Identification of micro-RNA expression profile related to recurrence in women with ESMO low-risk endometrial cancer

Endometrial cancer (EC) is the most common gynecologic cancer in women in developed countries. The highest estimated incidences in 2012 are in the USA and Canada (19.1/100,000) and northern (12.9/100,000) and Western Europe (15.6/100,000) [1, 2]. It is also the 14th cancer in terms of mortality accounting for 76,000 deaths/year worldwide [1]. The most frequently occurring histologic subtype is endometrioid adenocarcinoma, with a good prognosis [3]. Women are diagnosed at an early stage (stage I of the International Federation of Gynecology and Obstetrics (FIGO) classification) in 75% of the cases, when the disease is still confined to the uterus.

Among various risk stratification systems, the European Society for Medical Oncology (ESMO) system, with its four risk groups classification for early stage EC, provides the highest discrimination for both recurrence free survival (RFS) and the risk of nodal metastases [4]. However, it shows insufficient accuracy to accurately stratify recurrence risk in women with early-stage EC, as their recurrence rates range from 2 to 26% depending on several epidemiological and histological factors [5]. Thus, additional tools are needed to promote a more individualized management of early-stage EC women. In this perspective, a growing number of studies are focusing on the key role of genetics and molecular biology in EC [6].

Micro-RNAs are short non-coding RNA, ~ 22 nucleotides in length, which work at post-transcriptional level as epigenetic regulators. They can inhibit the translation process, or initiate the process of mRNA degradation [7]. They are implied in numerous process including carcinogenesis [8], where they act as tumor suppressors or activators (onco-miRs). Several micro-RNA profiles have been described in EC, in particular related to histopathological characteristics and prognosis of the tumor [9, 10]. Especially, down regulation of micro-RNA-184, and -34c-5p, involved in epithelial to mesenchymal transition (EMT), have recently been linked to the lymph node involvement in women with early-stage grade 1 and 2 EC [11]. They can be easily detected in FFPE carcinogenic specimens [12], which opens new opportunities to develop accurate diagnosis tools for tumor staging. However, few data are available about micro-RNA profiles and recurrences in low-risk EC (LREC).

The aim of this study was to evaluate whether micro-RNA profile from LREC women can be correlated with the occurrence of recurrences and used as a tool to adapt therapeutic management.

Many of the methods related to our patient cohort, micro-RNA extraction, microarray hybridization and analysis, qRT-PCR, and minimal p-value approach have been previously published [10, 11].

Our results show that in LREC, R+ women have different micro-RNA profiles from R− women. Moreover, we found that recurrence status can be accurately assessed using micro-RNA expression levels on primary FFPE tumor.

The crucial issue in managing patients with early-stage EC is to determine the risk of recurrence to better adapt the therapeutic strategy. Currently, higher risk of recurrence has been described when early stage EC was associated with histological settings, including myometrial involvement > 50% [16], histological type 2 and grade 3 tumors [17, 18] and presence of LVSI [19]. However, histological factors such as histological type or LVSI status have been criticized for their lack of reproducibility [20]. Guidelines for therapies widely differ from one country to another [21‐25], as they rely on various risk stratification systems (RSS) [3, 19, 21, 22, 26]. Bendifallah et al. in their comparison of five major RSS applied to a multicenter population with early-stage EC, showed that they all have a poor-to-moderate discrimination for recurrence, and are also heterogeneous in terms of classification performance [4]. Over or under estimation of these risks lead to inadequate surgical and adjuvant managements of EC women. This underlines the need for new biological markers, so as to enable a better stratification of women with early stage EC, and thus a better optimization of their treatments and follow up patterns.

