, using extensive genetic and proteomic analyses of clinically annotated fibroids and adjacent normal myometrium (ANM), we have identified the composition and expression patterns of ECM proteins in MED12 mutation–positive and mutation–negative uterine fibroids. Material and Methods Patient recruitment and tissue samples To study the genetic basis of uterine fibroids, we determined the mutation status in exon 1 and exon 2 of the MED12 gene in women from the Hunter New England region of New South Wales, Australia. Uterine fibroids from 2 to 14 cm in diameter and ANM tissues were collected from individuals who underwent hysterectomies at the John Hunter Hospital, Newcastle, New South Wales, Australia. Human tissue collection and experimentation were conducted in accordance with the guidelines of the Institutional Human Research Ethics Committee at the University of Newcastle. The age range of the patients was 42 to 72 years, with an average age of 47 years. Sixty-five uterine fibroids of different sizes from 14 patients (most patients had multiple fibroids) and 14 ANM samples from the same patients (1 ANM sample per patient), were collected. We classified fibroids into three groups according to their tumor size: small (diameter < 2.0 cm), medium (diameter 2 to 4 cm), and large (diameter > 4 cm). Histopathological analysis by a pathologist validated ANM and fibroid tissue samples. The collected tissues were immediately transferred to the laboratory, washed with phosphate-buffered saline to remove excessive blood, snap-frozen, and stored in liquid nitrogen until further analysis. Tissue arrays from ANM and fibroid tissue samples were generated with a core diameter of 2 mm, with two cores per patient. Genomic DNA extraction from tissue samples To evaluate the status of MED12 mutations in fibroids and ANM, 79 tissue samples were selected for DNA isolation. To isolate the genomic DNA, approximately 25 mg of frozen tissue (fibroid or ANM) was cut into smaller pieces, homogenized in tissue lysis buffer (Buffer ATL; Qiagen) and treated with proteinase K to aid protein degradation and tissue lysis. Genomic DNA was extracted using the QIAamp DNA Mini Kit (Qiagen) according to the manufacturer’s instructions. DNA concentrations were determined by using a NanoDropTM 2000/2000c spectrophotometer (Thermo Fisher Scientific). Polymerase chain reaction amplification and Sanger sequencing DNA amplification and sequencing were performed at the Hunter North Pathology sequencing core facility, Newcastle. The DNA from fibroid and ANM tissue samples was amplified with ImmolaseTM DNA polymerase (Bioline) using specific primers (Sigma-Aldrich) (sense 5′-CCTCCGGAACGTTTCATAGAT-3′ and antisense 5′-TTCGGGACTTTTGCTCTCAC-3′) targeting exon 1 of the MED12 gene (sense 5′-GCCCTTTCACCTTGTTCCTT-3′ and antisense 5′-TGTCCCTATAAGTCTTCCCAACC-3′) for exon 2 of the MED12 gene, as described previously (8). Briefly, a total of 24 µL master mix consisting of genomic DNA (3 µL), forward and reverse primers (each 0.4 µL), ImmolaseTM DNA polymerase (0.2 µL), buffer D (12 µL), and nuclease free water (8.4 µL) was prepared for each reaction. Samples were subjected to the following thermal cycling conditions on a T100TM thermal cycler (Bio-Rad): denaturation at 95°C for 10 minutes followed by 30 cycles of 95°C for 30 seconds, 58°C (for exon 1) and 60°C (for exon 2) for 30 seconds, 72°C for 30 seconds, and extension at 72°C for 10 minutes, followed by a final soak step at 4°C. After polymerase chain reaction (PCR) amplification, the products were purified by using the QIAquick PCR Purification Kit (Qiagen) and eluted in 20 µL buffer EB (10 mM Tris·Cl, pH 8.5) according to the manufacturer’s instructions. Purified PCR products were quantified to determine the suitability of the DNA concentration (should be >30 ng/µL) for sequencing. PCR products were bidirectionally sequenced using the Sanger method and detected with capillary electrophoresis on an ABI 3730xl Automatic DNA Analyzer (Applied Biosystems). PCR products were sequenced using the BigDye Terminator v.3.1 Ready Reaction Premix and Sequencing Buffer (5X) (Applied Biosystems) using the original PCR primers specific to MED12 exon 1 and exon 2, respectively. (5X of Sequencing Buffer allows users to dilute BigDye Terminator Ready Reaction Premix. Use of 5X buffer maintains PH and magnesium optimallly for reaction.) Sequence chromatographs were analyzed for somatic mutations in exon 1 and exon 2 of MED12 using Mutation Surveyor software (SoftGenetics). Protein sample preparation and 4plex iTRAQ labeling and processing To perform proteomic analysis of genetically annotated fibroids and ANM, 75 to 100 mg of each tissue sample was cut into smaller pieces, resuspended in 100 µL of ice-cold lysis buffer (0.1 M Na2CO3, 10 mM Na3VO4) with protease inhibitor cocktail (Complete Mini; Roche), and phosphatase inhibitors PhosStop (Roche). Homogenization was performed by using the BeadBug homogenizer (Benchmark Scientific). The homogenates, which contained 150 µg of proteins, were dissolved in urea buffer (12 M urea, 4 M thiourea), reduced with 10 mM dithiothreitol, and heated at 56°C for 1 hour, followed by alkylation with 20 mM iodoacetamide and incubation in the dark at room temperature for 45 minutes, as previously described (20). Protein samples were then digested with trypsin [Mass Spectrometry Grade (Promega), 2 µg/µL in Milli-Q water (Millipore), and 50 µg of trypsin/sample] at room temperature for 3 hours. The digest was diluted further by using 100 mM triethylammonium bicarbonate, pH 7.8, to a final concentration of urea buffer (0.75 M urea, 0.25 M thiourea) and subjected to a second digestion using trypsin for 16 hours at room temperature. Following digestion, trypsin was inactivated by acidifying the samples to a pH of 2 using 10% trifluoroacetic acid (TFA). Lipid precipitation was achieved by centrifuging the acidified samples at 14,000g for 10 minutes at room temperature, and the supernatant containing the peptides was collected in a LoBind tube (Eppendorf) for further processing. Before fractionation, peptide samples were desalted using C18-SD 4 mm/1 mL Extraction Disk Cartridge (Empore). Cartridges were activated with acetonitrile (ACN) and then equilibrated with 0.1% TFA in MilliQ water, according to the manufacturer’s protocol. Peptides were loaded into the cartridge and washed three times with 500 mL of MilliQ water containing 0.1% TFA. Bound peptides were eluted into a LoBind tube in two steps using 100 µL of 60% ACN, 0.1% TFA and then 100 µL of 80% ACN, 0.1% TFA. Peptide concentrations were determined using a Qubit 2.0 Fluorometer assay (Invitrogen). Peptides (minimum quantity ∼100 µg) were dried by SpeedVac SC100 (Thermo Fisher Scientific) vacuum centrifugation and were reconstituted in the dissolution buffer. Reconstituted peptides were then processed according to the manufacturer’s protocol for 4plex iTRAQ reagent (AB SCIEX). The 4plex labeling of three fibroids (small, medium, and large) and one ANM tissue sample from each of the three patients was performed for 2 hours at room temperature. Labeled peptides were pooled into a single tube and dried with the SpeedVac. Fractionation of labeled peptides by hydrophilic interaction liquid chromatography Hydrophilic interaction liquid chromatography (HILIC) fractionation was performed to purify and fractionate the mixed peptides. The fractionation procedure was conducted at the Mass Spectrometry Core Facility of the Charles Perkins Centre, University of Sydney. Dried mixed peptides were reconstituted in 90% ACN incorporating 0.1% TFA. Any insoluble material was removed by centrifugation at 20,000g for 5 minutes at 4°C. Twenty micrograms of peptide material was resolved onto the HILIC column using an inverted organic gradient of solvent A (water, 0.1% TFA) and solvent B (ACN, 0.1% TFA). Fractions were collected in a deep 96-well plate, dried, and resuspended in 0.1% formic acid. Each HILIC fraction was subjected to liquid chromatography (LC)/tandem mass spectrometry (MS/MS). Mass spectrometry and data analysis Peptides were injected onto a trapping column for preconcentration (Acclaim Pepmap100 20 × 0.075 mm3 µm C18, Thermo Fisher Scientific), followed by nanoflow LC (Thermo Dionex, Ultimate 3000 RSLCnano, Thermo Fisher Scientific). Peptide separation was achieved using a 500 × 0.075-mm ID, PepMap 2-µm EasySpray C18 column (Thermo Fisher Scientific) with the following mobile phases: 0.1% formic acid in high-performance LC water (solvent A) and 80% ACN combined with 0.1% formic acid (solvent B). Peptides were resolved using a linear gradient from 2% to 35% solvent B, over 120 minutes, with a constant flow of 250 nL/min. The peptide eluent flowed into a nano-electrospray emitter at the sampling region of a Q-Exactive Plus Orbitrap mass spectrometer (Thermo Fisher Scientific). The electrospray process was initiated by applying 2.0 kV to the liquid junction of the emitter, and data were acquired under the control of Xcalibur (Thermo Fisher Scientific) in data-dependent mode. The mass spectrometry (MS) survey scan was performed using a resolution of 35,000. The 10 most intense multiply charged precursors were selected for high-energy collisional dissociation fragmentation with a normalized collision energy of 30.0, then measured in the Orbitrap (Thermo Fisher Scientific) at a resolution of 17,500. Automatic gain control targets were 3E6 ions for Orbitrap scans and 17,000 for MS/MS scans. The raw MS data were processed with the Proteome Discoverer software package, version 2.0.0.802 (Thermo Fisher Scientific). Proteins and peptides were identified by searching against the UniProt Human reference proteome database (downloaded 21 October 2016, with a total of 48,140 entries). The following search parameters were used: mass tolerances in MS and MS/MS modes of 10 ppm and 20 ppm, respectively; trypsin designated as the digestion enzyme, with up to two missed cleavages allowed; S-carbamidomethylation of cysteine residues; oxidation on methionine; phosphorylation on serine, threonine, and tyrosine; and acetylation and methylation of lysine and deamidation of asparagine and glutamine set as variable modifications. The false discovery rate was set to 1% for positive identification of proteins, peptides, and phosphorylation sites. Data were analyzed by using Gitools 2.3.1 software (Biomedical Genomics Group) to generate heat maps and Venny 2.1.0 (BioinfoGP–CSIC) to create Venn diagrams illustrating the number of shared and differentially expressed proteins between patients. Proteins represented by at least two unique peptides were included in the analysis. We classified a protein as upregulated if the level of expression was 1.5-fold higher than the corresponding ANM sample and downregulated if it was 0.66-fold lower than the corresponding ANM sample. A q value must be below 0.05 to identify significant changes in expression. Western blotting Lysed soluble tissue extracts containing 30 µg of proteins were mixed in Laemmli buffer containing 10% β-mercaptoethanol (Sigma-Aldrich), heated for 5 minutes, and separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (10% Mini-PROTEAN® TGX™ Gels, Bio-Rad), then transferred to nitrocellulose membrane (AmershamTM ProtranTM 0.45 µM NC) for western blotting. After blocking with 5% milk (weight-to-volume ratio) in Tris-buffered saline (0.1% Tween-20), the membrane was incubated overnight at 4°C with primary antibodies (Table 1). Horseradish peroxidase–conjugated secondary antibodies (Jackson ImmunoResearch Laboratories) were applied to the membrane for 1 hour at room temperature. LAS-3000 Imager (Fujifilm) was used to capture the image. The band intensity of western blot was quantified using an ImageJ plugin (National Institutes of Health). Table 1. Antibody Dilutions, Conditions, and Providers With Catalog Numbers Antibody  Provider  Catalog No.  Dilution  RRID  POSTN  Sigma-Aldrich  HPA012306  WB: 1:1000  AB_1854827  IHC: 200  pSMAD2  Merck Millipore  566415  WB: 1:1000  AB_565143  IHC: 1:200  GAPDH  Cell Signaling Technology  5174  WB: 1:5000  AB_10622025  Actin  DSHB  JLA20  1: 4000  AB_528068  Antibody  Provider  Catalog No.  