shRNA targeting Bmi-1 sensitizes CD44+ nasopharyngeal cancer stem-like cells to radiotherapy

Introduction

Nasopharyngeal cancer (NPC) is currently the most lethal head and neck neoplasm in Southeast Asia, especially in the Cantonese region, and radiotherapy is a cornerstone in the treatment of this malignancy. However, the efficacy of radiotherapy, particularly for advanced patients, is limited due to the development of radioresistance. Evidence has suggested that cancer stem cells (CSCs) are involved in resistance to various forms of therapies, including radiotherapy, and represent potentially useful pharmacologic targets (1–3). However, the interaction between radioresistance and CSCs and its underlying mechanisms have not been previously explored. B-cell-specific Moloney murine leukemia virus integration site 1 (Bmi-1), a member of the Polycomb family of transcriptional repressors, is essential for maintaining the self-renewal and differentiation of human stem cells (4–7). Bmi-1 has been demonstrated to be overexpressed in various of tumors (8), such as lung (9), breast (10) and prostate cancer (11), esophageal carcinoma (12) and NPC (13). Bmi-1 inhibition has been shown to sensitize those tumor cells to radiation (10,14,15). However, whether Bmi-1 inhibition can potentiate the cytotoxic effects of radiation on nasopharyngeal CSCs remains to be elucidated.

Our previous study identified that the CD44+ NPC cells, derived from the human NPC cell line SUNE-1 5–8F, with high capacity of self-renewal, differentiation abilities, tumorigenesis and radiochemoresistance, may be assumed to be one of the markers of nasopharyngeal carcinoma cancer stem cell-like cells (CSC-LCs) and Bmi-1 is overexpressed in CD44+ NPC (16). To further explore the functional role of Bmi-1 in NPC-CD44+ cells, we used a lentiviral vector expressing shRNA to knock down Bmi-1 expression (sh-Bmi-1) in NPC-CD44+ cells and evaluated the effects of Bmi-1 inhibition by shRNA on the sensitivity of NPC-CD44+ cells to radiation, DNA damage and repair, cell cycle distribution and apoptosis, and Bmi-1 downstream related protein for the purpose of improving the radiosensitivity of nasopharyngeal cancer stem-like cells, which may provide a broad potential therapeutic paradigm against NPC.

Materials and methods

Cell culture and reagents

The human NPC cell line SUNE-1 5–8F was purchased from Xiang-Ya Central Experiment Laboratory, and was maintained in RPMI-1640 medium supplemented with 10% fetal bovine serum in a humidified incubator (37°C, 5% carbon dioxide). Purified CD44+ nasopharyngeal cancer stem-like cells were cultured in DMEM/F12 medium supplemented with 20 ng/ml EGF, 10 ng/ml bFGF, 5 μg/ml insulin and 0.4% bovine serum albumin (BSA).

Flow cytometric analysis and cell sorting (FACS)

The cell sorting procedures were the same as previously described (16). After sorting, an aliquot of sorted cells was always reanalyzed to check for purity, which was generally >95%.

Generation of stable Bmi-1 knockdown (KD) cell lines

Sequence-specific oligonucleotide stretch shRNA designed to target the Bmi-1 (gene sequence no. NM_005180) were synthesized as follows: Bmi1-RNAi (19426-1), 5′-TAATACTT TCCAGATTGAT-3′; Bmi1-RNAi (19427-1), 5′-GAAAGTAA ACAAAGACAAA-3′; Bmi1-RNAi (19428-1), 5′-AGAACAG ATTGGATCGGAA-3′; loop sequence CTCGAG. Retroviral vector Bmi-1 short hairpin RNA (shRNA) was constructed by GeneChem (Shanghai, China). Bmi-1 gene was introduced into CD44+ NPC cells by infecting cells with a retroviral vector LV-BMI1-RNAi (19428-1) KD. Negative control cells (NC) were infected with the empty retroviral vector GV118-U6-MCS-Ubi-EGFP (GeneChem). The lentivirus stock [LV-BMI1-RNAi (19428-1)] was added to CD44+ cells at a multiplicity of infection (MOI) of 30. The cells without transfection were used as a blank control (CON). Transfection rates were monitored with fluorescence microscopy. The silencing of Bmi-1 was confirmed by western blotting and real-time PCR analysis.