Recently, Levine et al. described four molecular EC categories based on integrated genomic, transcriptomic, and proteomic analysis: p53-mutant, microsatellite instability (MSI), POLE-mutant, and no specific molecular profile [27]. Confirmation of the prognostic capacity of these four subgroups in a large randomized trial was provided by Stelloo et al. who studied 834 women from the PORTEC cohorts [28]. However, to our knowledge, there’s no study specifically focusing on biomolecular prognosis profile for LREC women. To address this issue, we assessed micro-RNA level expression in primary FFPE specimen of LREC women according to recurrence status. Microarray analysis performed after RNA extraction from 7 FFPE specimens from LREC R+ women matched with 14 FFPE specimens from LREC R− women showed a significant difference in the expression of 6 micro-RNAs: -184, -497-5p, 195-5p, -7162-3p, -196b-3p, and -6080. When studying the validated target genes of these micro-RNAs by enrichment analysis, they all appeared as exclusively involved in carcinogenic pathways, thus reinforcing our analysis. According to literature, microRNAs-184, -497-5p, 195-5p, and -196b-3p had already been described as involved in various cancers, including EC [9, 29, 30]. Furthermore, they have been described associated to the epithelial-to-mesenchymal transition (EMT), which play a key role in cell motility and metastatic spread. The recent discovery of micro-RNA-7162-3p and micro-RNA-6080 could explain why there are no currently available articles about them. In their review of the current state of science regarding microRNA functionality in EC progression, Stope et al. note that, interestingly, most micro-RNAs in EC cells have been classified as tumor suppressors, with cellular expression being suppressed by malignant processes [31]. This is consistent with our results as 5 micro-RNAs are under expressed in LREC R+ women with a FC > 2 vs. one over expressed.

After validation by RT-qPCR, we found that micro-RNA-184, -497-5p, and -196b-3p emerged as being particularly relevant to determine the occurrence of recurrence in women with LREC.

Micro-RNA-184 has been described as a tumor suppressor in various cancers (lymphoma [32], renal carcinoma [33], breast cancer [34, 35], lung cancer [36, 37], ovarian cancer [38], glioma [39], otorhinolaryngologic cancer [40‐42], gastric cancer [43] or adrenocortical tumors [44]). Micro-RNA184 acts by repressing oncogenes such as TNFAIP2 [39, 43], SND1 [45], CDC25A, c-MYC and Bcl-2 [36], and by regulating the AKT/mTORC1 pathway. In EC, micro-RNA-184 was first described as under expressed by Chung et al. who profiled micro-RNA expression of 30 EC specimen FIGO stage from IA to IIIC, and 22 normal counterparts [29]. More recently, Canlorbe et al. found a strong correlation between level expression of micro-RNA-184 extracted from FFPE specimen of women with grade 1–2 EC and lymph node status [11]. In our study, all included women underwent pelvic nodal staging. They were excluded in case of positive lymph node, as it represents a major bias in the analysis of recurrences. In women with LREC, micro-RNA-184 emerges as an interesting prognostic tool, as its under expression is significantly associated with pelvic lymph node metastasis and/or recurrences.

Similarly, a decreased expression of micro-RNA-497-5p is associated to multiple myeloma [46], thyroid cancer [47], colorectal cancer [48, 49], cervical cancer [50], glioma [51], ovarian cancer [52, 53], breast cancer [54‐56], bladder cancer [57, 58], hepatocellular carcinoma [59, 60], otorhinolaryngologic cancer [61, 62], pancreas cancer [63], prostate cancer [64], lung cancer [65], renal cancer [66], gastric cancer [67], melanoma [68] or astrocytoma [69]. Various studies showed that over expression of micro-RNA-497 can suppress cell proliferation and induces apoptosis through targeting paired box 3 [46], paired box 2 [52] and YAP1 [59], suppress EMT and metastasis by targeting fos related antigen-1 [48] or vascular endothelial growth factor-A [49], and reduce tumor growth and invasion by suppressing BDNF [47].

Finally, micro-RNA-196 is a known emerging biomarker for digestive tract cancers, where it plays the role of an onco-miR, as it is consistently found over expressed in digestive tract cancer tissues [70‐73]. It has also been described as up regulated in otorhinolaryngologic cancer [74], glioblastoma [75], lung cancer [76], ovarian cancer [77] or osteosarcoma [78]. On the other hand, micro-RNA-196 down regulation is implicated in homeobox B7-vascular endothelial growth factor pathway, which plays an important role in cervical cancer progression [79]. It was also described as down regulated among 54 other micro-RNAs in endometrial serous carcinoma compared to normal endometrial tissue [30]. Abe et al., in their study on the global expression pattern of micro-RNA in endometriotic stroma cells, found that expression of micro-RNA-196b in endometriotic cyst stromal cells is repressed by DNA hyper methylation of the micro-RNA-196b gene, which may be involved in the development of proliferative and anti-apoptotic characteristics of endometriosis [80].

