The endometrial cancer dataset, including mRNA expression and clinical information, was obtained from The cancer genome atlas (TCGA, https://portal.gdc.cancer.gov/) database. The transcriptome data from TCGA was normalized and analyzed using the Limma package. Student t-test and Kruskal-Wallis test were applied to calculate the significance of expression differences between two or more groups, respectively. The univariate cox regression analysis was used to calculate the association between the expression level of TET2 and patient’s overall survival (OS). Hazard ratios (HRs) and 95% confidence intervals (CIs) were computed based on the Cox regression analysis. Survival curves were estimated using the Kaplan-Meier method and were compared using the log-rank test. The significance was defined as a P value of<0.05.

Approval for patient sample analyses was obtained from the Ethics Committee of Xuzhou Central Hospital Affiliated Xuzhou Medical University. The studies were conducted in accordance with the Declaration of Helsinki. All samples were collected from Xuzhou Central Hospital (Xuzhou, Jiangsu, China). In all, 88 EC tissues and 20 normal endometrial tissues were included in the study. Normal endometrial tissues were obtained from women who were undergoing a hysterectomy (for conditions such as uterine fibroids or prolapse).

Immunohistochemical staining was performed using 4-μm thick paraffin-embedded tissue blocks. Blocking with 3% hydrogen peroxide was performed for 30 min to quench endogenous peroxidases. Tissue sections on slides were incubated with primary antibody (1:200 dilution) at 4°C overnight. Then, secondary antibody was added and incubated with the tissues at 37°C for 30 min. Diaminobenzidine tetrahydrochloride was used as a chromogen. As a negative control, phosphate-buffered saline (PBS) was used instead of primary antibody. TET2 positivity was visible as yellow-brown granules in the nucleus and cytoplasm, and 5hmC-positive staining was visible as a brownish yellow color in the nucleus. Each section was independently assessed by two pathologists without prior knowledge of patient data. The samples were assigned a mean score considering both the intensity of staining and the proportion of cells with an unequivocal positive reaction in the immunohistochemical analysis. Positive reactions were defined as those showing brown signals mainly in the cell nucleus. The staining index (range, 0-3) was determined according to the staining intensity and positive area. Scores of 0-3 were defined as follows: 0, negative; 1, weak; 2, moderate; and 3, strong. For statistical analysis, scores of 0-1 were considered to indicate low expression, and scores of 2-3 were considered to indicate high expression.

The EC cell lines Ishikawa and HEC-1-A were purchased from The Cell Bank of Type Culture Collection of Chinese Academy of Sciences (Shanghai, China). The cell lines were cultured in RPMI-1640 (Thermo Fisher Scientific, Inc., Waltham, MA, USA) and McCoy’s 5A (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany) medium containing 10% fetal bovine serum (FBS; Thermo Fisher Scientific, Inc.) at 37°C in an atmosphere of 5% CO₂. The cells were passaged every 3-5 days. Metformin was purchased from Sigma-Aldrich; Merck KGaA. Primers were purchased from Sangon Biotech Co., Ltd. (Shanghai, China). The anti-phosphorylated (p)-AMPK (cat. no. BS4457P), and anti-AMPK (cat. no. BS4457) antibodies were purchased from Bioworld Technology, Inc. (St. Louis Park, MN, USA). TET2 primary antibodies were purchased from Abcam (USA) (ab94580, ab214728).

The endometrial carcinoma cell lines were transfected with siTET2 or siControl via reverse transfection using Lipofectamine RNAiMAX (Invitrogen, USA). In parallel, 1.5 μL of Lipofectamine RNAiMAX was mixed with 50 μL of Opti-MEM. The solution mixture was mixed by gentle pipetting and incubated for 10-20 min at room temperature to allow siRNA/lipid complexes to form. EC cells were suspended in complete growth medium without antibiotics at 50,000 cells/mL, gently mixed with 100 μL of the transfection solution, and plated. The cells were incubated for 24-72 h at 37°C and then assayed for gene knockdown.

The experiment was conducted according to the protocol of the Cell Counting Kit-8 Reagent Kit (Dojindo Molecular Technologies, Inc., Kumamoto, Japan). Cells transfected with siRNAs were seeded at 5,000 cells per well in 96-well plates and the TET2 knockdown cells and control cells were treated with 5mM metformin. They were both incubated in medium containing 10% FBS for 24 h, 48 h, 72 h and 96 h. After changing the medium without metformin, CCK-8 was added to each well, and the plates were incubated at 37°C for 1 h. Absorbance was measured at 450 nm using an automated microplate reader (Infinite 200; Tecan, Männedorf, Switzerland).

