Ascites in ovarian cancer: MicroRNA deregulations and their potential roles in ovarian carcinogenesis

3.1.1 Let-7b

The down-regulation of miRNA let-7b has been observed in three independent studies that focused on ascites [17, 21, 24], which tends to suggest its tumor suppressing role in ovarian cancer. Other literature data on the expression of let-7b in ovarian cancer is both scarce and controversial. Supporting the tumor suppressing view, previous investigations determined the down-regulation of let-7b in tumor tissues [25, 26]. Recently, we identified let-7b as the sixth most down-regulated miRNA (-25-fold) within a specific data set (Ct ⩽ 40) from the high-throughput screening of cancer tissue samples compared with the controls, thus indicating potential tumor suppressing functions in ovarian carcinomas [22]. Conversely, the earlier metanalysis of studies that addressed the clinicopathological functions and prognostic values of individual let-7 family members identified elevated levels of let-7b as a poor prognostic factor in high-grade serous ovarian cancer [27].

Nevertheless, strong recent experimental and clinical evidence has demonstrated the opposite pattern of the functioning of let-7b in ovarian cancer [28]. This study revealed the negative association of let-7b with the histological type and grade, FIGO stage and lymph node metastatic status. In vitro, let-7b mimics acted to slow down A2780 and SKOV3 cell proliferation and migration, while let-7b inhibitors served to increase the numbers of migrated cells. Finally, the results suggested that high levels of KDM2B in ovarian cancer may inhibit the expression of let-7b, while let-7b inhibition acted to accelerate xenografted tumor growth in mice [28].

Regarding other cancers, let-7b has been repeatedly determined down-regulated in cancer tissues and cell lines compared to the controls. With concern to hepatocellular carcinoma, the decreased expression of let-7b was associated with a poor prognosis, whereas this miRNA was observed to inhibit cell proliferation through the upregulation of p21 in a recent study [29]. Similarly, Wang et al. [30] observed the remarkable down-regulation of let-7b in hepatocellular cancer (HCC) tissues. Experimentally overexpressed let-7b acted to decrease the expression of β -catenin and c-Myc, and inhibited HCC cell proliferation [30]. Han et al. [31] demonstrated the tumor suppressive effects of let-7b in gastric cancer. Using a cisplatin-resistant gastric cancer cell line (SGC7901/DDP), the authors discovered that the low expression of let-7b was associated with cisplatin resistance, and they experimentally confirmed this finding using let-7b mimics that increased the cytotoxicity of DDP to SGC7901/DDP cells. Conversely, the knockdown of let-7b resulted in reduced sensitivity. Primary tumor growth was also observed to be reduced by let-7b mimic transfection [31]. A further study [32] revealed that let-7b was down-regulated in castration-resistant prostate cancer tissues and the C4-2 cell line. The overexpression of let-7b repressed cell proliferation and invasion in vitro and tumor growth in vivo [32]. Moreover, the overexpression of let-7b inhibited the proliferation of A375 and A2058 cells in melanoma via the targeting of UHRF1 [33]. Taken together, the vast majority of results indicate the tumor suppressing effects of let-7b in cancer.

3.1.2 miR-26a ( = miR-26a-5p)

