The miR-200 family in ovarian cancer

Disease stage

As regards disease stage and metastatic potential, patients with OC are usually diagnosed at an advanced stage, with metastatic disease already present within the peritoneal cavity [15]. OC has a characteristic metastatic model, since it rarely metastasizes through the vasculature. Instead, it metastasizes either by direct extension from the ovarian tumor to neighboring organs (urinary bladder/colon) or through exfoliated tumor cells shedding into the peritoneal cavity and subsequent cancer cell seeding to distant abdominal sites [15]. Extensive seeding of the peritoneal cavity by tumor cells is often associated with ascites, particularly in advanced, high-grade serous carcinomas [15].

Cao et al. [16] investigated the association of the expression of miR-200 family members with the clinicopathological characteristics of patients with epithelial ovarian cancer (EOC). Their study involved 100 patients with EOC and 50 healthy subjects, and demonstrated that tumors with high a expression of miR-200a and miR-200b were more likely to be associated with an advanced stage and higher grade of the disease, while a high miR-200c expression was only significantly associated with an advanced stage of the disease [16]. The aberrant expression of these three miRNAs was also suggested to be associated with aggressive clinicopathological characteristics in patients with EOC [16]. Furthermore, Cittelly et al. [17] reported on miR-200c expression in an intraperitoneal xenograft model of human OC, and found that its restoration resulted in decreased tumor formation and tumor burden. In addition, Yang et al. [18] studied 10 primary serous ovarian tumors and 10 normal ovarian surface epithelium (OSE) cell lines, and reported that an increased miR-200a expression was associated with tumors of a high-grade and late-stage. The authors suggested that the deregulation of miR-200a may contribute to ovarian tumor progression rather than initiation, despite the small number of specimens. Interestingly, Xu et al. investigated miR-200a expression in EOC tumors and compared it to benign ovarian epithelial cysts and normal ovaries, and reported contradictory findings [19]. This group’s results demonstrated that the expression of miR-200a was significantly downregulated in late-stage disease (FIGO III–V and grade 3 groups) compared to early-stage (FIGO I–II and grade 1, 2 groups) tumors. Additionally, the authors found that relatively low miR-200a levels were observed in the “lymph node-positive” compared to the “lymph node-negative” group [19]. In a study on miR-200 family expression in serum, which included a relatively large number of patients with EOC (70 EOC samples vs 70 controls), Zuberi et al. [20] showed that elevated miR-200a and miR-200b expression levels were significantly associated with disease progression and in particular, miR-200a was associated with tumor stage. Additionally, the significant upregulation of miR-200c was found in patients with lymph node metastasis [20]. Further evidence was provided by Marchini et al. [21], who showed that miR-200c expression in patients with stage I EOC correlated with disease outcome. Ιn a blood cohort by Meng et al. [22] involving serum samples from 180 patients with EOC and 66 healthy women, the expression of miR-429 was found to be significantly upregulated in the EOC group. Additionally, miR-429 levels positively correlated with serum CA125 values and differed significantly between early-stage (FIGO I–II) and late-stage (FIGO III–IV) disease [22]. The same researchers, in a subgroup of 163 patients with EOC, demonstrated that circulating exosomal miR-200a, miR-200b and miR-200c concentrations were higher in patients with EOC than in healthy women [23]. Moreover, while the levels of miR-200a were found to be increased in all FIGO/lymph node stages, the levels of miR-200b and miR-200c were higher in patients with late-stage disease (FIGO III–IV) and with lymph node metastasis, than in those with early-stage (FIGO I–II) disease, suggesting that these exosomal miRNAs could discriminate between patients with EOC and healthy women, as well as between malignant and benign ovarian tumors [23]. The researchers concluded that the increased levels of exosomal miR-200b and miR-200c were associated with lymph node metastasis, an advanced FIGO stage (III–IV), the tumor marker, CA125, and with a poor overall survival, demonstrating that these miRNAs may be involved in tumor progression [23].

