Endometrial cancer (EC) is the most common gynecological cancer, with an increasing incidence in Western countries ranging between 15 and 25 per 100,000 women (1, 2).

Historically, three distinct but overlapping EC classifications have been proposed to define prognosis (3): pathogenetic, histopathological, and molecular classification. However, despite their merits, the three classifications are not able to uniquely define the complexity and heterogeneity of the disease and present various issues. Histopathological parameters used alone to identify risk factors are not always easily reproducible, particularly in high-grade carcinomas and those with intratumoral heterogeneity (4–8).

To overcome this kind of problems, recently, The Cancer Genome Atlas (TCGA) endometrial collaborative project identified four distinct prognostic EC groups based on molecular alterations: (i) the ultramutated subtype that encompassed POLE exonuclease domain mutated (POLE) cases (excellent prognosis); (ii) the hypermutated subtype, characterized by MisMatch Repair deficiency (MMRd) (intermediate prognosis); (iii) the copy-number high subtype, with p53 abnormal/mutated features (p53abn) (poor prognosis); and (iv) the copy-number low subtype, also known as No Specific Molecular Profile (NSMP) (intermediate prognosis) (9). However, in order to translate the proposed TCGA scheme into clinical practice, two independent research groups proposed and validated surrogate markers (POLE mutation, microsatellite instability, and p53abn to improve the integration of the TCGA classification in the routine clinical practice (9–12). In 2021, the European Society of Gynecological Oncology (ESGO), the European Society for Radiotherapy and Oncology (ESTRO), and the European Society of Pathology (ESP) published updated guidelines for the determination of the risk group in EC, integrating both molecular and clinicopathological diagnostic variables, with the aim of improving patients’ treatment (13).

The ESGO/ESTRO/ESP guidelines consider all POLE mutant ECs up to stage II to be at low risk of recurrence regardless of other ESMO risk parameters, and adjuvant treatments are deemed unnecessary in these cases. The p53abn group benefits from adjuvant therapy which is also expected to be aggressive independently of histology and different managements for these two entities may not be recommended. Within the MMRd and NSMP groups, which have an intermediate prognosis, FIGO grade, LVSI, and depth of myometrial invasion are critical factors for risk stratification. Thus, in these two groups, treatment continues to rely largely on conventional risk factors, and the choice of the most appropriate adjuvant therapies remains a challenge. MMRd ECs represent a heterogeneous group of cancers. A better stratification within this group to select therapies is needed in consideration that a subgroup of MMRd ECs exhibiting MLH1 promoter methylation presents a worse prognosis than other MMRd ECs (14). In the NSMP group, the absence of specific molecular signatures appears to be accompanied by a more heterogeneous biological behavior than the other TCGA groups. This group mainly includes low-grade endometrioid-type ECs characterized by alterations in PI3K/AKT and Wnt/β-catenin signaling pathways often with mutations on exon 3 of CTNNB1 (52%). Although a few studies have characterized the CTNNB1 alterations in EC (15, 16), its prognostic significance among NSMP ECs is not completely understood; moreover, it is not clear if CTNNB1 mutations could indicate a distinct molecular group (17). This shows that integration of different prognostic factors may serve to accurately define the risk classes, leading to a more precise characterization of ECs which are extremely heterogeneous and sometimes unclassifiable.

MicroRNAs (miRNAs) are small non-coding RNAs spanning between 18 and 25 nucleotides in length, which are able to regulate specific target genes at the posttranscriptional level by inhibiting their expression (18). Consequently, miRNAs deregulation can have a deep impact on key cellular processes which, in turn, could drive carcinogenesis or promote cancer progression. In recent years, many studies investigating miRNA expression level in EC have been published, shedding light on this heterogeneous disease (19). However, identifying novel potential diagnostic biomarkers remains an unmet clinical need, particularly within the EC patients with intermediate prognosis who represent the most difficult category to manage from a clinical point of view (i.e., NSMP and MMRd groups).

