The human EC KLE and Ishikawa (ISH) cell lines, which have already been authenticated, were purchased from the American Type Culture Collection, which had already been authenticated. Both cell lines were maintained in DMEM medium. For DNA demethylation treatment, EC cell lines were treated with 3 µM 5-aza-2-deoxycytidine for 72 h at 37°C.

A total of 68 consecutive paired EC tissues with cancerous and non-cancerous tissues were obtained from patients who underwent primary resection without neoadjuvant therapy at the Fudan University Shanghai Cancer Center (FUSCC) between January 2014 and April 2014 and who were pathologically diagnosed with endometrioid adenocarcinoma or serous carcinoma of the endometrium. All specimens were frozen in liquid nitrogen immediately following surgical removal and were then stored at −80°C. Subsequently, tissue samples from 42 patients were randomly selected for further research. Informed consent was obtained from the patients and the study was approved by the Committee for the Ethical Review of Research at FUSCC. The characteristics of the patients are summarized in Table SI.

Total RNA was extracted from freshly-frozen tissue samples and cell lines using TRIzol^® reagent (Thermo Fisher Scientific, Inc.). miRNAs were isolated using an mirVana™ miRNA isolation kit (Thermo Fisher Scientific, Inc.). For the microarray analysis, purified total RNAs from EC cells lines treated with 5-aza-2-deoxycytidine or saline, and then hybridized with a microchip using an miRNA Complete Labeling and Hyb kit (Agilent Technologies, Inc.). The signals were then detected using the Agilent Microarray Scanner and the images were analyzed using the Agilent Feature Extraction Software version 10.5.1.1 (Agilent Technologies, Inc.). Significantly upregulated miRNAs following 5-aza-2-deoxycytidine treatment were identified. TaqMan miRNA assays (Applied Biosystems; Thermo Fisher Scientific, Inc.) were used to quantify the relative expression of miR-638. Semi-quantitative PCR with SYBR green I was used to compare the relative expression of specific mRNAs using a SYBR^® Premix Ex Taq™ kit (Takara Biotechnology Co., Ltd.). The thermocycling conditions for miR-638 were as follows: An initial step at 95°C for 15 sec followed by 40 cycles of 95°C for 5 sec and 60°C for 30 sec. Amplified fragments were then detected on 4% agarose gel electrophoresis containing ethidium bromide using a ChemiDoc XRS system and Quantity One 1.0 software (Bio-Rad Laboratories, Inc.). The primers used are listed in Table SII.

Genomic DNA was extracted from patient tissues and cell lines using the QIAamp DNA formalin-fixed paraffin-embedded Tissue kit (Qiagen, Inc.). Sodium bisulfite modification of DNA was performed using the EpiTect Bisulfite kit (Qiagen, Inc.). PCR products were recovered using Qiagen gel DNA kits (Qiagen, Inc.) and then sequenced at GeneTech (Shanghai) Co., Ltd. The primer sequences are listed in Table SII.

miR-638 mimic (5′-AGG GAT CGC GGG CG GGT GGC GGC CT-3′), miR-638 mimic negative control (5′-AGG TAC GAA ACG CTA AGA AT-3′), short hairpin RNA targeting MEF2C (shMEF2C; 5′-GGA CAA GGA ATG GGA GGA TAT-3′) and shMEF2C negative control (5′-UCU CCG AUG CAG GCU CAAC-3′) were synthesized by Shanghai GenePharma Co., Ltd. For transfection, 100 ng miR-638 mimic, miR-638 mimic negative control, shMEF2C or shMEF2C negative control was transfected using Lipofectamine™ 2000 Transfection Reagent (Invitrogen; Thermo Fisher Scientific, Inc.). In addition, a specific vector containing the entire coding sequence of MEF2C but without its 3′-untranslated region (UTR), as well as its negative control, were designed by Shanghai GenePharma Co., Ltd. The lentiviral packing kit used to transfect the various vectors were purchased from the Open Biosystems, Inc. and transfection was performed at a final concentration of 50 nM. 293T cells were purchased from the Cell Bank of Chinese Academy of Sciences. Generally, EC cell lines with 5×10⁴ cells were infected with 1×10⁸ lentivirus-transducing units. Following transfection for 24 h, the cells were cultured with 1 µg/ml puromycin for 2 weeks, and then stably transduced EC cell lines were selected.

