To predict the miRNAs targeting CSTB, miRWalk (version 1.0; http://zmf.mm.uni-heidelberg.de/apps/zmf/mirwalk/predictedmirnagene.html) was applied. To analyze the expression of CSTB mRNA and miR-143-3p between normal ovarian tissues and ovarian malignant tumors, three gene expression datasets (GSE36668, GSE40595, GSE63885) for CSTB and one microRNA expression dataset (GSE47841) for miR-143-3p were downloaded from the public Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/geo/). All three CSTB mRNA expression profiles from 12 normal ovary samples and 108 ovarian serous carcinoma samples were generated by the Affymetrix Human Genome U133 Plus 2.0 Array and a miR-143-3p expression profile from 9 normal ovary samples and 12 ovarian serous carcinoma samples was generated by the Affymetrix Multispecies miRNA-2 Array. These transcriptome microarray datasets were integrated and analyzed by R software version 3.4.2 (https://www.r-project.org/). Additionally, to analyze the differential expression of CSTB between ovarian surface epithelium and ovarian serous carcinoma, the online Oncomine tool (www.oncomine.org) was used. In this analysis, the median levels of CSTB mRNA expression were calculated from two datasets: Lu (26) and Bonome (27). To analyze the overall survival (OS) of patients with OC, the Kaplan-Meier Plotter database (www.kmplot.com) was employed. Patients with OC selected for OS analysis were divided into two groups according to the cutoff value set as the median expression levels of CSTB. The OS data were presented as a survival plot, and analyzed with a log rank test.

A total of 35 fresh ovarian tissues samples were obtained from patients (age range: 24-87 years) who underwent surgery at Jinshan Hospital from January 2012 to December 2015 (11 normal ovarian samples from patients with non-ovarian tumor and 24 ovarian serous tumor samples, including 10 benign, 3 borderline, and 11 malignant tumors). All tissue samples were immediately snapped frozen in liquid nitrogen and stored at -80°C until use. None of the patients had received chemotherapy or radiotherapy prior to surgery. The present study was approved by the Ethics Committee of Jinshan Hospital, Fudan University.

The human OC cell lines OVCAR-3 and SK-OV-3 cells were obtained from American Type Culture Collection (ATCC) and maintained in RPMI-1640 and Dulbecco's modified Eagle's media (Corning Inc.), respectively, with 10% (v/v) fetal bovine serum (Invitrogen; Thermo Fisher Scientific, Inc.). Cells were cultured as described previously (24). For TGF-β1 treatment, after seeding cells for 24 h, 10 ng/ml recombinant human TGF-β1 (cat. no. 240-B, R&D Systems, Inc.) was added and incubated at 37°C for 3 h. Cells without any treatment were used as a blank control. For blocking of the TGF-β1 signaling pathway, cells were pre-treated with 10 µM SB431542, a TGF-β type I receptor inhibitor (Sigma-Aldrich; Merck KGaA) for 30 min, followed by 10 ng/ml TGF-β1 treatment at 37°C for 3 h.

Total RNA was extracted from cells or tissue samples using TRIzol reagent (Thermo Fisher Scientific, Inc.) according to the manufacturer's protocols. Primary miRNA, mature miRNA and mRNA were reversely transcribed using a Transcriptor First Strand cDNA Synthesis Kit (Roche Applied Science) according to the manufacturer's instructions. The primer sequences were listed in Table S1. RT-qPCR was performed at 95°C for 10 min, followed by 40 cycles of 95°C for 10 sec and 60°C for 30 sec using the FastStart Universal SYBR-Green Master (Roche Applied Science) on an ABI PRISM 7300 Real Time PCR System (Applied Biosystems; Thermo Fisher Scientific, Inc.). The relative expression of mRNA, primary miRNA or mature miRNA was evaluated using the 2^−∆∆Cq method (28) and all experiments were conducted in triplicate. 18S and U6 was used as an internal control for mRNA and miRNA, respectively.