Both micro-RNA-184 and -196 have been described as tumor genesis marker. Indeed, MicroRNA 184 was also described as significantly downregulated in tissues from patients with colorectal cancer, ovarian cancer or prostate carcinoma, compared with matched normal adjacent tissues [38, 81, 82]. It was also differentially expressed in glioma tissues compared to normal brain tissues [83], and in tongue carcinomas tissues compared to paired normal tissues [84]. Micro-RNA-196 was differentially expressed in pancreatic resection specimens from carcinoma cases and its precursors vs. benign pancreatic specimens [85], in oral squamous cell carcinoma tissues (OSCC) vs. contral mucosa [86], in breast cancer samples vs. normal breast samples [87] and in colonic cancer specimens vs. para-cancerous specimens [88]. But both of them have also been described as associated with tumor recurrence, in granulosa tumors for micro-RNA-184 [89] and in OSCC for micro-RNA-196 [86]. Unfortunately, the design of our study didn’t include normal adjacent tissues nor blood samples. To confirm that these selected microRNAs are involved not only in tumor genesis but also in recurrences, further studies comparing their expression in LREC and normal sample should be completed.

In our study, we focused on post-operative FFPE EC specimen, since it is known that FFPE tumors can be used for RT-PCR-based quantitative micro-RNA profiling [9]. This enabled us to work on samples taken between 2000 and 2014, and selected among the entire specimen collections of the Pathology Departments. In the aim of optimizing the adjuvant management and the follow up of the women, working on post-operative specimen is not a problem. However, to provide a preoperative diagnostic tool so as to better stratify women and to optimize their surgical procedures, our analysis should be validated before initial surgery. To do so, two options can be considered: the first one is to assess micro-RNA expression in pre-operative tumor biopsy samples, the second one is to work on pre-operative blood samples. Studying micro-RNA expression profiles on biopsy samples comes with a risk of contamination of normal endometrial or myometrial tissue, which could lead to a wrong analysis. Interestingly, several authors have demonstrated that circulating micro-RNA profiles extracted from plasma samples were correlated with profiles extracted from FFPE specimen [90]. In EC, various studies have demonstrated that circulating micro-RNA could be used as diagnostic tools for endometrioid EC (micro-RNA-15b, -27a, and -233, and micro-RNA-9/micro-RNA-1228 and micro-RNA-9/micro-RNA-92a respectively [91, 92]). Micro-RNA-184, -497, and -196b-3p, have been described in various cancer as promising diagnosis tools preoperatively assessed in blood samples [34, 44, 50, 58, 69, 73, 93, 94]. Owing to their significant implication in EC progression, they may be promising tools of diagnosis, prognosis and treatment response. Micro-RNA 184, who was studied in numerous cancers, appears as a serious candidate for future researches. Although recent technology improvements may allow micro-RNA-184 detection and quantification easily from clinical samples, further investigations are still needed first to assess the exact role of this micro-RNA in endometrial carcinogenesis, second to precise its interactions with the other selected microRNA and third to prove the correlation between blood samples and tumor samples in patients with endometrial cancer.

Some limitations of the study should be underlined. Because of the use of FFPE specimens with insufficient quality and RNA integrity, we could not assess the expression of exact targets for the micro-RNAs validated in this study. Further studies are thus needed to precisely determine the different pathways affected by the selected micro-RNAs. Second, we included patients with both local and distance recurrence that have different long-term prognosis. Indeed, recently, Ouldamer et al. demonstrated that patients recurring under the form of distant metastasis or carcinomatosis had poorer prognosis [95]. Moreover, from clinical point of view this difference is relevant, as patients with local recurrence could have benefited from external beam radiotherapy while patients with distant metastasis or carcinomatosis could have benefited of adjuvant chemotherapy. Third, due to the sample size, it was impossible to determine a combination of micro-RNAs allowing the detection of all R+ patients. Future research is needed to resolve this problem. Finally, the retrospective nature of our data collection, which was focused on the occurrence of recurrence, should be underlined as a possible bias.

This is, to our knowledge, the first study investigating a correlation between micro-RNA signatures in tumor tissues from LREC women and the occurrence of recurrences.

Our results may provide a basis for further study of the micro-RNA function EC, and be used as a valuable tool for better diagnostic and therapeutic management.