Deregulation of miR-100, miR-99a and miR-199b in tissues and plasma coexists with increased expression of mTOR kinase in endometrioid endometrial carcinoma

Endometrial cancer is the most common malignant tumor of the female reproductive tract and the fourth most often occurring malignancy in the female population of developed countries [1, 2]. The incidence of endometrial cancer has increased in the Great Britain, as it was reported by Evans et al. [3]. Similar trend was also observed in Poland basing on the data provided by Polish Cancer Registry [4]. Most endometrial cancers are diagnosed in early clinical stages and can be managed surgically with good results, however there is still a number of cases with bad prognosis due to lack of effective treatment for advanced, recurrent or disseminated disease. Approximately 80% of endometrial cancers are of endometrioid histology [5]. Etiology of endometrioid endometrial tumors has been connected with alteration of several genes [6, 7]. In particular alterations of PTEN (phosphatase and tensin homologue) gene and its product were widely investigated and loss of PTEN activity was proven to induce increased AKT (v-akt murine thymoma viral oncogene homolog; rac-beta serine/threonine protein kinase) and mTOR activity [8‐10].

mTOR (mechanistic target of rapamycin) gene is located on chromosome 1 (locus 1p36.2) and encodes for the serine/threonine kinase, which plays a role in cellular response to various stress factors including decreased ATP level, hypoxia and DNA damage. mTOR is involved in the regulation of translation process, cellular proliferation and differentiation [11, 12]. It is of particular interest that mTOR kinase, which integrates several cellular pathways activated in the response to changes in ATP and growth factors levels as well as those responsible for cell growth and proliferation, becomes deregulated in many malignant diseases including endometrial cancer [13‐16]. It has been proven that it is activated via AKT phosphorylation both directly and indirectly through inactivation of TSC2 (tuberous sclerosis complex 2) [17]. A negative regulation of mTOR activity involves AMPK (5’-AMP-activated protein kinase), which inhibits mTOR actions through phosphorylation and activation of TSC1/TSC2 complex [18, 19]. Increased expression and activity of mTOR kinase in endometrial cancer has been connected with the loss of PTEN, LKB1 (Liver Kinase B1) and TSC2 activity as well as increased expression of Phospholipase D1 [20]. Nevertheless, such alterations were not uniformly reported in all EEC cases therefore other mechanisms could be involved in the regulation of mTOR pathway activity including posttranscriptional regulation of mTOR expression, which involves microRNAs (miRNAs). miRNAs are small, non-coding RNA molecules composed of 19 to 25 nucleotides which were proven to regulate gene expression during posttranscriptional stages [21]. miRNAs play important roles in development, cell proliferation and differentiation as well as cell cycle control. Several miRNAs were also associated with the pathogenesis of various malignancies including endometrial cancer [22]. Guo et al. have recently proven that in mammalian cells miRNAs act predominantly through decreasing target mRNA levels [23]. Therefore down-regulation of miRNAs targeting genes involved in carcinogenesis is likely to favor initiation of malignant transformation. Alterations of genes, which are upstream to mTOR and could be responsible for its activation, are not uniformly present in all EEC cases [5]. Therefore in the presented study we aimed to assess mTOR kinase mRNA levels and to explore possible association between altered expression of mTOR kinase transcripts and its targeting miRNAs, miR-99a, miR-100 and miR-199b, which were chosen basing on the three open-source in silico algorithms (Target Scan, DIANA microT v. 3.0, miRanda). Expression of the three miRNAs was also investigated in plasma of EEC patients and was assessed in terms of their diagnostic and prognostic utility.

Several studies reported increased expression of mTOR protein and its phosphorylated form (p-mTOR) in endometrial cancer tissues and cell lines. Study by Darb-Eshafani et al. was one of the first to describe activation of mTOR pathway in a set of human endometrial cancer tissues [16]. Authors found that the activation occurred predominantly in high-grade, high-stage tumors [16]. Similar observation was made by Lu et al., who observed mTOR pathway activation in primary tumors and endometrial cancer cell lines, which was associated with decreased expression of mTOR upstream regulators like TSC2, LKB1 and PTEN [39]. In line with those studies Li et al. reported on activation of mTOR pathway in HEC-1A and Ishikawa cell lines, and found increased expression of mTOR mRNA and higher expression of its downstream substrates in PTEN-deficient Ishikawa cell line [40]. Yoshida et al. pointed at nuclear localization of p-mTOR as an important indicator of poor prognosis and tumor progression, whereas most other authors reported cytoplasmic localization of that kinase [40‐42].

A number of studies linked mTOR pathway activation with loss of PTEN and increased AKT expression [39, 42]. However Mori et al. did not observe any correlation between expression of phosphorylated AKT and PTEN mutations and expression of its downstream effectors including p-mTOR [43]. Moreover, in the study performed by Yoshida et al. there was no association between activated AKT and p-mTOR expression along with no correlation between AKT expression and cancer progression or outcome [41]. In accordance with those findings Darb-Esfahani et al. observed activation of mTOR in PTEN willed-type cells, in the absence of activated AKT [16].