Lentiviral vectors containing shRNA for CSTB (CSTB-shRNA, sh-CSTB) and negative control (NC-shRNA, sh-NC) were obtained from Shanghai GenePharma Co., Ltd. Three specific CSTB-siRNAs (5′-GUCCCAGCUUGAAGAGAAATT-3′ for si-CSTB-1, 5′-GGACAAACUACUUCAUCAATT-3′ for si-CSTB-2, and 5′-CCCUUGACCUUAUCUAACUTT-3′ for si-CSTB-3) and a negative control (NC-siRNA, si-NC) were synthesized by Shanghai GenePharma Co., Ltd. The miR-143-3p mimics, miR-negative control (miR-NC), miR-143-3p inhibitor (anti-miR-143), and the negative control of inhibitor (anti-miR-NC) was obtained from Guangzhou RiboBio Co., Ltd. (Table S1). After seeding at 2×10⁵ cells/well in 6-well plates and incubating for 24 h, cells were transfected with 50 nM si-CSTB or si-NC, miR-143-3p mimics or miR-NC, 200 nM miR-143-3p inhibitors or miR-NC using X-tremeGENE siRNA Transfection Reagent (Roche Applied Science). Cyanine 3 (Cy3) dye-labeled miR-NC was used to evaluate the transfection efficiency detected by fluorescence microscopy. The sequence of si-CSTB-2 was used for subsequent experimentation. Knockdown of CSTB was performed in OC cells by infecting cells with CSTB-shRNA lentivirus according to the manufacturer's protocols. Briefly, after cells were plated in 6-well plates and incubated at 37°C for 24 h, the amount of lentiviral particle solution (multiplicity of infection=20) was added into the culture media containing 8 µg/ml/well Polybrene for 24 h. The media was replaced with fresh media containing 2 µg/ml puromycin for selection every few days. Following puromycin selection for several passages, the detection of protein knockdown or the phenotypic assay were performed.

Cells were lysed in radioimmunoprecipitation assay lysis buffer (Thermo Fisher Scientific, Inc.) with 1% phenylmethanesulfonyl fluoride (Beyotime Institute of Biotechnology) and 1% phosphatase inhibitor (Nanjing KeyGen Biotech Co., Ltd.), followed by sonication. The protein was extracted from the supernatant after 20 min of 18,000 × g centrifugation at 4°C. Proteins were quantified via a BCA assay. Equal amount proteins were separated on 15% SDS-PAGE and transferred to a polyvinylidene difluoride membrane (EMD Millipore). After blocking with 5% non-fat milk in Tris-buffered saline with Tween 20 at room temperature for 1 h, the membrane was incubated with a primary antibody at 4°C overnight and subsequently incubated with horseradish peroxidase-conjugated goat anti-rabbit or anti-mouse IgG (cat. nos. 7074 and 7076, respectively; 1:5,000 dilution; Cell Signaling Technology, Inc.) for 1 h at room temperature. The following primary antibodies were used: Rabbit anti-CSTB (1:5,000 dilution; Abcam) and mouse anti-β-actin (1:5,000 dilution; ProteinTech Group., Inc.). Signals were detected using Immobilon™ Western Chemiluminescent HRP Substrate (EMD Millipore) and quantified using Tanon-4500 Gel Imaging System with GIS ID Analysis Software v4.1.5 (Tanon Science and Technology Co., Ltd.).

The whole 3'-UTR of CSTB that contains the binding site of miR-143-3p predicted by the miRWalk program (http://mirwalk.umm.uni-heidelberg.de/) was amplified from genomic DNA using the Pfu Ultra II Fusion HS DNA Polymerase (Stratagene; Agilent Technologies, Inc.) with well-designed primers (Table S1). After the dual-luciferase reporter vector pEZX-FR2 (GeneCopoeia, Inc.) was linearized by EcoRI and XhoI restriction enzymes, the PCR product was inserted into pEZX-FR2 using the EasyGeno Assembly Cloning Kit (Tiangen Biotech Co., Ltd.) to construct a wild-type clone named as CSTB-3UTR-wt. A mutation was induced in the consensus sequence of the miR-143-3p binding site in the 3'-UTR of CSTB using the QuikChange II Site-Directed Mutagenesis Kit (Stratagene; Agilent Technologies, Inc.) to construct a mutated clone named as CSTB-3UTR-mut. All clones were verified by restriction enzymes EcoRI and XhoI digestion and DNA sequencing with primer (5′-GATCCG CGAGATCCTGAT-3′) by GENEWIZ, Inc. For the dual-luciferase reporter assay, 293T cells (ATCC) were cultured in 24-well plates. Once subconfluent, the cells were transfected with CSTB-3UTR-wt or CSTB-3UTR-mut plasmid (0.5 µg), plus miR-143-3p mimics, inhibitors or their negative controls using Roche X-tremeGENE siRNA Transfection Reagent (2 µl). After transfection for 24 h, the cells were lysed and luciferase activities were determined using the Luc-Pair™ Duo-Luciferase Assay Kit (GeneCopoeia, Inc.) according to the manufacturer's instructions. Renilla luciferase activity was used for the normalization of Firefly luciferase activity.

