Introduction
Epithelial ovarian cancer (EOC) is the third most common gynecological malignancy worldwide, and it is also one of the most fatal gynecological carcinomas in women (1), causing an estimated 125,000 deaths each year globally (2) and 14,240 deaths in the United States in 2015 alone (3). Currently, platinum-based chemotherapy acts as the primary treatment for this cancer (4). Although patient prognosis has improved significantly for various types of solid cancers, women suffering from ovarian cancer exhibited only a slight change in survival rate since platinum-based treatment was introduced more than 30 years ago (5,6). Furthermore, recent studies have revealed that some genetic and epigenetic alterations contribute to the survival of EOC (7). Thus, further understanding of the molecular basis of ovarian cancer is needed.
The anti-apoptotic effect resulting from the augmented expression of certain proteins involved in the apoptosis pathways is one of the most significant hallmarks for cancer cells. X-linked inhibitor of apoptosis protein (XIAP), BCL2-like 2 (BCL2L2) and Baculoviral IAP repeat-containing protein 5 (BIRC5) are three important anti-apoptotic proteins in EOC development. Overexpression of XIAP, the most potent mammalian IAP, contributes to the abnormal survival of EOC cells via inhibiting caspase activity (8). Previous studies indicated that by downregulating XIAP, apoptosis is induced both in vitro and in vivo (9). Furthermore, XIAP has been found to be the target of miR-24 leading to the decrease of apoptotic threshold in cancer cells (10). BCL2L2, a pro-survival member of the BCL2 family of proteins, is associated with the ability of cancer cells to evade apoptotic signals (11,12). Additionally, researchers have found that miR-15a can induce cell apoptosis by targeting BCL2L2 in non-small cell lung cancer (13), and miR-214 is able to enhance cisplatin-induced cytotoxicity via downregulation of BCL2L2 in cervical cancer cells (14).
Furthermore, a higher level of BCL2L2 also contributes to cancer cell resistance to drugs, such as camptothecin, cisplatin, etoposide (VP-16), adriamycin, or 1-D-arabinofuranosylcyto-sine (4). BIRC5 (survivin) is one of the eight well-studied members of the inhibitor of apoptosis protein (IAP) family, sharing a baculovirus IAP repeat (BIR), and has important roles in apoptosis (15). IAPs act as endogenous inhibitors of caspases, which are evolutionarily conserved (16,17). BIRC5 has a cell cycle-dependent expression pattern during mitosis, but it is also regulated by other non-cell cycle-dependent mechanisms (18). Previous studies suggested that BIRC5 counteracts apoptosis through interactions with caspases or initiators (19,20). BIRC5 is strongly expressed in embryonic and fetal organs but not in most differentiated normal tissues (21,22). Augmented expression of BIRC5 has been identified in tumors of breast, stomach, lung, pancreas and liver (23–27). Importantly, the aberrant expression of BIRC5 also correlates with human EOC in different aspects. BIRC5 promoter polymorphisms are found to be associated with age of onset of the disease (28). Furthermore, the overexpression of BIRC5 desensitizes EOC cell response to S-allylmercaptocysteine, a drug that inhibits the proliferation and metastasis of ovarian cancer cells (29). These findings make BIRC5 a key molecule in EOC research.
MicroRNA (miRNA), a type of small (20–25 nucleotides), noncoding RNA, acts as the main regulatory factor of gene expression by binding to the 3′-untranslated regions (3′UTRs) of their targeted mRNAs (30), and this has been found to be relevant to various diseases in humans (31). To date, more than 2588 human miRNAs have been described in the miRBase database (http://www.mirbase.org/) with each miRNA possibly regulating the expression pattern of hundreds of target genes (31,32). Moreover, numerous studies have shown that miRNA functions in oncogenic or cancer suppressor activities, including apoptosis (33). The downregulation of some miRNAs has been observed in a variety of cancers such as hepatocellular carcinoma and ovarian cancer, and others (33,34). MicroRNA-146a-5p (miR-146a-5p) is a highly important miRNA, the dysregulation of which underlies the pathogenesis of peripartum cardiomyopathy, osteoarthritis, and the development or metastasis of various cancers, including papillary thyroid carcinoma and breast cancer (31,35–38). Additionally, miR-146a-5p is also critical for the suppressor function of Foxp3+ regulatory T cells, indicating the indispensability of a single miRNA in immune homeostasis (39). In ovarian cancer, Cui et al discovered that augmented expression of miR-146a-5p prohibits cell proliferation, enhances apoptosis, and increases sensitivity to chemotherapy drugs by reducing SOD2 (40).