EC cells (1×10⁵/dish) were plated in 10-cm dishes and treated with 1, 5, or 15 mM metformin for 24 h. The cells were collected, resuspended in cell lysis buffer for western blotting, incubated on ice for 30 min, and centrifuged at 12,000× g for 10 min at 4°C. The supernatant was collected, and the protein content was quantified using a bicinchoninic acid (BCA) protein assay kit (Beyotime). Protein samples (20 µg) were separated in a 10% sodium dodecyl sulfate (SDS)-polyacrylamide gel and transferred to a polyvinylidene difluoride membrane (Millipore). After washing with PBST three times, the membranes were blocked with 5% nonfat milk for 30 min and then incubated with primary antibodies overnight at 4°C. The blots were then washed with PBST three times and incubated with the appropriate secondary antibodies. After washing with PBST three times, the protein bands were detected using an Odyssey Infrared Imaging system (Li-COR Biosciences). Primary antibodies against TET2 and AMPK were used at a dilution of 1:1,000, and secondary antibodies were used at a dilution of 1:2,000. The relative band intensity was analyzed with ImageJ software (version 1.47) (Schneider et al., 2012) and calculated as a ratio relative to the glyceraldehyde 3-phosphate dehydrogenase (GAPDH) band intensity. For evaluation of different blots, each band of the replicates was normalized to the GAPDH band intensity and then averaged. The averaged intensities were used for comparisons.

Ishikawa and HEC-1-A cells were plated at a concentration of 10⁵ cells/well in 6-well plates for 24 h at 37°C and subsequently treated with metformin (0, 1, 5 and 15 mM). Total RNA was extracted from the harvested EC cells according to the manufacturer’s instructions using TRI reagent (Sigma). The RNA concentration was determined by measuring the OD at 260 nm. First-strand complementary DNA (cDNA) was synthesized with a SuperScript II First-Strand Synthesis System for quantitative reverse transcription-polymerase chain reaction (qRT-PCR; Invitrogen). qPCR amplification was carried out using actin as an endogenous control. SYBR Green probes for each gene were used. The primers are listed in Table S1 . Real-time PCR was carried out with 50 ng of cDNA and SYBR PCR master mix (TaKaRa) in an Agilent Mx3000P Real-time PCR System using the two-step procedure (95°C 2 min, 1 cycle; 95°C 15 s, 60°C 1 min, 30 cycles). Relative quantitation of the expression of each single gene was performed using the comparative threshold cycle method.

The extracted genomic DNA was stored at −80°C. Global DNA hydroxymethylation (indicated by 5hmC) was assessed using a MethylFlash Global DNA Hydroxymethylation ELISA Easy kit (colorimetric) from EpiGentek according to the instructions provided by the manufacturer.

All assays were repeated independently a minimum of three times (n ≥ 3), and three wells per assay were used for each treatment in each cell line. The experimental data are expressed as the mean ± standard deviation (SD). One-way analysis of variance was used for statistical analyses and was performed using SPSS software (version 22.0). Data were compared between the two groups using a least significant difference test. The log-rank test was used to compare differences in the overall survival rate. The Cox proportional hazard regression model was used for multivariate analysis. Statistical significance is indicated by * for P < 0.05 and ** for P < 0.01.

The differential expression of TET2 between normal endometrial tissue and endometrial cancer tissue was observed using the endometrial cancer data in TCGA database. TET2 expression was significantly down-regulated in cancer samples (P = 3.250346E-05)( Figure 1A ). We also compared TET2 expression between different histopathological samples (Endometrioid endometrial adenocarcinoma, Serous endometrial adenocarcinoma, Mixed serous and endometrioid). The results showed that the expression of TET2 was also significantly different among different histopathological samples, and it was significantly over-expressed in Endometrioid endometrial adenocarcinoma (P< 0.05)( Figure 1B ). With the increase of stage, the expression of TET2 was continuously down-regulated, with significant differences in different stages(P=0.00495) ( Figure 1C ). Besides, TET2 was significantly down-regulated in the high-grade group(P=2.197738E-05) ( Figure 1D ). Comparison of TET2 expression between patients with and without recurrence showed that TET2 expression was significantly lower in patients with recurrence (P=0.0425) ( Figure 1E ). The correlation between TET2 expression and survival was observed by univariate Cox. The results showed that TET2 expression was a protective factor, but there was no significant correlation with overall survival (P =0.378; HR = 0.7257). Using the median expression value of TET2 as the dividing line, patients were divided into the high and low expression group (lower than the median is the green line, and higher than the median is the red line). There was no significant difference in survival between the two groups, but the high-expression group tended to have a better prognosis. (log-rank p = 0.206) ( Figure 1F ).