Tumor histology

As regards tumor histology, Iorio et al. [24] were the first to report the possibility of “histotype-specific” miRNAs in OC. Their study involved 69 cases of epithelial OC of various histotypes and 15 normal ovaries. The authors described a miRNA expression signature that discriminated OC tissues from normal ovaries. Among the most overexpressed miRNAs within the signature were miR-200a, miR-200b, miR-200c and miR-141 [24]. When correlated to the clinicopathological tumor features, the miRNA expression profiles were found to differ in distinct OC histotypes. In particular, miR-200a and miR-200c were overexpressed in serous, endometrioid and clear cell histotypes, whereas miR-200b and miR-141 overexpression was only detected in endometrioid and serous histotypes. This upregulation was suggested to be due to the amplification of the miRNA genes [25]. The authors further commented that different OC histotypes presumably represent biologically and pathogenetically distinct entities of EOCs, although they are currently treated with identical therapeutic strategies [26]. Microarray analysis has confirmed that different histotypes (serous, mucinous, endometrioid and clear cell) show alterations of different pathways, probably reflecting the gene expression pattern of the organ of origin (fallopian tubes, colonic mucosa and endometrium) [26]. Another study by Wyman et al. [27] involving serous, clear cell and endometrioid EOC tumors, were compared to a normal human ovarian surface epithelium (HOSE) cell line. Using massive parallel sequencing, 8 miRNAs were identified as most overexpressed in OC tissue, including miR-200a, miR-200b and miR-200c. While all 8 miRNAs were found to be over-expressed in serous and clear cell cancer subtypes, miR-200(a/b/c) was found to be consistenty up-regulated in the endometrioid subtype as well. The authors considered a competitive advantage of their study, the fact that they used primary cultures of HOSE as the normal control group, instead of whole ovary tissue. On the basis that OC has an epithelial origin [28, 29], performing comparisons of OC tissue to an appropriate normal comparison group is a challenging problem in OC research. The monolayer ovarian surface epithelium comprises less than 1% of the cellular content of the whole ovary [29], and the use of whole ovary tissue inevitably leads to study bias deriving from sample contamination by nearby stromal cells. A serum cohort conducted by Zuberi et al. [20], involved blood serum samples from 70 patients with EOC of various histotypes compared to 70 age-matched, cancer-free female volunteers as controls. Using qRT-PCR, the researchers showed that miR-200a overexpression was strongly associated with the tumor histological subtype. Importantly, the authors found that miR-200a expression was significantly upregulated in the mucinous subtype of EOC.

The above-mentioned studies demonstrate the possibility that distinct miR-200 family member (mainly miR-200a and miR-200c) expression signatures are able to discriminate between OC histological subtypes (Table ​(Table1).1). Histotype-predictive miRNAs belonging to the miR-200 family have been reported in other solid malignancies, such as renal cancer. Working on renal cell tumors, Silva-Santos et al. [30] reported that th expression analysis of miR-141 or miR-200b could accurately distinguish renal cell tumors from normal renal tissues, and even between histological subtypes (oncocytoma, clear cell, papillary).

In the past, miRNA expression profiling has been successfully used for classifying human cancers [25, 31, 32]. miRNAs have been proven to be surprisingly informative in reflecting the developmental lineage and differentiation state of poorly differentiated tumors [32]. A gene expression study showed that genes expressed in different ovarian carcinomas were concordantly expressed in the normal tissues they resemble histologically [26]. Three anatomical sites are the potential origin of high grade serous carcinomas: the ovarian surface epithelium, the fallopian tube epithelium and the mesothelium covering the surface of the peritoneal cavity [26]. Similarly, a plausible scenario regarding tumor histology and miR-200 family expression would be that distinct patterns of miRNA overexpression could represent the developmental ancestry of distinct ovarian tumors.

Chemotherapy response

Despite the improvement in the management of OC over the past 30 years, with treatments such as cytoreductive (debulking) surgery and optimized combinational chemotherapy, usually with platinum-based drugs combined with taxanes, the overall cure rate is only 30%, mainly due to OC resistance to chemotherapy [33, 34]. In general, clear-cell carcinomas, which are associated with a worse prognosis in early-stage disease, and mucinous carcinomas do not respond as well as serous and endometrioid tumors to platinum- and taxane-based chemotherapy [35–37]. Original evidence regarding the association of miRNA expression and cancer chemosensitivity was firstly raised by Meng et al. in cholangiocarcinoma [38]. It was found that the inhibition of miR-21 and miR-200b enhanced the sensitivity of cholangiocarcinoma cells to gemcitabine, probably by the modulation of CLOCK, PTEN and PTPN12, and downstream oncogene products, such as c-Abl, Src and Ras [38, 39]. Furthermore, mounting evidence indicates the existence of similarities between drug-resistant and metastatic cancer cells in terms of resistance to anoikis and enhanced invasiveness [17, 40].