Dilution  RRID  POSTN  Sigma-Aldrich  HPA012306  WB: 1:1000  AB_1854827  IHC: 200  pSMAD2  Merck Millipore  566415  WB: 1:1000  AB_565143  IHC: 1:200  GAPDH  Cell Signaling Technology  5174  WB: 1:5000  AB_10622025  Actin  DSHB  JLA20  1: 4000  AB_528068  Abbreviations: DSHB, Developmental Studies Hybidoma Bank; GAPDH, glyceraldehyde 3–phosphate dehydrogenase; POSTN, periostin; pSMAD2, phosphorylated form of mothers against decapentaplegic homolog 2; RRID, Research Resource Identifier; WB, western blot. View Large Immunohistochemistry Immunohistochemistry (IHC) was performed as we described previously (21). Briefly, fresh tissues were washed three times with phosphate-buffered saline for 15 minutes per wash and fixed overnight at 4°C in 10% buffered formalin. After fixation, tissues were embedded in paraffin and 5-μm sections were prepared. Tissue sections were deparaffinized, quenched to eliminate internal peroxidase activity, and incubated with primary antibodies (Table 1) or normal rabbit immunoglobulin G (negative control; Jackson ImmunoResearch Laboratories) overnight at 4°C. Following three washes in phosphate-buffered saline with 0.1% Tween-20, tissue sections were incubated for 1 hour in biotinylated secondary antibody, followed by incubation in horseradish peroxidase streptavidin (1:250; Jackson ImmunoResearch Laboratories). Tissue sections were exposed to diaminobenzidine (Sigma-Aldrich) to develop antibody signal. Nuclei were counterstained using hematoxylin. Human endometrium was used as positive control (Supplemental Fig. 1). Images were captured using an Aperio AT2 slide scanner (Leica Biosystems). Periostin (POSTN) staining was quantified using a HaloTM image analysis platform (Indica Laboratories). An area quantification algorithm was used to quantify the pixel intensities of diaminobenzidine staining. The H-score was then calculated using pixel intensities, according to the following formula:  H-score=(3×percentage of pixels with strong staining)+(2×percentage of pixels with intermediate staining)+(1×percentage of pixels with weak staining) Statistical analysis Data were analyzed and graphed with Prism 6.0 software (GraphPad). Values are presented as mean ± standard error of the mean. Statistical significance was calculated using the Student t test. P values <0.05 were considered to indicate statistically significant differences. For MS data analysis, the Student t test value comparing protein expression differences between ANM and fibroid across patients were corrected to P values using the Benjamini-Hochberg method (22). Pearson correlation analysis was used to determine the correlation between two groups. Results Analysis of mutations in exon 1 and exon 2 of the MED12 gene in human uterine fibroids Previous studies have established that approximately 70% of uterine fibroids harbor specific mutations in MED12, either missense mutations or small in-frame insertions and deletions in exon 2 (8, 23, 24). Additional mutations in exon 1 have also been reported (9) and presented similar tumorigenic mechanism as exon 2 mutations (25, 26). To determine the genetic basis of the uterine fibroids used in our study, we performed Sanger sequencing and screened for somatic mutations in 14 Australian patients. Altogether, 39 of 65 fibroids from 14 different patients had MED12 mutations [39 of 65 (60%)] (Table 2). More specifically, 27 of 37 fibroids (72%) from 11 patients displayed missense mutations, affecting codon 44 in exon 2 of MED12. These include G44V, G44C, G44R, G44A, G44S, and G44D mutations. Another hot spot was found in codons V41Q (sample ULM 5.4) and Q43P (sample ULM 5.7) and, interestingly, these were identified in only one patient (Table 2; Supplemental Table 1). As opposed to the missense mutations, we also identified 10 tumors (27%) from four patients with deletion mutations in MED12 exon 2 (Table 2). We next sought to analyze the mutation in exon 1 and observed that only two missense mutations were detected in two patients [namely, codons E10K (sample ULM3.3) and E33K (sample ULM10.4)] (Table 2; Supplemental Table 1). As expected, we did not detect any mutations in MED12 in any of the ANM samples (Supplemental Table 1). The remaining 26 of 65 fibroids from 14 patients were found to be MED12 mutation negative (Supplemental Table 1). Our genetic analysis highlighted that multiple fibroids within the same patient can harbor diverse genetic profiles (Supplemental Table 1). For example, eight variously sized tumors from one of the patients harbored five different hot spots, all of which affected glycine at codon 44 in exon 2 of MED12 (Fig. 1). This analysis also confirmed that mutations in exon 2 of MED12 are common in uterine fibroids in Australian women. On the basis of the mutation screening of 14 patients, we selected three patients (n = 12 tissue samples: 9 fibroid and 3 ANM samples) harboring both MED12 mutation–positive and mutation–negative tumors and performed proteomic analysis. Table 2. Summary of Somatic MED12 Mutations Observed in Fibroids of Various Sizes Location per Mutation Type  Nucleotide Change  Predicted Protein Change  No. of Mutations Out of 65 Fibroids (%)  Missense         Exon 1  c.28G>A  p.E10K  1 (1.5)   Exon 1  c.296G>A  p.E33K  1 (1.5)   Exon 2  c.122T>A  p.V41Q  1 (1.5)   Exon 2  c.128A>C  p.Q43P  1 (1.5)   Exon 2  c.130G>A  p.G44S  5 (7.7)   Exon 2  c.130G>C  p.G44R  3 (4.6)   Exon 2  c.130G>T  p.G44C  2 (3.0)   Exon 2  c.131G>A  p.G44D  9 (13.8)   Exon 2  c.131G>C  p.G44A  1 (1.5)   Exon 2  c.131G>T  p.G44V  5 (7.7)  Deletion         Exon 2  c.100_131del  p.D34_G44del  1 (1.5)   Exon 2  c.114_134del  p.L39_F45del  2 (3.0)   Exon 2  c.115_135del  p.L39_F45del  3 (4.6)   Exon 2  c.120_165del  p.N40_E55del  1 (1.5)   Exon 2  c.121_125del  p.V41_K42del  1 (1.5)   Exon 2  c.126_132del  p.K42_G44del  1 (1.5)   Exon 2  c.127_143del  p.Q43_Q48del  1 (1.5)  Location per Mutation Type  Nucleotide Change  Predicted Protein Change  No. of Mutations Out of 65 Fibroids (%)  Missense         Exon 1  c.28G>A  p.E10K  1 (1.5)   Exon 1  c.296G>A  p.E33K  1 (1.5)   Exon 2  c.122T>A  p.V41Q  1 (1.5)   Exon 2  c.128A>C  p.Q43P  1 (1.5)   Exon 2  c.130G>A  p.G44S  5 (7.7)   Exon 2  c.130G>C  p.G44R  3 (4.6)   Exon 2  c.130G>T  p.G44C  2 (3.0)   Exon 2  c.131G>A  p.G44D  9 (13.8)   Exon 2  c.131G>C  p.G44A  1 (1.5)   Exon 2  c.131G>T  p.G44V  5 (7.7)  Deletion         Exon 2  c.100_131del  p.D34_G44del  1 (1.5)   Exon 2  c.114_134del  p.L39_F45del  2 (3.0)   Exon 2  c.115_135del  p.L39_F45del  3 (4.6)   Exon 2  c.120_165del  p.N40_E55del  1 (1.5)   Exon 2  c.121_125del  p.V41_K42del  1 (1.5)   Exon 2  c.126_132del  p.K42_G44del  1 (1.5)   Exon 2  c.127_143del  p.Q43_Q48del  1 (1.5)  View Large Figure 1. View largeDownload slide Somatic MED12 exon 2 mutations in one patient with uterine leiomyomata showing different genomic profiles. Sequence chromatograms of various size (large, medium, small) fibroids and corresponding ANM tissues are shown. Codon 44 in MED12 exon 2 is highlighted by the horizontal bars above the traces. Mutated bases are indicated by arrows. Figure 1. View largeDownload slide Somatic MED12 exon 2 mutations in one patient with uterine leiomyomata showing different genomic profiles. Sequence chromatograms of various size (large, medium, small) fibroids and corresponding ANM tissues are shown. Codon 44 in MED12 exon 2 is highlighted by the horizontal bars above the traces. Mutated bases are indicated by arrows. Proteome analysis reveals changes in ECM expression profile in genetically annotated uterine fibroids To determine whether fibroid size and MED12 mutation status influence fibroid protein expression, we compared the protein composition of genetically annotated fibroids from each size category with ANM control tissue samples. We first performed proteomic analysis of three fibroids (small, medium, and large) and one ANM tissue sample from each of the three patients using the 4plex iTRAQ labeling system (n = 3 normal and n = 9 fibroids) (Fig. 2). This system enabled comparisons between different-size fibroids and ANM samples in the same LC/MS run, thereby minimizing experimental variation. From the three patients, a combined total of 1061, 1081, and 1116 upregulated proteins in small, medium, and large fibroids, respectively, compared with ANM, were identified using the iTRAQ experiment (Fig. 3D). In contrast, a total of 466, 659, and 592 downregulated proteins in small, medium, and large fibroids, respectively, were identified (Supplemental Fig. 2). Quantitative LC-MS/MS results from these three patients were then compared using Venn diagrams to identify the number of common and differentially expressed proteins within each fibroid size category. A representative Venn diagram for upregulated proteins in small fibroids is shown in Fig. 3D. The complete set of Venn diagrams, classified into downregulated proteins, is provided in Supplemental Fig. 2. We set our predefined criteria of q < 0.05 with relative expression levels at least >1.5-fold or <0.66-fold compared with ANM for upregulated and downregulated proteins, respectively; proteins represented by at least two unique iTRAQ-labeled peptides were included in the analysis. The P values were then calculated by Student t test comparing the protein expression differences between ANM and fibroid across the three patients and adjusted for multiple testing by the Benjamini-Hochberg method (22). (P < 0.05 is considered to indicate statistical significance after the Benjamini-Hochberg correction for multiple testing) (Fig. 3A–3C andFig. 4A–4C; Supplemental Table 2). Figure 2. View largeDownload slide Schematic representation of the proteomics workflow. Twelve samples (nine fibroids and three ANM) from three patients were prepared in parallel, digested, and labeled with 4plex iTRAQ and mixed. To reduce sample complexity, pooled peptides were subjected to HILIC fractionation before LC-MS analysis. Experimental procedures are described in the Materials and Methods section. m/z 114–117 are the iTRAQ reporter ions. Figure 2. View largeDownload slide Schematic representation of the proteomics workflow. Twelve samples (nine fibroids and three ANM) from three patients were prepared in parallel, digested, and labeled with 4plex iTRAQ and mixed. To reduce sample complexity, pooled peptides were subjected to HILIC fractionation before LC-MS analysis. Experimental procedures are described in the Materials and Methods section. m/z 114–117 are the iTRAQ reporter ions. Figure 3. View largeDownload slide Upregulated proteins in uterine fibroids. Heat map represents the protein expression patterns in the various sizes—(A) small, (B) medium, and (C) large fibroids—against the corresponding ANM tissues, which were commonly identified as shown in Table 2. Ratio is mapped from red (increase) to blue (decrease) or black (no change); see color key inset. ECM proteins are highlighted in green and POSTN is indicated in a box. *P < 0.05, adjusted by the Benjamini-Hochberg method. (D) Venn diagram illustrating the number of shared and uniquely differentially upregulated (>1.5-fold) expressed proteins identified by 4plex iTRAQ in the fibroid (small, medium, and large) vs ANM. +, MED12 mutation–positive; −, MED12 mutation–negative. Figure 3. View largeDownload slide Upregulated proteins in uterine fibroids. Heat map represents the protein expression patterns in the various sizes—(A) small, (B) medium, and (C) large fibroids—against the corresponding ANM tissues, which were commonly identified as shown in Table 2. Ratio is mapped from red (increase) to blue (decrease) or black (no change); see color key inset. ECM proteins are highlighted in green and POSTN is indicated in a box. *P < 0.05, adjusted by the Benjamini-Hochberg method. (D) Venn diagram illustrating the number of shared and uniquely differentially upregulated (>1.5-fold) expressed proteins identified by 4plex iTRAQ in the fibroid (small, medium, and large) vs ANM. +, MED12 mutation–positive; −, MED12 mutation–negative. Figure 4. View largeDownload slide Differentially downregulated (<0.66-fold) proteins identified by 4plex iTRAQ in uterine fibroids. Heat map depicts the protein expression patterns in fibroids of different sizes—(A) small, (B) medium, and (C) large fibroids—against the corresponding ANM tissues. Ratio is mapped from red (increase) to blue (decrease) or black (no change); see color key inset. ECM proteins are highlighted in green. *P < 0.05, adjusted by the Benjamini-Hochberg method. +, MED12 mutation–positive; −, MED12 mutation–negative. Figure 4. View largeDownload slide Differentially downregulated (<0.66-fold) proteins identified by 4plex iTRAQ in uterine fibroids. Heat map depicts the protein expression patterns in fibroids of different sizes—(A) small, (B) medium, and (C) large fibroids—against the corresponding ANM tissues. Ratio is mapped from red (increase) to blue (decrease) or black (no change); see color key inset. ECM proteins are highlighted in green. *P < 0.05, adjusted by the Benjamini-Hochberg method. +, MED12 mutation–positive; −, MED12 mutation–negative. On the basis of on our quantitative results, we extracted the list of proteins that were up- and downregulated in fibroids and performed pathway analysis using STRING, version 10.5 (a protein-protein interaction network database), to identify the associated protein-protein interaction pathways. These analyses highlighted several protein networks, including extracellular matrix receptor and focal adhesion interactions, and PI3K-Akt signaling (Supplemental Fig. 3). Analysis of small fibroids vs ANM control revealed that 18 proteins were upregulated in all patients (Fig. 3A and 3D). Of these 18 proteins, 10 ECM proteins, including collagen type II α 1 (COL2A1), collagen type III α 1 (COL3A1), collagen type VII α 1 (COL7A1), matrix metalloproteinase 2 (MMP2), POSTN, serpin H1 (SERPINH1), secreted protein acidic and rich in cysteine (SPARC)–related modular calcium-binding protein 2 (SMOC2), transforming growth factor-β–induced protein ig-h3 (TGFBI), tenascin (TNC), and versican core protein (VCAN) were identified. In the analysis of medium fibroid vs ANM tissue, 15 common proteins were detected in the three patients (Fig. 3B and 3D). Thirteen of these proteins—asporin, collagen type XII α1 (COL12A1), COL2A1, COL3A1, collagen type V α 2 (COL5A2), COL7A1, fibromodulin (FMOD), POSTN, α-1-antitrypsin (SERPINA1), SMOC2, SPARC-like 1 (SPARCL1), TGFBI, and TNC—were of ECM origin. In addition, we also detected 22 upregulated proteins present in large fibroids that were commonly shared among patients (Fig. 3C and 3D). Of these proteins, 19 were ECM proteins: COL12A1, collagen, type I α 1 (COL1A1), collagen type I α 2 (COL1A2), COL2A1, COL3A1, collagen type V α 1 (COL5A1), COL5A2, COL7A1, FMOD, laminin subunit α-2 (LAMA2),lumican (LUM), osteoglycin (OGN), POSTN, SERPINA1, SMOC2, SPARCL1, TGFBI, TNC, and VCAN. We further examined proteins that were commonly downregulated in each of the different-sized fibroids (27 proteins in small fibroids, 25 proteins in medium fibroids, and 32 proteins in large fibroids) (Fig. 4A–4C; Supplemental Fig. 2). The ECM-related proteins that were downregulated across the different-sized fibroids were annexin A1 (ANXA1), annexin A2 (ANXA2), collagen type XIV α 1 (COL14A1), collagen type XVIII α 1 (COL18A1), laminin subunit α 5 (LAMA5), laminin subunit β 2 (LAMB2), protein S100-A6 (S100A6), serpin peptidase inhibitor clade C (SERPINC1), and tubulointerstitial nephritis antigen-like (TINAGL1) in small fibroids; ANXA1, collagen type XV α 1 (COL15A1), fibrillin-1 (FBN1), LAMA5, LAMB2, and TINAGL1 in medium fibroids; and ANXA1, ANXA2, LAMA5, LAMB2, S100A6, and TINAGL1 in large fibroids. The proteomic data (accession number, gene ID, number of unique peptides, coverage, fold change in abundance of fibroid compared with ANM, q values, and P< 0.05 adjusted by the Benjamini-Hochberg method) of the commonly identified proteins found in small, medium, and large fibroids from the three patients are shown in Supplemental Table 2. POSTN is a significantly upregulated ECM protein during early stages of uterine leiomyogenesis From the list of ECM proteins identified as upregulated in fibroids, COL2A1, COL3A1, COL7A1, POSTN, SMOC2, TGFBI, and TNC were the only ones detected across the fibroid size range. POSTN was one of the significantly upregulated ECM proteins observed in small fibroids, suggesting that it is upregulated at the earliest stages of uterine leiomyogenesis. One small fibroid containing MED12 mutation–positive or mutation–negative and corresponding ANM tissue (total n = 24) from each of the eight patients were selected and analyzed for POSTN expression by western blotting and IHC. Western blot results revealed higher levels of POSTN in small fibroids than in corresponding ANM in the eight patients regardless of MED12 mutations (Fig. 5A) [eight ANM vs eight fibroids with MED12-positive mutation (P = 0.02); eight ANM vs eight fibroids with MED12-negative mutation (P = 0.04)]. Phospho-SMAD2, which indicates the phosphorylated form of mothers against decapentapligic homolog 2 (pSMAD2), is one of the key components for transforming growth factor-β (TGF-β) signaling, which is an upstream regulator of POSTN (27). We next monitored the expression level of pSMAD2 in these patients to determine whether any correlation existed with POSTN. We observed greater levels of pSMAD2 in small fibroids compared with corresponding ANM in these patients irrespective of MED12 mutations (Fig. 5A). We observed a significant linear correlation between the expression of pSMAD2 and POSTN (r = 0.48; P = 0.0068) (Fig. 5B). IHC of small fibroids with MED12-positive mutations also exhibited upregulation of POSTN compared with ANM, supporting western blot results (Fig. 5B). Quantitative analysis of IHC POSTN expression using H-Score (Fig. 5B) confirmed higher staining intensity and, therefore, POSTN upregulation in fibroids compared with ANM (n = 15 fibroids, n = 4 ANM; P = 0.00083). Figure 5. View largeDownload slide Differential expression of POSTN in human uterine fibroids. (A) Greater upregulation of POSTN and pSMAD2 in fibroid than ANM. Western blot showing the overexpression of POSTN and pSMAD2 in small MED12 mutation–positive or mutation–negative fibroids compared with ANM controls (n = 8 patients; *P< 0.05). Levels of β-actin and glyceraldehyde 3–phosphate dehydrogenase (GAPDH) were monitored as a quantitative control in western blot analysis. (B) IHC detection of POSTN showing elevated levels staining (in brown) in fibroid (III, IV) compared with ANM (I, II; n = 4 normal and n = 15 fibroids). Correlation analysis of pSMAD2 and POSTN expression levels in human fibroid tissues (n = 30). Scale bars: 100 µm. +, MED12 mutation–positive; −, MED12 mutation–negative; wt, wild-type. Figure 5. View largeDownload slide Differential expression of POSTN in human uterine fibroids. (A) Greater upregulation of POSTN and pSMAD2 in fibroid than ANM. Western blot showing the overexpression of POSTN and pSMAD2 in small MED12 mutation–positive or mutation–negative fibroids compared with ANM controls (n = 8 patients; *P< 0.05). Levels of β-actin and glyceraldehyde 3–phosphate dehydrogenase (GAPDH) were monitored as a quantitative control in western blot analysis. (B) IHC detection of POSTN showing elevated levels staining (in brown) in fibroid (III, IV) compared with ANM (I, II; n = 4 normal and n = 15 fibroids). Correlation analysis of pSMAD2 and POSTN expression levels in human fibroid tissues (n = 30). Scale bars: 100 µm. +, MED12 mutation–positive; −, MED12 mutation–negative; wt, wild-type. Discussion Uterine fibroids affect up to 70% of women during their reproductive years and are an important source of gynecological and obstetrical problems (1, 2). Despite the major health care burden posed by fibroids, very little is known about their etiology and pathogenesis. Treatment strategies for these benign tumors are limited to invasive surgical procedures and hormonal therapies. Therefore, identification of novel drug targets and improved treatment strategies are required. Our research data showed that fibroid clusters taken from the same patient possess diverse genomic profiles. In the literature, mutations in exon 2 of the MED12 gene are the most common genetic anomalies in fibroids (8), and our genomic data supported this observation. Our study determined the MED12 status of fibroids in the Australian population. We identified that 27 tumors from 11 patients that displayed missense mutations affecting codon 44 in exon 2 of MED12. Glycine residue at codon 44 appears to be essential for normal MED12 function as all six base substitutions lead to changes in this amino acid. Excessive ECM deposition is a major hallmark of this disease (5), which contributes to fibroid growth and bulk-type symptoms. Despite the significance of ECM in fibroids, there has been limited research on ECM characterization and expression in fibroids. Our study investigated whether MED12 status and fibroid size influence ECM composition and expression patterns. After genetic analysis, we generated heat maps showing protein upregulation and downregulation in fibroids compared with ANM control tissues. Interestingly, we observed that, overall, the set of proteins (ECM-related and otherwise) that were upregulated or downregulated were the same in all three patients, irrespective of MED12 mutation status and fibroid size. Because a copious amount of ECM is characteristic of fibroids, we next focused on the upregulated ECM proteins. Comparison of expression levels revealed that POSTN, an ECM protein, was one of the most upregulated proteins, with an average threefold higher expression in fibroids than in ANM. Because fibroids grow as they progress, small fibroids indicate early stages of disease. Therefore, because POSTN had upregulated expression levels in small fibroids, this ECM protein may have potential as a novel drug target to hinder the progression of uterine leiomyogenesis. We further validated the expression of POSTN in our fibroid samples using western blot and IHC. Our proteomic and western blot analysis demonstrated that in all cases of MED12 mutation–positive and mutation–negative fibroids, POSTN was upregulated compared with ANM. In contrast, in the absence of mutations in MED12, POSTN expression was not consistent (Fig. 5A). In the case of MED12-negative fibroids, the expression of POSTN may be affected by the presence of other genetic mutations. Consequently, the effect of other genetic mutations on fibroid ECM protein expression requires further investigation. Our study focused on MED12 mutations because in the literature these are the most commonly detected mutations in fibroids. POSTN is a 90-kDa secreted ECM protein with a multidomain structure that contains an amino-terminal EMI domain, a tandem repeat of four fasciclin (FAS1) domains, and a carboxyl-terminal domain (28–30). Each domain interacts with specific ECM proteins and cell-surface integrins to promote the assembly of extracellular architectures (31–34), which govern the biomechanical properties of connective tissues. Previous studies have shown that POSTN interacts with collagens, laminins, and tenascins (31–34). In fibroids of all sizes (small, medium, and large), we observed upregulation of POSTN, along with upregulation of collagens, laminins, and tenascins (Fig. 3A–3C). This suggests that these proteins are important contributors to the excessive ECM build-up that leads to fibroid expansion. Further studies are required to target the binding sites on the multidomain structure of POSTN to investigate the relationship between POSTN and the interacting proteins. POSTN plays a central role in normal tissue homeostasis and disease development (34). Previous studies have demonstrated that high levels of POSTN expression are associated with a variety of cancers, including head and neck (35, 36), oral (37), lung (38), breast (39), ovarian (40), colon (41), pancreatic (42), and liver (43, 44) cancer. POSTN participates in many biological processes involved in cancer, including cell adhesion, invasion, metastasis, and tumor angiogenesis (36, 37, 41, 45). Our study compared the expression level of POSTN in fibroids relative to ANM. POSTN is a candidate for further functional and clinical investigations given its abundance in uterine fibroids. The molecular mechanism of POSTN is still unclear, bu a recent study revealed that it interacts with integrins and activates focal adhesion kinase and PI3K-Akt–mediated signaling pathways, promoting tumor angiogenesis, invasion, and metastasis (46). Our STRING analysis also identified the involvement of these signaling networks in fibroids (Supplemental Fig. 3). Our proteomic analysis revealed that TGF-β was also upregulated in fibroids of all sizes compared with ANM (Fig. 3A–3C). Deletion of the Periostin (Postn) gene in a mouse model of muscular dystrophy altered TGF-β signaling, resulting in enhanced tissue regeneration and reduced levels of fibrosis, thereby providing evidence for interaction between POSTN and TGF-β (47). Therefore, we can also speculate that POSTN interacts with TGF-β signaling pathway to promote ECM formation in leiomyogenesis. Further research is warranted to define the functional role of POSTN and other ECM-related proteins in fibroid initiation and progression. In summary, our study defined the global protein expression patterns of fibroids vs ANM in human patients. Furthermore, we used genomic and proteomic analysis to study the relationship among exon 2 MED12 mutation status, fibroid size, and ECM protein expression. This analysis revealed that the same group of ECM proteins was upregulated or downregulated in all patients, despite variations in fibroid size and the MED12 gene. Furthermore, we validated the expression level of POSTN because it was one of the significantly upregulated ECM proteins identified in small fibroids and these fibroids indicate early stages of leiomyogenesis. Because excessive ECM is a prominent feature of fibroids, targeting upregulated ECM proteins is a rational approach for overcoming this disease. Investigation of the role of the ECM in cancer and other abnormalities has illustrated that the ECM is an active participant in disease progression and is responsive to surrounding cell types and signaling molecules. Further studies are essential for understanding the mechanisms and mediators responsible for the overproduction of specific ECM components. This understanding will provide an opportunity to develop intervening strategies that will curb the production of these ECM proteins, thereby reducing fibroid size and removing the associated disease burden. Abbreviations: ACN acetonitrile ANM adjacent normal myometrium ANXA1 annexin A1 ANXA2 annexin A2 COL12A1 collagen type XII α 1 COL2A1 collagen type II α 1 COL3A1 collagen type III α 1 COL5A2 collagen type V α 2 COL7A1 collagen type VII α 1 ECM extracellular matrix FMOD fibromodulin HILIC hydrophilic interaction liquid chromatography IHC immunohistochemistry LAMA5 laminin subunit α 5 LAMB2 laminin subunit β 2 LC liquid chromatography MED12 mediator complex subunit 12 gene MS mass spectrometry MS/MS tandem mass spectrometry PCR polymerase chain reaction POSTN periostin pSMAD2 phosphorylated form of mothers against decapentaplegic homolog 2 S100A6 protein S100-A6 SERPINA1 α-1-antitrypsin SMOC2 secreted protein acidic and rich in cysteine–related modular calcium-binding protein 2 SPARC secreted protein acidic and rich in cysteine SPARCL1 secreted protein acidic and rich in cysteine–like 1 TFA trifluoroacetic acid TGF-β transforming growth factor-β TGFBI transforming growth factor-β–induced protein ig-h3 TINAGL1 tubulointerstitial nephritis antigen-like TNC tenascin VCAN versican core protein. Acknowledgments The authors thank Dr. Ben Crossett and Dr. Trisha Al Mazi for help with HILIC and Nathan Smith for help with LC-MS/MS. Financial Support: Work in the Tanwar laboratory was in part supported by funding from the National Health and Medical Research Council (Grant APP1081461 to P.S.T.), the Australian Research Council, the Cancer Institute New South Wales, and the John Hunter Hospital Charitable Trust. Y.B., P.B., and P.B.N. are recipients of the University of Newcastle Postgraduate Research Fellowship. M.D.D. is supported by a Cancer Institute New South Wales, Australia Early Career Fellowship. Author Contributions: M.F.B.J. and P.S.T. designed the research. M.F.B.J., Y.