Based on the above, we aimed to characterize the miRNA expression level in these groups defined as at intermediate risk, considering the CTNNB1 status, in order to better stratify these patients. To do that, we first focused on the NSMP and MMRd EC patients comparing patients harboring, or not, CTNNB1 mutations; secondly, we tested our findings in the TCGA EC cohort to corroborate them.

In the last decade, several miRNAs have been described as potential biomarkers of EC prognosis or diagnosis. For example, among others, miR-34a was identified as able to stratify patients at high risk of recurrence (26) and reduced miR-497-5p levels were reported in high-grade ECs compared with low-grade Ecs (27). However, to the best of our knowledge, this is the first study to analyze miRNA expression in EC with the aim to obtain prognostic information in the intermediate-risk groups (i.e., NSMP and MMRd) of the molecular classification, trying to create subclassifications while taking into account the CTNNB1 mutational status.

In the attempt to better define this latter group, we first evaluated the miRNA expression profile in CTNNB1^mut ECs compared with the CTNNB1^wt NSMP ones. To this purpose, we analyzed the miRNA expression profile in 23 EC cases, identifying miR-499a-3p and 5p as significantly upregulated miRNAs in the CTNNB1^mut cases when compared with the wt ones. However, considering that in literature the NSMP EC subgroup appears to be similar to the MMRd group at the prognosis level, we included seven additional MMRd patients.

Based on the results, we selected four miRNAs to validate in 72 EC samples and we observed a high statistical significance for miR-499a-3p and -5p. Finally, to corroborate our data, we analyzed the TCGA EC cohort, stratifying the patients based on the CTNNB1 mutation. In agreement with our results, the CTNNB1^mut patients showed a higher expression of both miR-499a-3p and -5p in NSMP ECs as well as among the NSMP/MMRd patients. We further analyzed the survival rate in terms of OS and PFI with regard to the two-miRNA level. We did not perform this analysis in our study cohort due to the limited number of samples and events which, from a statistical point of view, will not give strong results. In fact, considering the MSI and NSMP groups, the time needed for the development of such events is a limiting factor and requires a statistical power that we did not have; therefore, at this time, validation of our results in the TCGA cohort appeared more appropriate.

Based on miRNA expression, the patients were divided into two groups (high and low expression) using the median of their expression level as a cutoff. MiR-499a-5p was significantly associated with OS, showing that NSMP patients with a lower expression have a prolonged survival. Conversely, the high level of miR-499a-5p was an independent predictor of poor survival.

The same correlation was observed when miRNA level and CTNNB1 status were combined. Indeed, NSMP patients with no CTNNB1 mutations and high miR-499a-5p expression had an increased risk of death.