Three other miRNAs (miR-26a, miR-145, miR-143) have been consistently reported to be downregulated in ascitic fluid-based samples (see above). Several novel reports support the role of miR-26a ( = miR-26a-5p) as an ovarian cancer inhibitor. We determined miR-26a down-regulated (-6-fold) within a specific data set (Ct ⩽ 40) from the screening of ovarian cancer tissue samples compared to the controls [22]. Gao et al. [34] analyzed the expression of miR-26a in ovarian cancer patient tissues and the SK-OV-3 and A2780 cell lines, where low miR-26a and high TCF12 expressions were observed. TCF12 was identified as the direct target of miR-26a, while the experimental upregulation of miR-26a inhibited the proliferation, migration and invasion of, and induced apoptosis in, ovarian cancer cells [34]. Sun et al. [35] demonstrated the reduced expression of miR-26a in ovarian cancer tissues and a negative association with the expression of Cdc6. Further, the expression of miR-26a was associated with the differentiation and stage of cancer cells. Using miRNA-26a mimics, the proliferation and colony formation rate was reduced in SKOV3 cells and ES2 cells, while apoptosis was induced following transfection. Interestingly, contrasting findings were reported by Shen et al. [36] who revealed the overexpression of miR-26a in ovarian cancer and increased miR-26a plasma levels that distinguished the patients from the healthy controls. Experimentally, the ectopic expression of miR-26a in SKOV3 and ES2 cells enhanced both cell proliferation and clonal formation. An in vivo experiment using mice and the observation of the formation of xenograft tumors following the injection of SKOV3 cells transfected with either miR-26a or anti-miR-26a also indicated that the inhibition of miR-26a acted to suppress tumor formation [36].

The decreased expression and inhibitory roles of miR-26a that have been reported with respect to the progression of tumor/cell lines in tongue squamous cell carcinoma [37], non-small cell lung cancer [38] and hepatocellular carcinoma serve to support its tumor suppressing functions [39]. With respect to hepatocellular carcinoma, it should be noted, however, that opposite findings have also been reported [40]. This study revealed that on the one hand miR-26a promoted invasion/metastasis by inhibiting PTEN and, on the other, inhibited cell proliferation by repressing EZH2, thus suggesting the complexity of the regulation roles of miR-26a in cancer [40]. However, again with concern to hepatocellular cancer, miR-26a has recently been reported to be downregulated in carcinoma tissues and cells lines; miR-26a mimics served to inhibit cell proliferation and enhanced the doxorubicin sensitivity of hepatocellular carcinoma cells by targeting aurora kinase A [41]. Similarly, the activity of miR-26a as a tumor suppressor that targets FBXO11 acted to inhibit the proliferation, migration and invasion of hepatocellular carcinoma cells [42].

The ambiguity of the findings on miR-26a has also been demonstrated with respect to prostate and breast cancers. Concerning prostate cancer cells, miR-26a was observed to be overexpressed, the expression of PTEN (the target of miR-26a) was lower, and the experimental inhibition of miR-26a resulted in enhanced cell apoptosis, weakened proliferation ability, and an arrested cell cycle at the G0/G1 phase, thus suggesting the oncogenic roles of miR-26a in prostate cancer [43]. Huang et al. [44] determined the opposite findings for breast cancer, concerning which miR-26a-5p was downregulated in cancer tissues and cells lines and was associated with a poor prognosis. Moreover, this miRNA acted to inhibit breast cancer cell growth by suppressing the expression of RNF6. miR-26a serum levels were found to be reduced for gastric cancer [45]. Other reports have been published that identified the relatively tumor suppressing functions of miR-26a in a variety of other cancers such as colorectal cancer [46], lung cancer [47], bladder cancer [48], pancreatic ductal adenocarcinoma [49], cervical cancer [50], melanoma [51] and nasopharyngeal cancer [52].

3.1.3 miR-145 ( = miR-145-5p)