Given the significant role of miR-146a-5p in EOC, we hypothesized that miR-146a-5p accelerates cisplatin-induced apoptosis via targeting certain anti-apoptotic genes. We found that miR-146a-5p lowers the IC50 value of cisplatin in OVCAR3 and SKOV3 ovarian cancer cell lines and promotes apoptosis in OVCAR3, SKOV3 and primary EOC cells. By computational predictions, two genes in the IAP family (XIAP and BIRC5) along with one gene in BCL2 family (BCL2L2) were predicted to be targets. Since miRNAs usually work by binding to their target mRNAs, a dual-luciferase assay was used to validate targets in our study. The results showed miR-146a-5p downregulates XIAP, BCL2L2 and BIRC5 via their 3′UTRs. After their 3′UTRs mutated, no differences were observed between the control group and miR-146a-5p-treated group, demonstrating that the 3′UTRs are the regulatory site of miR-146a-5p. For the most strongly inhibited target XIAP, we further investigated the mechanism using a lentivirus packaging system, western blotting and DAPI staining. Overexpression of XIAP rescued the apoptosis-promoting effect of a miR-146a-5p mimic, and suppression of XIAP rescued the apoptosis-inhibiting effect of a miR-146a-5p inhibitor in a dose-dependent fashion. These data together suggest a pivotal role for miR-146a-5p in targeting several anti-apoptotic genes in ovarian cancer cells, and this suggests a mechanism that promotes apoptosis induced by cisplatin.
Materials and methods
Vector construction
To overexpress XIAP, human XIAP cDNA without its native 3′UTR was cloned downstream of the CMV promoter in the lentiviral expression vector pCDH-CMV-MCS-EF1-copGFP (pCDH; System Biosciences, Mountain View, CA, USA), and the construct was named LV-XIAP. To construct the luciferase reporter, the wild-type 3′UTRs of XIAP, BIRC2, BIRC5 and BCL2L2, containing the putative miR-146a-5p binding sites as well as the seed region mutated 3′UTR of XIAP, BIRC5 and BCL2L2, were fused to the Renilla luciferase reporter gene at the 3′UTR in the psiCHECK2 vector. The primers used for developing the constructs above are listed in Table I. DNA sequencing was performed to confirm all constructs.
Cell lines and cell culture
OVCAR3 and SKOV3 human ovarian cancer cell lines were purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA) and were maintained in a 37°C, 5% CO2 incubator in DMEM or RPMI (Invitrogen, Carlsbad, CA, USA) with 10% fetal bovine serum (FBS) and penicillin/streptomycin (Invitrogen) (100 U/ml). To isolate primary epithelial ovarian cancer cells from freshly collected malignant ascites from one patient, a previously described method was used (41). The pathological type of ovarian cancer of the patient was ovarian serous adenocystic carcinoma. The patient signed an informed consent form, and the use of the sample and study protocol were approved by the ethics committee of the Second Affiliated Hospital of Guangzhou Medical University (no. 2013034).