-A. K., M.K., Y.B., P.B., P.B.N, and D.A.S.-B. performed the research. M.F.B.J., H.H., M.A.B., M.D.D., R.J.S., P.N., and P.S.T. analyzed the data. M.F.B.J., Y.B., and P.S.T. wrote the paper. P.S.T. supervised the study, provided financial support, and edited and had final approval of the manuscript. All authors reviewed, approved, and commented on the manuscript. Disclosure Summary: The authors have nothing to disclose. References 1. Cramer SF, Patel A. The frequency of uterine leiomyomas. Am J Clin Pathol . 1990; 94( 4): 435– 438. Google Scholar CrossRef Search ADS PubMed  2. Baird DD, Dunson DB, Hill MC, Cousins D, Schectman JM. High cumulative incidence of uterine leiomyoma in black and white women: ultrasound evidence. Am J Obstet Gynecol . 2003; 188( 1): 100– 107. Google Scholar CrossRef Search ADS PubMed  3. Arslan AA, Gold LI, Mittal K, Suen TC, Belitskaya-Levy I, Tang MS, Toniolo P. Gene expression studies provide clues to the pathogenesis of uterine leiomyoma: new evidence and a systematic review. Hum Reprod . 2005; 20( 4): 852– 863. Google Scholar CrossRef Search ADS PubMed  4. Bulun SE. Uterine fibroids. N Engl J Med . 2013; 369( 14): 1344– 1355. Google Scholar CrossRef Search ADS PubMed  5. Stewart EA, Laughlin-Tommaso SK, Catherino WH, Lalitkumar S, Gupta D, Vollenhoven B. Uterine fibroids. Nat Rev Dis Primers . 2016; 2: 16043. Google Scholar CrossRef Search ADS PubMed  6. Mehine M, Kaasinen E, Mäkinen N, Katainen R, Kämpjärvi K, Pitkänen E, Heinonen HR, Bützow R, Kilpivaara O, Kuosmanen A, Ristolainen H, Gentile M, Sjöberg J, Vahteristo P, Aaltonen LA. Characterization of uterine leiomyomas by whole-genome sequencing. N Engl J Med . 2013; 369( 1): 43– 53. Google Scholar CrossRef Search ADS PubMed  7. Mehine M, Kaasinen E, Heinonen H-R, Mäkinen N, Kämpjärvi K, Sarvilinna N, Aavikko M, Vähärautio A, Pasanen A, Bützow R, Heikinheimo O, Sjöberg J, Pitkänen E, Vahteristo P, Aaltonen LA. Integrated data analysis reveals uterine leiomyoma subtypes with distinct driver pathways and biomarkers. Proc Natl Acad Sci USA . 2016; 113( 5): 1315– 1320. Google Scholar CrossRef Search ADS PubMed  8. Mäkinen N, Mehine M, Tolvanen J, Kaasinen E, Li Y, Lehtonen HJ, Gentile M, Yan J, Enge M, Taipale M, Aavikko M, Katainen R, Virolainen E, Böhling T, Koski TA, Launonen V, Sjöberg J, Taipale J, Vahteristo P, Aaltonen LA. MED12, the mediator complex subunit 12 gene, is mutated at high frequency in uterine leiomyomas. Science . 2011; 334( 6053): 252– 255. Google Scholar CrossRef Search ADS PubMed  9. Kämpjärvi K, Park MJ, Mehine M, Kim NH, Clark AD, Bützow R, Böhling T, Böhm J, Mecklin JP, Järvinen H, Tomlinson IP, van der Spuy ZM, Sjöberg J, Boyer TG, Vahteristo P. Mutations in Exon 1 highlight the role of MED12 in uterine leiomyomas. Hum Mutat . 2014; 35( 9): 1136– 1141. Google Scholar CrossRef Search ADS PubMed  10. Taatjes DJ. The human Mediator complex: a versatile, genome-wide regulator of transcription. Trends Biochem Sci . 2010; 35( 6): 315– 322. Google Scholar CrossRef Search ADS PubMed  11. Kämpjärvi K, Mäkinen N, Mehine M, Välipakka S, Uimari O, Pitkänen E, Heinonen HR, Heikkinen T, Tolvanen J, Ahtikoski A, Frizzell N, Sarvilinna N, Sjöberg J, Bützow R, Aaltonen LA, Vahteristo P. MED12 mutations and FH inactivation are mutually exclusive in uterine leiomyomas. Br J Cancer . 2016; 114( 12): 1405– 1411. Google Scholar CrossRef Search ADS PubMed  12. Mittal P, Shin YH, Yatsenko SA, Castro CA, Surti U, Rajkovic A. Med12 gain-of-function mutation causes leiomyomas and genomic instability. J Clin Invest . 2015; 125( 8): 3280– 3284. Google Scholar CrossRef Search ADS PubMed  13. Moravek MB, Yin P, Ono M, Coon JS V, Dyson MT, Navarro A, Marsh EE, Chakravarti D, Kim JJ, Wei J-J, Bulun SE. Ovarian steroids, stem cells and uterine leiomyoma: therapeutic implications. Hum Reprod Update . 2015; 21( 1): 1– 12. Google Scholar CrossRef Search ADS PubMed  14. Peddada SD, Laughlin SK, Miner K, Guyon JP, Haneke K, Vahdat HL, Semelka RC, Kowalik A, Armao D, Davis B, Baird DD. Growth of uterine leiomyomata among premenopausal black and white women. Proc Natl Acad Sci USA . 2008; 105( 50): 19887– 19892. Google Scholar CrossRef Search ADS PubMed  15. Stewart EA, Shuster LT, Rocca WA. Reassessing hysterectomy. Minn Med . 2012; 95( 3): 36– 39. Google Scholar PubMed  16. Mammoto T, Ingber DE. Mechanical control of tissue and organ development. Development . 2010; 137( 9): 1407– 1420. Google Scholar CrossRef Search ADS PubMed  17. Bonnans C, Chou J, Werb Z. Remodelling the extracellular matrix in development and disease. Nat Rev Mol Cell Biol . 2014; 15( 12): 786– 801. Google Scholar CrossRef Search ADS PubMed  18. Pickup MW, Mouw JK, Weaver VM. The extracellular matrix modulates the hallmarks of cancer. EMBO Rep . 2014; 15( 12): 1243– 1253. Google Scholar CrossRef Search ADS PubMed  19. Malik M, Norian J, McCarthy-Keith D, Britten J, Catherino WH. Why leiomyomas are called fibroids: the central role of extracellular matrix in symptomatic women. Semin Reprod Med . 2010; 28( 3): 169– 179. Google Scholar CrossRef Search ADS PubMed  20. Dun MD, Chalkley RJ, Faulkner S, Keene S, Avery-Kiejda KA, Scott RJ, Falkenby LG, Cairns MJ, Larsen MR, Bradshaw RA, Hondermarck H. Proteotranscriptomic profiling of 231-BR breast cancer cells: identification of potential biomarkers and therapeutic targets for brain metastasis. Mol Cell Proteomics . 2015; 14( 9): 2316– 2330. Google Scholar CrossRef Search ADS PubMed  21. Bajwa P, Nielsen S, Lombard JM, Rassam L, Nahar P, Rueda BR, Wilkinson JE, Miller RA, Tanwar PS. Overactive mTOR signaling leads to endometrial hyperplasia in aged women and mice. Oncotarget . 2017; 8( 5): 7265– 7275. Google Scholar CrossRef Search ADS PubMed  22. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc B . 1995; 57: 289– 300. 23. Markowski DN, Bartnitzke S, Löning T, Drieschner N, Helmke BM, Bullerdiek J. MED12 mutations in uterine fibroids--their relationship to cytogenetic subgroups. Int J Cancer . 2012; 131( 7): 1528– 1536. Google Scholar CrossRef Search ADS PubMed  24. McGuire MM, Yatsenko A, Hoffner L, Jones M, Surti U, Rajkovic A. Whole exome sequencing in a random sample of North American women with leiomyomas identifies MED12 mutations in majority of uterine leiomyomas. PLoS One . 2012; 7( 3): e33251. Google Scholar CrossRef Search ADS PubMed  25. Turunen M, Spaeth JM, Keskitalo S, Park MJ, Kivioja T, Clark AD, Mäkinen N, Gao F, Palin K, Nurkkala H, Vähärautio A, Aavikko M, Kämpjärvi K, Vahteristo P, Kim CA, Aaltonen LA, Varjosalo M, Taipale J, Boyer TG. Uterine leiomyoma-linked MED12 mutations disrupt mediator-associated CDK activity. Cell Reports . 2014; 7( 3): 654– 660. Google Scholar CrossRef Search ADS PubMed  26. Kämpjärvi K, Järvinen TM, Heikkinen T, Ruppert AS, Senter L, Hoag KW, Dufva O, Kontro M, Rassenti L, Hertlein E, Kipps TJ, Porkka K, Byrd JC, de la Chapelle A, Vahteristo P. Somatic MED12 mutations are associated with poor prognosis markers in chronic lymphocytic leukemia. Oncotarget . 2015; 6( 3): 1884– 1888. Google Scholar CrossRef Search ADS PubMed  27. Kudo A. Introductory review: periostin-gene and protein structure. Cell Mol Life Sci . 2017; 74( 23): 4259– 4268. Google Scholar CrossRef Search ADS PubMed  28. Horiuchi K, Amizuka N, Takeshita S, Takamatsu H, Katsuura M, Ozawa H, Toyama Y, Bonewald LF, Kudo A. Identification and characterization of a novel protein, periostin, with restricted expression to periosteum and periodontal ligament and increased expression by transforming growth factor beta. J Bone Min Res . 1999; 14( 7): 1239– 1249. Google Scholar CrossRef Search ADS   29. Kudo A. Periostin in fibrillogenesis for tissue regeneration: periostin actions inside and outside the cell. Cell Mol Life Sci . 2011; 68( 19): 3201– 3207. Google Scholar CrossRef Search ADS PubMed  30. Sugiura T, Takamatsu H, Kudo A, Amann E. Expression and characterization of murine osteoblast-specific factor 2 (OSF-2) in a baculovirus expression system. Protein Expr Purif . 1995; 6( 3): 305– 311. Google Scholar CrossRef Search ADS PubMed  31. Kii I, Nishiyama T, Li M, Matsumoto K, Saito M, Amizuka N, Kudo A. Incorporation of tenascin-C into the extracellular matrix by periostin underlies an extracellular meshwork architecture. J Biol Chem . 2010; 285( 3): 2028– 2039. Google Scholar CrossRef Search ADS PubMed  32. Maruhashi T, Kii I, Saito M, Kudo A. Interaction between periostin and BMP-1 promotes proteolytic activation of lysyl oxidase. J Biol Chem . 2010; 285( 17): 13294– 13303. Google Scholar CrossRef Search ADS PubMed  33. Snider P, Hinton RB, Moreno-Rodriguez RA, Wang J, Rogers R, Lindsley A, Li F, Ingram DA, Menick D, Field L, Firulli AB, Molkentin JD, Markwald R, Conway SJ. Periostin is required for maturation and extracellular matrix stabilization of noncardiomyocyte lineages of the heart. Circ Res . 2008; 102( 7): 752– 760. Google Scholar CrossRef Search ADS PubMed  34. Kii I, Ito H. Periostin and its interacting proteins in the construction of extracellular architectures. Cell Mol Life Sci . 2017; 74( 23): 4269– 4277. Google Scholar CrossRef Search ADS PubMed  35. Chang Y, Lee TC, Li JC, Lai TL, Chua HH, Chen CL, Doong SL, Chou CK, Sheen TS, Tsai CH. Differential expression of osteoblast-specific factor 2 and polymeric immunoglobulin receptor genes in nasopharyngeal carcinoma. Head Neck . 2005; 27( 10): 873– 882. Google Scholar CrossRef Search ADS PubMed  36. Kudo Y, Ogawa I, Kitajima S, Kitagawa M, Kawai H, Gaffney PM, Miyauchi M, Takata T. Periostin promotes invasion and anchorage-independent growth in the metastatic process of head and neck cancer. Cancer Res . 2006; 66( 14): 6928– 6935. Google Scholar CrossRef Search ADS PubMed  37. Siriwardena BS, Kudo Y, Ogawa I, Kitagawa M, Kitajima S, Hatano H, Tilakaratne WM, Miyauchi M, Takata T. Periostin is frequently overexpressed and enhances invasion and angiogenesis in oral cancer. Br J Cancer . 2006; 95( 10): 1396– 1403. Google Scholar CrossRef Search ADS PubMed  38. Ouyang G, Liu M, Ruan K, Song G, Mao Y, Bao S. Upregulated expression of periostin by hypoxia in non-small-cell lung cancer cells promotes cell survival via the Akt/PKB pathway. Cancer Lett . 2009; 281( 2): 213– 219. Google Scholar CrossRef Search ADS PubMed  39. Shao R, Bao S, Bai X, Blanchette C, Anderson RM, Dang T, Gishizky ML, Marks JR, Wang XF. Acquired expression of periostin by human breast cancers promotes tumor angiogenesis through up-regulation of vascular endothelial growth factor receptor 2 expression. Mol Cell Biol . 2004; 24( 9): 3992– 4003. Google Scholar CrossRef Search ADS PubMed  40. Ismail RS, Baldwin RL, Fang J, Browning D, Karlan BY, Gasson JC, Chang DD. Differential gene expression between normal and tumor-derived ovarian epithelial cells. Cancer Res . 2000; 60( 23): 6744– 6749. Google Scholar PubMed  41. Bao S, Ouyang G, Bai X, Huang Z, Ma C, Liu M, Shao R, Anderson RM, Rich JN, Wang XF. Periostin potently promotes metastatic growth of colon cancer by augmenting cell survival via the Akt/PKB pathway. Cancer Cell . 2004; 5( 4): 329– 339. Google Scholar CrossRef Search ADS PubMed  42. Baril P, Gangeswaran R, Mahon PC, Caulee K, Kocher HM, Harada T, Zhu M, Kalthoff H, Crnogorac-Jurcevic T, Lemoine NR. Periostin promotes invasiveness and resistance of pancreatic cancer cells to hypoxia-induced cell death: role of the beta4 integrin and the PI3k pathway. Oncogene . 2007; 26( 14): 2082– 2094. Google Scholar CrossRef Search ADS PubMed  43. Lv Y, Wang W, Jia WD, Sun QK, Huang M, Zhou HC, Xia HH, Liu WB, Chen H, Sun SN, Xu GL. High preoparative levels of serum periostin are associated with poor prognosis in patients with hepatocellular carcinoma after hepatectomy. Eur J Surg Oncol . 2013; 39( 10): 1129– 1135. Google Scholar CrossRef Search ADS PubMed  44. Lv Y, Wang W, Jia WD, Sun QK, Li JS, Ma JL, Liu WB, Zhou HC, Ge YS, Yu JH, Xia HH, Xu GL. High-level expression of periostin is closely related to metastatic potential and poor prognosis of hepatocellular carcinoma. Med Oncol . 2013; 30( 1): 385. Google Scholar CrossRef Search ADS PubMed  45. Gillan L, Matei D, Fishman DA, Gerbin CS, Karlan BY, Chang DD. Periostin Secreted by Epithelial Ovarian Carcinoma Is a Ligand for αVβ3 and αVβ5 integrins and promotes cell motility. Cancer Res . 2002; 62: 5358– 5364. Google Scholar PubMed  46. Utispan K, Sonongbua J, Thuwajit P, Chau-In S, Pairojkul C, Wongkham S, Thuwajit C. Periostin activates integrin α5β1 through a PI3K/AKT‑dependent pathway in invasion of cholangiocarcinoma. Int J Oncol . 2012; 41( 3): 1110– 1118. Google Scholar CrossRef Search ADS PubMed  47. Lorts A, Schwanekamp JA, Baudino TA, McNally EM, Molkentin JD. Deletion of periostin reduces muscular dystrophy and fibrosis in mice by modulating the transforming growth factor-β pathway. Proc Natl Acad Sci USA . 2012; 109( 27): 10978– 10983. Google Scholar CrossRef Search ADS PubMed  Copyright © 2018 Endocrine Society http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Endocrinology Oxford University Press

Proteomic Profiling of Human Uterine Fibroids Reveals Upregulation of the Extracellular Matrix Protein Periostin

Loading next page...
 
/lp/ou_press/proteomic-profiling-of-human-uterine-fibroids-reveals-upregulation-of-efYQZ0mVeb
Publisher
Endocrine Society
Copyright
Copyright © 2018 Endocrine Society
ISSN
0013-7227
eISSN
1945-7170
D.O.I.
