A total of 60 cases (age range 38–79 years, median age 59 years) of patients with endometrial adenocarcinoma, who had undergone total hysterectomy or pelvic and para-aortic lymphadenectomy simultaneously, were recruited for the present study. Cancer tissue samples were collected during the operation. A total of 40 normal endometrial specimens (age range, 38–77 years; median age, 58 years), from patients with abnormal uterine bleeding who had undergone an endometrium biopsy where the pathology results found proliferation or secretion, were collected. All tissue specimen were collected at Yantai Affiliated Hospital of Binzhou Medical University between January 2007 and December 2008. The tumors of all patients with endometrial adenocarcinoma were staged according to the 2009 International Federation of Gynecology and Obstetrics (FIGO) staging system (24). All of the specimens were fixed in 10% buffered formalin solution at room temperature for 48 h, embedded in paraffin and consecutive 4-µm-thick sections were cut. None of the patients received preoperative radiotherapy, chemotherapy, hormone therapy or treatment with other medications, and had no previous history of other types of cancer. Regular follow-ups began at the day of surgery and ended after 120 months, or upon death. The present study was approved by the Ethics Committee of the Yantai Affiliated Hospital of Binzhou Medical University (Approval no. 2018-016) and the study adhered to the principles of the Declaration of Helsinki (25). Oral informed consent was obtained from each patient prior to collection of tissues.

Tissue sections were stained as described previously (15). Briefly, following deparaffinization in dimethylbenzene and rehydration in a series of decreasing concentrations of alcohol (100% for 5 min, 95% for 2 min, 80% for 2 min and 75% for 2 min) the sections were heated in an antigen retrieval buffer pH 6.0 (EDTA, 1:300 dilution; cat. no. C1034; Beijing Solarbio Science & Technology Co., Ltd.) at 120°C for 5 min. After endogenous peroxidases were quenched with 3% hydrogen peroxide for 30 min, the sections were incubated at 4°C for 24 h with a primary antibody against BMI-1 (1:200; cat. no. ab126738; rabbit anti-human monoclonal antibody; Abcam). The samples were subsequently incubated with biotinylated secondary goat-anti-rabbit antibodies (1:5,000; cat. no. A0208; Beyotime Institute of Biotechnology) and horseradish peroxidase labelled avidin, and the staining was developed using DAB (OriGene Technologies, Inc.).

BMI-1 staining was analyzed by two investigators who were blinded to the clinical and prognostic data. The proportion of stained cells were scored as follows: 0, <5% of cells stained; 1, 5–25% of cells stained; 2, 26–50% of cells stained; 3, 51–75% of cells stained; or 4 >75% of cells stained. The staining intensity was scored as follows: 0, no staining; 1, weak staining; 2, medium staining; or 3, strong staining. The overall staining score was calculated as follows: Staining intensity × proportion of stained cells. Representative examples of the tissues with different staining intensities are presented as follows: -, 0–2 points; +, 3–4 points; ++, 5–8 points; and +++, 9–12 points (Fig. 1).

The human endometrial cancer cell lines, Ishikawa cells were obtained from The Cell Bank of Type Culture Collection of the Chinese Academy of Sciences; cat. no. EB081). JEC cells were purchased from Shanghai Fusheng Industrial Co., Ltd.; cat. no. FS-0129). Both cell lines were validated using short tandem repeat DNA profiling. Both cell lines were cultured in RPMI 1640 medium (HyClone; GE Healthcare Life Sciences) with 10% FBS (Gibco; Thermo Fisher Scientific, Inc.), 100 µg/ml penicillin, and 100 mg/ml streptomycin and cultured at 37°C in 5% CO₂ and 95% atmospheric air.

siRNA targeting the BMI-1 gene and scrambled siRNA control were purchased from Shanghai GenePharma Co., Ltd. The si-BMI-1 sequences were as follows: Sense, 5′-CCAGAUUGAUGUCAUGUAUTT-3′ and antisense, 5′-AUACAUGACAUCAAUC-UGGTT-3′; and scramble siRNA sense 5′-UUCUCCGAACGUGCACGUTT-3′, and antisense, 5′-ACGUGACAGGUUCGGAGAATT-3′. Cells were plated into 6-well plates overnight, to reach 50–60% confluency, and transfected with the BMI-1 siRNA/NC using Lipofectamine^® 2000 reagent (Invitrogen; Thermo Fisher Scientific, Inc.), according to the manufacturer's protocol. After 6 h of incubation with the transfection reagent and DNA, the cells were incubated in fresh supplemented medium for a further 24 or 48 h.