Basing on the literature indicating mTOR pathway activation in endometrial tumors several mTOR inhibitors have been studied in this cancer [44, 45]. Interestingly in the recent, phase II study of temsirolimus in management of patients with recurrent or metastatic endometrial cancer mTOR inhibition did not correlate with PTEN status. In addition authors did not found any relation between expression of phosphorylated AKT, p-mTOR, or phosphorylated S6 and response to treatment or disease progression [46].

Taken together such observations suggest that alternative mechanisms, not related to PTEN/AKT alterations, could play a role in mTOR pathway activation. Based on that assumption we hypothesized that increased expression of mTOR kinase could be caused by alterations in posttranscriptional phases of its biogenesis, which might occur due to disturbed expression of its targeting miRNAs.

Using a quantitive PCR and three experimentally validated endogenous controls we found that mTOR transcripts level was elevated in EEC tissues. In addition we did not observe significant alteration of PTEN mRNA expression, which is a negative upstream regulator of mTOR kinase pathway. Such observation is in accordance with the previous study reporting increased expression of mTOR transcripts and protein in endometrial cancer cell lines as well as with findings presented by Yoshida et al. and Mori et al. [40, 41, 43]. In addition we have also observed that mTOR mRNA expression tended to be higher in histologically more advanced tumors (G2 and G3 of histological grading). Such finding is in line with the results presented by Darb-Eshafani et al., who observed that mTOR pathway activation was more prevalent in high grade, high stage tumors [16]. Yoshida et al. reported on similar observation, which indicated more pronounced p-mTOR expression in endometrial cancer and correlation between nuclear expression of phosphorylated mTOR and worse prognosis [41]. Contrary to those findings, Choi et al. described significantly decreased p-mTOR expression in deep infiltrating endometrial tumors and indicated that expression of p-mTOR predicted better disease-free survival in endometrial cancer patients [47]. No et al. reported on infrequent mTOR expression (7.1%) in endometrial cancer and did not reveal any correlation between mTOR expression and histological type, grade, stage, myometrial invasion, lymph node metastasis and survival in a large group of 141 patients [48]. In opposition to those observations Wahl et al. observed cytoplasmic p-mTOR expression in 89.2% of type 1 endometrial carcinomas and Mori et al. found moderate and strong expression of phosphorylated mTOR in 81% of endometrioid endometrial cancer samples [42, 43]. Such discrepant findings presented by different authors are difficult to interpret. To some extent they could be explained by histological heterogeneity of sample populations used in some of those studies [48]. However, relations between mTOR kinase expression and EEC biology warrant further evaluation and are particularly important in the context of only moderate responses observed in endometrial cancer patients treated with mTOR pathway inhibitors.

In our study increased level of mTOR mRNA was accompanied by down-regulation of miR-99a, miR-100 and miR-199b, which were found to target this gene in the analysis performed using three in silico algorithms. No previous study reported on the connection between expression of those miRNAs and mTOR transcripts level in endometrial cancer. However, miR-100 was found strongly down-regulated in clear cell ovarian carcinoma and its altered expression was associated with increased expression of mTOR mRNA and protein. Moreover, a direct link between miR-100 and mTOR expression was proven experimentally in the same study. The authors also found that the increase of intracellular miR-100 level enhanced therapeutic effect of mTOR inhibitor in clear cell ovarian carcinoma cell line [49]. Li et al. described down-regulation of miR-99a in hepatocellular carcinoma, which was accompanied by increased expression of its two target genes mTOR and IGF1R. The authors demonstrated, that transfection with miR-99a resulted in mTOR pathway inhibition and cell cycle arrest in that cancer [50]. Chen at al. obtained similar results and observed down-regulation of both mTOR and IGF1R after transfection with miR-99 family members in head and neck squamous cell carcinoma [51]. A connection between decreased expressions of miR-99a, miR-99b and miR-100 and elevated level of mTOR was also described in prostate cancer and was associated with more advanced tumor stage [52]. Important role of miR-199b-5p was found in medulloblastoma. Garzia et al. reported its lower expression in non-metastatic versus metastatic medulloblastomas and observed that over-expression of miR-199b-5p reduced proliferation of medulloblstoma cells in vitro and depleted tumor growth in vivo[53].

Although a number of studies reported alterations in miRNAs expression in endometrial cancer only three studies described down-regulation of miR-199b and miR-99 family members [54‐56]. Wu et al. compared miRNA expression in ten EEC samples and ten matching endometrial tissues and found decreased expression of miR-99b [54]. Chung et al. reported down-regulation of miR-199b and the fold change of its expression was similar to the one observed in our study [56]. In the study performed by Snowdon et al. the authors compared normal endometrium, atypical hyperplasia (AH) and EEC samples and found that expression of miR-100 and miR-199b was decreased both in AH and in EEC tissues [55]. Moreover down-regulation of those miRNAs was grater in EEC comparing to AH samples [55]. In our study a correlation was found between higher histological grade and lower expression of miR-199b and miR-99a. Although Snowdon at al. were not able to perform correlation analysis between histological grade and clinical stage and miRNA expression due to the small size of studied population, the difference between AH and EEC groups indicated that expression of miR-100 and miR-199b tended to be higher in more advanced stages of endometrial neoplasia [55].