For the cell proliferation assay, NC-shRNA- or CSTB-shRNA-infected OVCAR-3 cells were plated in 96-well plates at a density of 3×10³ cells/well. Cell proliferation was measured at 24, 48, and 72 h post-infection by the Cell Proliferation Reagent (Cell Counting Kit-8; Dojindo Molecular Technologies, Inc.) according to the manufacturer's protocols. The signal in optical density was read by a microplate reader (BioTek Instruments, Inc.) at 450 nm.

The cell cycle was detected by flow cytometry. Briefly, CSTB-shRNA-infected OVCAR-3 cells were cultured in 6-well plates for 48 h. The cells were harvested, washed twice with PBS, and fixed in cold 70% ethanol for 2 h at -20°C. After washing with PBS twice, the cells were resuspended in 500 µl of propidium iodide (PI)/RNase Staining Buffer (BD Pharmingen) and incubated in the dark for 15 min at room temperature. Cells (15,000) were then detected by flow cytometry (Beckman Coulter, Inc.). The data were analyzed using ModFit LT software v4.1.7 (Verity Software House, Inc.), and presented as the percentage of the cell population at the G0/G1, S and G2/M phases. Scramble-shRNA-infected OVCAR-3 cells were used as a control.

Flow cytometry and western blotting were conducted to detect apoptotic cells. Briefly, cells were cultured in 6-well plates for 48 h. After washing with PBS, the cells were collected and stained with 1 µl Annexin V conjugated with allophycocyanin (BD Pharmingen) and 5 µl PI for 15 min at room temperature. Apoptotic cells were detected by flow cytometry with FlowJo X software v10.0.7r2 (BD Biosciences). The expression of pro-apoptotic protein Bcl-2-associated X protein (Bax) was analyzed by western blotting as aforementioned using an antibody against Bax (1:2,000 dilution; Cell Signaling Technology, Inc.).

All experiments were repeated at least three times and all analyses were performed with SPSS 21 for Windows (IBM Corp.). For comparisons between two groups in an experiment, a Student's t-test was applied. For multiple comparisons, one-way analysis of variance was used, followed by a Tukey's post-hoc test. For the correlation analysis between CSTB and miR-143-3p, nonparametric Spearman rank correlation was used. The results were presented as the mean ± the standard error of the mean. P<0.05 was considered to indicate a statistically significant difference.

miRNAs which can bind to the CSTB 3'-UTR were screened using web-based programs in the miRWalk database. For targeting CSTB, miR-143-3p was proposed as a potential candidate by 5 out of 10 available prediction programs (Fig. 1A). Additionally, the expression data of transcripts were extracted from GSE36668, GSE40595 and GSE63885 datasets to obtain the mRNA expression profile of CSTB, and GSE47841 dataset for miR-143-3p in an online GEO database using integrated bioinformatics analyses. Upregulated expression of CSTB mRNA was detected in ovarian serous carcinoma (n=108) compared with normal ovary controls (n=12) (Fig. 1B), while miR-143-3p expression was lower in ovarian serous carcinoma (n=12) than normal ovary controls (n=9) (Fig. 1C). As the patients with CSTB mRNA data available in the online datasets differed from patients with miR-143-3p information, the expression of CSTB mRNA and miR-143-3p was validated in the same freshly isolated ovarian tissue from the same patient; correlation analysis of these transcripts was performed.

CSTB mRNA expression levels were significantly increased in malignant tumors than in normal ovarian tissue, and benign and borderline tumors (P<0.01; Fig. 1D), which supported the findings of our previous study (3). On the contrary, miR-143-3p was downregulated in malignant tumors compared with normal ovarian tissues and benign tumors (P<0.05; Fig. 1E). Spearman correlation analysis revealed a significant negative correlation between the expression levels of miR-143-3p and CSTB mRNA (r=-0.362, P<0.05; Fig. 1F).

To determine whether miR-143-3p directly binds to the 3'-UTR of CSTB (Fig. 2A), we cloned two CSTB 3'-UTR plasmids: CSTB-3UTR-wt and CSTB-3UTR-mut. The predicted binding site and mutant site of 3'-UTR (Fig. 2B) were confirmed by sequencing analysis. The luciferase activity in 293T cells was measured after co-transfection of CSTB-3'UTR-containing plasmids with miR-143-3p mimics or miR-NC by a dual-luciferase reporter assay. Transfection of cells with CSTB 3'-UTR-wt and miR-143-3p revealed a significant reduction in the luciferase activity compared with the corresponding control (P<0.01). However, the cells expressing the mutated CSTB 3'-UTR-mut demonstrated no marked alterations in luciferase activity, suggesting that mutations in the 3'UTR of CSTB abrogated the binding to the miR-143-3p (Fig. 2C). These findings indicate that miR-143-3p can directly target CSTB.