Reports on a further consistently expressed miRNA in ascites, i.e. miR-145 ( = miR-145-5p), indicate its down-regulation in ovarian cancer and, thus, its tumor suppressive role. In two of the first studies to be conducted on miRNAs in ovarian cancer, Iorio et al. [53] and Nam et al. [25] determined miR-145 to be one of the most down-regulated miRNAs in ovarian tumors versus normal ovary tissue; moreover, a large number of recent reports that consider the cell lines, tumors or serum of affected patients have indicated a similar pattern. In addition, several mRNA targets associated with sensitivity or resistance to chemotherapy and clinical outcomes have been identified. Wu et al. [54] reported miR-145 to be downregulated in ovarian cancer tissues, cell lines and in serum samples associated with ovarian cancer. Further, the same study demonstrated that this miRNA suppresses p70S6K1 and MUC1 [54]. In subsequent investigations, the down-regulation of miR-145 in cancer-associated samples was observed to be linked with enhanced proliferation, migration and invasion, the in vivo tumorigenicity of EOC cells and the overexpression of other direct miR-145 targets, including Cdk6 and Sp1 [55], c-Myc [56], metadherin (MTDH, [57]), HMGA2 [58], TRIM2 [59], MUC1 [60], D-type cyclin 2, cyclin D2 (CCND2), E2F transcription factor 3 (E2F3) [61], SMAD4 [62] and YTHDF2 [63]. Several reports have indicated that a higher miR-145 expression may be linked to improved survival rates as observed for both tissues [64, 65] and for serum [66]. The elevated expression of miR-145 has also been associated with increased sensitivity to paclitaxel [55]. We observed that miR-145 was consistently down-regulated (-9-fold) in ovarian cancer tumors compared with the controls [22]. Only two investigations have reported opposite or controversial findings for miR-145 in ovarian cancer. Wang et al. [67] observed upregulated levels of miR-145 in plasma and its down-regulated expression in tumors. Focusing on exosomes, Kim et al. [68] reported the higher expression of miR-145 in serum exosomes in a cancer group than in a non-cancer group (benign and borderline tumors). The investigation of other cancers imply the tumor suppressing roles of miR-145 with respect to non-small cell lung cancer (NSCLC, [69]), breast cancer [70] and pancreatic cancer [71].

3.1.4 miR-143 ( = miR-143-3p)

The potential tumor suppressor miR-143 ( = miR-143-3p) has been investigated in a number of studies, one of the most recent of which [72] focused on the roles of miR-143 in chemoresistance using chemoresistant (A2780/CDDP) and non-resistant (A2780) ovarian cancer cell lines. In line with its presumed tumor suppressor roles, decreased cisplatin resistance in A2780/CDDP has been associated with the overexpression of miR-143 while the inhibition of miR-143 resulted in decreased A2780 sensitivity to cisplatin.

According to recent investigations, miR-143 may interact with competing endogenous (ce)RNA such as lncRNA CDKN2B-AS1 [73], thus acting to reduce its expression. The decreased expression of miR-143 may, in turn, eventually lead to the derepression of the miR-143-3p target SMAD3 and, finally, to the progression of ovarian cancer. A similar interaction was noted for lncRNA MCM3AP-AS1 that acts as an oncogenic lncRNA by binding to miR-143-3p and, thereby, promoting the expression of transforming growth factor-beta-activated kinase 1 (TAK1) [74]. Another lncRNA, i.e. lncRNA urothelial carcinoma-associated 1 (UCA1) was observed to be upregulated in cisplatin-resistant patient tissues and cell lines, and its repression promoted the expression of miR-143 [75]. Lin et al. [76] observed that the inhibition of MALAT1 resulted in an increase in the expression of miR-143-3p. Guan et al. [77] explored the mechanisms underlying the transforming growth factor-beta (TGF-beta) 1-mediated regulation of the cystatin B (CSTB) progression marker and links to the expression of miR-143. By exploring the TGF-beta/miR-143-3p/CSTB axis of the regulatory interactions, the authors proved that the downregulated expression of miR-143 was associated with the overexpression of its target CSTB. Moreover, higher levels of CSTB expression were associated with the poor overall survival rate of patients with ovarian cancer. TGF-beta 1 treatment induced the upregulation of miR-143, while the knockdown of CSTB acted to decrease ovarian cancer cell proliferation [77]. Via the identification of transforming growth factor (TGF)-ss-activated kinase 1 (TAK1) as a further potential miR-143 target, Shi et al. [78] observed a similar miR-143 deregulation pattern. The decreased expression of miR-143 was noted in ovarian cancer tissues and in three cell lines, i.e. SKOV3, ES2 and OVCAR3, while its upregulation acted to reduce cell proliferation, migration and invasion in vitro, and inhibited the growth of ovarian tumors in vivo in a xenograft experiment [78].