Total RNA was extracted using TRIzol^® (Invitrogen; Thermo Fisher Scientific, Inc.), treated with DNase I (Takara Bio, Inc.) to eliminate contaminating genomic DNA, and then reverse transcribed using a PrimeScript™ RT reagent kit (Perfect Real Time kit; Dalian Meilun Biology Technology Co., Ltd.). RT-qPCR was performed in a reaction volume of 20 µl containing SYBR green PCR mix, according to the manufacturer's protocol. Each sample was run in triplicate. GAPDH was used as the internal control, to which all samples were normalized. Results were calculated using the 2^−ΔΔCq method (26). The primer sequences for BMI-1 and GAPDH used were: BMI-1 forward, TCATGGTCATCCTTCTGCTGATGCTG and reverse, GCATGAGCATCACAGTCATTGCTGCT; and GAPDH forward, CATATGCAAGGTCATCCATGACAACTTTG and reverse, AAGCTTGTCCACCACCCTGTTGCTGTAG.

A total of 2×10³ cells were plated per well in 200 ml medium in a 96-well plate, with six wells per a group, and the cells were transfected with si-BMI-1 or negative control-siRNA as described above. After 24, 48, 72 or 96 h, 20 µl 1 mg/ml MTT solution (Sigma-Aldrich; Merck KGaA) was added to each well, and cells were incubated for a further 4 h at 37°C with 5% CO₂ and 95% air. Subsequently, the medium was removed, and the precipitated formazan was dissolved in 200 µl DMSO. After agitation for 15 min, the absorbance of the medium was measured at 495 nm using a microplate reader (Omega Bio-Tek, Inc.).

Briefly, 24-well transwell chambers, with 8 µm pores (Costar; Corning, Inc.) were used to assess the migratory and invasive properties. For the transwell migration assays, frozen Matrigel^® (Corning, Inc.) was dissolved at room temperature and was diluted to a working solution of 1:8 Matrigel: Serum-free medium. The upper chambers were coated with 100 µl Matrigel solution per well, and incubated for 24 h to allow the Matrigel to polymerize. For the invasion assays, the chambers were not coated prior to use. A total of 24 h after transfection, the cells were seeded in the upper chamber of the transwell inserts in 200 µl serum-free medium (1×10⁵ cells/ml). After 24 h of incubation, cells which had not migrated or invaded were removed. The cells on the underside of the chambers were fixed using methanol and 3.7%formaldehyde solution at room temperature, each for 5 min. Subsequently, the cells were stained with 0.1% crystal violet at room temperature for 30 min, imaged using an Olympus IX51 inverted microscope (Olympus Corporation) with a UIS2 optical system and phase contrast objectives, and the number of cells in five randomly chosen fields of view (magnification, ×200) were counted. A total of three independent experiments were performed for statistical analysis.

After transfection for 48 h, proteins were extracted, and western blotting was performed. Proteins were extracted using RIPA lysis buffer (Beyotime Institute of Biotechnology) with 1% PMSF (Thermo Fisher Scientific Inc.) and 1% NaF (Beyotime Institute of Biotechnology). Protein concentrations were determined using a bicinchoninic acid protein assay kit (Beyotime Institute of Biotechnology). A total of 40 mg of protein was loaded per lane on 10% SDS-gel, resolved by SDS-PAGE and subsequently transferred to a PVDF membrane. Membranes were blocked with 5% fat-free dry milk at room temperature for 2 h, followed by incubation with primary rabbit anti-BMI-1 monoclonal antibodies (1:1,000 dilution; cat. no. ab126738; Abcam), and rabbit monoclonal antibodies against E-cadherin, N-cadherin, vimentin, Keratin and slug (1:1,000 dilution; cat. nos. 3195, 4061, 5741s, 4546p and 58613, respectively; Cell Signaling Technology, Inc.) overnight at 4°C. After washing with TBS-Tween three times 5 min each, the membranes were incubated with secondary goat anti-rabbit IgG antibody conjugated to horseradish peroxidase (1:5,000; cat. nos. ab6721; Abcam) for 2 h at room temperature. GAPDH was used as the loading control with a rabbit anti-GAPDH antibody (1:1,000; cat. no. cst2118; Cell Signaling Technology, Inc.) overnight at 4°C. Densitometry analysis was performed using ImageJ software version 1.46 (National Institutes of Health). A total of three independent experiments were performed for statistical analysis.

Experiments were performed at least three times for statistical analysis. SPSS version 23.0 (IBM Corp.) and GraphPad version 6.0 (GraphPad Software, Inc.) were used for statistical analysis. Results are presented as the mean ± standard error of the mean. Comparison among groups was performed using a one-way ANOVA with a post hoc Tukey's test for multiple comparisons. Unpaired nominal-scale data was analyzed using a χ² test. Pearson's correlation coefficient tests were used to analyze the correlation between BMI-1 expression and the clinicopathological parameters. A log-rank test was used for Kaplan-Meier survival analysis. The significance of the clinicopathological characteristics on survival were analyzed using a Cox proportional hazards model in univariate and multivariate analysis. P<0.05 was considered to indicate a statistically significant difference.