Real-time PCR

Total RNA was isolated using TRIzol reagent (Invitrogen). Reverse transcription (RT) was performed using 1 μg of total RNA and the MML-V reverse transcriptase (Invitrogen). Real-time PCR was performed using the Platinum SYBR-Green Super Mix (Invitrogen) and a real-time PCR apparatus (ABI Prism 7000). Primer sets used were: Bmi-1 (F), 5′-CCACCTGATGTGTGTGCTTTG-3′ and (R), 5′-TTCAGTAGTGGTCTGGTCTTGT-3′; p16 (F), 5′-GGAGT TTTCAGAAGGGGTTTGT-3′ and (R), 5′-CCTCATTCCTCT TCCTTGGTTT-3′; p14 (F), 5′-TTCTGCCTTTTCACTGTGT TGGA-3′ and (R), 5′-CTCAAGAGAAGCCAGTAACCC-3′; GAPDH (F), 5′-TGACTTCAACAGCGACACCCA-3′ and (R), 5′-CACCCTGTTGCTGTAGCCAAA-3′; β-actin (F), 5′-GTCC ACCGCAAATGCTTCTA-3′ and (R), 5′-TGCTGTCACCTTC ACCGTTC-3′. GAPDH and β-actin were used as internal standard for data calibration, which was for Bmi-1, and p16 and p14, respectively. The 2ΔΔCt formula was used for the calculation of differential gene expression.

Western blot analysis

Cells were lysed in lysis buffer [25 mM Tris-HCl pH 7.5, 2.5 mM EDTA, 137 mM NaCl, 2.7 mM KCl, 1% sodium deoxycholic acid, 0.1% SDS, 1% Triton X-100, 2 mM PMSF and protease inhibitor cocktail (Nacalai Tesque)] for 30 min at 4°C. Lysates were clarified by centrifugation at 12,000 rpm for 15 min at 4°C, and the supernatants were collected. Protein concentrations were measured using a DC Protein Assay kit (Bio-Rad Laboratories, Hercules, CA, USA). An equal amount of the protein from each sample was separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to a polyvinylidene difluoride (PVDF) membrane (Immobilon; Millipore). The membranes were developed using Immobilon Western Detection Reagents (Millipore) according to the manufacturer’s instructions. The chemiluminescence of the membrane was detected using VersaDoc (Bio-Rad). Densitometric analyses of the band intensities were performed using ImageJ software (version 1.38x; National Institutes of Health). The following antibodies were used: Bmi-1 (1:1,000; rabbit anti-human); p16 (1:200; rabbit anti-human) (both from Epitomics); p14 (1:200; rabbit anti-human) and p53 (1:1,000; rabbit anti-human) (both from Cell Signaling Technology). β-actin protein levels were used as a control to verify equal protein loading.

Immunofluorescence staining

Cells were seeded on coverslips in 24-well plates and allowed to grow overnight. Cells were irradiated with a single dose of 2 Gy. Then, cells were washed and fixed with 4% paraformaldehyde at specific measuring times (0.5, 6, 12 and 24 h) and cells were rinsed in phosphate-buffered saline (PBS). After blocking in 5% BSA at room temperature for 30 min, slides were incubated with γH2AX antibody, mouse anti-human (Abcam) at 4°C overnight and then incubated with goat anti-mouse IgG-conjugated with PE (Abbkine). Images were captured by Olympus laser scanning confocal microscopy (Olympus Optical Co., Tokyo, Honshu, Japan). Cells were judged as ‘positive’ for γH2AX foci when they displayed 10 or more discrete dots of brightness. For quantitation of foci, a minimum of 100 cells were analyzed for each time point. All data points represent means ± SD of three experiments.

