All evaluated HGSOC and fallopian tube (FT) tissues were collected from the Department of Obstetrics and Gynecology, Qilu Hospital, Shandong University from 2006 to 2016. The 4 µm tissue microarrays (TMAs), including 210 cases of HGSOC and 42 normal FT tissues, were designed in our laboratory. Fresh tissues of 28 cases of HGSOC and 14 normal FT tissues were also collected. Moreover, 18 specimens with chemosensitivity information for second-line chemotherapy were obtained from the specimen library. HGSOC tissues were obtained from patients who underwent primary surgery and no neo-adjuvant chemotherapy was performed. The normal FT tissues were obtained from patients who received hysterectomy and salpingo-oophorectomies due to benign disease. The clinicopathological characteristics of the patients are presented in Table I. The accurate assessment of the disease response was based on RECIST or GCIG standards (11). Ethics approval was obtained from the Ethics Committee of Shandong University, and written informed consent was obtained from each patient.

The A2780 human ovarian cancer cell line was originally established from the tumor tissue of an untreated patient, and the cells grew as a monolayer and in suspension in spinner cultures. The SKOV3 cell line was originally isolated from the ascites of patients with ovarian tumors, and tumorigenesis in nude mice resulted in a moderately differentiated adenocarcinoma consistent with carcinoma in situ of the ovary. The above-mentioned cells were cultured in RPMI-1640 medium and McCoy's 5A medium, respectively. 293T cells, which could continuously express SV40 antigen, was used in the transfection experiments with a high transfection efficiency, and was cultured in DMEM. The A2780 cell line was originally purchased from the European Collection of Authenticated Cell Cultures (ECACC; Cat. no. 93112519). The SKOV3 and 293T cells were purchased from the American Type Culture Collection (ATCC; Cat. nos. HTB-77 and CRL-11268, respectively). All the culture media were supplemented with 10% fetal bovine serum (FBS). All these cells were cultured in a humidified incubator under standard culture conditions (37°C, 5% CO₂).

Total RNA from cells and tissues was isolated using TRIZOL reagent (Invitrogen; Thermo Fisher Scientific) and cDNA was synthesized using the PrimeScript RT reagent kit (Takara). qPCR was performed using SYBR-Green qPCR master mix (Takara). The conditions for PCR indluced 3 stages: Hold stage (95°C, 30 sec), PCR stage for 40 cycles (95°C, 5 sec and 60°C, 34 sec), melt curve stage (95°C, 15 sec; 60°C, 1 min and 95°C, 15 sec).GAPDH served as the endogenous control. The primer sequences of PRC1 for RT-qPCR were as follows: Forward primer, ACA CTC TGT GCA GCG AGT TAC; reverse primer, TTC GCA TCA ATT CCA CTT GGG. The primer sequences of GAPDH were as follows: Forward, ACA ACT TTG GTA TCG TGG AAG G and reverse, GCC ATC ACG CCA CAG TTT C. The method of quantification was relative quantification and ΔΔcq was calculated to analyze the relative gene expression (12).

A lentivirus vector expressing shPRC1 (TRCN0000280715) was purchased from Sigma-Aldrich. siRNA was synthesized by the GenePharma. The sequences were as follows: PRC1, 5′-CGC UGU UUA CUC AUA CAG U-3′; forkhead box protein M1 (FOXM1), 5′-GGA CCA CUU UCC CUA CUU UUU-3′ and negative control (NC), 5′-UUC UCC GAA CGU GUC ACG UdT dT-3′.

Lentivirus was produced in 293T cells packaged by psPAX2 and pMD2.G. The cells were infected with the 1 ml lentivirus liquid for 24 h in the presence of polybrene (8 µg/ml), and was selected with puromycin (2 µg/ml) for 1 week to acquire stable expressing cell lines. siRNA manipulation was performed in accordance with the instructions of the manufacturer of Lipofectamine 2000 (Invitrogen™). The time duration between transfection subsequent experimentation was approximately 10-14 h.

