Gain-of-function microRNA screens identify miR-193a regulating proliferation and apoptosis in epithelial ovarian cancer cells

MicroRNAs (miRNAs) are a small class of non-coding RNAs that negatively regulate gene expression, and are considered as new therapeutic targets for treating cancer. In this study, we performed a gain-of-function screen using miRNA mimic library (319 miRNA species) to identify those affecting cell proliferation in human epithelial ovarian cancer cells (A2780). We discovered a number of miRNAs that increased or decreased the cell viability of A2780 cells. Pro-proliferative and anti-proliferative miRNAs include oncogenic miR-372 and miR-373, and tumor suppressive miR-124a, miR-7, miR-192 and miR-193a, respectively. We found that overexpression of miR-124a, miR-192, miR-193a and miR-193b inhibited BrdU incorporation in A2780 cells, indicating that these miRNAs affected the cell cycle. Overexpression of miR-193a and miR-193b induced an activation of caspase 3/7, and resulted in apoptotic cell death in A2780 cells. A genome-wide gene expression analysis with miR-193a-transfected A2780 cells led to identification of ARHGAP19, CCND1, ERBB4, KRAS and MCL1 as potential miR-193a targets. We demonstrated that miR-193a decreased the amount of MCL1 protein by binding 3′UTR of its mRNA. Our study suggests the potential of miRNA screens to discover miRNAs as therapeutic tools to treat ovarian cancer.

Based on these findings, the clinical potential of miRNAs as cancer biomarkers and/or therapeutic agents is widely recognized and accepted (15). A single miRNA can regulate multiple mRNA transcripts that cooperatively work in cellular differentiation and function (16)(17)(18)(19). The use of miRNA mimics or anti-miRNAs may represent powerful therapeutic tools to accomplish regression and/or re-differentiation of cancer by effectively targeting tumor suppressive or oncogenic genes with less toxicity (15,20). Indeed, a number of pre-clinical trials of miRNAs are currently in progress (21). In this study, we performed a gain of function screen using miRNA mimics library containing 319 miRNAs to identify miRNAs that can affect cell proliferation in A2780 ovary cancer cells. We found several anti-proliferative miRNAs including miR-124, miR-192 and miR-193 in A2780, suggesting that the potential of miRNA screens for discovering miRNAs as therapeutic tools to treat ovarian cancer.

