MicroRNA MiR-214 Regulates Ovarian Cancer Cell Stemness by Targeting p53/Nanog*

Background: Ovarian cancer stem cells (OCSC) play a critical role in chemoresistance and relapse. Results: Expression of miR-214 induces, whereas knockdown of miR-214 decreases, OCSC and Nanog. MiR-214 targets p53, a repressor of Nanog. Conclusion: miR-214 targets p53 to induce OCSC and Nanog. Significance: MiR-214 is a target for OCSC. Previous studies have shown aberrant expression of miR-214 in human malignancy. Elevated miR-214 is associated with chemoresistance and metastasis. In this study, we identified miR-214 regulation of ovarian cancer stem cell (OCSC) properties by targeting p53/Nanog axis. Enforcing expression of miR-214 increases, whereas knockdown of miR-214 decreases, OCSC population and self-renewal as well as the Nanog level preferentially in wild-type p53 cell lines. Furthermore, we found that p53 is directly repressed by miR-214 and that miR-214 regulates Nanog through p53. Expression of p53 abrogated miR-214-induced OCSC properties. These data suggest the critical role of miR-214 in OCSC via regulation of the p53-Nanog axis and miR-214 as a therapeutic target for ovarian cancer.

The ovarian cancer cell lines were grown in RPMI 1640 medium supplemented with 10% fetal bovine serum at 37°C. Transfections used either siPORT NeoFX transfection agent (Ambion) for oligonucleotides or Lipofectamine 2000 (Invitrogen) for the expression plasmid.
Cell Proliferation and Cell Survival-Cells were plated in 96-well cell culture plates (1 ϫ 10 4 cells/well) and transfected with indicated oligonucleotides and plasmids. Following incubation for 24 h, cells were treated with and without cisplatin (CDDP, Sigma) or doxorubicin (Sigma) for 24 h. Cell growth and survival were examined with cell counting and an MTT assay according to the instructions of the manufacturer (Sigma). In addition, the programmed cell death was determined by cleavage of PARP and caspase 3.
Luciferase Assay-Cells were transfected with 0.2 g of the reporter plasmids, 0.1 g of pCMV-␤-gal, and, where applicable, 5 nm of miR-214 precursor or control or 50 nm of ASO miR-214 or control/well on 96-well plates. Following 48 h of incubation, cells were subjected to a luciferase reporter assay using the luciferase assay system (Promega). Luciferase activities were normalized by ␤-galactosidase activities. Each experiment was repeated at least three times in triplicate.
ALDEFLUOR Assay and Sphere Growth-ALDH1 activity was detected using the ALDEFLUOR assay kit (StemCell Technologies) as described by the manufacturer (19). Briefly, cells were suspended in ALDEFLUOR assay buffer containing an ALDH1 substrate, bodipy-aminoacetaldehyde, at 1.5 M and incubated for 1 h at 37°C. A specific inhibitor of ALDH1, diethylaminobenzaldehyde, at a 10-fold molar excess, was used as a negative control. Flow cytometry data were analyzed by BD FACSDiva software V6.1.3 (BD Biosciences) or FlowJo software (TreeStar).
Sphere culture was carried out as described previously (19). Briefly, cells were plated in ultra-negative attachment 6-well plates (Corning) at a density of 5000 viable cells/well. Cells were grown in a serum-free sphere culture medium (MammoCult, StemCell Technologies) supplemented with MammoCult proliferation supplements for 12 days. Sphere numbers were counted under microscopy.
Western Blot and Immunofluorescence-A Western blot analysis was performed as described previously (18). For immu-nofluorescence, cells were transfected with Nanog promoter-phrGFP together with and without pre-miR-214 or antisense (ASO) miR-214. Following incubation for 48 h, cells were examined with a fluorescence microscope.
Statistic Analysis-Statistical significance was analyzed by unpaired Student's t test and one-way analysis of variance and Duncan's multiple range test using SAS statistical software package version 6.12 (SAS Institute, Cary, NC). p Յ 0.05 was considered to be statistically significant.