Cell viability was investigated using Cell Counting Kit-8 and colony formation assays. Cell migration and invasion assays were performed using Transwell cell migration assay kits and BioCoat™ Matrigel^® Invasion Chambers (Corning Incorporated), respectively. For the migration assay, 5×10⁴ cells were seeded into the upper chamber and incubated for 24 h at 37°C in 5% CO₂. For the invasion assay, 1×10⁵ cells were plated on chambers pre-coated with 1.6 mg Matrigel (Corning, Inc.) and incubated for 36 h at 37°C in 5% CO₂. Subsequently, cells were fixed, stained with 0.5% crystal violet solution at room temperature for 10 min, photographed and counted under a light microscope (magnification, ×200).

Cell apoptosis was analyzed with the Annexin V-fluorescein isothiocyanate/propidium iodide Apoptosis Detection kit. Briefly, 1×10⁵ cells were re-suspended in 500 µl binding buffer and stained with 5 µl Annexin V and 5 µl PI in the dark at room temperature for 20 min. The samples were then analyzed using a FACScan flow cytometer and the CellQuest™ Pro 1.0 software (BD Biosciences), according to the manufacturer's instructions.

Cells were harvested and lysed in ice-cold radioimmunoprecipitation assay lysis buffer. Protein concentrations were determined by the BCA Protein assay reagent and measured using a NanoDrop2000 Spectrophotometer (Thermo Fisher Scientific, Inc.). Subsequently, 10 µg protein/lane was separated by 10% SDS-PAGE gels and then transferred to PVDF membranes. The membranes were blocked with 5% non-fat milk for 2 h at room temperature and then incubated with primary antibodies against MEF2C (cat. no. 5030; Cell Signaling Technology, Inc.), matrix metalloproteinase (MMP)2 (cat. no. 40994; Cell Signaling Technology, Inc.), MMP9 (cat. no. 13667; Cell Signaling Technology, Inc.), vascular endothelial growth factor (VEGF; cat. no. ab69479; Abcam), cyclin D1 (cat. no. 2922; Cell Signaling Technology, Inc.), cyclin-dependent kinase (CDK)2 (cat. no. 2546; Cell Signaling Technology, Inc.), CDK4 (cat. no. 12790; Cell Signaling Technology, Inc.) and tubulin (cat. no. 5568; Cell Signaling Technology, Inc.) overnight at 4°C. All aforementioned antibodies were used at 1:1,000. Then, the membranes were incubated with the corresponding secondary antibodies overnight at 4°C. The secondary antibodies used in the present study included: Horseradish peroxidase-conjugated secondary antibodies against rabbit (cat. no. SA00001-2; 1:10,000) and mouse (cat. no. SA00001-1; 1:10,000), purchased from Proteintech Group, Inc. Finally, protein bands were visualized using a chemiluminescence detection system (Bio-Rad Laboratories, Inc.). Protein expression was quantified using Image-Pro^® Plus software (version 6.0; Media Cybernetics, Inc.).

A total of 100 4-week-old female nude mice were purchased from the Shanghai Institute of Materia Medica, Chinese Academy of Sciences and placed in laminar flow cabinets under specific pathogen-free conditions for 1 week. Then, 5-week-old mice were randomly grouped (6 mice/treatment group) and 1×10⁶ cells were injected subcutaneously into each mouse. The tumor size was recorded weekly. A total of 50 days following cancer cell injection, the mice were euthanized and the tumor weight was determined. All of the mouse experiments were performed under the guidelines of the National Institutes of Health for the Care and Use of Laboratory Animals. The study protocol was approved by the Committee on the Use of Live Animals in Teaching and Research of Fudan University.