3-(4,5-Dimethyl-2-thiazolyl)-2,5-diphenyl-2-H-tetrazolium bromide (MTT) assay was firstly used to measure the proliferative ability of the A2780 cells subjected to PRC1 knockdown or overexpression compared to the cells trans-fected with empty plasmid. The A2780 cells were incubated in RPMI-1640 medium supplemented with 10% FBS, and were seeded in 96-well plates at concentrations of 1,000 cells per well. Cell proliferation was monitored at different time points (1-5 days). At fixed time points, 20 µl of MTT (Sigma-Aldrich) solution were added to each well. Following incubation for 4 h at 37°C, the supernatant was removed and 100 µl of DMSO (Sigma-Aldrich) were added to each well. The absorbance at 490 nm was evaluated using a microplate reader (Bio-Rad).

In addition, the changes in the IC₅₀ values of cisplatin, taxol and pegylated liposomal doxorubicin (LPD) were evaluated in the A2780 cells subjected to PRC1 knockdown or overexpression compared to the controls. The A2780 cells were seeded in 96-well plates at concentrations of 4,000 cells per well and were treated with gradually increasing concentrations of the drugs (as shown in Fig. 5C-F). Following incubation for 48 h at 37°C, 20 µl of MTT (Sigma-Aldrich) solution were added to each well. The remaining steps were the same as mentioned above.

Colony formation assay was used to measure the proliferative ability of the A2780 cells subjected to PRC1 knockdown or overexpression compared to the controls. The A2780 cells were seeded in a 6-well plate with 1,000 cells per well, and cultured at 37°C for 10-14 days. Colonies were fixed with methanol for 15 min and stained with 0.5% crystal violet (Solarbio) for 20 min at room temperature. The number of colonies with >50 cells was counted under a microscope (IX71, Olympus).

Cell cycle analysis was performed to verify whether PRC1 regulates the cell cycle. The A2780 cells transfected with PRC1 siRNA and the controls were harvested and fixed with 75% ice-cold ethanol overnight at 4°C. Cell suspensions were stained with 1 ml propidium iodide stock solution (MultiSciences Lianke Biotech Co.) and analyzed using a FACScan flow cytometer (BD Biosciences).

Transwell assay was performed to confirm whether RPC1 knockdown and overexpression affected the invasive and migratory abilities of the A2780 and SKOV3 cells. In addition, it was used to determine whether the silencing of PRC1 can reverse the FOXM1-mediated enhancement of the metastatic abilities of the A2780 cells. The cells were re-suspended in serum-free medium, and 200 µl suspension containing 1-2×10⁵ cells was seeded into the upper chambers, and 700 µl culture medium containing 20% fetal bovine serum was added to the lower compartment. The filter membrane of invasion assays was coated with diluted Matrigel (BD Biosciences). The cells that penetrated to the bottom were fixed with methanol and stained with 0.5% crystal violet (Solarbio) at room temperature for >30 min following incubation under standard culture conditions (37°C, 5% CO₂). The incubation times for the migration of the A2780 and SKOV3 cells were 24 and 10 h, and the times for invasion was 36 and 18 h, respectively. A light microscope (IX71, Olympus) was used to examine the cells and obtain images (×100 magnification).

To confirm whether PRC1 overexpression in A2780 cells facilitates the migratioryability compared with the control, a wound healing assay was performed. The A2780 cells were seeded in 24-well plate at 3.0×10⁵ per well. A 20 µl pipette tip was used to scratch a line when the cells reached 100% confluency as a monolayer. The cells were then cultured in serum-free medium. The distance between the two edges was monitored at 0 and 48 h following wound formation using a microscope (IX71, Olympus).