MiR-214 Regulates Ovarian Cancer Stem-like Cells-
The findings that miR-214 plays a critical role in chemoresistance and metastasis in ovarian cancer and melanoma (2,5) prompted us to examine the role of miR-214 in regulating OCSC. Supplemental Fig. S1 shows miR-214 levels and p53 status in a panel of ovarian cancer cell lines and two immortalized ovarian surface epithelial cell lines, T80 and MCC3-HOSE. Because the endogenous miR-214 level is high in A2780S and OV8 and low in OV2008 and SKOV3 cells, we knocked down miR-214 in A2780 and OV8 cells and ectopically expressed miR-214 in OV2008 and SKOV3 cells (Fig. 1A). OCSCs were detected and isolated by flow cytometry, and sorting of the cells that were labeled by ALDH1, a common marker of OCSC (19 -21). Fig. 1B shows that depletion of miR-214 reduced the ALDH1-positive cell population in wild-type p53 A2780S but not in p53-mutant OV8 cells. Enforcing expression of miR-214 significantly increased OCSC in wild-type p53 OV2008 but not in p53-mutant SKOV3 (Fig. 1C). These findings were also confirmed by another OCSC marker, CD133, and in four more cell lines (e.g. OV433, OV3, OV10, and A2780CP (supplemental Fig. S2)). Furthermore, sphere growth was evaluated in these cell lines. We observed that knockdown of miR-214 decreased, whereas ectopic expression of miR-214 enhanced, sphere growth in A2780S/OV433 and OV2008/OV10 cells, respectively, with minimal effect on OV8 and SKOV3 cells ( Fig. 1D and supplemental Fig. S3). Because CSCs closely relate to chemoresistance (22), we further investigated miR-214 effects on chemosensitivity of OCSC. As shown in supplemental Fig.  S4, ALDH1-positive A2780S and OV2008 cells expressed high levels of miR-214 and resisted cisplatin and doxorubicin treatment when compared with ALDH1-negative cells. Knockdown of miR-214 in ALDH1-positive cells largely restored their sensitivity to chemotherapeutic agents. These results suggest that miR-214 regulates OCSC in a subset of ovarian cancers.
Nanog Is Regulated by miR-214 in a Subset of Ovarian Cancers-Because accumulating evidence indicates that transcription factors Nanog, Sox-2, and Oct-4 and polycomb ring finger oncogene Bmi-1 are essential to maintain the pluripotent stem cell phenotype (17,23), we examined the effect of miR-214 on their expression. Western blot analysis and semiquantitative RT-PCR revealed that protein and mRNA levels of Nanog were induced by expression of miR-214 in wild-type p53 OV2008 and OV433 cells ( Fig. 2A) but not in p53-mutant SKOV3 and OV3 cells (supplemental Fig. S5A). Knockdown of miR-214 decreased Nanog expression in wild-type p53 A2780S and OV10 cells (Fig. 2B) but not in p53-mutant OV8 and A2780CP cells (supplemental Fig. S5B). We did not observe significant change of expression in the other three genes examined ( Fig. 2 and supplemental Fig. S5).