Materials and methods
Cell culture. Human ovarian cancer cell line A2780 was obtained from Dr T. Tsuruo (22), and human colorectal cancer cell line DLD-1 was obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). A2780 and DLD-1 were cultured in RPMI-1640 (Gibco, Life Technologies, Carlsbad, CA, USA) containing 50 IU/ml penicillin and 50 µg/ml streptomycin (Gibco, Life Technologies), supplemented with 5% (A2780) or 10% (DLD-1) fetal bovine serum (FBS, JRH Biosciences, Lenexa, KS, USA) at 37˚C in an atmosphere of 5% CO 2 . miRNA library screening. A gain of function miRNA screen on cell viability was performed using A2780 as previously described (23). A2780 was seeded at 2,500 cells per well in 96-well plates the day before transfection. Synthetic miRNA mimic library (human Pre-miR™ miRNA precursor libraryver.1, Ambion, Applied Biosystems, Foster City, CA, USA) was screened using 50 nM in a duplicate. The library contained 319 miRNAs registered in miRBase ver. 7.1 (http://www.mirbase. org/). miRNA mimics were transfected using Lipofectamine 2000 (Life Technologies) according to the manufacturer's protocols. Pre-miR miRNA precursor molecule-negative control (13) (Ambion, Applied Biosystems) was used as a negative control for miRNA mimics. We confirmed transfection efficiency (>90%) using siControl TOX transfection control (50 nM, Dharmacon, Lafayette, CO, USA). After 3 days of transfection, the cell viability was measured using the Cell Titer-Glo Luminescent Cell Viability Assay (Promega, Madison, WI, USA) according to the manufacturer's instructions. Data were expressed as percentage of the negative control. Several miRNA hits were selected to assess reproducibility and dose-dependency (5, 25 and 50 nM).
BrdU incorporation assay. miRNA (25 nM)-transfected cells in 96-well format were harvested for one day, and then were incubated with 10 µM of 5'-bromo-2-deoxy-uridine (BrdU) for 2-4 h. The cells were fixed with cold ethanol/HCl, and the incorporated BrdU was detected using BrdU labeling and detection kit III (Roche Diagnostics GmbH, Mannheim, Germany) according to the manufacturer's instruction.
RNA isolation and whole genome microarray. A2780 cells were transfected with miR-193a (Pre-miR miRNA precursor molecules, hsa-miR-193a-3p, Ambion, Applied Biosystems) or negative control miRNA (25 nM), and allowed to grow in the medium (RPMI-1640) for 10 h before RNA isolation. Total RNA was isolated using the RNeasy mini RNA isolation kit (Qiagen). The integrity of the RNA was verified using an Agilent 2100 Bioanalyzer (1.8-2.0: Agilent Technologies, Palo Alto, CA, USA). Transcriptome microarray analysis was carried out using the 44K Whole Human Genome Microarray chip (Agilent Technologies) according to the manufacturer's instructions. Scanning microarray chips and processing data were done by Pharmafrontier Co., Ltd, Kyoto, Japan. Differentially expressed probe sets were identified with a fold change >1.5. Gene ontology (GO) pathway enrichment analysis was performed among genes differentially expressed after miR-193a transfection by SigTerm software (24). The downregulated genes with miR-193a transfection were compared with predicted miR-193a target genes searched by TargetScan (http://www.targetscan.org/). Over-representation of predicted miR-193a target genes within downregulated gene sets was assessed by SigTerm software.
qRT-PCR. Total RNA was prepared from miRNA or siRNA (25 nM)-transfected cells 2 days after transfection using RNeasy mini kit (Qiagen), and then first strand cDNA was synthesized using SuperScript III (Life Technologies) according to the manufacturer's instruction. Real-time RT-PCR was performed using 7900 HT fast real-time PCR system (Applied Biosystems Inc., Foster City, CA, USA) with SYBR-Green as a reporter. The following primers were used for detection: MCL1 (222 bp) forward: TCTAAGTGCTGAC TGGCTACG, reverse: CCTGGCACAGCTATCAAAAG; GAPDH (137 bp) forward: ACTTTGTCAAGCTCATTTCCTG, reverse: CTCTCTTCCTCTTG GCTCTTG.
DLD-1 cells were cultured in 96-well formats and co-transfected with 100 ng of pGL3 Firefly luciferase reporter vector, 20 ng of pRL-TK Renilla luciferase control vector (Promega) and 25 nM miRNA or negative control miRNA using Lipofectamine 2000. Firefly and Renilla luciferase activities were measured consecutively using the Dual-Luciferase Reporter Assay System (Promega) 24 h after transfection. All the experiments were done in triplicate and repeated at least twice on different days.

Results
Effects of miRNA mimic library transfection on cell proliferation of A2780 cell line. To identify miRNAs that affect cell proliferation of ovarian cancer cells, we performed a gain of function screen using synthetic miRNA mimic library (319 miRNAs) for human epithelial ovary cancer cells (A2780). The library consists of miRNAs registered in early version of miRBase (ver. 7.1 in October, 2005, http://www.mirbase.org/), and many of them were expressed in ovarian normal and cancer tissues and cell lines (5). We detected cellular ATP to assess cell viability in miRNA (50 nM)-transfected cells 3 days after transfection. Frequency distribution indicated that broad ranges of miRNA mimic transfections affected the cell viability of A2780 (Fig. 1A). A total of 46 out of 319 miRNAs induced more than 50% changes in the cell viability of A2780 after 3 days transfection. Table I shows top 10 miRNAs that increased or decreased the cell viability of A2780. They included known oncogenic miRNAs such as miR-372 (cell viability, 187%) and miR-373 (165%), and tumor suppressive miRNAs such as miR-124a (28.3%), miR-7 (37.1%), miR-192 (36.6%) and miR-193a (29.7%) in several different cancer types (18,(25)(26)(27). The seed family miRNAs that have the same sequences in seed region (2nd to 8th nucleotide) of miRNAs showed similar effects on cell viability   . We are interested in miRNA mimics that decreased the cell viability of A2780 since these miRNA mimics themselves could have therapeutic potential to treat ovarian cancer. To further evaluate miRNA mimics on the inhibition of cell proliferation in A2780, we selected top 10 anti-proliferative miRNAs (miR-7, miR-124a, miR-192, miR-193a, miR-193b, miR-199a * , miR-432 * , miR-497, miR-506 and miR-517c) from the first screen, and examined the cell viability in A2780 cells transfected with different concentrations of miRNAs (5, 25, 50 nM). We confirmed results of our first screening at 50 nM, and found that the transfection of miR-124a, miR-192, miR-193a and miR-193b induced a large decrease in the cell viability of A2780 even at 5 nM (Fig. 1B), indicating that these miRNAs had a profound anti-proliferative effect in A2780 cells. We examined whether miR-124a, miR-192, miR-193a and miR-193b affected DNA synthesis to inhibit cell proliferation in A2780 cells.
One day after miRNA transfection, BrdU incorporation was examined to evaluate DNA synthesis in transfected cells. As shown in Fig. 2A, miR-124a, miR-192, miR-193a and miR-193b decreased an incorporation of BrdU compared with the negative control, indicating that these miRNAs induced the inhibition of DNA synthesis in A2780 cells. We next examined whether these miRNAs affected apoptotic pathway to inhibit cell proliferation in A2780 cells. We found that miR-193a and miR-193b but not miR-124a and miR-192 induced more than twofold increase in an activity of caspase 3/7, the effector of apoptotic pathway, in A2780 cells (Fig. 2B). The result indicated that miR-193a and miR-193b could induce the apoptotic cell death in A2780 cells. Actually, apoptotic cell debris was frequently observed in miR-193a-transfected A2780 cells (Fig. 2C, arrows).