Transfection of RNA oligonucleotides
A miR-146a-5p mimic and its control RNA, a miR-146a-5p inhibitor and its control RNA, and an XIAP siRNA (si-XIAP) and its control RNA were synthesized, purified and annealed by GenePharma (Shanghai, China). The RNA sequences were as follows. miR-146a-5p mimic: sense 5′-UGAGAACUGAAUUCCAUGG GUU-3′ and antisense 5′-CCCAUGGAAUUCAGUUCUC AUU-3′; mimic control: sense 5′-UUCUCCGAACGUGUCACGUTT-3′ and antisense 5′-ACGUGACACGUUCGGAGAATT-3′; miR-146a-5p inhibitor: 5′-AACCCAUGGAAUUCAGUUCUCA-3′; inhibitor control: 5′-CAGUACUUUUGUGUAGUACAA-3′; si-XIAP: sense 5′-CAUGCAGCUGUAGAUAGAUGGCAAU-3′ and antisense 5′-AUUGCCAUCUAUCUACAGCUGCAUG-3′; siRNA control: sense 5′-UUCUCCGAACGUGUCACGUTT-3′, antisense 5′-ACGUGACAC GUUCGGAGAATT-3′. RNA oligonucleotides (100 or 50 nM) were used with the X-tremeGENE siRNA Transfection Reagent (Roche, Basel, Switzerland) to transfect the miRNA mimic or siRNA into OVCAR3 and SKOV3 cells.
Western blotting
SKOV3 cells were seeded in 12-well plates at 2×105 cells/well and lysed using RIPA lysis buffer (BioTeke Corp., Beijing, China) 48 h post-transfection, and a bicinchoninic acid protein assay kit (Beyotime, Shanghai, China) was used to measure protein concentration. The protein sample (20 µg per lane) was first heat-denatured and then separated by 12% SDS-polyacrylamide gel electrophoresis. Protein was transferred to a PVDF membrane (Millipore, Bedford, MD, USA) and blocked with 3% BSA. After blocking, the PVDF membrane was incubated for 2 h at room temperature with a mouse polyclonal antibody against human XIAP (1:10,000; Santa Cruz Biotechnology, Santa Cruz, CA, USA) or a mouse monoclonal antibody against human β-actin (1:3000; Abcam, Cambridge, MA, USA) and incubated for 1 h with a goat anti-mouse (1:5000; Abcam) secondary antibody. To detect the bound antibodies, enhanced chemiluminescence detection reagents (Pierce, Rockford, IL, USA) were used. Band intensities were quantified with a Kodak Image Station 4000 MM Pro (Kodak, Tokyo, Japan).
Dual-luciferase assays
The dual-luciferase reporter assay was performed as previously described (42). First, 293T cells were seeded in a 96-well plate at 1×104 cells/well, and cells were co-transfected with 50 ng of luciferase reporter vector and 6 ng of synthetic miR-146a-5p mimic or mimic control. The cells were harvested 48 h post-transfection, and the luciferase activity was measured using a dual-luciferase reporter assay system (Promega, Madison, WI, USA) following the manufacturer's instructions. Renilla luciferase activities were normalized to firefly luciferase activities.
Assay of IC<sub>50</sub> value of cisplatin
The 3-(4,5)-dimethylthiazol(-2-y1) -3,5-di-phenyl-tetrazolium bromide (MTT) assay was used to measure the IC50 value of cisplatin. OVCAR3 and SKOV3 cells were transfected with miR-146a-5p mimic or mimic control one day after seeding into a 96-well plate at 1×104 cells/well. One day post-transfection, cisplatin was added to each well with a concentration gradient of 0.625, 1.25, 2.50, 5.00, 10.0, 20.0, and 40.0 µM, 48 h later, 5 mg/ml of MTT reagent was added, and the plate was incubated in a 5% CO2 incubator at 37°C for 4 h. DMSO (100 µl) was added to each well, and the plate was mixed for 10 min. The absorbance at 490 nm was detected with a BioTek microplate reader (Winooski, VT, USA).
4′,6-Diamidino-2-phenylindole (DAPI) staining
OVCAR3, SKOV3 or primary EOC cells were seeded in a 48-well plate with a cell density of 1.5×104 cells per well and subsequently transfected with miR-146a-5p mimic or mimic control. To achieve 10–30% apoptosis, 5 µM of cisplatin was added to the medium 16–24 h after transfection. After another 24–48 h, the cells were stained with 1 µg/ml DAPI (Sigma, St. Louis, MO, USA). A fluorescence microscope was used to observe apoptotic cells, and the apoptotic ratio was calculated from average 200 cellular nuclei in one image, and three images without overlap were taken randomly in each experiment. The average values ± SD of three separate experiments were plotted.