DIANA-microT-CDS (http://diana.imis.athena-innovation.gr/DianaTools/index.php), TargetScan (http://www.targetscan.org/) (15) and miRNA (www.miRNA.org) were used for the prediction of miR-638 target genes. Luciferase constructs were generated by ligating oligonucleotides containing the wild-type or a 4 bp mutation in the putative target site of the MEF2C 3′-UTR into the Psi-CHECK2 vector (Promega Corporation) downstream of a luciferase gene. The wild-type/mutant type vectors were co-transfected with miR-638 mimics/negative control into 293T cells (Cell Bank of Chinese Academy of Sciences), using Lipofectamine (Thermo Fisher Scientific, Inc.). After 48 h, the cell lines were harvested and the luciferase activity of each group was detected using the dual-luciferase reporter assay system (Promega Corporation), with Renilla luciferase selected as the internal control.

All results are presented as the mean ± standard deviation. All statistical analyses were performed using GraphPad Prism software 7.0 (GraphPad Software, Inc.) and SPSS v.22.0 (IBM Corp.). For comparisons, one-way analysis of variance, t-test, and Pearson chi-square test were performed as indicated. Tukey's post hoc test was used for multiple comparisons test. Disease free survival (DFS), defined as the time interval between surgery to disease recurrence or mortality due to any cause, was estimated using the Kaplan-Meier method and differences among subgroups were investigated by the log-rank test. The correlations between relative miR-638 expression and DNA methylation level, as well as between relative miR-638 expression and relative MEF2C mRNA expression, were determined using the Spearman correlation coefficient. P<0.05 was considered to indicate a statistically significant difference.

To identify potential cancer-associated miRNAs whose promoters were highly methylated in EC, the EC ISH and KLE cell lines were treated with demethylation agents for 72 h, and the expression levels of 1,347 common human miRNAs were measured using micro-array analysis (Fig. 1A). With the cut-off criteria of P<0.05 and fold change >2, miR-638, miR-3665 and miR-210 were revealed to be significantly upregulated in both cell lines (Fig. 1B). miR-638 was selected for further study as it was a newly identified cancer-associated miRNA, and data concerning its specific biological roles in different types of human cancer are conflicting (9). A total of 2 CpG-rich regions at/near the promoter of miR-638 gene were identified using CpG Island Searcher (15). The first CpG island overlapped with the open reading frame sequence of the miR-638 gene and the second CpG island was located about 5kb upstream of the miR-638 gene (Fig. 1C). The methylation statuses of consecutive CpG sites in the second CpG island were highly concordant (Fig. 1D) and the mean value was used as representative of the methylation level of the miR-638 gene. A significantly higher degree of DNA methylation was identified in the cancerous tissues compared with the non-cancerous tissues among clinical samples from 42 patients with EC (Fig. 1E). The associations between the DNA methylation level of miR-638 in cancerous tissues with various clinicopathological parameters, including age, primary tumor size, Federation of Gynecology and Obstetrics (FIGO) stage (2014 version) (16), tumor differentiation and tumor histology were also analyzed. A significant positive correlation between miR-638 promoter methylation level and FIGO stage was identified (Fig. 1F). Of note, no statistical significance was identified when comparing the methylation (P=0.679) or the relative expression (P=0.523) levels of miR-638 in the cancerous tissue from the 36 patients with endometrial adenocarcinoma with that from the 6 patients with serous carcinoma.

To examine the effect of miR-638 promoter region methylation on miR-638 expression, the relative miR-638 expression levels were measured in 42 paired EC samples. A significant downregulation of miR-638 expression in cancerous tissues was detected (Fig. 2A). Furthermore, a significant inverse correlation between miR-638 DNA methylation and miR-638 expression was demonstrated (Fig. 2B). Additionally, DNA methylation and miR-638 expression levels were investigated in ISH and KLE cell lines treated with 5-aza-2′-deoxycitidine. Both cell lines exhibited hypermethylation of the miR-638 promoter region prior to treatment and the DNA methylation levels decreased significantly (Fig. 2C), whereas miR-638 expression was increased significantly following treatment (Fig. 2D).