Colony formation assay

The exponential growth cells were plated on six-well plates with 200–6,000 cells/well, irradiated the next day at the distinct doses (0, 2, 4, 6 and 8 Gy). The cells were maintained in culture for an additional 14 days to allow colony formation, then fixed with methanol, stained with 0.5% crystal violet (Sigma) and colonies containing at least 50 cells in size were counted. Each treatment was carried out in triplicate. Cell survival fraction (SF), radiation dose (D), the bottom of the natural logarithm (e), the mean death dose (D0)and extrapolate number (N) were used for cell survival curves. Finally, the sensitization enhancing ratio (SER) was calculated (a ratio of SF2).

Cell cycle analysis

The cells were harvested and fixed in 70% ethanol at 4°C overnight. Next day, cells were suspended in 200 μl PBS and 2 μl RNase A (5 mg/ml). After 30 min incubation, 10 μl PI (1 mg/ml) was added and the cells were incubated for 30 min in the dark for analysis. The flow cytometry data was acquired using FACSCanto II flow cytometer (BD Biosciences, San Jose, CA, USA) and the results were analyzed by ModFit LT2.0 software (Coulter Electronics).

Cell apoptosis analysis

The cells were exposed to 6 Gy X-rays, 24 h after IR and cells were treated according to the Annexin V-PE/7-AAD apoptosis detection kit (KeyGen, Nanjing, Jiangsu, China). Cells were harvested and resuspended in 50 μl of binding buffer and 7-AAD (5 μl) in the dark for 10 min, at room temperature. Then, binding buffer (450 μl) and 1 μl Annexin V-PE were added in the dark for 10 min, at room temperature, and cells were analyzed by flow cytometry. PE-positive cells were regarded as apoptotic cells. This experiment was repeated three times.

Statistical analysis

Results are expressed as the means ± SD. Analysis was performed using a two-way Student’s t-test. P<0.05 was considered to indicate a statistically significant difference.

Results

FACS detects CD44 expression in the SUNE-1 5–8F cells, and the sorting purity

We examined CD44 expression in the human NPC SUNE-1 5–8F cell line by flow cytometry. As shown in Fig. 1, in the NPC cell line SUNE-1 5–8F, CD44+ cells occupied ~49.3% of the total cells, which was in agreement with our previous study (43–52%) (16). CD44+ cells were collected for subsequent experiments. The purity of these cells was 99.6%.

Figure 1 Cell sorting results and sorting purity. (A) In the NPC cell line, SUNE-1 5–8F, CD44+ cells occupied ~49.3% of the total cells. (B) The sorting purity of the freshly sorted CD44+ cells was 99.6%. NPC, nasopharyngeal cancer.

Knockdown of Bmi-1 by lentivirus-mediated RNA interference in CD44+ NPC cells

Previous studies in other systems suggested that Bmi-1 expression was associated with significantly reduced patient survival (17–19) after RT with or without chemotherapy, indicating that Bmi-1 overexpression may be mediating radioresistance. To address this possibility, CD44+ NPC-CLCs cells were transfected with lentivirus-mediated RNA targeting Bmi-1 [LV-BMI1-RNAi (19428-1), KD]. Control cells were infected with the empty retroviral vector GV118-U6-MCS-Ubi-EGFP, NC. To ensure the transfection efficiency of lentivirus into CD44+ NPC CLCs cells, GFP expression was monitored by fluorescence microscopy and FACS analysis. It showed a high-efficiency infection that the majority of cells displayed green fluorescence 72 h after lentivirus transfection, with the efficiency reaching >90% in both KD and NC-treated cells (Fig. 2A). LV-Bmi-1-RNAi showed 90% reduction in Bmi-1 transcript level, observed at 72 h post-transfection by real-time PCR (Fig. 2B), which resulted in nearly undetectable Bmi-1 protein expression (Fig. 2C).