Lentivirus packaging and infection
293T cells were seeded in a 6-cm dish at 8×105 cells/dish and co-transfected with 1.8 µg of packaging plasmid pPAX2, 0.6 µg of envelope plasmid pMD2.G and 2.5 µg of the XIAP expression vector LV-XIAP or empty vector pCDH (LV-control) with the transfection reagent Lipofectamine 2000 (Invitrogen). Viral supernatants were harvested and stored at −80°C 48 h post-transfection. Before infection, 0.01 µg/ml of polybrane (Sigma) was added in order to improve the infection efficiency.
Statistical analysis
The data were analyzed using Student's t-test and are presented as the mean ± standard deviation (± SD) of three separate experiments. P-values <0.05, 0.01, or 0.001 indicate statistical significance. Prism software version 5.0 (GraphPad Software, La Jolla, CA, USA) was used to analyze the data.
Results
miR-146a-5p decreases the IC<sub>50</sub> values of cisplatin in OVCAR3 and SKOV3 ovarian cancer cells
To investigate the function of miR-146a-5p in ovarian cancer cells, we determined the IC50 values of cisplatin in OVCAR3 and SKOV3 ovarian cancer cells transfected with miR-146a-5p mimic or mimic control. After transfection, cisplatin was added to each well with a concentration gradient of 40, 20, 10, 5, 2.5, 1.25, 0.625 and 0 µM to evaluate the IC50 values. We found that for the two cell lines, the IC50 values decreases with transfection of miR-146a-5p, and this is most significant in the SKOV3 cell line (Fig. 1A and B). These results reveal that miR-146a-5p can effectively increase the sensitivity of OVCAR3 and SKOV3 ovarian cancer cells to cisplatin-induced apoptosis.
miR-146a-5p promotes apoptosis in OVCAR3, SKOV3 and primary ovarian cancer cells
We hypothesized that miR-146a-5p lowers IC50 values of cisplatin in ovarian cancer cells by promoting apoptosis. To test this effect, two ovarian cancer cell lines (OVCAR3 and SKOV3) and primary EOC cells were used. After seeding onto a 48-well plate, OVCAR3 cells were transfected with miR-146a-5p mimic, mimic control, miR-146a-5p inhibitor or inhibitor control. Cisplatin treatment and DAPI staining were performed as previously described (43). Our results show that more apoptotic nuclei were observed in OVCAR3 cells transfected with miR-146a-5p mimic compared to mimic control (Fig. 2A). Apoptosis was markedly suppressed in OVCAR3 cells transfected with miR-146a-5p inhibitor compared to the inhibitor control (Fig. 2B). These results indicate that miR-146a-5p is the driving force of the observed higher apoptosis rate.
In order to define the function of miR-146a-5p, SKOV3 cells were transfected with mimic control and miR-146a-5p mimic and treated with cisplatin. DAPI staining revealed a higher apoptosis ratio in SKOV3 cells transfected with miR-146a-5p mimic compared to mimic control (Fig. 2C). Primary EOC cells isolated from ascites were also transfected with miR-146a-5p mimic or mimic control and treated with 5 µM cisplatin for 24 or 48 h. We found that the apoptosis rate in transfected primary cultured EOC cells was increased approximately 50% after 24 h, and this rate almost doubled after another 24 h (Fig. 3). Together, these robust results strongly suggest that miR-146a-5p accelerates apoptosis by sensitizing EOC cells to cisplatin.