Transcriptome analysis to assess target genes regulated by miR-193a.
We further characterized the anti-proliferative effect of miR-193a in A2780 cells. To examine target genes regulated by miR-193a, we performed genome wide gene expression analysis using miR-193a (25 nM)-transfected cells compared with the negative control miRNA-transfected ones. We identified 518 genes that were downregulated more than 1.5-fold by miR-193a transfection after 10 h. To evaluate the potential functional significance of the genes downregulated after miR-193a transfection, we subjected the gene expression data to gene ontology (GO) pathway enrichment analysis. The 20 most significantly over-represented pathways listed in Table II include small GTPase signaling and vesicular transport. We compared these downregulated 518 genes with predicted miR-193a target genes (142 genes) obtained by TargetScan (Fig. 3). This resulted in the match of 34 candidate miR-193a target genes, and they were significantly over-represented in the downregulated gene sets by using the SigTerm software (24).     4A). We demonstrated that overexpression of miR-193a decreased MCL1 proteins in A2780 cells (Fig. 4A). We next performed qRT-PCR with miR-193a-transfected cells to examine whether miR-193a affected MCL1 mRNA expression. We found that miR-193a induced about 50% decrease in MCL1 mRNA expression in A2780 cells (Fig. 4B). These results indicated that miR-193a affected MCL1 expression at both protein and mRNA levels. To validate whether miR-193a can directly regulate the translation of MCL1 mRNAs, we constructed a luciferase reporter plasmid that inserted MCL1 3'UTR (around 1.5 kb) at the downstream of Firefly luciferase gene, and tested the luciferase activity. As shown in Fig. 4C, co-transfection of miR-193a and MCL1 3'UTR reporter vector induced around 40% reduction of the luciferase activity compared with co-transfection of the negative control miRNA and the reporter vector. The decrease of the luciferase activity was attenuated by using the mutant reporter vector deleting miR-193a seed region complementary sites in MCL1 3'UTR (Fig. 4C, MCL1-3'UTR-MU). These results indicated that MCL1 would be a direct target of miR-193a. We further examined whether the downregulation of endogenous MCL1 could induce apoptosis in A2780 cells. As shown in Fig. 4D, the transfection of MCL1 siRNA (25 nM) induced caspase 3/7 activation comparable with miR-193a transfection in A2780 cells (Fig. 4D), indicating that the downregulation of MCL1 by miR-193a could contribute to miR-193a-induced apoptosis in A2780 cells.
Exogenous expression of a single miRNA mimic can coordinately regulate gene expression on cellular function, which encourages the therapeutic use of miRNA to direct cancer cell death and/or re-differentiation without undesirable side-effects (15,20). One of the challenges to the therapeutic use of miRNA is to predict precisely molecular consequences induced by modulating cellular miRNAs. Transcriptome analysis by microarray has been widely used for miRNA target identification at the transcription level. Protein-profiling techniques have been applied to miRNA-transfected cells for the identification of miRNA targets at the translational level (38)(39)(40)(41).
In summary, we performed a gain-of-function miRNA screen and discovered several miRNAs affecting cell proliferation and death in A2780 ovary cancer cells. Among them, we identified miR-193a as strong anti-proliferative miRNAs in A2780 cells. miR-193a induced the inhibition of DNA synthesis and apoptosis by targeting genes including ARHGAP19, CCND1, ERBB4, KRAS, MCL1, indicating a tumor suppressive role of this miRNA in epithelial ovarian cancer cells. Our study suggests the potential of miRNA screens to discover miRNAs as therapeutic tools to treat ovarian cancer.