Our study revealed that greater degree of miR-199b and miR-99a down-regulation was attributed to high–grade endometrial tumors. At the same time expression of mTOR kinase mRNA was higher in more histologically advanced tumors, which further supported our hypothesis on the connection between mTOR kinase and studied miRNAs. In addition, there was a significant reverse correlation between expression of all three studied miRNAs and mTOR mRNA expression in matched EEC tissues.

In order to further evaluate the role of miR-99a, miR-100 and miR-199b in EEC we assessed their expression in matched plasma samples obtained from the same group of EEC patients. Surprisingly we found that plasma expression levels of all studied miRNAs were significantly increased in EEC patients in comparison to healthy controls and that expression of miR-99a was higher in patients with clinically more advanced disease. Although the discrepant expression of miR-99a, miR-100 and miR-199b between EEC tissues and plasma is difficult to explain at the time, it could be an additional hallmark of this disease and warrants further investigation. Discrepant expression of another miRNA – miR-92a, was observed in tissue and plasma samples obtained from patients with leukemia, lymphoma and hepatocellular carcinoma [57‐59]. In the study performed by Tanaka et al. authors observed that miR-92a, which was attributed oncogenic properties, was highly increased in leukemic cells and down-regulated in the plasma of leukemia patients. Authors hypothesized, that apart from transcribing essential miRNAs, cancer cell might specifically take in exosomes containing miR-92a from the blood, which would result in observed decrease of that miRNA in plasma [57]. Ohyashiki et al. reported similar discordant expression of miR-92a in plasma and tumor tissues in non-Hodgkin’s lymphoma [58]. Shigoka et al. also found that miR-92a was highly up-regulated in tumors and significantly decreased in plasma of hepatocellular carcinoma patients [59].

Another important aim of our study was to evaluate diagnostic and prognostic values of miR-99a, miR-100 and miR-199b expression in tissues and plasma of EEC patients. For this purpose we constructed ROC curves and developed miRNA signatures, which could classify cancer and control samples with high sensitivity and specificity. The 3-miRNA tissue signature developed using multivariate logistic regression was more accurate in discriminating EEC samples in comparison to single miRNAs. Utility of such signature could potentially facilitate diagnosis in poorly differentiated, disseminated and metastatic tumors. Our study also revealed that the 2-miRNA plasma signature (miR-99a/miR-199b) was very accurate in discriminating EEC samples with sensitivity of 88% and specificity of 93%. Survival analysis indicated that both miR-99a and miR-100 were significantly associated with prognosis of overall survival and were independent prognostic factors in a multivariate analysis. One of the great challenges for performing qPCR based miRNA expression analysis in tissues and especially in plasma samples is to use proper normalization strategy, which in most cases bases on reference miRNAs or small RNAs. In the presented study normalization was performed using carefully chosen endogenous controls both for tissue and plasma analysis, which increases validity of our results.

To our best knowledge only two previously published studies attempted to evaluate diagnostic and prognostic values of miRNA expression in endometrial cancer tissues.

Corresponding with our results, Chung et al. reported that expression of miR-199b yielded AUC value of 0.837 in detection of EEC tissue samples [56]. Cohn et al. found that miR-199a was more commonly overexpressed in a subset of patients who did not experience recurrence or death from the disease and its down-regulation in endometrial cancer tissues correlated significantly with shorter progression-free and overall survival [60].

Early diagnosis of endometrial cancer is of immense importance as it facilitates radical surgical treatment and prevents the necessity of adjuvant therapy, which can be connected with severe side effects. However, at present no reliable screening strategy is available for this malignancy. We suggest, that after proper validation tissue and plasma miRNA signatures developed in the current study could have a potential to be applied in a clinical setting. Although plasma miRNA signature was highly accurate in distinguishing between EEC and control samples the number of samples used for its development can limit the value of this finding. Therefore we are planning to further investigate and validate our results in a larger group of patients and in the context of other gynecological pathologies.

We conclude that increased expression of mTOR kinase in endometrioid endometrial cancer coexists with down-regulation of its targeting miRNAs, miR-99a, miR-100 and miR-199b, which could suggest a new mechanism of mTOR pathway activation in this malignancy. Although functional studies are necessary to directly prove mTOR regulation by studied miRNAs in EEC, presented results along with studies performed in other tumors seem to support this novel mechanism of altered mTOR kinase pathway. In addition, our findings implicate that miR-99a, miR-100 and miR-199b signatures are strongly associated with the diagnosis of endometrioid endometrial carcinoma and after validation in larger groups of patients could be considered promising non-invasive biomarkers for detection and prognosis of this cancer.