The fresh tissues and cells were lysed with RIPA buffer on ice. The BCA Protein Assay kit (Merck Millipore) was used to detect the protein concentration of the samples. The mass of protein loaded per lane was 50-90 µg. The sample proteins were separated by SDS-PAGE (separating gel, 10%; stacking gel, 5%) and electrotransferred onto a PVDF membrane (Merck Millipore). After blocking with 5% non-fat milk, the membrane was incubated overnight with primary antibodies at 4°C followed by appropriate HRP-conjugated secondary antibodies for 2 h at room temperature the following day. The ECL system (GE Healthcare) was used to detect the protein bands, and β-actin was used as an endogenous control. The primary antibodies in this study included: Anti-β-actin (1:5,000, F3022, Sigma-Aldrich), anti-PRC1 (1:500, ab51248, Abcam), anti-cyclin B1 (CCNB1; 1:1,000, ab32053, Abcam), anti-cyclin D1 (CCND1; 1:1,000, ab16663, Abcam), anti-Aurora B kinase (AURKB; 1:1,000, ab2254, Abcam), anti-p21 (1:1,000, ab109520, Abcam), anti-E-Cadherin (1:1,000, #3195, Cell Signaling Technology), anti-N-Cadherin (1:1,000, #13116, Cell Signaling Technology), anti-matrix metalloproteinase (MMP)9 (1:1,000, ab38898, Abcam), anti-Slug (1:1,000, ab27568, Abcam), anti-breast cancer 1 (BRCA1; 1:1,000, #9010, Cell Signaling Technology), anti-RAD51 (1:1,000, ab113534, Abcam), anti-poly(ADP-Ribose) polymerase 1 (PARP1; 1:1,000, ab32138, Abcam), anti-c-Myc (1:1,000, ab32072, Abcam) and anti-FOXM1 (1:1,000, ab207298, Abcam). The secondary antibodies were purchased from KPL. Anti-rabbit IgG (5220-0336) was diluted at 1:4,000 and anti-mouse IgG (5220-0341) was diluted at 1:6,000. ImageJ software (National Institutes of Health, 1.48v) was used to calculate the relative density.

Immunohistochemical staining was performed in 210 cases of HGSOC and 42 normal FT tissues to determine the expression and clinical significance of PRC1 in HGSOC. In addition, the immunohistochemical staining of PRC1 and FOXM1 in corresponding HGSOC samples was analyzed to verify the association between PRC1 and FOXM1. The TMAs were incubated at 60°C for 1 h, and subsequently deparaffinized in xylene and hydrated in a degraded series of ethanol. The heat-mediated antigen retrieval in EDTA buffer (pH 9.0), the inactivation of endogenous peroxidase activity and the blocking of non-specific antigens were then performed gradually. PRC1 antibody (1:100 dilution, HPA034521, Sigma-Aldrich) was incubated with the slides at 4°C overnight followed by secondary antibodies (SP-9000, ZSBG-BIO) incubation for 20 min at room temperature. Then the sections were incubated with streptavidin-peroxidase for 15 min at room temperature. The staining was detected with the DAB detection system (Zhongshan Biotechnology Co.). Two investigators blinded to the clinical data evaluated the staining. Each sample had two duplicates and the average scores were used as the final result.

To determine whether FOXM1 can regulate the expression of PRC1 and identify the binding site, a dual-luciferase reporter assay was performed. The 293T cells were plated in 96-well plates and cultured for 24 h at 37°C, and transiently co-transfected with PCMV-FOXM1C (RC202540, Origene), PGL4.26-PRC1 (E8441, Promega) and pRL-TK (E2241, Promega) using Lipofectamine 2000 (Invitrogen™). The luciferase activity was evaluated using the Dual-Glo^® Lu ciferase Assay System E2920 which was supplied by Promega. After 24 h, 75 µl fresh medium were added to the 96-well plates after the medium was removed. This was followed by the addition of 75 µl lysis reagent and Firefly luminescence was measured using a microplate reader (Bio-Rad) after 10 min. Subsequently, 75 µl Stop buffer were added to each well and the Renilla luminescence was measured after 10 min. The relative luciferase activity was determined by the ratio of values between Firefly luminescence and Renilla luminescence.

In order to verify whether PRC1 expression is associated with germline BRCA mutation, germ-line BRCA genetic testing was performed in 52 patients. Fresh blood of 6 ml was extracted and sequenced using the NGS platform from Shanghai Topgen Bio-pharm Co. Ltd. The BRCA1/2 panel (Morgen, China) was used which covers the entire coding sequences of BRCA1 and BRCA2, including 10-50 bases of adjacent intronic sequence of each exon. The variants were classified based on a highly accepted 5-class classification (13).

Oncomine (www.oncomine.org) was used to visualize the differential expression of PRC1 in ovarian cancer and control samples. TCGA RNA expression data of ovarian serous cystadenocarcinoma were analyzed by the Cancer Genomics Browser (https://genome-cancer.ucsc.edu). Kaplan Meier-plotter (http://kmplot.com/analysis/) was used to analyze overall survival and the progression-free survival of patients as regards PRC1 expression in ovarian cancer. Gene regulation website (www.gene-regulation.com) was used to analyze the promoter of PRC1. Pearson's correlation analysis was used to analyze the correlation of PRC1 and FOXM1 expression in TCGA cohort.