Figure 2 Lentivirus-mediated RNA targeting Bmi-1 (KD) and empty retroviral vector (NC) were constructed and transfected into CD44+ NPC cells. Untreated CD44+ NPC cells (CON) served as a control. (A) The infection efficiency was determined by fluorescent microscopy. (B) Real-time PCR verified the efficiency of silencing of Bmi-1 at 72 h after transfection and Bmi-1 gene downstream gene of p16 and p14 relative mRNA levels. (C) The expression of Bmi-1 protein was detected by western blotting 72 h after transfection. β-actin was a loading control. KD, knockdown; NC, negative control; NPC, nasopharyngeal cancer; CON, control.

Real-time PCR and western blot detection of the Bmi-1 downstream related genes p16, p14 and p53 mRNA and protein changes

Emerging evidence indicates that Bmi-1 is upregulated in various malignancies and promotes tumor progression by inhibiting the transcription of tumor suppressors, such as p53, p16INK4a and p14Arf (15,20–23). As shown in Fig. 3, knockdown of Bmi-1 increased the downstream related genes p16, p14 and p53 mRNA and protein levels.

Figure 3 (A) Real-time PCR quantification of p16 and p14 mRNA expression levels in each groups. (B) Western blot analysis of Bmi-1 gene downstream related gene of p16, p14 and p53 proteins changes at 24 h following treatment with lentivirus (72 h) and subsequent 6 Gy irradiation. β-actin was a loading control. Data are displayed as one representative out of three independent experiments. Lane 1 and 4, CON; Lane 2 and 5, NC; Lane 3 and 6, KD. CON, control; NC, negative control; KD, knockdown.

Bmi-1 knockdown sensitizes CD44+ NPC CLCs to radiation

To determine the radiosensitizing effect of Bmi-1 depletion on the CD44+ NPC CLCs, a clonogenic formation assay was performed. As depicted in Fig. 4, the shoulder area of the survival curves was significantly narrowed and the surviving fractions (SFs) at each dose (2, 4, 6 and 8 Gy) decreased in Bmi-1-depleted cells. The values of SF2, D0, Dq and N were all lower in Bmi-1 KD cells (Table I). SF2 was reduced to 44% in KD cells from 78% in CON cells and the enhancement ratio was 1.77. Thus, we concluded that downregulation of Bmi-1 could radiosensitize CD44+ NPC CLCs cells.

Figure 4 Downregulation of Bmi-1 increases the sensitivity of CD44+ NPC cells to irradiation. P<0.001; ANOVA test. NPC, nasopharyngeal cancer.

Table I The main parameters of cell survival curves after ionizing radiation.a

Table I

The main parameters of cell survival curves after ionizing radiation.a

Parameters	CON	NC	KD
D0	2.89	2.83	2.14
Dq	3.36	2.42	0.34
N	2.16	1.85	1.16
SF2	0.78	0.71	0.44
SER	1.10	1.77

a Survival curves with surviving fractions (SFs) normalized to the plating efficiency were fitted according to multi-target click model: SF = 1 - (1 − e^−D/D0) N.

{ label (or @symbol) needed for fn[@id='tfn2-or-32-02-0764'] } D₀, mean lethal dose; D_q, quasi-threshold dose; N, extrapolation number; SF₂, surviving fraction at 2 Gy; SER, radiation sensitizing enhancement ratio. SER=1 suggests an additive radiation effect and SER >1, a supra-additive effect as against a sub-additive effect in the case of SER <1. CON, control; NC, negative control; KD, knockdown.