MiR-499a-3p and 499a-5p belong to the miR-499 family, together with miR-499b. Human miR-499 is located in intron 19 of the MYH7B gene and it is composed of two genes, miR-499a and miR-499b. Those are located in sense (miR-499a) and anti-sense (miR-499b) DNA chains of the same region and transcribed in antiparallel directions (28). Data about miR-499b are missing, whereas several reports on miR-499a dysregulation have recently been published. Interestingly, a previous work by Zhang et al. showed a link between miR-499a and the Wnt/β-catenin signaling pathway. In particular, the authors demonstrated that overexpression of miR-499a activated β-catenin in cardiac differentiation. With regard to cancer, a few recent studies have tried to clarify the roles of miR-499a. However, it is not surprising that the results were contradictory among the analyzed tumor types because, as well established, the precise function of a single miRNA is strongly dependent on the specific biological context. Li et al. demonstrated that overexpressed miR-499 inhibited non-small cell lung cancer growth by suppressing cell proliferation and promoting apoptosis (29). On the contrary, Liu et al. showed that in colorectal cancer (CRC), miR-499 acts as an onco-mir by targeting two tumor suppressor genes, FOXO4 and PDCD4. Thus, overexpression of miR-499 is associated with advanced CRC stage and with cellular invasion and tumor metastasis in in vitro CRC models. In our results, a lower expression of the miRNA was associated with better survival in NSMP patients. Even if a clear mechanism has not been investigated, an interesting explanation could be related to the APC gene. Physiologically, APC acts as a scaffold in the β-catenin destruction complex, so it tightly regulates the Wnt pathway activity (30). APC is a target of miR-499a-5p based on TargetScan prediction, and its upregulation could promote inhibition of APC which, in turn, is less efficient in regulating the β-catenin level (31). Consequently, β-catenin can translocate in the nucleus and activate the Wnt pathway which fosters carcinogenesis. Thus, the higher miR-499a-5p expression could raise the APC inhibition which, consecutively, will be less effective in regulating the cytoplasmatic β-catenin destruction. Figure 6 shows the proposed mechanism. The CTNNB1 mutations by themselves are classically associated with nuclear translocation and activation of Wnt/β-catenin signaling (15). Interestingly, these mutations are usually found in heterozygosity (17, 32, 33) and, thus, it is reasonable to suppose that miR-499a-5p could play a role in CTNNB1 wild-type cells. Basically, higher miR-499a-5p expression represents an advantage for the tumor cells which may activate transcription of pro-tumorigenic genes besides normal CTNNB1. This could explain why a lower expression of miR-499a-5p resulted to be associated with a better outcome in NSMP patients; however, functional studies are needed in order to test this hypothesis. Our work has several strengths; first, the sample size of NSMP and MMRd patients is numerically important considering the incidence of this tumor. Second, the data were corroborated in the TCGA cohort, an independent and well-described group of EC patients. In the TCGA data, the Illumina platform was used to assess the miRNA expression, whereas we used the RT-PCR-based array and assays; accordingly, two different techniques were exploited to investigate miRNA levels. Third, our cohort was highly homogeneous from a molecular point of view, and this allowed us to limit potential biases; finally, all the patients were homogeneously treated according to ESMO risk classifications. However, there are some limits to consider; among these, we have to include the retrospective design of the study which, however, was essential to creating a model that could be prospectively validated in larger and independent cohorts and possibly on blood samples to assess the validity of the miRNA as a liquid prognostic biomarker. We did not analyze the two miRNAs in the other TCGA groups (i.e., POLE and p53abn groups) because we focused our attention on the NSMP and MMRd ECs in which TCGA had the smallest impact as regards the definition of therapy. Indeed, the classic pathological parameters with all their limitations, defined by ESGO/ESTRO/ESP guidelines, continue to be the gold standard for the therapy selection. Unfortunately, we could not perform any functional validations of our model due to the lack of commercial cell lines with no mutations on p53 or without microsatellite instability.

In conclusion, to the best of our knowledge, we showed, for the first time, that specific miRNAs are associated with the molecular subgroups proposed by the TCGA. miR-499a-5p, in particular, may represent a novel independent prognostic biomarker in NSMP ECs and, if validated in an independent cohort of patients, could improve the current classification and promote a more personalized management of those patients.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding authors.

The studies involving human participants were reviewed and approved by 189/2021/Oss/AOUBo, ClinicalTrials.gov Identifier: NCT04845425. The patients/participants provided their written informed consent to participate in this study.

Conceptualization, GR, ADL, and AMP. Methodology, GR, ADL, FG, and DdB. Validation, GR, FG, ADL, and AMP. Formal analysis, GR, CaC, and ADL. Investigation, ADL, DdB, ClC, AMP, and PDI. Resources, AMP and PDI. Data curation, GR, ADL, FG, DdB, AMP, and ClC. Writing—original draft preparation, GR, AMP, ADL, and DdB. Writing—review and editing, GR, CaC, GT, DdB, PDI, AMP, GD, MT, DS, EDC, AGC, and PH. Supervision, GR, AMP, SA, and PDI. All authors contributed to the article and approved the submitted version.

The work was supported by Fondazione Cassa di Risparmio in Bologna (Carisbo): project number #19094 to AMP.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.