Statistical analysis was carried out using SPSS 23 software. The differences between continuous data were analyzed using a Student's t-test, and the comparisons between multiple groups were performed by one-way ANOVA, and Fishers' Least Significant Difference (LSD) was used as a post hoc test. The association between PRC1 expression and the clinical characteristics of the patients were analyzed using the Chi-square test. Multivariate cox regression analysis was used to analyze the association between clinical prognostic markers and overall survival. Overall survival analysis was performed by Kaplan-Meier and the log-rank test. A value of P<0.05 was considered to indicate a statistically significant difference.

To determine the expression of PRC1 in HGSOC, the publicly accessible database Oncomine and TCGA cohort were employed to analyze PRC1 mRNA expression, and it was observed that PRC1 mRNA expression in serous ovarian carcinoma (SOC) was significantly higher compared with that in normal ovary tissues or normal peritoneum tissues (Fig. 1A-C). It was also found that the expression of PRC1 was positively associated with the malignancy of ovarian tumors (Fig. 1E and F). To confirm PRC1 overexpression in HGSOC, RT-qPCR was performed to measure PRC1 expression in normal FT (n=14) and HGSOC (n=28) tissues, and it was observed that the PRC1 mRNA level was significantly higher in HGSOC (Fig. 1D). Subsequently, western blot analysis was performed to determine the protein expression levels and the relative protein expression of PRC1 was calculated in normal FT (n=7) and HGSOC (n=8) tissues. The results obtained were similar to those for mRNA expression (Fig. 1G and H). All these results verified that PRC1 was markedly overexpressed in HGSOC tissues at both the mRNA and protein level, and suggested that PRC1 plays an important role in ovarian cancer development and biological characteristics.

To further determine the expression and clinical significance of PRC1 in HGSOC, an immunohistochemistry assay was conducted to examine PRC1 protein expression in 210 HGSOC tissues and 42 normal FT tissues. As shown in Fig. 2E, PRC1 staining was mainly distributed in the cytoplasm, and the expression of the PRC1 was significantly upregulated in HGSOC tissues. To better understand the significance of PRC1 in HGSOC, the association between the expression of PRC1 and the clinicopathological parameters of 210 patients was analyzed. The results revealed that the overexpression of PRC1 was significantly associated with platinum-based chemotherapy sensitivity and the secondary recurrence intervals (P<0.05). Among the 95 patients with FIGO stage III-IV disease who received satisfactory cytoreductive surgery, 7 patients were resistant to platinum-based chemotherapy, and the staining of these 7 patients, surprisingly, revealed a high expression of PRC1. For second-line chemotherapy, 5 patients had a direct progression of the disease following partial remission among the 18 drug-resistant cases with PRC1 high expression. The primary and secondary recurrence interval and the mean recurrence interval were 12.1 vs. 10.2 months (P>0.05) and 7.3 vs. 4.9 months (P<0.05), respectively. In total, 52 cases of germline BRCA mutations were summarized; 17 cases were BRCA mutation carriers, and 13 of these 17 patients had a low expression of PRC1, with a P-value of 0.019 (Table I). However, no significant association was observed between the expression of PRC1 and the age of the patients, FIGO stage, the CA-125 level, ascites volume, residual lesions, family history and the primary recurrence interval (P>0.05). All the above-mentioned data confirmed that PRC1 played a major role in drug resistance and recurrence of HGSOC.

Kaplan Meier-plotter was used to investigate the effects of PRC1 expression on clinical prognosis, and the results revealed that the expression of PRC1 was negatively associated with the survival time (Fig. 2A-D). In total, 143 patients with HGSOC with complete survival information were analyzed to evaluate the importance of PRC1 overexpression in predicting HGSOC clinical outcomes according to immunohistochemical staining. It is found that patients with an elevated PRC1 expression had extremely significant poor outcomes in the 3-year survival rate, 5-year survival rate (Table I) and overall survival (Fig. 2F). Notably, multivariate Cox regression analysis revealed that a high expression of PRC1 protein was a unique independent prognostic factor for patients with HGSOC (Table II). These results provide evidence that a high expression of PRC1 indicates a worse prognosis of patients with HGSOC.