A further study that determined similar results [79] observed lower levels of the expression of miR-143 and identified RalA-binding protein 1 (RALBP1) as another potential target of this miRNA. Moreover, transfection with miR-143-3p mimics and RALBP1 siRNA was observed to promote the apoptosis of ovarian cancer cells [79]. Similar findings were also reported by Wang et al. [80] in a study in which downregulated miR-143 was associated with stage, grade and lymph node metastasis in patients. The investigation of the behavior of ovarian cancer cell lines in vitro revealed the elevated expression of miR-143-inhibited cell proliferation, migration and invasion, while connective tissue growth factor (CTGF) was shown to comprise a direct miR-143 target [80]. Our screening study revealed the consistent down-regulation (-10-fold) of miR-143 in ovarian cancer tissues compared to the controls [22]. Similar findings concerning other cancers also suggest the tumor-suppressing role of miR-143 in e.g. colorectal cancer [81] and lung cancer [82].

3.1.5 miR-95 ( = miR-95-3p)

Two studies on ovarian cancer that employed ascites samples, concluded that miR-95 ( = miR-95-3p) may function as an oncogenic miRNA [19, 21]. However, no other reports on miR-95-3p expression alterations and its potential functional biological roles have been published with respect to ovarian cancer. Interestingly, we did not determine any significant alterations during the screening of ovarian cancer tumors and the controls [22]. Reports on other cancers generally provide supporting results. Huang et al. [83] proposed miR-95 as an oncogenic miRNA in colorectal cancer, the evidence for which comprised high levels of expression in tumor tissues and its in vitro/in vivo behavior. Other reports also proved the oncogenic roles of miR-95-3p in colorectal cancer [84, 85, 86, 87], gastric cancer [88] and prostate cancer [89]. However, controversial results have recently been reported with respect to colorectal carcinomas (CRC) suggesting miR-95-3p as potential tumor suppressor that is downregulated in CRC tissues with low levels that are associated with a poor prognosis. In vitro, overexpressed miR-95-3p was found to inhibit cell proliferation, colony formation and metastasis; moreover, this behavior was also observed in vivo [90]. These results are in direct contradiction to previous investigations on this type of cancer. Overall, it is presumed that miR-95-3p exerts oncogenic functions as noted in ovarian cancer based on the research of ascites; however, it is important to exercise caution in view of the current lack of proof with concern to ovarian tumor tissues.

3.1.6 miR-182-5p

The overexpression of miR-182-5p in ascitic fluid indicates its oncogenic roles in ovarian cancer. This view has generally been supported via a number of studies that determined that miR-182-5p comprises one of the most upregulated miRNAs. Elgaaen et al. [91] identified miR-200 family members and miR-182-5p as the most overexpressed miRNAs in high-grade serous ovarian carcinomas and clear cell ovarian carcinomas as compared to ovarian surface epithelium. Similarly, miR-182-5p was found to be among the eight most overexpressed miRNAs and was used for discriminating ovarian cancer tissues from normal tissues with 97% sensitivity and 92% specificity [92]. Moreover, Marzec-Kotarska [93] identified miR-182-5p as the sixth most overexpressed miRNA compared to normal ovarian tissue and its expression correlated with significantly shorter overall survival rates. In line with these studies, Wang et al. [94] suggested that miR-182-5p is upregulated in ovarian cancer tissues and cell lines and that it promotes cell proliferation and invasion. The long non-coding RNA ADAMTS9-AS2 that evinces a decreased expression in cancer samples has been identified as a miRNA sponge, and FOXF2 as a direct target for miR-182-5p in ovarian cancer [94]. The most recent support for the oncogenic view of miR-182-5p is provided by Wang et al. [95] who determined that circMTO1 inhibits the proliferation and invasion of ovarian cancer cells and sponged miR-182-5p, while supporting the expression of KLF15. Nevertheless, according to a recent investigation [96], miR-182-5p is downregulated in ovarian cancer tissues and cells and its overexpression acts to sensitize ovarian cancer cells to cisplatin. Similarly, the roles of miR-182-5p in other cancers appear to be ambivalent. For example, it was observed to be downregulated in colon cancer tissues and cells, whereas in vitro, the overexpression of miR-182-5p acted to repress the proliferation of colon cancer cells, colony formation, migration and invasion, and triggered G1 arrest and apoptosis [97]. Conversely, it was observed to be upregulated in non-small cell lung cancer [98]. These contradictions thus suggest the need for the further research of miR-182-5p and its various roles in carcinogenesis.