10.1210/en.2017-03018
Publisher site
See Article on Publisher Site

Abstract

Abstract The central characteristic of uterine fibroids is excessive deposition of extracellular matrix (ECM), which contributes to fibroid growth and bulk-type symptoms. Despite this, very little is known about patterns of ECM protein expression in fibroids and whether these are influenced by the most common genetic anomalies, which relate to MED12. We performed extensive genetic and proteomic analyses of clinically annotated fibroids and adjacent normal myometrium to identify the composition and expression patterns of ECM proteins in MED12 mutation–positive and mutation–negative uterine fibroids. Genetic sequencing of tissue samples revealed MED12 alterations in 39 of 65 fibroids (60%) from 14 patients. Using isobaric tagged–based quantitative mass spectrometry on three selected patients (n = 9 fibroids), we observed a common set of upregulated (>1.5-fold) and downregulated (<0.66-fold) proteins in small, medium, and large fibroid samples of annotated MED12 status. These two sets of upregulated and downregulated proteins were the same in all patients, regardless of variations in fibroid size and MED12 status. We then focused on one of the significant upregulated ECM proteins and confirmed the differential expression of periostin using western blotting and immunohistochemical analysis. Our study defined the proteome of uterine fibroids and identified that increased ECM protein expression, in particular periostin, is a hallmark of uterine fibroids regardless of MED12 mutation status. This study sets the foundation for further investigations to analyze the mechanisms regulating ECM overexpression and the functional role of upregulated ECM proteins in leiomyogenesis. Uterine leiomyomas, also known as fibroids, are the most prevalent tumors in women, occurring in >70% of premenopausal women (1–3). Although fibroids are benign, depending on their size, number, and location, they can cause substantial morbidity and burden on the health care system (4). The clinical signs and symptoms of fibroids include heavy or prolonged menstrual bleeding, discomfort, abdominal pain, pregnancy complications, and infertility (4, 5). Fibroids are also a major reason for hysterectomy (4). Treatment options for patients with fibroids are predominantly limited to surgery and hormonal agents because of an incomplete understanding of the etiology and pathogenesis of these tumors. Whole genome sequencing has provided insight into the genetic basis of fibroids (6). Fibroids harbor four main types of driver mutations: high mobility group AT-hook 2 (HMGA2) reorganizations; biallelic inactivation of fumarate hydratase; deletions in collagen, type IV α 5 (COL4A5) and collagen, type IV α 6 (COL4A6); and mutations in the mediator complex subunit 12 (MED12) (7). However, mutations in exon 1 and exon 2 of MED12 represent the most common genetic aberrations in fibroids (8, 9). MED12 is a subunit of the mediator complex that participates in genome-wide regulation as well as gene-specific transcription (10). The mediator complex connects transcription factors to RNA polymerase (10). MED12 mutations or deletions are mutually exclusive to certain other genetic alterations, such as fumarate hydratase inactivation, suggesting that distinct mechanisms operating downstream contribute to the pathogenesis of this disease (11). Conditional expression of a Med12 missense variant in a mouse model, a liaison commonly present in human patients, led to the development of uterine fibroids, providing further evidence that Med12 genetic aberrations are drivers of uterine leiomyogenesis (12). In contrast, fibroids did not form in another mouse model in which the Med12 was deleted, indicating that fibroids associated with Med12 defects arise from a gain of function mechanism (12). However, the downstream signaling pathways involved in the development of fibroids remain unknown. Fibroids are thought to arise from a single myocyte that undergoes hyperplastic transformation (13). Unlike malignant tumors, fibroids are characterized by modest rates of cell proliferation (5, 14). However, fibroids can expand to the size of a grapefruit because of their abundant extracellular matrix (ECM), which is a defining feature of uterine leiomyomas (4, 5, 13, 15). The ECM is a component of all organs and plays an integral role in tissue development, wound healing, and maintenance of normal tissue homeostasis (16). This three-dimensional matrix structure is composed of a complex network of proteins belonging to the collagen, proteoglycan, and glycoprotein families (17, 18). The ECM not only supports cell organization but also regulates cell response and behavior through ligand-integrin interactions and the release of bound soluble factors, including growth factors and Wnt signaling molecules (19). It is well established that aberrant remodeling of the ECM contributes to the development and progression of a range of diseases, including cancer and fibrosis (17). Despite disorganized and copious ECM being a hallmark feature of fibroids, no studies to date have investigated whether there is an association between underlying genetic anomalies of fibroids and ECM expression patterns. In this study, using extensive genetic and proteomic analyses of clinically annotated fibroids and adjacent normal myometrium (ANM), we have identified the composition and expression patterns of ECM proteins in MED12 mutation–positive and mutation–negative uterine fibroids. Material and Methods Patient recruitment and tissue samples To study the genetic basis of uterine fibroids, we determined the mutation status in exon 1 and exon 2 of the MED12 gene in women from the Hunter New England region of New South Wales, Australia. Uterine fibroids from 2 to 14 cm in diameter and ANM tissues were collected from individuals who underwent hysterectomies at the John Hunter Hospital, Newcastle, New South Wales, Australia. Human tissue collection and experimentation were conducted in accordance with the guidelines of the Institutional Human Research Ethics Committee at the University of Newcastle. The age range of the patients was 42 to 72 years, with an average age of 47 years. Sixty-five uterine fibroids of different sizes from 14 patients (most patients had multiple fibroids) and 14 ANM samples from the same patients (1 ANM sample per patient), were collected. We classified fibroids into three groups according to their tumor size: small (diameter < 2.0 cm), medium (diameter 2 to 4 cm), and large (diameter > 4 cm). Histopathological analysis by a pathologist validated ANM and fibroid tissue samples. The collected tissues were immediately transferred to the laboratory, washed with phosphate-buffered saline to remove excessive blood, snap-frozen, and stored in liquid nitrogen until further analysis. Tissue arrays from ANM and fibroid tissue samples were generated with a core diameter of 2 mm, with two cores per patient. Genomic DNA extraction from tissue samples To evaluate the status of MED12 mutations in fibroids and ANM, 79 tissue samples were selected for DNA isolation. To isolate the genomic DNA, approximately 25 mg of frozen tissue (fibroid or ANM) was cut into smaller pieces, homogenized in tissue lysis buffer (Buffer ATL; Qiagen) and treated with proteinase K to aid protein degradation and tissue lysis. Genomic DNA was extracted using the QIAamp DNA Mini Kit (Qiagen) according to the manufacturer’s instructions. DNA concentrations were determined by using a NanoDropTM 2000/2000c spectrophotometer (Thermo Fisher Scientific). Polymerase chain reaction amplification and Sanger sequencing DNA amplification and sequencing were performed at the Hunter North Pathology sequencing core facility, Newcastle. The DNA from fibroid and ANM tissue samples was amplified with ImmolaseTM DNA polymerase (Bioline) using specific primers (Sigma-Aldrich) (sense 5′-CCTCCGGAACGTTTCATAGAT-3′ and antisense 5′-TTCGGGACTTTTGCTCTCAC-3′) targeting exon 1 of the MED12 gene (sense 5′-GCCCTTTCACCTTGTTCCTT-3′ and antisense 5′-TGTCCCTATAAGTCTTCCCAACC-3′) for exon 2 of the MED12 gene, as described previously (8). Briefly, a total of 24 µL master mix consisting of genomic DNA (3 µL), forward and reverse primers (each 0.4 µL), ImmolaseTM DNA polymerase (0.2 µL), buffer D (12 µL), and nuclease free water (8.4 µL) was prepared for each reaction. Samples were subjected to the following thermal cycling conditions on a T100TM thermal cycler (Bio-Rad): denaturation at 95°C for 10 minutes followed by 30 cycles of 95°C for 30 seconds, 58°C (for exon 1) and 60°C (for exon 2) for 30 seconds, 72°C for 30 seconds, and extension at 72°C for 10 minutes, followed by a final soak step at 4°C. After polymerase chain reaction (PCR) amplification, the products were purified by using the QIAquick PCR Purification Kit (Qiagen) and eluted in 20 µL buffer EB (10 mM Tris·Cl, pH 8.5) according to the manufacturer’s instructions. Purified PCR products were quantified to determine the suitability of the DNA concentration (should be >30 ng/µL) for sequencing. PCR products were bidirectionally sequenced using the Sanger method and detected with capillary electrophoresis on an ABI 3730xl Automatic DNA Analyzer (Applied Biosystems). PCR products were sequenced using the BigDye Terminator v.3.1 Ready Reaction Premix and Sequencing Buffer (5X) (Applied Biosystems) using the original PCR primers specific to MED12 exon 1 and exon 2, respectively. (5X of Sequencing Buffer allows users to dilute BigDye Terminator Ready Reaction Premix. Use of 5X buffer maintains PH and magnesium optimallly for reaction.) Sequence chromatographs were analyzed for somatic mutations in exon 1 and exon 2 of MED12 using Mutation Surveyor software (SoftGenetics). Protein sample preparation and 4plex iTRAQ labeling and processing To perform proteomic analysis of genetically annotated fibroids and ANM, 75 to 100 mg of each tissue sample was cut into smaller pieces, resuspended in 100 µL of ice-cold lysis buffer (0.1 M Na2CO3, 10 mM Na3VO4) with protease inhibitor cocktail (Complete Mini; Roche), and phosphatase inhibitors PhosStop (Roche). Homogenization was performed by using the BeadBug homogenizer (Benchmark Scientific). The homogenates, which contained 150 µg of proteins, were dissolved in urea buffer (12 M urea, 4 M thiourea), reduced with 10 mM dithiothreitol, and heated at 56°C for 1 hour, followed by alkylation with 20 mM iodoacetamide and incubation in the dark at room temperature for 45 minutes, as previously described (20). Protein samples were then digested with trypsin [Mass Spectrometry Grade (Promega), 2 µg/µL in Milli-Q water (Millipore), and 50 µg of trypsin/sample] at room temperature for 3 hours. The digest was diluted further by using 100 mM triethylammonium bicarbonate, pH 7.8, to a final concentration of urea buffer (0.75 M urea, 0.25 M thiourea) and subjected to a second digestion using trypsin for 16 hours at room temperature. Following digestion, trypsin was inactivated by acidifying the samples to a pH of 2 using 10% trifluoroacetic acid (TFA). Lipid precipitation was achieved by centrifuging the acidified samples at 14,000g for 10 minutes at room temperature, and the supernatant containing the peptides was collected in a LoBind tube (Eppendorf) for further processing. Before fractionation, peptide samples were desalted using C18-SD 4 mm/1 mL Extraction Disk Cartridge (Empore). Cartridges were activated with acetonitrile (ACN) and then equilibrated with 0.1% TFA in MilliQ water, according to the manufacturer’s protocol. Peptides were loaded into the cartridge and washed three times with 500 mL of MilliQ water containing 0.1% TFA. Bound peptides were eluted into a LoBind tube in two steps using 100 µL of 60% ACN, 0.1% TFA and then 100 µL of 80% ACN, 0.1% TFA. Peptide concentrations were determined using a Qubit 2.0 Fluorometer assay (Invitrogen). Peptides (minimum quantity ∼100 µg) were dried by SpeedVac SC100 (Thermo Fisher Scientific) vacuum centrifugation and were reconstituted in the dissolution buffer. Reconstituted peptides were then processed according to the manufacturer’s protocol for 4plex iTRAQ reagent (AB SCIEX). The 4plex labeling of three fibroids (small, medium, and large) and one ANM tissue sample from each of the three patients was performed for 2 hours at room temperature. Labeled peptides were pooled into a single tube and dried with the SpeedVac. Fractionation of labeled peptides by hydrophilic interaction liquid chromatography Hydrophilic interaction liquid chromatography (HILIC) fractionation was performed to purify and fractionate the mixed peptides. The fractionation procedure was conducted at the Mass Spectrometry Core Facility of the Charles Perkins Centre, University of Sydney. Dried mixed peptides were reconstituted in 90% ACN incorporating 0.1% TFA. Any insoluble material was removed by centrifugation at 20,000g for 5 minutes at 4°C. Twenty micrograms of peptide material was resolved onto the HILIC column using an inverted organic gradient of solvent A (water, 0.1% TFA) and solvent B (ACN, 0.1% TFA). Fractions were collected in a deep 96-well plate, dried, and resuspended in 0.1% formic acid. Each HILIC fraction was subjected to liquid chromatography (LC)/tandem mass spectrometry (MS/MS). Mass spectrometry and data analysis Peptides were injected onto a trapping column for preconcentration (Acclaim Pepmap100 20 × 0.075 mm3 µm C18, Thermo Fisher Scientific), followed by nanoflow LC (Thermo Dionex, Ultimate 3000 RSLCnano, Thermo Fisher Scientific). Peptide separation was achieved using a 500 × 0.075-mm ID, PepMap 2-µm EasySpray C18 column (Thermo Fisher Scientific) with the following mobile phases: 0.1% formic acid in high-performance LC water (solvent A) and 80% ACN combined with 0.1% formic acid (solvent B). Peptides were resolved using a linear gradient from 2% to 35% solvent B, over 120 minutes, with a constant flow of 250 nL/min. The peptide eluent flowed into a nano-electrospray emitter at the sampling region of a Q-Exactive Plus Orbitrap mass spectrometer (Thermo Fisher Scientific). The electrospray process was initiated by applying 2.0 kV to the liquid junction of the emitter, and data were acquired under the control of Xcalibur (Thermo Fisher Scientific) in data-dependent mode. The mass spectrometry (MS) survey scan was performed using a resolution of 35,000. The 10 most intense multiply charged precursors were selected for high-energy collisional dissociation fragmentation with a normalized collision energy of 30.0, then measured in the Orbitrap (Thermo Fisher Scientific) at a resolution of 17,500. Automatic gain control targets were 3E6 ions for Orbitrap scans and 17,000 for MS/MS scans. The raw MS