To further confirm that CSTB is a target of miR-143-3p, gain-of-function and loss-of-function approaches were applied. The transfection efficiency of miR-143-3p mimics and inhibitors in OVCAR-3 cells was evaluated. A significant increase and decrease in miR-143-3p expression were detected by RT-qPCR after cells were transfected with miR-143-3p mimics of inhibitors (anti-miR-143) for 12 h, respectively, compared with the corresponding NC group (Fig. S1A and B). The transfection efficiency of mimics was further detected by fluorescence microscopy at 24 h post-transfection and ~90% of cells were Cy3-mimic positive (Fig. S1C). Additionally, the expression of CSTB mRNA and protein were examined by RT-qPCR and western blot analyses of OVCAR-3 cells transfected with miR-143-3p mimics or inhibitors. CSTB expression was significantly downregulated at the mRNA (Fig. 3A) and protein (Fig. 3B and C) levels after miR-143-3p mimics transfection. Silencing of miR-143-3p by anti-miR-143 resulted in a significant increase in CSTB expression at the mRNA (Fig. 3D) and protein (Fig. 3C and D) levels in OVCAR-3 cells. These results indicated that miR-143-3p may negatively regulate CSTB expression in OC cells.

As our previous findings indicated that CSTB is mediated by the TGF-β signaling pathway in epithelial OC cells (3) and the present study revealed that miR-143-3p can negatively regulate CSTB expression, whether the TGF-β signaling pathway affects miR-143-3p expression was investigated. RT-qPCR revealed that primary miR-143-3p expression was significantly increased after treatment with 10 ng/ml TGF-β1 in OVCAR-3 cells compared with the control (P<0.01; Fig. 4A). The effects of TGF-β1 were partially abolished by the blocking of the TGF-β type I receptor using its inhibitor SB431542. In addition, an increase in the levels of mature miR-143-3p was observed following TGF-β1 treatment, which was significantly abolished in the presence of SB431542 (P<0.05; Fig. 4B).

Overexpression of CSTB in OC was detected in two microarray datasets from the Oncomine database. The median value of CSTB mRNA expression was calculated from two datasets. CSTB mRNA levels were higher in ovarian serous carcinoma than normal ovarian surface epithelium in the Lu dataset (Reporter ID: 35816_at; Fig. 5A) and in the Bonome dataset (Reporter ID: 201201_at; Fig. 5B). Furthermore, OS analysis was performed using the Kaplan-Meier Plotter dataset. Increased CSTB mRNA expression levels were associated with poor OS in patients with OC (Fig. 5C).

To determine the functional effects of CSTB on the biological behaviors of OC cells, loss-of-function experiments were applied. A total of three siRNAs specific to human CSTB (CSTB-siRNA) were synthesized; transfection efficiency was determined prior to subsequent experiments. Knockdown of CSTB expression at the mRNA (Fig. S2A and B) and protein (Fig. S2C and D) levels was confirmed by RT-qPCR and western blotting in OVCAR-3 (Fig. S2A and C) and SK-OV-3 (Fig. S2B and D) cells, respectively. Increased transfection efficiency was determined in cells transfected with the second CSTB-siRNA (si-CSTB-2), which was then used for the subsequent experiments and to generate shRNAs. Additionally, a cell viability assay was performed after OVCAR-3 cells were transfected with si-CSTB-2. The results revealed that knockdown of CSTB expression by CSTB-siRNA significantly decreased cell proliferation at 48 and 72 h compared with the NC group (Fig. S3A). The suppression of CSTB by CSTB-siRNA appeared to have arrested the cell cycle in OVCAR-3 cells at G2/M phase as detected by flow cytometry (Fig. S3B and C). Similar results were observed in OVCAR-3 cells infected with CSTB-shRNA lentivirus. Knockdown of CSTB protein was confirmed by western blotting (Fig. 6A and B). Cell proliferation was significantly reduced following lentiviral transduction of CSTB-shRNA compared with the NC group at 72 h (Fig. 6C). In addition, the suppression of CSTB notably increased the number of OVCAR-3 cells at G2/M phase compared with negative controls (Fig. 6D-F).

Apoptotic cells were detected by flow cytometry after infecting OVCAR-3 cells with CSTB-shRNA or sh-NC via lentivirus for 48 h (Fig. 7A and B). The cell population in the bottom right of quadrant graphs represented the early apoptotic cells which were significantly increased in the CSTB-shRNA group compared with the NC group (Fig. 7C). Additionally, an increase in the expression of pro-apoptotic protein Bax was observed by western blotting in OVCAR-3 cells infected with CSTB-shRNA lentivirus (Fig. 7D).