IHC was performed to detect BMI-1 expression levels in endometrial adenocarcinoma and normal endometrial tissues. BMI-1 expression levels were significantly increased in endometrial adenocarcinoma tissues compared with the normal endometrial tissue (χ²=25.0, P<0.001). A total of 10% (4/40) of normal endometrial tissues exhibited high expression of BMI-1 (representative example in Fig. 1C and D), and the remaining tissues (90%, 36/40) had no or low expression levels of BMI-1 (representative example in Fig. 1A and B). However, 60% (36/60) of the cancer tissues exhibited high BMI-1 expression levels (representative example in Fig. 1G and H) and the remainder (40%, 24/60) exhibited low expression (representative example in Fig. 1E and F). The intensity of staining observed in endometrial adenocarcinoma tissue was higher compared with the normal endometrial tissue (Table II). Interestingly, BMI-1 expression was detected in both the nucleus and cytoplasm in normal and cancer cells (Fig. 1C, D, G and H). These results suggest that BMI-1 expression is increased in endometrial adenocarcinoma tissues compared with normal endometrial tissues (Table I).

The possible correlations between BMI-1 expression and clinicopathological characteristics were assessed. As shown in Tables II and III, BMI-1 expression levels were correlated with the FIGO stage, myometrial invasion and lymph node metastasis (P<0.05). However, there was no correlation between BMI-1 expression levels and age or grade.

As shown in Table IV, in the 120-month follow-up study, Kaplan-Meier analyses showed that 36 patients with high expression levels of BMI-1 had a survival rate of 52.8% and 24 patients with low expression levels of BMI-1 had a survival rate of 91.7%. There was a significant difference in survival rate in patients with endometrial adenocarcinoma with high and low expression levels of BMI-1 (P=0.040, HR=1.596). Kaplan-Meier analyses revealed that the overall survival time of patients with high expression levels of BMI-1 was significantly shorter compared with patients with low expression levels (Fig. 2).

Cox regression proportional hazard analyses were used to determine the risk factors associated with death. As shown in Table V, univariate Cox regression analyses revealed that the risk of death with high BMI-1 expression levels was significantly increased compared with low BMI-1 expression levels in patients with endometrial adenocarcinoma (P=0.040, HR=1.596). Additionally, late-stage (III, P=0.006, HR=1.67), myometrial invasion (P=0.006, HR=1.509) and lymph node metastasis (P=0.004, HR=1.703) were also significantly associated with less favorable prognosis. However, a high grade was not associated with increased risk of death (P=0.234). After adjustment for confounding factors, BMI-1 expression levels were still shown to predict a less favorable prognosis using multivariate Cox regression analysis (P=0.037, HR=1.698; Table VI). Furthermore, late-stage (P=0.017, HR=1.645), myometrial invasion (P=0.010, HR=1.305) and lymph node metastasis (P=0.016, HR=1.352; Table VI) were still shown to predict a less favorable prognosis.

To determine the potential functional roles of BMI-1 in the Ishikawa and JEC endometrial cancer cell lines, si-BMI-1 was transfected into cells to knockdown the expression levels of BMI-1, and RT-qPCR was used to assess the transfection efficiency. The results showed that the quantitative expression of BMI-1 mRNA decreased 37.45±2.7% in Ishikawa cells and 39.95±5.0% in JEC cells following si-BMI-1 transfection (Fig. 3).

The MTT assay results showed that cell growth and proliferation were decreased in the si-BMI-1 transfected Ishikawa and JEC cells compared with the respective controls (Fig. 4).

After the cells were transfected with si-BMI-1, the effects of BMI-1 on cell migration and invasion were assessed using transwell assays. The results showed that knockdown of BMI-1 expression significantly reduced the migratory and invasive capacities of Ishikawa (Fig. 5A and B) and JEC cells (Fig. 5C and D; P<0.001).

To further investigate the functional roles of BMI-1 expression on the migratory and invasive capacities of Ishikawa and JEC cells, the expression of EMT-associated proteins was assessed in cells transfected with si-BMI-1. Western blotting showed that the expression levels of the mesenchymal markers N-cadherin, vimentin and the downstream EMT transcription factor, Slug, were decreased in the si-BMI-1 transfected cells, whereas the expression levels of the epithelial markers E-cadherin and keratin were increased relative to the respective control cells (Fig. 6).