data were processed with the Proteome Discoverer software package, version 2.0.0.802 (Thermo Fisher Scientific). Proteins and peptides were identified by searching against the UniProt Human reference proteome database (downloaded 21 October 2016, with a total of 48,140 entries). The following search parameters were used: mass tolerances in MS and MS/MS modes of 10 ppm and 20 ppm, respectively; trypsin designated as the digestion enzyme, with up to two missed cleavages allowed; S-carbamidomethylation of cysteine residues; oxidation on methionine; phosphorylation on serine, threonine, and tyrosine; and acetylation and methylation of lysine and deamidation of asparagine and glutamine set as variable modifications. The false discovery rate was set to 1% for positive identification of proteins, peptides, and phosphorylation sites. Data were analyzed by using Gitools 2.3.1 software (Biomedical Genomics Group) to generate heat maps and Venny 2.1.0 (BioinfoGP–CSIC) to create Venn diagrams illustrating the number of shared and differentially expressed proteins between patients. Proteins represented by at least two unique peptides were included in the analysis. We classified a protein as upregulated if the level of expression was 1.5-fold higher than the corresponding ANM sample and downregulated if it was 0.66-fold lower than the corresponding ANM sample. A q value must be below 0.05 to identify significant changes in expression. Western blotting Lysed soluble tissue extracts containing 30 µg of proteins were mixed in Laemmli buffer containing 10% β-mercaptoethanol (Sigma-Aldrich), heated for 5 minutes, and separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (10% Mini-PROTEAN® TGX™ Gels, Bio-Rad), then transferred to nitrocellulose membrane (AmershamTM ProtranTM 0.45 µM NC) for western blotting. After blocking with 5% milk (weight-to-volume ratio) in Tris-buffered saline (0.1% Tween-20), the membrane was incubated overnight at 4°C with primary antibodies (Table 1). Horseradish peroxidase–conjugated secondary antibodies (Jackson ImmunoResearch Laboratories) were applied to the membrane for 1 hour at room temperature. LAS-3000 Imager (Fujifilm) was used to capture the image. The band intensity of western blot was quantified using an ImageJ plugin (National Institutes of Health). Table 1. Antibody Dilutions, Conditions, and Providers With Catalog Numbers Antibody  Provider  Catalog No.  Dilution  RRID  POSTN  Sigma-Aldrich  HPA012306  WB: 1:1000  AB_1854827  IHC: 200  pSMAD2  Merck Millipore  566415  WB: 1:1000  AB_565143  IHC: 1:200  GAPDH  Cell Signaling Technology  5174  WB: 1:5000  AB_10622025  Actin  DSHB  JLA20  1: 4000  AB_528068  Antibody  Provider  Catalog No.  Dilution  RRID  POSTN  Sigma-Aldrich  HPA012306  WB: 1:1000  AB_1854827  IHC: 200  pSMAD2  Merck Millipore  566415  WB: 1:1000  AB_565143  IHC: 1:200  GAPDH  Cell Signaling Technology  5174  WB: 1:5000  AB_10622025  Actin  DSHB  JLA20  1: 4000  AB_528068  Abbreviations: DSHB, Developmental Studies Hybidoma Bank; GAPDH, glyceraldehyde 3–phosphate dehydrogenase; POSTN, periostin; pSMAD2, phosphorylated form of mothers against decapentaplegic homolog 2; RRID, Research Resource Identifier; WB, western blot. View Large Immunohistochemistry Immunohistochemistry (IHC) was performed as we described previously (21). Briefly, fresh tissues were washed three times with phosphate-buffered saline for 15 minutes per wash and fixed overnight at 4°C in 10% buffered formalin. After fixation, tissues were embedded in paraffin and 5-μm sections were prepared. Tissue sections were deparaffinized, quenched to eliminate internal peroxidase activity, and incubated with primary antibodies (Table 1) or normal rabbit immunoglobulin G (negative control; Jackson ImmunoResearch Laboratories) overnight at 4°C. Following three washes in phosphate-buffered saline with 0.1% Tween-20, tissue sections were incubated for 1 hour in biotinylated secondary antibody, followed by incubation in horseradish peroxidase streptavidin (1:250; Jackson ImmunoResearch Laboratories). Tissue sections were exposed to diaminobenzidine (Sigma-Aldrich) to develop antibody signal. Nuclei were counterstained using hematoxylin. Human endometrium was used as positive control (Supplemental Fig. 1). Images were captured using an Aperio AT2 slide scanner (Leica Biosystems). Periostin (POSTN) staining was quantified using a HaloTM image analysis platform (Indica Laboratories). An area quantification algorithm was used to quantify the pixel intensities of diaminobenzidine staining. The H-score was then calculated using pixel intensities, according to the following formula:  H-score=(3×percentage of pixels with strong staining)+(2×percentage of pixels with intermediate staining)+(1×percentage of pixels with weak staining) Statistical analysis Data were analyzed and graphed with Prism 6.0 software (GraphPad). Values are presented as mean ± standard error of the mean. Statistical significance was calculated using the Student t test. P values <0.05 were considered to indicate statistically significant differences. For MS data analysis, the Student t test value comparing protein expression differences between ANM and fibroid across patients were corrected to P values using the Benjamini-Hochberg method (22). Pearson correlation analysis was used to determine the correlation between two groups. Results Analysis of mutations in exon 1 and exon 2 of the MED12 gene in human uterine fibroids Previous studies have established that approximately 70% of uterine fibroids harbor specific mutations in MED12, either missense mutations or small in-frame insertions and deletions in exon 2 (8, 23, 24). Additional mutations in exon 1 have also been reported (9) and presented similar tumorigenic mechanism as exon 2 mutations (25, 26). To determine the genetic basis of the uterine fibroids used in our study, we performed Sanger sequencing and screened for somatic mutations in 14 Australian patients. Altogether, 39 of 65 fibroids from 14 different patients had MED12 mutations [39 of 65 (60%)] (Table 2). More specifically, 27 of 37 fibroids (72%) from 11 patients displayed missense mutations, affecting codon 44 in exon 2 of MED12. These include G44V, G44C, G44R, G44A, G44S, and G44D mutations. Another hot spot was found in codons V41Q (sample ULM 5.4) and Q43P (sample ULM 5.7) and, interestingly, these were identified in only one patient (Table 2; Supplemental Table 1). As opposed to the missense mutations, we also identified 10 tumors (27%) from four patients with deletion mutations in MED12 exon 2 (Table 2). We next sought to analyze the mutation in exon 1 and observed that only two missense mutations were detected in two patients [namely, codons E10K (sample ULM3.3) and E33K (sample ULM10.4)] (Table 2; Supplemental Table 1). As expected, we did not detect any mutations in MED12 in any of the ANM samples (Supplemental Table 1). The remaining 26 of 65 fibroids from 14 patients were found to be MED12 mutation negative (Supplemental Table 1). Our genetic analysis highlighted that multiple fibroids within the same patient can harbor diverse genetic profiles (Supplemental Table 1). For example, eight variously sized tumors from one of the patients harbored five different hot spots, all of which affected glycine at codon 44 in exon 2 of MED12 (Fig. 1). This analysis also confirmed that mutations in exon 2 of MED12 are common in uterine fibroids in Australian women. On the basis of the mutation screening of 14 patients, we selected three patients (n = 12 tissue samples: 9 fibroid and 3 ANM samples) harboring both MED12 mutation–positive and mutation–negative tumors and performed proteomic analysis. Table 2. Summary of Somatic MED12 Mutations Observed in Fibroids of Various Sizes Location per Mutation Type  Nucleotide Change  Predicted Protein Change  No. of Mutations Out of 65 Fibroids (%)  Missense         Exon 1  c.28G>A  p.E10K  1 (1.5)   Exon 1  c.296G>A  p.E33K  1 (1.5)   Exon 2  c.122T>A  p.V41Q  1 (1.5)   Exon 2  c.128A>C  p.Q43P  1 (1.5)   Exon 2  c.130G>A  p.G44S  5 (7.7)   Exon 2  c.130G>C  p.G44R  3 (4.6)   Exon 2  c.130G>T  p.G44C  2 (3.0)   Exon 2  c.131G>A  p.G44D  9 (13.8)   Exon 2  c.131G>C  p.G44A  1 (1.5)   Exon 2  c.131G>T  p.G44V  5 (7.7)  Deletion         Exon 2  c.100_131del  p.D34_G44del  1 (1.5)   Exon 2  c.114_134del  p.L39_F45del  2 (3.0)   Exon 2  c.115_135del  p.L39_F45del  3 (4.6)   Exon 2  c.120_165del  p.N40_E55del  1 (1.5)   Exon 2  c.121_125del  p.V41_K42del  1 (1.5)   Exon 2  c.126_132del  p.K42_G44del  1 (1.5)   Exon 2  c.127_143del  p.Q43_Q48del  1 (1.5)  Location per Mutation Type  Nucleotide Change  Predicted Protein Change  No. of Mutations Out of 65 Fibroids (%)  Missense         Exon 1  c.28G>A  p.E10K  1 (1.5)   Exon 1  c.296G>A  p.E33K  1 (1.5)   Exon 2  c.122T>A  p.V41Q  1 (1.5)   Exon 2  c.128A>C  p.Q43P  1 (1.5)   Exon 2  c.130G>A  p.G44S  5 (7.7)   Exon 2  c.130G>C  p.G44R  3 (4.6)   Exon 2  c.130G>T  p.G44C  2 (3.0)   Exon 2  c.131G>A  p.G44D  9 (13.8)   Exon 2  c.131G>C  p.G44A  1 (1.5)   Exon 2  c.131G>T  p.G44V  5 (7.7)  Deletion         Exon 2  c.100_131del  p.D34_G44del  1 (1.5)   Exon 2  c.114_134del  p.L39_F45del  2 (3.0)   Exon 2  c.115_135del  p.L39_F45del  3 (4.6)   Exon 2  c.120_165del  p.N40_E55del  1 (1.5)   Exon 2  c.121_125del  p.V41_K42del  1 (1.5)   Exon 2  c.126_132del  p.K42_G44del  1 (1.5)   Exon 2  c.127_143del  p.Q43_Q48del  1 (1.5)  View Large Figure 1. View largeDownload slide Somatic MED12 exon 2 mutations in one patient with uterine leiomyomata showing different genomic profiles. Sequence chromatograms of various size (large, medium, small) fibroids and corresponding ANM tissues are shown. Codon 44 in MED12 exon 2 is highlighted by the horizontal bars above the traces. Mutated bases are indicated by arrows. Figure 1. View largeDownload slide Somatic MED12 exon 2 mutations in one patient with uterine leiomyomata showing different genomic profiles. Sequence chromatograms of various size (large, medium, small) fibroids and corresponding ANM tissues are shown. Codon 44 in MED12 exon 2 is highlighted by the horizontal bars above the traces. Mutated bases are indicated by arrows. Proteome analysis reveals changes in ECM expression profile in genetically annotated uterine fibroids To determine whether fibroid size and MED12 mutation status influence fibroid protein expression, we compared the protein composition of genetically annotated fibroids from each size category with ANM control tissue samples. We first performed proteomic analysis of three fibroids (small, medium, and large) and one ANM tissue sample from each of the three patients using the 4plex iTRAQ labeling system (n = 3 normal and n = 9 fibroids) (Fig. 2). This system enabled comparisons between different-size fibroids and ANM samples in the same LC/MS run, thereby minimizing experimental variation. From the three patients, a combined total of 1061, 1081, and 1116 upregulated proteins in small, medium, and large fibroids, respectively, compared with ANM, were identified using the iTRAQ experiment (Fig. 3D). In contrast, a total of 466, 659, and 592 downregulated proteins in small, medium, and large fibroids, respectively, were identified (Supplemental Fig. 2). Quantitative LC-MS/MS results from these three patients were then compared using Venn diagrams to identify the number of common and differentially expressed proteins within each fibroid size category. A representative Venn diagram for upregulated proteins in small fibroids is shown in Fig. 3D. The complete set of Venn diagrams, classified into downregulated proteins, is provided in Supplemental Fig. 2. We set our predefined criteria of q < 0.05 with relative expression levels at least >1.5-fold or <0.66-fold compared with ANM for upregulated and downregulated proteins, respectively; proteins represented by at least two unique iTRAQ-labeled peptides were included in the analysis. The P values were then calculated by Student t test comparing the protein expression differences between ANM and fibroid across the three patients and adjusted for multiple testing by the Benjamini-Hochberg method (22). (P < 0.05 is considered to indicate statistical significance after the Benjamini-Hochberg correction for multiple testing) (Fig. 3A–3C andFig. 4A–4C; Supplemental Table 2). Figure 2. View largeDownload slide Schematic representation of the proteomics workflow. Twelve samples (nine fibroids and three ANM) from three patients were prepared in parallel, digested, and labeled with 4plex iTRAQ and mixed. To reduce sample complexity, pooled peptides were subjected to HILIC fractionation before LC-MS analysis. Experimental procedures are described in the Materials and Methods section. m/z 114–117 are the iTRAQ reporter ions. Figure 2. View largeDownload slide Schematic representation of the proteomics workflow. Twelve samples (nine fibroids and three ANM) from three patients were prepared in parallel, digested, and labeled with 4plex iTRAQ and mixed. To reduce sample complexity, pooled peptides were subjected to HILIC fractionation before LC-MS analysis. Experimental procedures are described in the Materials and Methods section. m/z 114–117 are the iTRAQ reporter ions. Figure 3. View largeDownload slide Upregulated proteins in uterine fibroids. Heat map represents the protein expression patterns in the various sizes—(A) small, (B) medium, and (C) large fibroids—against the corresponding ANM tissues, which were commonly identified as shown in Table 2. Ratio is mapped from red (increase) to blue (decrease) or black (no change); see color key inset. ECM proteins are highlighted in green and POSTN is indicated in a box. *P < 0.05, adjusted by the Benjamini-Hochberg method. (D) Venn diagram illustrating the number of shared and uniquely differentially upregulated (>1.5-fold) expressed proteins identified by 4plex iTRAQ in the fibroid (small, medium, and large) vs ANM. +, MED12 mutation–positive; −, MED12 mutation–negative. Figure 3. View largeDownload slide Upregulated proteins in uterine fibroids. Heat map represents the protein expression patterns in the various sizes—(A) small, (B) medium, and (C) large fibroids—against the corresponding ANM tissues, which were commonly identified as shown in Table 2. Ratio is mapped from red (increase) to blue (decrease) or black (no change); see color key inset. ECM proteins are highlighted in green and POSTN is indicated in a box. *P < 0.05, adjusted by the Benjamini-Hochberg method. (D) Venn diagram illustrating the number of shared and uniquely differentially upregulated (>1.5-fold) expressed proteins identified by 4plex iTRAQ in the fibroid (small, medium, and large) vs ANM. +, MED12 mutation–positive; −, MED12 mutation–negative. Figure 4. View largeDownload slide