TET2 expression among different histopathological samples (Endometrioid endometrial adenocarcinoma, Serous endometrial adenocarcinoma, Mixed serous and endometrioid) from TCGA public database showed that the expression of TET2 was also significantly different among different histopathological samples, and it was significantly over-expressed in Endometrioid endometrial adenocarcinoma ( Figure 1B ). Endometrioid adenocarcinoma accounts for 80-90% of the pathological types of endometrial cancer. Thus, we examined the levels of TET2 and 5hmC in endometrioid endometrial adenocarcinoma tissues. The conversion of 5mC to 5hmC occurs through an oxidative reaction catalyzed by the TET protein family of dioxygenases. We analyzed TET2 expression in a series of 88 endometrial carcinoma samples and 20 normal proliferative endometrium samples via immunohistochemistry. TET2 expression was observed in the nucleus and cytoplasm of cells as yellow-brown granules. 5hmC-positive staining was indicated by a brownish yellow color in the nucleus. We found that TET2- and 5hmC-positive staining was present at a higher level in patients with a proliferative endometrium than in patients with EC ( Figure 2 ). These data suggest that TET2 may be associated with EC.

Correlations between TET2 and 5hmC expression and the clinicopathologic characteristics of endometrial carcinoma are shown in Table 1 . TET2 expression in endometrial adenocarcinoma was correlated with the degree of differentiation (P < 0.05). The positive TET2 expression rate in poorly differentiated tissues was lower than that in highly differentiated tissues (P < 0.05). The 5hmC level in endometrial adenocarcinoma was associated with clinical stage, differentiation, depth of myometrial invasion, and lymph node metastasis (P < 0.05). The positive rate of 5hmC staining decreased with tumor malignancy ( Table 1 ).

Among the 88 patients with endometrial adenocarcinoma, 58 (65.9%) survived, 12 (13.6%) died, and 18 (20.5%) were lost to follow-up. Surviving patients were followed up for 60 to 89 months, and the patients who died had survival times ranging from 6 to 45 months. The mean survival time of patients was 58.00 ± 18.11 months, and the median was 66 months. The mean and median survival times of patients with positive TET2 expression were 66.57 and 68.22 months, respectively, and those of patients with negative TET2 expression were 46.74 and 44 months, respectively. The mean and median survival times of patients with positive staining for 5hmC were 64.47 and 68.33 months, respectively, and those for patients with negative staining for 5hmC were 48.09 and 46.50 months, respectively. The mean survival time of patients with positive TET2 and 5hmC staining was 63.46 months, and the median survival time was 67.50 months. The mean survival time of patients with negative staining TET2 and 5hmC was 41.35 months, and the median survival time was 36.50 months. The mean survival time of patients with negative TET2 and 5hmC staining was significantly shorter than that of patients with positive staining (P < 0.01). The 5-year survival rates of TET2-positive and TET2-negative patients were 97.7% and 57.7%, respectively, and those of 5hmC-positive and TET2-negative patients were 93.5% and 62.5%, respectively ( Figure 3 ).

Multivariate Cox regression analysis revealed that stage, lymph node metastasis, and TET2 expression were associated with prognosis in patients with endometrial adenocarcinoma (P < 0.05). After adjusting for possible confounders, stage, lymph node metastasis, and TET2 expression may be independent prognostic factors in patients with endometrial adenocarcinoma [hazard ratio (HR) = 13.553, 95% confidence interval (CI): 1.509–121.677, P = 0.020; HR = 15.359, 95% CI: 1.284–183.783, P = 0.031; HR = 14.520, 95% CI: 1.060–198.843, P = 0.045] ( Table 2 ).

To explore the role of TET2 in EC cell lines, we knocked down TET2 using siRNA. The TET2 knockdown efficiency of the gene-specific siRNA was confirmed using real-time PCR and western blotting ( Figure 4A ). Knockdown of TET2 gene expression with siRNA significantly increased the proliferation rate of Ishikawa (P<0.01) and HEC-1-A (P<0.01) cells compared with that of cells transfected with nontargeting siRNA ( Figure 4B ). In addition, we further examined the effect of TET2 on metformin-mediated inhibition of EC cell proliferation. TET2 knockdown cells were treated with 5 mM metformin for 72h, and the results showed that knockdown of TET2 attenuated the inhibitory effect of metformin on EC cell proliferation ( Figure 4C ).

To examine the potential regulation of the expression and activation of TET2 and its substrates by metformin in endometrial carcinoma, two types of EC cells were treated with metformin at different concentrations for 24 h. In our study, western blotting and real-time PCR results showed that metformin treatment resulted in a potent increase in TET2 protein and mRNA expression, which occurred in a dose-dependent manner ( Figure 5A ). Metformin also increased the 5hmC level in a dose-dependent manner ( Figure 5B ).

Metformin is known as a traditional AMPK activator. Therefore, we further tested whether AMPK is involved in the regulation of TET2 by metformin. We knocked down AMPK using siRNA. The AMPK knockdown efficiency of the gene-specific siRNA was confirmed using real-time PCR and western blotting ( Figure 6A ). At 24 h after siRNA knockdown of AMPK gene expression, TET2 expression and 5hmC level were significantly reduced ( Figure 6B ). Western blotting and real-time PCR results showed that metformin did not significantly increase TET2 or 5hmC expression in these cells after 72 hours of metformin treatment ( Figure 6C ). Thus, we speculated that metformin could regulate TET2 expression through the AMPK pathway.