3.1.7 miR-200c-3p

miR-200c-3p is a member of the miR-200 family of miRNAs, the various members of which (miR-200a, -200b, -200c, -141 and -429) were initially considered to comprise oncogenic miRNAs in ovarian cancer due to their overexpression in a variety of types of samples. Our studies revealed the remarkably elevated expression of miR-200 family members in ovarian cancer tissues as well as in ascites [21, 22]. Similar evidence from initial reports [53, 25] has indicated that miR-200 family members evince important roles in terms of promoting ovarian carcinogenesis while being upregulated in cancer samples. However, other data (obtained mainly from experimental studies) has suggested the opposite expression pattern; thus, the roles of miR-200c and/or other family members will require further attention (e.g. [99, 100]).

Though findings based on other cancers may serve to support the conclusion that miR-200 family members function mainly as tumor suppressors and metastatic inhibitors evincing suppressive effects on cell transformation, cancer cell proliferation, migration, invasion, tumor growth and metastasis [101], the reported effects of the miR-200 family with concern to ovarian cancer are less consistent. The somewhat controversial results have triggered a plethora of research that both provides additional insight into miR-200 family regulations in cancer and suggests further inconsistencies and doubts. The results of diagnostic metanalysis have revealed that miR-200c-3p (along with miR-200a) comprises a reliable diagnostic marker with an upregulated expression for ovarian cancer that evinces a pooled sensitivity of 0.75 and specificity of 0.66. Moreover, it is suspected that both of these miRNAs contribute to EOC progression via their effect on cellular adhesion processes [102]. Nevertheless, the complex roles of the miR-200 family in ovarian carcinogenesis could be linked to differing expression in various stages of disease progression and different ovarian cancer subtypes. Moreover, the reproducibility of the various results obtained on cell lines requires further investigation. Meanwhile, the supporting and conflicting findings published to date serve to reinforce the confusing view of the roles of the miR-200 family in ovarian cancer.

The results of experimental research suggest that the miR-200 family suppresses epithelial to mesenchymal transition and enhances mesenchymal to epithelial transition processes. For example, Bendoraite et al. [103] focused on the expression of the miR-200 family and ZEB1/2 in ovarian cancer. Along with the up-regulation of the miR-200 family, the authors observed the diminished expression of ZEB1/2, suggested reciprocal regulation between the miR-200 family miRNAs and ZEB transcription factors and proposed a model for mesothelial-to-epithelial transition (MET) in the course of the progression of ovarian cancer. Jabbari et al. [104] determined that the transient overexpression of miR-200 family members in metastatic ovarian cell lines induced MET, with the increased expression of epithelial markers (KRT8, KRT18, KRT7) and a reduction in ZEB1/2 mesenchymal markers. FN1, another mesenchymal marker, was also observed to be down-regulated (at both the mRNA and protein levels) in cells transfected with miR-200b, miR-200c and miR-429 [104]. When comparing exosomes originating from two ovarian cancer cell lines with high invasion potential (SKOV-3) and low invasion potential (OVCAR-3), Kobayashi et al. [105] discovered differences between the number of released exosomes and the expression of let-7 and miR-200 family members. While the let-7 family transcripts (let-7a-f) were up-regulated in exosomes of the SKOV-3 cells, miR-200 family members were expressed only in the OVCAR-3 cells and their exosomes [105].