Screening of anti-apoptotic genes targeted by miR-146a-5p
Next, we wanted to identify genes targeted by miR-146a-5p. Given its role in accelerating apoptosis, we predicted that the target genes would be anti-apoptotic. Scores of studies have confirmed the tight connection between proteins of the IAP family and BCL2 family with carcinogenesis, including studies of ovarian cancer (9,15,44–46). We hypothesized that miR-146a-5p could possibly regulate these two families, by decreasing expression of these proteins which are often overexpressed in chemoresistant EOC cells. To elucidate which proteins could be targeted by miR-146a-5p, a computational method was utilized to predict potential miR-146a-5p-binding anti-apoptotic genes (Table II). Four genes (XIAP, BIRC2 and BIRC5 from the IAP family and BCL2L2 from the BCL2 family) out of 13 candidate genes were predicted to have miR-146a-5p binding sites in their 3′UTRs (Fig. 4A). XIAP and BIRC2 had one predicted site, while BCL2L2 and BIRC5 had two potential target sites.
A dual-luciferase reporter assay was used to test if miR-146a-5p regulates XIAP, BCLCL2, BIRC2 and BIRC5 by binding to their 3′UTRs. 293T cells were seeded in a 96-well plate 24 h before transfection, and cells were co-transfected with miR-146a-5p mimic or mimic control and the psiCHECK2 vector containing a luciferase reporter gene fused with the 3′UTR of each the four genes. After 48 h cells were harvested, and the luciferase activity was measured. The luciferase activity data showed that the miR-146a-5p mimic inhibited the three 3′UTRs of XIAP, BCL2L2 and BIRC5 to different extents (Student's t-test, p<0.05) (Fig. 4B). This effect was most significant with XIAP. The results for BCL2L2 showed a moderate influence, which was followed by BIRC5. There was no interaction between miR-146a-5p and the 3′UTR of BIRC2.
miR-146a-5p targets XIAP, BCL2L2 and BIRC5
To determine whether miR-146a-5p recognizes the predicted sites (XIAP 3′UTR 2781-2803; BCL2L2 3′UTR 394-420, 567-586; BIRC5 3′UTR 918-940, 1831-1852), we constructed the seed region-mutated Renilla luciferase reporter for XIAP 3′UTR, both seed region-mutated reporters for BCL2L2 and BIRC5 3′UTRs (Fig. 4A). For each gene, both wild-type and mutant reporters were co-transfected into 293T cells with miR-146a-5p mimic or mimic control, respectively (Fig. 5A). By testing the luciferase activity, we found that the ability of miR-146a-5p to inhibit XIAP, BCL2L2 and BIRC5 were abrogated through a mutation for XIAP, and mutations of both sites for BCL2L2 and BIRC5 (Fig. 5A). Since XIAP is the most significantly inhibited gene, we chose XIAP for further experiments.
To confirm whether miR-146a-5p could also influence endogenous target expression, SKOV3 cells were transfected with miR-146a-5p mimic or mimic control. Western blot analysis indicated that the level of endogenous XIAP decreased, and this was ascribed to the transfection of miR-146a-5p mimic compared to the control group. The level of the internal reference GADPH was consistent between the two groups (Fig. 5B). Taken together, our data demonstrated that miR-146a-5p targets XIAP, BCL2L2 and BIRC5 via their 3′UTRs in 293T and ovarian cancer cells.
XIAP rescues the effects of miR-146a-5p on apoptosis
We showed that miR-146a-5p promoted cisplatin-induced apoptosis in ovarian cancer cells (Fig. 2), and it was also able to inhibit the expression of XIAP. To further validate that miR-146a-5p indeed accelerates cisplatin-induced apoptosis via down-regulating XIAP, a rescue experiment was performed. OVCAR3 cells were co-transfected with 50 nM si-XIAP and different amounts of miR-146a-5p inhibitor (100 and 200 nM). DAPI staining was used to allow quantification of the state of apoptosis. When 50 nM si-XIAP and 100 nM miR-146a-5p inhibitor was co-transfected, there was no significant difference between the 100 nM miR-146a-5p inhibitor experimental group and control group, but there was weaker apoptotic induction in the 200 nM miR-146a-5p inhibitor experimental group compared to the control group (Fig. 6A). This suggests that XIAP siRNA can counteract the effect of miR-146a-5p inhibitor in suppressing apoptosis, and there is decreased apoptosis when a higher amount of miR-146a-5p inhibitor is transfected. Similarly (Fig. 6B), OVCAR3 cells were infected with constant quantities of XIAP without its 3′UTR (multiplicities of infection equal to 1) and were then transfected with varying amounts of miR-146a-5p mimic (50 and 100 nM), and there was no obvious difference between 50 nM miR-146a-5p mimic group and control group and a higher apoptosis ratio in 100 nM miR-146a-5p mimic group compared to control group (Fig. 6B). These results indicate that XIAP can counteract the effect of miR-146a-5p in promoting apoptosis, and there is enhanced apoptosis when miR-146a-5p is expressed excessively. The western blot results from our previous study indicated that XIAP was overexpressed with lentiviral infection, and si-XIAP decreased XIAP expression in OVCAR3 cells (47). In summary, our results indicate that these four anti-apoptotic proteins are targets of miR-146a-5p, providing insights into how miR-146a-5p can promote apoptosis in EOC cells.