Differentially downregulated (<0.66-fold) proteins identified by 4plex iTRAQ in uterine fibroids. Heat map depicts the protein expression patterns in fibroids of different sizes—(A) small, (B) medium, and (C) large fibroids—against the corresponding ANM tissues. Ratio is mapped from red (increase) to blue (decrease) or black (no change); see color key inset. ECM proteins are highlighted in green. *P < 0.05, adjusted by the Benjamini-Hochberg method. +, MED12 mutation–positive; −, MED12 mutation–negative. Figure 4. View largeDownload slide Differentially downregulated (<0.66-fold) proteins identified by 4plex iTRAQ in uterine fibroids. Heat map depicts the protein expression patterns in fibroids of different sizes—(A) small, (B) medium, and (C) large fibroids—against the corresponding ANM tissues. Ratio is mapped from red (increase) to blue (decrease) or black (no change); see color key inset. ECM proteins are highlighted in green. *P < 0.05, adjusted by the Benjamini-Hochberg method. +, MED12 mutation–positive; −, MED12 mutation–negative. On the basis of on our quantitative results, we extracted the list of proteins that were up- and downregulated in fibroids and performed pathway analysis using STRING, version 10.5 (a protein-protein interaction network database), to identify the associated protein-protein interaction pathways. These analyses highlighted several protein networks, including extracellular matrix receptor and focal adhesion interactions, and PI3K-Akt signaling (Supplemental Fig. 3). Analysis of small fibroids vs ANM control revealed that 18 proteins were upregulated in all patients (Fig. 3A and 3D). Of these 18 proteins, 10 ECM proteins, including collagen type II α 1 (COL2A1), collagen type III α 1 (COL3A1), collagen type VII α 1 (COL7A1), matrix metalloproteinase 2 (MMP2), POSTN, serpin H1 (SERPINH1), secreted protein acidic and rich in cysteine (SPARC)–related modular calcium-binding protein 2 (SMOC2), transforming growth factor-β–induced protein ig-h3 (TGFBI), tenascin (TNC), and versican core protein (VCAN) were identified. In the analysis of medium fibroid vs ANM tissue, 15 common proteins were detected in the three patients (Fig. 3B and 3D). Thirteen of these proteins—asporin, collagen type XII α1 (COL12A1), COL2A1, COL3A1, collagen type V α 2 (COL5A2), COL7A1, fibromodulin (FMOD), POSTN, α-1-antitrypsin (SERPINA1), SMOC2, SPARC-like 1 (SPARCL1), TGFBI, and TNC—were of ECM origin. In addition, we also detected 22 upregulated proteins present in large fibroids that were commonly shared among patients (Fig. 3C and 3D). Of these proteins, 19 were ECM proteins: COL12A1, collagen, type I α 1 (COL1A1), collagen type I α 2 (COL1A2), COL2A1, COL3A1, collagen type V α 1 (COL5A1), COL5A2, COL7A1, FMOD, laminin subunit α-2 (LAMA2),lumican (LUM), osteoglycin (OGN), POSTN, SERPINA1, SMOC2, SPARCL1, TGFBI, TNC, and VCAN. We further examined proteins that were commonly downregulated in each of the different-sized fibroids (27 proteins in small fibroids, 25 proteins in medium fibroids, and 32 proteins in large fibroids) (Fig. 4A–4C; Supplemental Fig. 2). The ECM-related proteins that were downregulated across the different-sized fibroids were annexin A1 (ANXA1), annexin A2 (ANXA2), collagen type XIV α 1 (COL14A1), collagen type XVIII α 1 (COL18A1), laminin subunit α 5 (LAMA5), laminin subunit β 2 (LAMB2), protein S100-A6 (S100A6), serpin peptidase inhibitor clade C (SERPINC1), and tubulointerstitial nephritis antigen-like (TINAGL1) in small fibroids; ANXA1, collagen type XV α 1 (COL15A1), fibrillin-1 (FBN1), LAMA5, LAMB2, and TINAGL1 in medium fibroids; and ANXA1, ANXA2, LAMA5, LAMB2, S100A6, and TINAGL1 in large fibroids. The proteomic data (accession number, gene ID, number of unique peptides, coverage, fold change in abundance of fibroid compared with ANM, q values, and P< 0.05 adjusted by the Benjamini-Hochberg method) of the commonly identified proteins found in small, medium, and large fibroids from the three patients are shown in Supplemental Table 2. POSTN is a significantly upregulated ECM protein during early stages of uterine leiomyogenesis From the list of ECM proteins identified as upregulated in fibroids, COL2A1, COL3A1, COL7A1, POSTN, SMOC2, TGFBI, and TNC were the only ones detected across the fibroid size range. POSTN was one of the significantly upregulated ECM proteins observed in small fibroids, suggesting that it is upregulated at the earliest stages of uterine leiomyogenesis. One small fibroid containing MED12 mutation–positive or mutation–negative and corresponding ANM tissue (total n = 24) from each of the eight patients were selected and analyzed for POSTN expression by western blotting and IHC. Western blot results revealed higher levels of POSTN in small fibroids than in corresponding ANM in the eight patients regardless of MED12 mutations (Fig. 5A) [eight ANM vs eight fibroids with MED12-positive mutation (P = 0.02); eight ANM vs eight fibroids with MED12-negative mutation (P = 0.04)]. Phospho-SMAD2, which indicates the phosphorylated form of mothers against decapentapligic homolog 2 (pSMAD2), is one of the key components for transforming growth factor-β (TGF-β) signaling, which is an upstream regulator of POSTN (27). We next monitored the expression level of pSMAD2 in these patients to determine whether any correlation existed with POSTN. We observed greater levels of pSMAD2 in small fibroids compared with corresponding ANM in these patients irrespective of MED12 mutations (Fig. 5A). We observed a significant linear correlation between the expression of pSMAD2 and POSTN (r = 0.48; P = 0.0068) (Fig. 5B). IHC of small fibroids with MED12-positive mutations also exhibited upregulation of POSTN compared with ANM, supporting western blot results (Fig. 5B). Quantitative analysis of IHC POSTN expression using H-Score (Fig. 5B) confirmed higher staining intensity and, therefore, POSTN upregulation in fibroids compared with ANM (n = 15 fibroids, n = 4 ANM; P = 0.00083). Figure 5. View largeDownload slide Differential expression of POSTN in human uterine fibroids. (A) Greater upregulation of POSTN and pSMAD2 in fibroid than ANM. Western blot showing the overexpression of POSTN and pSMAD2 in small MED12 mutation–positive or mutation–negative fibroids compared with ANM controls (n = 8 patients; *P< 0.05). Levels of β-actin and glyceraldehyde 3–phosphate dehydrogenase (GAPDH) were monitored as a quantitative control in western blot analysis. (B) IHC detection of POSTN showing elevated levels staining (in brown) in fibroid (III, IV) compared with ANM (I, II; n = 4 normal and n = 15 fibroids). Correlation analysis of pSMAD2 and POSTN expression levels in human fibroid tissues (n = 30). Scale bars: 100 µm. +, MED12 mutation–positive; −, MED12 mutation–negative; wt, wild-type. Figure 5. View largeDownload slide Differential expression of POSTN in human uterine fibroids. (A) Greater upregulation of POSTN and pSMAD2 in fibroid than ANM. Western blot showing the overexpression of POSTN and pSMAD2 in small MED12 mutation–positive or mutation–negative fibroids compared with ANM controls (n = 8 patients; *P< 0.05). Levels of β-actin and glyceraldehyde 3–phosphate dehydrogenase (GAPDH) were monitored as a quantitative control in western blot analysis. (B) IHC detection of POSTN showing elevated levels staining (in brown) in fibroid (III, IV) compared with ANM (I, II; n = 4 normal and n = 15 fibroids). Correlation analysis of pSMAD2 and POSTN expression levels in human fibroid tissues (n = 30). Scale bars: 100 µm. +, MED12 mutation–positive; −, MED12 mutation–negative; wt, wild-type. Discussion Uterine fibroids affect up to 70% of women during their reproductive years and are an important source of gynecological and obstetrical problems (1, 2). Despite the major health care burden posed by fibroids, very little is known about their etiology and pathogenesis. Treatment strategies for these benign tumors are limited to invasive surgical procedures and hormonal therapies. Therefore, identification of novel drug targets and improved treatment strategies are required. Our research data showed that fibroid clusters taken from the same patient possess diverse genomic profiles. In the literature, mutations in exon 2 of the MED12 gene are the most common genetic anomalies in fibroids (8), and our genomic data supported this observation. Our study determined the MED12 status of fibroids in the Australian population. We identified that 27 tumors from 11 patients that displayed missense mutations affecting codon 44 in exon 2 of MED12. Glycine residue at codon 44 appears to be essential for normal MED12 function as all six base substitutions lead to changes in this amino acid. Excessive ECM deposition is a major hallmark of this disease (5), which contributes to fibroid growth and bulk-type symptoms. Despite the significance of ECM in fibroids, there has been limited research on ECM characterization and expression in fibroids. Our study investigated whether MED12 status and fibroid size influence ECM composition and expression patterns. After genetic analysis, we generated heat maps showing protein upregulation and downregulation in fibroids compared with ANM control tissues. Interestingly, we observed that, overall, the set of proteins (ECM-related and otherwise) that were upregulated or downregulated were the same in all three patients, irrespective of MED12 mutation status and fibroid size. Because a copious amount of ECM is characteristic of fibroids, we next focused on the upregulated ECM proteins. Comparison of expression levels revealed that POSTN, an ECM protein, was one of the most upregulated proteins, with an average threefold higher expression in fibroids than in ANM. Because fibroids grow as they progress, small fibroids indicate early stages of disease. Therefore, because POSTN had upregulated expression levels in small fibroids, this ECM protein may have potential as a novel drug target to hinder the progression of uterine leiomyogenesis. We further validated the expression of POSTN in our fibroid samples using western blot and IHC. Our proteomic and western blot analysis demonstrated that in all cases of MED12 mutation–positive and mutation–negative fibroids, POSTN was upregulated compared with ANM. In contrast, in the absence of mutations in MED12, POSTN expression was not consistent (Fig. 5A). In the case of MED12-negative fibroids, the expression of POSTN may be affected by the presence of other genetic mutations. Consequently, the effect of other genetic mutations on fibroid ECM protein expression requires further investigation. Our study focused on MED12 mutations because in the literature these are the most commonly detected mutations in fibroids. POSTN is a 90-kDa secreted ECM protein with a multidomain structure that contains an amino-terminal EMI domain, a tandem repeat of four fasciclin (FAS1) domains, and a carboxyl-terminal domain (28–30). Each domain interacts with specific ECM proteins and cell-surface integrins to promote the assembly of extracellular architectures (31–34), which govern the biomechanical properties of connective tissues. Previous studies have shown that POSTN interacts with collagens, laminins, and tenascins (31–34). In fibroids of all sizes (small, medium, and large), we observed upregulation of POSTN, along with upregulation of collagens, laminins, and tenascins (Fig. 3A–3C). This suggests that these proteins are important contributors to the excessive ECM build-up that leads to fibroid expansion. Further studies are required to target the binding sites on the multidomain structure of POSTN to investigate the relationship between POSTN and the interacting proteins. POSTN plays a central role in normal tissue homeostasis and disease development (34). Previous studies have demonstrated that high levels of POSTN expression are associated with a variety of cancers, including head and neck (35, 36), oral (37), lung (38), breast (39), ovarian (40), colon (41), pancreatic (42), and liver (43, 44) cancer. POSTN participates in many biological processes involved in cancer, including cell adhesion, invasion, metastasis, and tumor angiogenesis (36, 37, 41, 45). Our study compared the expression level of POSTN in fibroids relative to ANM. POSTN is a candidate for further functional and clinical investigations given its abundance in uterine fibroids. The molecular mechanism of POSTN is still unclear, bu a recent study revealed that it interacts with integrins and activates focal adhesion kinase and PI3K-Akt–mediated signaling pathways, promoting tumor angiogenesis, invasion, and metastasis (46). Our STRING analysis also identified the involvement of these signaling networks in fibroids (Supplemental Fig. 3). Our proteomic analysis revealed that TGF-β was also upregulated in fibroids of all sizes compared with ANM (Fig. 3A–3C). Deletion of the Periostin (Postn) gene in a mouse model of muscular dystrophy altered TGF-β signaling, resulting in enhanced tissue regeneration and reduced levels of fibrosis, thereby providing evidence for interaction between POSTN and TGF-β (47). Therefore, we can also speculate that POSTN interacts with TGF-β signaling pathway to promote ECM formation in leiomyogenesis. Further research is warranted to define the functional role of POSTN and other ECM-related proteins in fibroid initiation and progression. In summary, our study defined the global protein expression patterns of fibroids vs ANM in human patients. Furthermore, we used genomic and proteomic analysis to study the relationship among exon 2 MED12 mutation status, fibroid size, and ECM protein expression. This analysis revealed that the same group of ECM proteins was upregulated or downregulated in all patients, despite variations in fibroid size and the MED12 gene. Furthermore, we validated the expression level of POSTN because it was one of the significantly upregulated ECM proteins identified in small fibroids and these fibroids indicate early stages of leiomyogenesis. Because excessive ECM is a prominent feature of fibroids, targeting upregulated ECM proteins is a rational approach for overcoming this disease. Investigation of the role of the ECM in cancer and other abnormalities has illustrated that the ECM is an active participant in disease progression and is responsive to surrounding cell types and signaling molecules. Further studies are essential for understanding the mechanisms and mediators responsible for the overproduction of specific ECM components. This understanding will provide an opportunity to develop intervening strategies that will curb the production of these ECM proteins, thereby reducing fibroid size and removing the associated disease burden. Abbreviations: ACN acetonitrile ANM adjacent normal myometrium ANXA1 annexin A1 ANXA2 annexin A2 COL12A1 collagen type XII α 1 COL2A1 collagen type II α 1 COL3A1 collagen type III α 1 COL5A2 collagen type V α 2 COL7A1 collagen type VII α 1 ECM extracellular matrix FMOD fibromodulin HILIC hydrophilic interaction liquid chromatography IHC immunohistochemistry LAMA5 laminin subunit α 5 LAMB2 laminin subunit β 2 LC liquid chromatography MED12 mediator complex subunit 12 gene MS mass spectrometry MS/MS tandem mass spectrometry PCR polymerase chain reaction POSTN periostin pSMAD2 phosphorylated form of mothers against decapentaplegic homolog 2 S100A6 protein S100-A6 SERPINA1 α-1-antitrypsin SMOC2 secreted protein acidic and rich in cysteine–related modular calcium-binding protein 2 SPARC secreted protein acidic and rich in cysteine SPARCL1 secreted protein acidic and rich in cysteine–like 1 TFA trifluoroacetic acid TGF-β transforming growth factor-β TGFBI transforming growth factor-β–induced protein ig-h3 TINAGL1 tubulointerstitial nephritis antigen-like TNC tenascin VCAN versican core protein. Acknowledgments The authors thank Dr. Ben Crossett and Dr. Trisha Al Mazi for help with HILIC and Nathan Smith for help with LC-MS/MS. Financial Support: Work in the Tanwar laboratory was in part supported by funding from the National Health and Medical Research Council (Grant APP1081461 to P.S.T.), the Australian Research Council, the Cancer Institute New South Wales, and the John Hunter Hospital Charitable Trust. Y.B., P.B., and P.B.N. are recipients of the University of Newcastle Postgraduate Research Fellowship. M.D.D. is supported by a Cancer Institute New South Wales, Australia Early Career Fellowship. Author Contributions: M.F.B.J. and P.S.T. designed the research. M.F.B.J., Y.