Identification of p53 as a Direct Target of miR-214-Because miRNAs negatively regulate their target genes (24), Nanog would not be a direct target of miR-214. Because miR-214 induces Nanog expression at the mRNA level, miR-214 could target a transcription factor(s) that represses Nanog transcription. Previous studies have shown that Nanog is negatively regulated by p53 and positively regulated by Oct4 and Sox2 (15,25,26). Our findings that miR-214 regulates OCSC preferentially in wild-type p53 (OV2008, OV433, A2780, and OV10) but not in p53-mutant (SKOV3, OV3, OV8, and A2780CP) cells (27,28) prompted us to examine whether miR-214 regulates p53.   Because OV2008 and OV433 cells express a low level of miR-214, we transfected the cells with pre-miR-214. This resulted in a reduction of about 60 -70% of p53 protein and mRNA (Fig. 3, A-C). In contrast, A2780 and OV10 cells express a high level of miR-214, so we proceeded to suppress miR-214 expression with miR-214 ASO. This resulted in an increase of p53 expression at protein and mRNA levels ( Fig. 4, A-C). Furthermore, p53 expression is inversely correlated with miR-214 levels in the majority of ovarian cancer cell lines examined especially in wild-type p53 lines. (Note: SKOV3 is p53-null; supplemental Fig. S6A). Immunofluorescence staining showed that the p53 signal was enhanced in A2780S cells transfected with ASO miR-214 but was decreased in OV2008 cells treated with pre-miR-214 (supplemental Fig. S6B). In addition, we stably transfected miR-214 into OV433 cells that express wild-type p53 (29) and a low level of miR-214 (supplemental Fig. S1). Immunoblot analysis revealed that the expression levels of p53 and its downstream targets MDM2 and BAX were significantly reduced in miR-214 stably transfected clonal cell lines (supplemental Fig.  S7). On the basis of these data, we conclude that miR-214 represses p53.
We next examined whether the 3Ј UTR of p53 interacts with miR-214. Because published algorithms (such as PicTar, microRNA.org and DIANA-MICRO) did not show p53 as a potential target of miR-214, we performed a sequence analysis with the RNA22 database and found that the 3Ј UTR of p53 contains a region that matches the "seed" sequence of miR-214 (Fig. 5A). To ascertain whether miR-214 directly regulates p53, we cloned the 746 bp-3ЈUTR of p53 containing the wild-type miR-214-p53 response element (MRE), or mutant (MUT) into the pMIR-REPORT plasmid downstream of luciferase. Basal levels of pMIR-p53 reporter activity were examined in two miR-214 low (OV2008 and MCC3-HOSE) and two miR-214 high (A2780 and OV8) cell lines. As shown in Fig. 5B, the reporter activity of WT but not mutated (MUT) pMIR-p53 is reversely correlated with miR-214 expression levels. Furthermore, expression of miR-214 represses WT-pMIR-p53 but not MUT-pMIR-p53 reporter activity in OV2008 (Fig. 5C). In contrast, cotransfection of ASO miR-214 and WT-pMIR-p53 in A2780S resulted in an increase of luciferase activity, but the same experiment carried out with the MUT-pMIR-p53 construct resulted in few changes (Fig. 5D). As control, expression, and knockdown of miR-199a, an unrelated miRNA, in hTERTimmortalized MCC3-HOSE cells were shown no effects on pMIR-p53 activity (Fig. 5E). These data indicate that miR-214 directly targets the p53-3Ј UTR to repress p53 expression.
Up-regulation of Nanog by miR-214 through Targeting p53-Because Nanog is transcriptionally repressed by p53 (25) and is induced by miR-214, we further investigated whether miR-214induced Nanog is mediated by p53. An 806-bp fragment of the Nanog promoter region, which contains the p53 binding site, was cloned into the pGL3-basic and phrGFP promoter-less vectors (Fig. 6A). A luciferase assay revealed that the basal level of Nanog promoter activity is much lower in wild-type p53 OV429 and OV10 than in p53-mutant OV8 and A2780CP cells (Fig.  6B). Ectopic expression of miR-214 induced, whereas knockdown of miR-214 decreased, Nanog promoter activity in wildtype p53 but not p53-mutant cells (Fig. 6C). Moreover, Nanog promoter-driven GFP expression was increased upon expression of miR-214 and was decreased following miR-214 knockdown in OV429 and OV10 but not OV8 and A2780CP cells (Fig.  6D). Accordingly, inverse expression of p53 and Nanog at mRNA and protein levels was detected in miR-214-manipulated OV429 and OV10 cells (Fig. 6E and supplemental Fig. S8). In addition, ectopic expression of p53 without the 3Ј UTR and knockdown of p53 abrogated the effects of miR-214 on Nanog expression (Fig. 6F). Taken together, these data indicate that miR-214 regulates Nanog expression via targeting p53.