Knockdown of Bmi-1 increases DNA double strand break (DSB) and decreases DSB repair

γH2AX is commonly utilized to assess DSBs inflicted by RT (24). When CD44+ NPC were depleted of Bmi-1, we monitored kinetics of DSB repair by immunofluorescent γH2AX foci at different time points after exposure to 2 Gy of X-rays. As shown in Fig. 5, γH2AX foci appeared immediately following IR treatment in the three groups, indicating that Bmi-1 is dispensable for γH2AX focus formation. The average number of γH2AX foci/cell in KD cells was significantly higher than in cells treated with NC or untreated/control cells at 0.5, 6, 12 and 24 h after radiation (P<0.05). This delayed clearance of γH2AX foci in Bmi-1-depleted cells demonstrated that DSB repair is severely impaired in the absence of Bmi-1 (25).

Figure 5 Bmi-1 knockout delays DNA damage repair in CD44+ NPC cells.(A) Graph showing γH2AX foci in CON, NC and Bmi-1-depleted KD cells at different points after 2 Gy radiation. Original magnification, ×400. (B) γH2AX foci positive cells at different time points after 2 Gy irradiation of the three types of cells (CON, NC and KD). *P<0.001; t-test. Means ± SD of three experiments are reported. NPC, nasopharyngeal cancer; CON, control; NC, negative control; KD, knockdown.

Effects of Bmi-1 downregulation on IR-induced cell cycle distribution and apoptosis

To further evaluate potent reasons conferring radiation sensitivity induced by Bmi-1 knockdown, we tested cell cycle distribution and cell apoptosis following treatment of lentivirus and/or IR by flow cytometry assays. As shown in Fig. 6A, compared with the blank CON group, cells treated with LV-BMI1-RNAi showed prolonged G1 from 64.40 to 73.21%. G2/M phase accumulation came to max within 24 h in the KD group after IR; however, CON and NC G2/M phase increased by 48 h after irradiation, the number of cells in G2/M in KD had to some extent decreased at 48 h after IR. The levels were still significantly greater than that for NC-treated or control cells. Depletion of Bmi-1 exhibited a markedly prolonged accumulation of cells in G2/M. In addition, there was an increase in IR-induced apoptosis in the absence of Bmi-1 (Fig. 6B). Based on these results, we concluded that knockdown of Bmi-1 leads to an extended G2/M accumulation and an increase of IR-induced apoptosis.

Figure 6 CD44+ NPC cells were treated with the indicated lentivirus for 72 h, irradiated (6 Gy). Cell cycle distribution and apoptosis were determined by flow cytometry assays. (A) Bmi-1 depletion increased G1 population and cells arrested in G2/M after 6 Gy irradiation. (B) Apoptotic rate, calculated as percent number of Annexin PE-positive cells. Annexin PE-positive cells were regarded as apoptotic cells. These experiments were repeated three times and similar results were obtained. (C) Representative scatter grams from flow cytometry profile represent Annexin V-PE staining in the x-axis and 7-AAD in the y-axis. *P<0.001. Columns, mean; error bars, SEM, from three independent experiments. NPC, nasopharyngeal cancer.

Discussion

Cancer stem cell theory is based on the hypothesis that cancers arise from a rare population of cells, which are responsible for the chemo- and radioresistance of the tumors. Therefore, the cancer stem cell population is thought to be an important determinant of treatment failure. Polycomb group proteins (PcG) are a family of proteins that is essential for proper embryonic development and for the specification of stem cell gene expression profiles, controlling self-renewal, pluripotency and differentiation (25), and PcG may confer radioresistance to stem cells (25). Bmi-1, a member of the Polycomb family of transcriptional repressors, is essential for maintaining the self-renewal abilities of adult stem cells. Facchino et al observed that Bmi-1 was responsible for the radioresistance in both glioma stem cells and normal stem cells (26). Our results demonstrate that the isolated CD44+ subpopulation from SUNE-1 5–8F human NPC cell line presents the key biological properties of CSCs and Bmi-1 was overexpressed in our previous study (16).