Discussion
Many studies have illustrated that miR-146a-5p is engaged in multiple diseases and the development of metastasis in various cancer types (31,35–38). Furthermore, augmented expression of miR-146a-5p can inhibit cell proliferation, enhance apoptosis, and increase chemosensitivity (40), which prompted us to speculate that miR-146a-5p may play a suppressive role in ovarian cancer. In support of this hypothesis, our results showed that overexpression of miR-146a-5p reduces the IC50 values of cisplatin in OVCAR3 and SKOV3 cells and enhances cisplatin-induced apoptosis in OVCAR3, SKOV3 and primary ovarian cancer cells.
XIAP, BCL2L2 and BIRC5 are three important anti-apoptotic proteins in a multitude of cancers (8,38). Numerous studies have reported that XIAP, whose expression is at a high level in ovarian cancer (48–50), is essential for suppressing apoptosis and promoting cell proliferation (8). Likewise, BIRC5 also have functions associated with apoptosis (19,51). Overexpression of BIRC5 can decrease the sensitivity of EOC cells to an anti-proliferative drug (29). BCL2L2 is also associated with overriding apoptotic signals in cancer cells (11,12). Bioinformatic predictions and luciferase assays demonstrated that these three proteins can be targeted by miR-146a-5p via their 3′UTRs, and the effect on XIAP was the most significant (Figs. 4 and 5). Furthermore, the experiments showed that XIAP functionally rescued the apoptotic effect of miR-146a-5p (Fig. 6). Consequently, we deduced that the potentially abnormal functions of these three proteins in ovarian cancer might be partially ascribed to the dysregulation of some miRNAs including miR-146a-5p. Several studies have illustrated that XIAP is also regulated by other miRNAs such as miR-519d, miR-509-3p, miR-155 and miR-7 (43,52–54), which may allow us to construct a XIAP-regulating network.
Furthermore, this study provided novel targets for EOC treatment. The five-year relative survival rate of ovarian cancer patients is 46% from 2005 to 2011, without much increase compared to that from 1987 to 1989 at 36% (3). Currently, standard platinum chemotherapy leads to a recurring drug-resistant cancer in approximately 25% of patients within six months (55). This relatively unsatisfactory prognosis means that further understanding of the molecular mechanisms of ovarian cancer is needed. Previous research revealed that overexpression of miR-146a-5p in EOC cells inhibits cell proliferation and enhances apoptosis and chemosensitivity through reduction of SOD2 (40). Herein, data from our work further showed that miR-146a-5p decreases the IC50 value of cisplatin in ovarian cancer cells by downregulating three anti-apoptotic genes including XIAP. Thus, augmenting the level of miR-146a-5p by gene therapy combined with cisplatin chemotherapy might be a more efficient treatment.
In conclusion, our study indicates that miR-146a-5p promotes cisplatin-induced apoptosis in ovarian cancer cells by repressing multiple anti-apoptotic genes, including XIAP, BCL2L2 and BIRC5, of which XIAP shows the strongest evidence. The miR-146a-5p/three protein axis should be further tested by evaluation of whether miR-146a-5p level correlates with prognosis using clinical specimens. These studies may illuminate new targets for treatment of ovarian cancer, which may facilitate cisplatin-induced apoptosis and improve chemotherapy.