Expression of p53 Overrides miR-214-induced OCSC and
Chemoresistance-Having demonstrated that p53 is a direct target of miR-214, we also examined whether expression of p53 will overrule the effect of miR-214 on OCSC property and chemoresistance. Stable miR-214/OV2008 clonal cells were transfected with p53 cDNA without the 3Ј UTR. Vector-transfected cells were used as a control (Fig. 7A). As shown in Fig. 7B, expression of p53 abrogated the miR-214-enhanced OCSC population. Furthermore, expression of miR-214 decreased, whereas knockdown of miR-214 increased, CDDP-induced cleavage of PARP and caspase 3 as well as cell death (Figs. 7, C  and D). However, coexpression of p53 significantly reduced the miR-214 protective effect on cell death induced by CDDP (Fig.  7E). These data further support the finding of p53 as a direct target of miR-214 and suggest miR-214 regulation of OCSC and chemosensitivity, at least to some extent, through targeting p53.

DISCUSSION
Up-regulation of miR-214 has been detected in various human malignancies, including pancreatic, prostate, gastric, breast, and ovarian cancers as well as malignant melanoma (2)(3)(4)(5)(6)(7)(8). Furthermore, miR-214 has been shown to play an important role in chemoresistance, tumor progression, and metastasis (2,5). In this study, we demonstrated that miR-214 regulates ovarian cancer stem cells. Enforcing expression of miR-214 induces, whereas depletion of miR-214 decreases, OCSC properties as well as expression of Nanog. MiR-214 represses p53 by directly interacting with the 3Ј UTR of p53. Moreover, we showed that p53 mediates miR-214-induced Nanog, OCSC, and chemoresistance. These findings are important for several reasons. First, they provide a mechanistic understanding of the miR-214 function in OCSC and chemoresistance. Second, these data further support the notion that miR-214 is an oncomiR in human ovarian cancer. Finally, restoration of p53 by inhibition of miR-214 could be a valuable therapeutic approach in ovarian cancer.
Previous studies have shown that miR-214 is deregulated in CSC and high metastatic tumors (11,30,31) and that miR-214 induces cell migration and invasion and chemoresistance (2,5), which are characteristics of CSC (10). These findings implicate an important role of miR-214 in the regulation of CSC. ALDH1 has been proved to be a useful marker for cancer stem cells and has been widely used to isolate CSC in various malignancies, including ovarian cancer (11, 19 -21). In ovarian carcinoma, ALDH1-positive cells exhibit sphere growth, highly tumorigenicity, and resistance to chemotherapy (19,(32)(33)(34). Moreover, the expression level of ALDH1 has been correlated with a poor prognosis in serous ovarian cancers (19). We showed in this  study that ALDH1-positive ovarian cancer cells express high levels of miR-214 (supplemental Fig. S4A) and CSC markers, including LIN28, Nanog, OCT4, and SOX2 (supplemental Fig.  S9). Moreover, we demonstrated that knockdown of miR-214 reduced Nanog expression, ALDH1-positive cell population, OCSC growth, and sensitization of OCSC cells to therapeutic agent-induced apoptosis. In contrast, expression of miR-214 had the opposite effect. Thus, we provide direct evidence that FIGURE 6. MiR-214 regulates Nanog expression via p53. A, diagram represents an 806-bp fragment of Nanog promoter region was cloned into pGL3-basic or phrGFP promoter-less vectors that are regulated by p53 and miR-214. Luc, luciferase. B, Nanog promoter activity is lower in wild-type p53 OV429 and OV10 than in p53-mutant OV8 and A2780CP cells. Cells were transfected with pGL3-Nanog-Luc and assayed for luciferase activity after incubation for 48 h. C and D, expression of miR-214 induces, but knockdown of miR-214 reduces, Nanog promoter activity only in