To evaluate the biological function of PRC1 in ovarian cancer, stable ovarian cancer cell lines with PRC1 overexpression and knockdown were established. The effect of PRC1 on the proliferative ability of the A2780 cells was then examined. Both growth curve analysis and colony formation assay demonstrated that PRC1 knockdown markedly inhibited cell growth and that PRC1 overexpression significantly enhanced cell proliferation (Fig. 3A and B). To further verify whether PRC1 regulates the cell cycle, PRC1 was knocked down in A2780 cells using siRNA and the changes in the cell cycle were then examined using fluorescence-activated cell sorting (FACS) analysis. As shown in Fig. 3E, PRC1 depletion significantly increased the percentage of cells in the G2 phase and decreased the percentage of cells in the G1 phase. In addition, the results of western blot analysis revealed that the CCNB1, CCND1 and AURKB expression levels were decreased and p21 expression was increased following PRC1 knockdown, while the corresponding expression levels were reversed with PRC1 overexpression in A2780 cells (Fig. 3C and D). All these results suggested that PRC1 promoted ovarian cancer proliferation in vitro.

A wound healing assay was first performed to determine the effects of PRC1 expression on the metastasis of ovarian cancer cells, and the results reveled that PRC1 overexpression in A2780 cells significantly facilitated the migratory ability of the cells compared with the control (Fig. 4E). Transwell assay also revealed that the migratory and invasive ability of the cells was significantly inhibited in the PRC1-depleted cells; on the contrary however, the metastatic ability was enhanced in the PRC1-overexpressing cells (Fig. 4A and B). Finally, the levels of epithelial-mesenchymal transition (EMT) markers were measured by western blot analysis, as shown in Fig. 4C and D. The expression of E-Cadherin, N-Cadherin, MMP9 and Slug was decreased following PRC1 knockdown. Taken together, all these data suggest an important role of PRC1 in the migratory and invasive ability of ovarian cancer cells via EMT.

Immunohistochemical staining was first applied to examine the expression of PRC1 in HGSOC samples with second-line treatment. As shown in Fig. 5A, the expression of PRC1 in the resistant samples was higher than in the sensitive ones. In total, 38.3% (18/47) patients with a high expression of PRC1 were resistant to platinum-based chemotherapy, while only 17.1% (7/41) of patients with a low expression of PRC1 were resistant (Table I). The mRNA expression level of PRC1 in the partial samples with second-line treatment described above was further investigated; it was found that 10 patients were sensitive to platinum, 8 were platinum-resistant, and the expression of PRC1 was evidently higher in the samples of resistant patients (Fig. 5B). Finally, MTT assay was performed to evaluate the changes in the IC₅₀ values of cisplatin, taxol and LPD, and the results revealed that PRC1 knockdown significantly enhanced the sensitivity of the chemotherapeutic drugs (Fig. 5C-F). In addition, it was also found that PRC1 knockdown downregulated BRCA1, RAD51, PARP1 and c-Myc expression in SKOV3 cells and the expression of c-Myc was also downregulated in the A2780 cells (Fig. 5G and H). These results confirmed that PRC1 played an important role in the drug resistance of ovarian cancer.

The promoter of PRC1 was analyzed using a gene regulation website (www.gene-regulation.com), and it was found that a number of transcription factors could bind to its promoter, including FOXM1. The expression of PRC1 was found to positively correlate with the expression of FOXM1 in the TCGA cohort analysis (Fig. 6B). It was found that the expression of PRC1 was markedly decreased at both the mRNA and protein level following FOXM1 knockdown (Fig. 6C-E). The results of immunohistochemical staining of corresponding HGSOC samples revealed that the expression of PRC1 was consistent with the expression of FOXM1 (Fig. 6A). The results of luciferase report assay revealed that FOXM1 activated the expression of PRC1 by directly binding to the promoter of PRC1 and PRC1 luciferase activity was decreased when the binding site was mutated (Fig. 6F and G). To determine whether PRC1 serves as a downstream target of FOXM1, PRC1 was knocked down in FOXM1-overexpressing cells. Transwell assay was then used to determine whether the silencing of PRC1 could reverse the FOXM1-mediated increase in the metastatic ability of A2780 cells. As shown in Fig. 6H and I, the silencing of PRC1 inhibited the FOXM1-mediated promotion of the migratory and invasive ability of the ovarian cancer cells. These results thus suggested that the overexpression of PRC1 was ascribed from the regulation of FOXM1, and it may serve as an important executor.