As regards the influence of the miR-200 family on the response to chemotherapy in OC, Cochrane et al. [41] demonstrated that the restoration of miR-200c in various ovarian cancerous cell cultures led to a marked decrease in cell migration and invasion, and to an up to 85% increase in sensitivity to microtubule-targeting chemotherapeutic agents, presumably through the downregulation of class III β-tubulin (TUBB3). Additionally, a low miR-200c expression strongly correlated with the lack of E-cadherin expression and the gain of mesenchymal markers, including ZEB1, while the restoration of miR-200c seemed to restore E-cadherin expression and reduce migration and invasion. Using mouse models, Mateescu et al. [42] showed that an enhanced miR-141 and miR-200a expression increased tumor growth, but at the same time rendered the tumor more susceptible to treatment. The authors suggested that since miR-141 and miR-200a both target p38α and modulate the oxidative stress response, their enhanced expression could mimic p38α deficiency and enhance chemosensitivity [42]. In accordance with previous studies are the findings of Leskelä et al. [43], who studied 72 cases of OC, and showed that a low tumoral miR-200 family content was significantly associated with a high β-tubulin III protein expression [43]. Additionally, in a panel of 57 advanced OC tumors, patients without a complete response had lower miR-200c levels than patients with a complete response. The authors indicated that a low miR-200c expression resulted in a high β-tubulin III expression, and thus increased resistance to paclitaxel-based therapies, and suggested that miR-200 downregulates β-tubulin III in ovarian tumors [43]. Similar findings were reported by Marchini et al. [21] in a series of 144 patients with stage I EOC, where the downregulation of miR-200c was observed in relapsers compared to non-relapsers. Studying OC cell lines overexpressing miR-200c, Cittelly et al. [17] demonstrated that the in vitro restoration of miR-200c in OC cells resulted in increased anoikis sensitivity and reduced adherence to components of the extracellular matrix. Additionally, the restoration of miR-200c in an intraperitoneal xenograft model of human OC resulted in decreased tumor formation and ultimately, tumor burden [17]. As previously shown, in established tumors where the restoration of miR-200c seems to delay tumor progression and increase sensitivity to paclitaxel, presumably by targeting the TUBB3 gene [41], Cittelly et al. [17] suggested that the restoration of miR-200c immediately before treatment could enhance the treatment response or allow for a lower initial dose. In a panel of ovarian adenocarcinoma cell lines, a direct correlation between miR200c overexpression and chemoresistance was also demonstrated by Prislei et al. [44]. The stable overexpression of miR-200c was noted in chemoresistant cell lines (A2780 and Hey), regardless of whether the resistance to paclitaxel and cisplatin was inherent or acquired.

The differential expression of miR-141 and miR-200c between cisplatin-sensitive and cisplatin-resistant OC cell lines was shown by van Jaarsveld et al. [45]. In their study, the overexpression of miR-141 resulted in enhanced resistance to cisplatin in OC cell lines. Additionally, in a setting of 132 primary ovarian tumors (108 serous and 24 non-serous), although no differences were observed in the serous tumors, miR-141 expression levels were higher in non-serous ovarian tumors that did not respond well to therapy [45]. The authors further demonstrated that miR-141 directly targets KEAP1 and that the downregulation of KEAP1 through miR-141 overexpression could induce cisplatin resistance, presumably through the activation of the NF-κB pathway.

Evidence regarding the role of miR-429 in the response to chemotherapy in OC was provided by Wang et al. [46]. The researchers generated a primary cell line deriving from free-floating OC ascites cells (OCI-984 cell line) and sought to determine whether metastasizing OC cells could be induced by miR-429 to undergo mesenchymal-to-epithelial transition (MET) and become sensitized to established first-line platinum-based therapies. Drug sensitivity assays showed that OCI-984 cells transfected with miR-429 exhibited morphological and functional features , as well as the “switching on” of epithelial markers consistent with MET, resulting in a concomitant significant increase in the sensitivity of the converted cells to cisplatin [46]. More recently, Brozovic et al. demonstrated that miR-200 family members differentially regulated sensitivity to paclitaxel and carboplatin in human OC cell lines [47]. Using two ovarian cancer cell lines, OVCAR-3 and MES-OV, and their paclitaxel-resistant variants, OVCAR-3/TP and MES-OV/TP, the authors showed a marked decrease in the expression of miR-200c and miR-141 in OVCAR-3/TP cells, while all five members of the miR-200 family were decreased in MES-OV/TP cells [47]. The retroviral transfection of both lines with miR-200 mimics resulted in a 4-fold increased resistance to carboplatin after miR-c141 cluster transfection, but not after transfection with the miR-ab429 cluster [47]. Additionally, miR-200c and miR-141 mimics treatment, although conferring resistance to carboplatin, sensitized MES-OV/TP cells to paclitaxel [47]. It is known that the expression of ZEB1, ZEB2 and SNAI2 inversely correlates with the expression of the miR-200 family [48–50]. In the study by Brozovic et al. [47], both resistant variants displayed a strong EMT phenotype, with a strong expression of ZEB1, ZEB2 and SNAI2 in miR-141 and miR-200c treated OVCAR-3 cells. The authors also examined the expression of TUBB3 and found that the inhibition of miR-200c and miR-141 resulted in the upregulation of this β-tubulin isotype in the paclitaxel-sensitive lines [47], verifying previous reports regarding the inverse correlation between miR-200 members and [41, 43] EMT markers.

The authors thank the Lega Italiana per la Lotta contro i Tumori – Catania for its support.