wild-type p53 cells. The cells were cotransfected with pGL3-Nanog-Luc (C) or Nanog promoter-driven GFP (D) together with indicated oligos. Ctr, control. Following incubation for 48 h, cells were subjected to luciferase assay (C) and fluorescence imaging (D). E, p53 expression is repressed by pre-miR-214 and induced by ASO miR-214. OV429 cells were transfected with indicated oligos and then p53 and Nanog levels were examined by RT-PCR (upper panels) and immunoblot analysis (lower panels) after 72 h of transfection. F, expression of p53 abrogates miR-214-induced Nanog, whereas knockdown of p53 inhibits depleting miR-214-repressed Nanog expression. OV433 cells were transfected with pre-miR-214 together with and without p53 (upper panels), and OV10 cells were treated with ASO miR-214 in combination with and without siRNA-p53 (lower panels). After incubation for 72 h, cells were immunoblotted with the indicated antibodies. OCTOBER  The p53 level, in addition to mutation, has been linked to control of stem cell and CSC properties. Loss of p53 diminishes spontaneous apoptosis and differentiation of embryonic stem cells. Several studies have shown that loss of p53 improves the generation of induced pluripotent stem cells from adult cells (35)(36)(37)(38). In hematopoietic stem cells, p53 negatively regulates their self-renewal (39). Mice deficient in p53 show enhanced HSC self-renewal and have an increased HSC pool size. Similarly, loss of p53 has been shown to increase CSC self-renewal, and expression of p53 represses CSC properties (40). Furthermore, p53 has been shown to repress expression of Nanog, a key molecule in maintaining pluripotency and self-renewal of stem cells, through binding and inhibiting its promoter (25,41). Because miRNA functions as a regulator of gene expression, it is plausible that miRNA plays an important influence on stem cells through modulation of p53. In fact, a recent study showed that miR-33 enhances HSC transplantation efficiency through down-regulation of p53 (42). We showed in this study that miR-214 represses p53 in ovarian cancer cells expressing wild-type or mutant p53 (Fig. 3 and supplemental Fig. S10). However, miR-214 had no effect on Nanog expression in p53-mutant cells (supplemental Fig. S10) because of the fact that mutated p53 loses its DNA-binding activity (43). Our data showed that miR-214 regulates OCSC primary in wild-type p53 cells even though a number of protein-coding genes such as PTEN, TFAP2C, Sufu, and Patched (PTCH) are targeted by miR-214 (2, 5, 44, 45). Nevertheless, these findings further suggest the impor-FIGURE 7. Expression of p53 abrogated the effects of miR-214 on OCSC property and chemoresistance. OV2008 cells were transfected with miR-214 together with and without p53 cDNA and then immunoblotted with the indicated antibodies (A), ALDEFLUOR, and sphere growth assay (B). SSC, side-scattered light; Ctr, control. C-E, A2780S and OV2008 cells were transfected with the indicated oligos and plasmid. After 24 h of CDDP treatment, cleaved PARP, cleaved caspase-3, and cell viability were examined. Experiments were repeated three times in triplicate. Note: p Ͻ 0.05 when comparing the untreated group with the CDDP, CDDP/pre-miR-214, CDDP/ASO miR-214, and CDDP/pre-miR-214/p53 groups. In addition, it is also statistically significant (p Ͻ 0.05) of CDDP versus CDDP/ASO miR-214 and CDDP versus CDDP/pre-miR-214/p53.
In summary, we identified p53 as a major target of miR-214. MiR-214 regulates OCSC preferentially in wild-type p53 cells. MiR-214 induces Nanog through p53. Because OCSC has been implicated in ovarian cancer recurrence and relapse, our study suggests that miR-214 could be a pivotal therapeutic target for OCSC, especially in wild-type p53 tumors.               Pre-miR-214 Ctr. Oligo