In the present study, we provided evidence that Bmi-1 was correlated with radiation response in CD44+ NPC CSC-LCs. We showed that Bmi-1 overexpression developed radiation resistance, whereas Bmi-1 inhibition significantly increased radiation sensitivity with a DER of 1.77 by radiation clonogenic survival assay. It suggested that Bmi-1 may play an important role in the regulation of cellular response to radiation in CD44+ NPC CSC-LCs. Thus, the results underscore the importance of Bmi-1 targeting in combination with irradiation in nasopharyngeal neoplasm therapy. This finding is consistent with the reports of Wang et al (13) and Alajez et al (15).

Bmi-1 is a transcriptional repressor of the Ink4a/Arf locus. This locus encodes 2 separate tumor suppressor genes, namely, p16Ink4a and p19Arf (p14Arf in humans) (27). The p16INK4a (pRb/p16INK4a/cyclin D1) and p53 (p14ARF/mdm2/p53) pathways are the two main cell cycle control pathways. p16INK4a is a potent inhibitor of the kinase activities of CDK4/6, resulting in Rb hypophosphorylation, binding to E2F1, the initiation of cell cycle arrest in the G1 phase (28). In addition, cell cycle arrest and apoptosis are promoted by p14 through its regulation of p53 stability stages, and p14 effectively prevents degradation of the tumor suppressor protein p53, which is required for cell cycle arrest. Hence, Bmi-1 knockdown inhibited cell cycle progression through derepression of the p16INK4a/p14ARF locus.

It is well known that DSBs are suggestive of critical lesions in DNA caused by ionizing radiation (29). DSB is the main mechanism of tumor cell death after irradiation (30), DNA DSB repair shows strong cell cycle dependency and radiation therapy may mediate cell cycle redistribution of tumor cells. Our findings showed that knockdown of Bmi-1 prolonged G1, after 6 Gy irritation, and the proportion of the cells in G2/M phase was increased, which was the most radiosensitive phase of the cell cycle. In addition, Gilbert et al showed that low-expression of p16 has been linked to decreased chemoradiotherapy responsiveness, p16Ink4a negative patients had significantly poorer overall survival (31). The present study showed that p16 increased after Bmi-1 inhibition and this may be one of the reasons resulting in increased DSBs and decreased DSB repair in Bmi-1 knockdown cells. Bmi-1 inhibition markedly increased DSB and significantly decreased the rate of DNA DSB repair induced by irradiation, suggesting a participation of Bmi-1 in DNA strand damage and repair (26,32). If complete DNA damage repair fails, apoptosis is triggered for the elimination of damaged cells. We noted that the number of apoptotic cells in Bmi-1 knockdown cells was apparently much higher compared with the controls using FACS analysis (Fig. 6B). The mechanism of Bmi-1 inhibition combined radiation induced apoptosis is complex. A possible reason is that the P53 pathway for apoptosis was activated in response to combined Bmi-1 inhibition with ionizing radiation (15).

In conclusion, we reported the enhanced radiosensitivity of CD44+ NPC CSCs after Bmi-1 inhibition using shRNA. The increased sensitivity was associated with increased DSBs and decreased DSB repair. We also showed that combination Bmi-1 inhibition with irradiation could induce apoptosis, elevated protein p53, p16 and p14 expression. These results suggest that Bmi-1 knockdown combined radiotherapy could induce a synergistic effect on increasing the radiosensitivity of CD44+ CSC-LCs. Bmi-1 inhibition in NPC stem cells may be an effective therapeutic strategy against NPC. However, further study is required to confirm these results using in vivo xenograft models.

Acknowledgements

This study was supported by The Foundation of Health Department of Hubei Province, China (no. JX4B52). We thank Jing-Hua Ren, Wei-Hong Chen, Hong-Xia Zhou and other personnel in the Laboratory of Cancer Center, the Union Hospital, the Tongji Medical College, the Huazhong University of Science and Technology (Wuhan, China) for their technical assistance.

References