-A. K., M.K., Y.B., P.B., P.B.N, and D.A.S.-B. performed the research. M.F.B.J., H.H., M.A.B., M.D.D., R.J.S., P.N., and P.S.T. analyzed the data. M.F.B.J., Y.B., and P.S.T. wrote the paper. P.S.T. supervised the study, provided financial support, and edited and had final approval of the manuscript. All authors reviewed, approved, and commented on the manuscript. Disclosure Summary: The authors have nothing to disclose. References 1. Cramer SF, Patel A. The frequency of uterine leiomyomas. Am J Clin Pathol . 1990; 94( 4): 435– 438. Google Scholar CrossRef Search ADS PubMed  2. Baird DD, Dunson DB, Hill MC, Cousins D, Schectman JM. High cumulative incidence of uterine leiomyoma in black and white women: ultrasound evidence. Am J Obstet Gynecol . 2003; 188( 1): 100– 107. Google Scholar CrossRef Search ADS PubMed  3. Arslan AA, Gold LI, Mittal K, Suen TC, Belitskaya-Levy I, Tang MS, Toniolo P. Gene expression studies provide clues to the pathogenesis of uterine leiomyoma: new evidence and a systematic review. Hum Reprod . 2005; 20( 4): 852– 863. Google Scholar CrossRef Search ADS PubMed  4. Bulun SE. Uterine fibroids. N Engl J Med . 2013; 369( 14): 1344– 1355. Google Scholar CrossRef Search ADS PubMed  5. Stewart EA, Laughlin-Tommaso SK, Catherino WH, Lalitkumar S, Gupta D, Vollenhoven B. Uterine fibroids. Nat Rev Dis Primers . 2016; 2: 16043. Google Scholar CrossRef Search ADS PubMed  6. Mehine M, Kaasinen E, Mäkinen N, Katainen R, Kämpjärvi K, Pitkänen E, Heinonen HR, Bützow R, Kilpivaara O, Kuosmanen A, Ristolainen H, Gentile M, Sjöberg J, Vahteristo P, Aaltonen LA. Characterization of uterine leiomyomas by whole-genome sequencing. N Engl J Med . 2013; 369( 1): 43– 53. Google Scholar CrossRef Search ADS PubMed  7. Mehine M, Kaasinen E, Heinonen H-R, Mäkinen N, Kämpjärvi K, Sarvilinna N, Aavikko M, Vähärautio A, Pasanen A, Bützow R, Heikinheimo O, Sjöberg J, Pitkänen E, Vahteristo P, Aaltonen LA. Integrated data analysis reveals uterine leiomyoma subtypes with distinct driver pathways and biomarkers. Proc Natl Acad Sci USA . 2016; 113( 5): 1315– 1320. Google Scholar CrossRef Search ADS PubMed  8. Mäkinen N, Mehine M, Tolvanen J, Kaasinen E, Li Y, Lehtonen HJ, Gentile M, Yan J, Enge M, Taipale M, Aavikko M, Katainen R, Virolainen E, Böhling T, Koski TA, Launonen V, Sjöberg J, Taipale J, Vahteristo P, Aaltonen LA. MED12, the mediator complex subunit 12 gene, is mutated at high frequency in uterine leiomyomas. Science . 2011; 334( 6053): 252– 255. Google Scholar CrossRef Search ADS PubMed  9. Kämpjärvi K, Park MJ, Mehine M, Kim NH, Clark AD, Bützow R, Böhling T, Böhm J, Mecklin JP, Järvinen H, Tomlinson IP, van der Spuy ZM, Sjöberg J, Boyer TG, Vahteristo P. Mutations in Exon 1 highlight the role of MED12 in uterine leiomyomas. Hum Mutat . 2014; 35( 9): 1136– 1141. Google Scholar CrossRef Search ADS PubMed  10. Taatjes DJ. The human Mediator complex: a versatile, genome-wide regulator of transcription. Trends Biochem Sci . 2010; 35( 6): 315– 322. Google Scholar CrossRef Search ADS PubMed  11. Kämpjärvi K, Mäkinen N, Mehine M, Välipakka S, Uimari O, Pitkänen E, Heinonen HR, Heikkinen T, Tolvanen J, Ahtikoski A, Frizzell N, Sarvilinna N, Sjöberg J, Bützow R, Aaltonen LA, Vahteristo P. MED12 mutations and FH inactivation are mutually exclusive in uterine leiomyomas. Br J Cancer . 2016; 114( 12): 1405– 1411. Google Scholar CrossRef Search ADS PubMed  12. Mittal P, Shin YH, Yatsenko SA, Castro CA, Surti U, Rajkovic A. Med12 gain-of-function mutation causes leiomyomas and genomic instability. J Clin Invest . 2015; 125( 8): 3280– 3284. Google Scholar CrossRef Search ADS PubMed  13. Moravek MB, Yin P, Ono M, Coon JS V, Dyson MT, Navarro A, Marsh EE, Chakravarti D, Kim JJ, Wei J-J, Bulun SE. Ovarian steroids, stem cells and uterine leiomyoma: therapeutic implications. Hum Reprod Update . 2015; 21( 1): 1– 12. Google Scholar CrossRef Search ADS PubMed  14. Peddada SD, Laughlin SK, Miner K, Guyon JP, Haneke K, Vahdat HL, Semelka RC, Kowalik A, Armao D, Davis B, Baird DD. Growth of uterine leiomyomata among premenopausal black and white women. Proc Natl Acad Sci USA . 2008; 105( 50): 19887– 19892. Google Scholar CrossRef Search ADS PubMed  15. Stewart EA, Shuster LT, Rocca WA. Reassessing hysterectomy. Minn Med . 2012; 95( 3): 36– 39. Google Scholar PubMed  16. Mammoto T, Ingber DE. Mechanical control of tissue and organ development. Development . 2010; 137( 9): 1407– 1420. Google Scholar CrossRef Search ADS PubMed  17. Bonnans C, Chou J, Werb Z. Remodelling the extracellular matrix in development and disease. Nat Rev Mol Cell Biol . 2014; 15( 12): 786– 801. Google Scholar CrossRef Search ADS PubMed  18. Pickup MW, Mouw JK, Weaver VM. The extracellular matrix modulates the hallmarks of cancer. EMBO Rep . 2014; 15( 12): 1243– 1253. Google Scholar CrossRef Search ADS PubMed  19. Malik M, Norian J, McCarthy-Keith D, Britten J, Catherino WH. Why leiomyomas are called fibroids: the central role of extracellular matrix in symptomatic women. Semin Reprod Med . 2010; 28( 3): 169– 179. Google Scholar CrossRef Search ADS PubMed  20. Dun MD, Chalkley RJ, Faulkner S, Keene S, Avery-Kiejda KA, Scott RJ, Falkenby LG, Cairns MJ, Larsen MR, Bradshaw RA, Hondermarck H. Proteotranscriptomic profiling of 231-BR breast cancer cells: identification of potential biomarkers and therapeutic targets for brain metastasis. Mol Cell Proteomics . 2015; 14( 9): 2316– 2330. Google Scholar CrossRef Search ADS PubMed  21. Bajwa P, Nielsen S, Lombard JM, Rassam L, Nahar P, Rueda BR, Wilkinson JE, Miller RA, Tanwar PS. Overactive mTOR signaling leads to endometrial hyperplasia in aged women and mice. Oncotarget . 2017; 8( 5): 7265– 7275. Google Scholar CrossRef Search ADS PubMed  22. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc B . 1995; 57: 289– 300. 23. Markowski DN, Bartnitzke S, Löning T, Drieschner N, Helmke BM, Bullerdiek J. MED12 mutations in uterine fibroids--their relationship to cytogenetic subgroups. Int J Cancer . 2012; 131( 7): 1528– 1536. Google Scholar CrossRef Search ADS PubMed  24. McGuire MM, Yatsenko A, Hoffner L, Jones M, Surti U, Rajkovic A. Whole exome sequencing in a random sample of North American women with leiomyomas identifies MED12 mutations in majority of uterine leiomyomas. PLoS One . 2012; 7( 3): e33251. Google Scholar CrossRef Search ADS PubMed  25. Turunen M, Spaeth JM, Keskitalo S, Park MJ, Kivioja T, Clark AD, Mäkinen N, Gao F, Palin K, Nurkkala H, Vähärautio A, Aavikko M, Kämpjärvi K, Vahteristo P, Kim CA, Aaltonen LA, Varjosalo M, Taipale J, Boyer TG. Uterine leiomyoma-linked MED12 mutations disrupt mediator-associated CDK activity. Cell Reports . 2014; 7( 3): 654– 660. Google Scholar CrossRef Search ADS PubMed  26. Kämpjärvi K, Järvinen TM, Heikkinen T, Ruppert AS, Senter L, Hoag KW, Dufva O, Kontro M, Rassenti L, Hertlein E, Kipps TJ, Porkka K, Byrd JC, de la Chapelle A, Vahteristo P. Somatic MED12 mutations are associated with poor prognosis markers in chronic lymphocytic leukemia. Oncotarget . 2015; 6( 3): 1884– 1888. Google Scholar CrossRef Search ADS PubMed  27. Kudo A. Introductory review: periostin-gene and protein structure. Cell Mol Life Sci . 2017; 74( 23): 4259– 4268. Google Scholar CrossRef Search ADS PubMed  28. Horiuchi K, Amizuka N, Takeshita S, Takamatsu H, Katsuura M, Ozawa H, Toyama Y, Bonewald LF, Kudo A. Identification and characterization of a novel protein, periostin, with restricted expression to periosteum and periodontal ligament and increased expression by transforming growth factor beta. J Bone Min Res . 1999; 14( 7): 1239– 1249. Google Scholar CrossRef Search ADS   29. Kudo A. Periostin in fibrillogenesis for tissue regeneration: periostin actions inside and outside the cell. Cell Mol Life Sci . 2011; 68( 19): 3201– 3207. Google Scholar CrossRef Search ADS PubMed  30. Sugiura T, Takamatsu H, Kudo A, Amann E. Expression and characterization of murine osteoblast-specific factor 2 (OSF-2) in a baculovirus expression system. Protein Expr Purif . 1995; 6( 3): 305– 311. Google Scholar CrossRef Search ADS PubMed  31. Kii I, Nishiyama T, Li M, Matsumoto K, Saito M, Amizuka N, Kudo A. Incorporation of tenascin-C into the extracellular matrix by periostin underlies an extracellular meshwork architecture. J Biol Chem . 2010; 285( 3): 2028– 2039. Google Scholar CrossRef Search ADS PubMed  32. Maruhashi T, Kii I, Saito M, Kudo A. Interaction between periostin and BMP-1 promotes proteolytic activation of lysyl oxidase. J Biol Chem . 2010; 285( 17): 13294– 13303. Google Scholar CrossRef Search ADS PubMed  33. Snider P, Hinton RB, Moreno-Rodriguez RA, Wang J, Rogers R, Lindsley A, Li F, Ingram DA, Menick D, Field L, Firulli AB, Molkentin JD, Markwald R, Conway SJ. Periostin is required for maturation and extracellular matrix stabilization of noncardiomyocyte lineages of the heart. Circ Res . 2008; 102( 7): 752– 760. Google Scholar CrossRef Search ADS PubMed  34. Kii I, Ito H. Periostin and its interacting proteins in the construction of extracellular architectures. Cell Mol Life Sci . 2017; 74( 23): 4269– 4277. Google Scholar CrossRef Search ADS PubMed  35. Chang Y, Lee TC, Li JC, Lai TL, Chua HH, Chen CL, Doong SL, Chou CK, Sheen TS, Tsai CH. Differential expression of osteoblast-specific factor 2 and polymeric immunoglobulin receptor genes in nasopharyngeal carcinoma. Head Neck . 2005; 27( 10): 873– 882. Google Scholar CrossRef Search ADS PubMed  36. Kudo Y, Ogawa I, Kitajima S, Kitagawa M, Kawai H, Gaffney PM, Miyauchi M, Takata T. Periostin promotes invasion and anchorage-independent growth in the metastatic process of head and neck cancer. Cancer Res . 2006; 66( 14): 6928– 6935. Google Scholar CrossRef Search ADS PubMed  37. Siriwardena BS, Kudo Y, Ogawa I, Kitagawa M, Kitajima S, Hatano H, Tilakaratne WM, Miyauchi M, Takata T. Periostin is frequently overexpressed and enhances invasion and angiogenesis in oral cancer. Br J Cancer . 2006; 95( 10): 1396– 1403. Google Scholar CrossRef Search ADS PubMed  38. Ouyang G, Liu M, Ruan K, Song G, Mao Y, Bao S. Upregulated expression of periostin by hypoxia in non-small-cell lung cancer cells promotes cell survival via the Akt/PKB pathway. Cancer Lett . 2009; 281( 2): 213– 219. Google Scholar CrossRef Search ADS PubMed  39. Shao R, Bao S, Bai X, Blanchette C, Anderson RM, Dang T, Gishizky ML, Marks JR, Wang XF. Acquired expression of periostin by human breast cancers promotes tumor angiogenesis through up-regulation of vascular endothelial growth factor receptor 2 expression. Mol Cell Biol . 2004; 24( 9): 3992– 4003. Google Scholar CrossRef Search ADS PubMed  40. Ismail RS, Baldwin RL, Fang J, Browning D, Karlan BY, Gasson JC, Chang DD. Differential gene expression between normal and tumor-derived ovarian epithelial cells. Cancer Res . 2000; 60( 23): 6744– 6749. Google Scholar PubMed  41. Bao S, Ouyang G, Bai X, Huang Z, Ma C, Liu M, Shao R, Anderson RM, Rich JN, Wang XF. Periostin potently promotes metastatic growth of colon cancer by augmenting cell survival via the Akt/PKB pathway. Cancer Cell . 2004; 5( 4): 329– 339. Google Scholar CrossRef Search ADS PubMed  42. Baril P, Gangeswaran R, Mahon PC, Caulee K, Kocher HM, Harada T, Zhu M, Kalthoff H, Crnogorac-Jurcevic T, Lemoine NR. Periostin promotes invasiveness and resistance of pancreatic cancer cells to hypoxia-induced cell death: role of the beta4 integrin and the PI3k pathway. Oncogene . 2007; 26( 14): 2082– 2094. Google Scholar CrossRef Search ADS PubMed  43. Lv Y, Wang W, Jia WD, Sun QK, Huang M, Zhou HC, Xia HH, Liu WB, Chen H, Sun SN, Xu GL. High preoparative levels of serum periostin are associated with poor prognosis in patients with hepatocellular carcinoma after hepatectomy. Eur J Surg Oncol . 2013; 39( 10): 1129– 1135. Google Scholar CrossRef Search ADS PubMed  44. Lv Y, Wang W, Jia WD, Sun QK, Li JS, Ma JL, Liu WB, Zhou HC, Ge YS, Yu JH, Xia HH, Xu GL. High-level expression of periostin is closely related to metastatic potential and poor prognosis of hepatocellular carcinoma. Med Oncol . 2013; 30( 1): 385. Google Scholar CrossRef Search ADS PubMed  45. Gillan L, Matei D, Fishman DA, Gerbin CS, Karlan BY, Chang DD. Periostin Secreted by Epithelial Ovarian Carcinoma Is a Ligand for αVβ3 and αVβ5 integrins and promotes cell motility. Cancer Res . 2002; 62: 5358– 5364. Google Scholar PubMed  46. Utispan K, Sonongbua J, Thuwajit P, Chau-In S, Pairojkul C, Wongkham S, Thuwajit C. Periostin activates integrin α5β1 through a PI3K/AKT‑dependent pathway in invasion of cholangiocarcinoma. Int J Oncol . 2012; 41( 3): 1110– 1118. Google Scholar CrossRef Search ADS PubMed  47. Lorts A, Schwanekamp JA, Baudino TA, McNally EM, Molkentin JD. Deletion of periostin reduces muscular dystrophy and fibrosis in mice by modulating the transforming growth factor-β pathway. Proc Natl Acad Sci USA . 2012; 109( 27): 10978– 10983. Google Scholar CrossRef Search ADS PubMed  Copyright © 2018 Endocrine Society

Journal

EndocrinologyOxford University Press

Published: Feb 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off