When examining the association between miR-638 expression with the aforementioned common clinical variables, a significant negative correlation between miR-638 expression and FIGO stage was identified (Fig. 2E). Furthermore, with a median follow-up of 48.0 months (range, 14-58), 9 patients exhibited disease recurrence, and those with increased expression levels of miR-638 exhibited a prolonged DFS (Fig. 2F).

EC cell lines transfected with miR-638 mimics (miR-638) or its corresponding negative control (miR-638-NC), were constructed (Fig. 3A). Overexpression of miR-638 repressed cell proliferation (Fig. 3B and C), migration (Fig. 3D) and invasion (Fig. 3E), and promoted cell apoptosis (Fig. 3F). Furthermore, miR-638 overexpression led to the downregulation of MMP2, MMP9, VEGF, cyclin D1, CDK2 and CDK4 (Fig. 3G).

EC cell lines stably transfected with miR-638 mimics or its negative control were subcutaneously implanted into nude mice. It was identified that the overexpression of miR-638 significantly decreased tumor growth compared with controls (Fig. 3H).

MEF2C was identified to be one of the targets of miR-638 using several commonly used microRNA database. To validate this result, a luciferase reporter assay was employed. The entire wild-type 3′-UTR of MEF2C or the mutant 3′-UTR with a 4 bp mutation in the seed region was cloned downstream of the luciferase gene open reading frame (Fig. 4A). Luciferase activity was significantly decreased upon transfection of miR-638 mimics in the wild type reporter. Conversely, luciferase activity of mutant reporter was unaffected following transfection with miR-638 mimics or the control vector (Fig. 4B). In addition, the MEF2C mRNA (Fig. 4C) and protein (Fig. 4D) expression levels were substantially decreased following miR-638 overexpression in both EC cell lines. Furthermore, in the 42 paired clinical samples, MEF2C mRNA levels were significantly increased in the cancerous tissues (Fig. 4E) and a significant inverse correlation between MEF2C mRNA level and miR-638 expression was demonstrated (Fig. 4F). Of note, no statistical significance was revealed when comparing the MEF2C mRNA levels in the cancerous tissue from the 36 patients with endometrial adenocarcinoma with that from the 6 patients with serous carcinoma (P=0.472).

EC cell lines were transfected with lentivirus encoding shMEF2C, which mediated the stable knockdown of endogenous MEF2C. siRNA-mediated knockdown of MEF2C inhibited cell proliferation, migration and invasion, but promoted cell apoptosis. It also downregulated the expression levels of MMP2, MMP9, VEGF, cyclin D1, CDK2 and CDK4 and repressed in vivo tumor growth in mouse models (Fig. 5).

Next, a rescue experiment was performed to examine the biological and molecular associations between MEF2C and miR-638 in EC. A specific lentiviral vector (MEF2C analog) that encoded the entire coding sequence of MEF2C but lacked its 3′-UTR was constructed, in order to ectopically express MEF2C without affecting the expression of miR-638 and its numerous other targets. EC cell lines stably transfected with miR-683 mimics and/or MEF2C analog were selected, and MEF2C expression levels in these cells is presented in Fig. S1. Overexpressing miR-638 led to significant downregulation of MEF2C, while co-transfection of miR-638 mimics and the MEF2C analog (miR-638 + MEF2C) rescued the expression of MEF2C. Using the 3 specifically designed EC cell lines (miR-638-NC, miR-638 and miR-638 + MEF2C), it was demonstrated that the ectopic expression of MEF2C significantly abrogated the tumor-suppressive effect induced by miR-638 (Fig. 6).