The investigation of clinical data based on samples from patients with ovarian cancer revealed the dysregulated expression of miR-200c. Pivotal studies that focused on tissues have identified miR-200c as comprising one of the most elevated miRNAs. Iorio et al. [53] determined that miR-200c along with other members of the miR-200 family comprise the most upregulated miRNA of the three histotypes investigated (serous, endometrioid and clear cell). Subsequent reports indicated the upregulation of miR-200c in ovarian cancer [25, 106, 107, 108]. A growing body of evidence also indicates differences between the various miR-200 family members in terms of their expression, presumed roles in carcinogenesis and association with the outcomes of patients. Regarding the latter, due to the publication of conflicting results, the issue remains open to question. The data available implies that the biological functions and the alterations in the expression of the miR-200 family are subject to changes and may depend on the cellular context, the stage of tumor progression and metastasis, and the nuclear or cytoplasmic localization of the interacting targets. This may explain why miR-200 family members have been identified as both tumor suppressors and oncogenes in cancer [109]. Moreover, the various members of the miR-200 family appear to differ in terms of their expression alterations. For example, Marchini et al. [110] compared the expression of miRNAs between tumors in relapsers and non-relapsers. Interestingly, the levels of miR-200c in the tumors were observed to be elevated in the non-relapsers, while miR-200b was upregulated in the relapsers versus the non-relapsers. Moreover, the higher expression of miR-200c was associated with improved overall survival and progression-free survival. Prislei et al. [111] identified a multidimensional miRNA regulation pattern while poor or good outcomes correlated to the cellular localization of HuR. The nuclear localization of HuR and the high miR-200c expression acted to inhibit TUBB3 and resulted in favorable outcomes. Conversely, poor outcomes were observed when HuR occurred in cytoplasm and miR-200c enhanced the expression of TUBB3. In another study, high miR-200c-3p expression was associated with poor progression-free and overall survival in high-grade serous ovarian carcinoma patients [91]. The increased expression of miR-200c has been associated with favorable outcomes for patients based on the metanalysis of 15 studies on ovarian cancer [112]. Interestingly, the opposite pattern, i.e. the association of higher levels of miR-200c with worse survival rates has been identified based on the results of 58 research articles on various cancer types [109]. Conversely, with respect to ovarian cancer, Gao et al. [113] observed a descending trend in miR-200c expression levels from the early stages to the advanced stages associated with a lower 2-year survival rate in the latter group.

Ibrahim et al. [114] determined miR-200c to be most upregulated in tissues in cases where there is evidence of metastasis. Experimental transfection using miR-200c mimics or inhibitors has demonstrated that overexpressed miR-200c acted to enhance cell proliferation and colony formation, but served to reduce the migration and invasion of ovarian cancer cells. The inhibition of miR-200c resulted in the opposite pattern [114].

Elevated miR-200c serum levels have also been observed [115] in connection with disease progression [116], even when using the exosomal serum fraction [117]. Low levels of plasma miR-200c have been associated with the 5-month prolongation of PFS when treated with bevacizumab compared to standard chemotherapy [118].

Several reports based on cell lines have suggested that miR-200c may enhance sensitivity to chemotherapy [100, 119, 120, 121, 122] or inhibit tumorigenicity and metastasis [123]. Recently, it has been determined that the upregulation of miR-200b and miR-200c may enhance the sensitivity of cancer cells to cisplatin, in addition to the finding that miR-200b and miR-200c were downregulated in ovarian tumors compared with normal tissues [100].