5-Azacytidine Enhances the Radiosensitivity of CNE2 and SUNE1 Cells In Vitro and In Vivo Possibly by Altering DNA Methylation

The radioresistance of tumor cells remains a major cause of treatment failure in nasopharyngeal carcinoma (NPC). Recently, several reports have highlighted the importance of epigenetic changes in radiation-induced responses. Here, we investigated whether the demethylating agent 5-azacytidine (5-azaC) enhances the radiosensitivity of NPC cells. The NPC cell lines CNE2 and SUNE1 were treated with 1 μmol/L 5-azaC for 24 h before irradiation (IR); clonogenic survival was then assessed. Tumor growth was investigated in a mouse xenograft model in vivo. The apoptosis, cell cycle progression and DNA damage repair were examined using flow cytometry, immunofluorescent staining and western blotting. Promoter methylation and the expression of four genes epigenetically silenced during the development of NPC were evaluated by pyrosequencing and real-time PCR. We found that pretreatment with 5-azaC significantly decreased clonogenic survival after IR compared to IR alone; the sensitivity-enhancement ratio of 5-azaC was 1.4 and 1.2 for CNE2 and SUNE1 cells, respectively. The combined administration of 5-azaC and IR significantly inhibited tumor growth in the mouse xenograft model, and enhanced radiation-induced apoptosis in vitro compared to 5-azaC alone or IR alone. 5-AzaC also decreased promoter methylation and upregulated the expression of genes which are epigenetically silenced both in vitro and in vivo in NPC. Thus, 5-azaC enhance the radiosensitivity of both the CNE2 and SUNE1 cell lines, possibly by altering DNA methylation levels and increasing the ability of irradiated cells to undergo apoptosis. The use of 5-azaC combined with IR maybe represent an attractive strategy for the treatment of NPC.


Introduction
Nasopharyngeal carcinoma (NPC) is prevalent in southeastern Asia, especially in southern China where the incidence is approximately 25-50 per 100,000 population per year [1]. Radiotherapy is the primary treatment modality for locally-and regionally-confined NPC. Despite recent significant advances in the treatment of NPC, local recurrence is frequently observed [2]. Radiation resistance is one of the major obstacles that leads to locoregional recurrence of NPC during treatment [3].Therefore, the identification of effective radiosensitizing agents to enhance the radiosensitivity of NPC cells may help to decrease both tumor recurrence and radiation-associated morbidity.
More recently, increasing evidence supports the suggestion that genome-wide changes in methylation levels are associated with the radiosensitivity of cancer cells [4,5,6]. Epigenetic modifications, specifically DNA hypermethylation that leads to the aberrant silencing of multiple tumor suppressor genes, are believed to play a pivotal role in variety of cellular events [7,8], including alterations in apoptosis, cell cycle progression, mitotic checkpoint regulation and DNA repair; all of these mechanisms have been considered to mediate radiosensitizing effects [4,9]. DNA hypermethylation has been frequently reported in NPC [10,11]. Aberrant promoter methylation of tumor suppressor genes, such as Ras association domain family member 1A (RASSF1A), Cyclin-dependent kinase inhibitor 2A (CDKN2A), Reprimo (RPRM) and Stratifin (14-3-3s), have been commonly detected in NPC [11,12,13]. These genes play important roles in cell cycle control, apoptosis, checkpoint activation, and translational regulation [14,15,16,17]. With the increasing awareness of epigenetic abnormalities in NPC, counteracting these changes using methyltransferase inhibitors such as 5-azacytidine (5-azaC) [18] may be a potential strategy of radiosensitizing NPC cells.
The relationship between DNA methylation changes and radiosensitivity in NPC, however, remain unknown. This study aimed to investigate the radiosensitizing effect and possible mechanism of the demethylating agent 5-azaC in NPC cells in vitro and in vivo. Our data demonstrated that 5-azaC enhances the level of cellular radiosensitivity and was associated with increased rates of apoptosis, potentially as a consequence of the altered methylation levels.

Methods and Materials
Cell culture and treatment The human NPC cell lines CNE2 and SUNE1 were obtained from the Cancer Center of Sun Yat-sen University. The cells were maintained in RPMI-1640 (Gibco, USA) containing 10% fetal bovine serum, 100 U/ml penicillin, and 100 mg/ml streptomycin at 37uC in 5% CO 2 . Cells were exposed to 5-azaC (Sigma, USA) and/or irradiated at 12.7 Gy/min at room temperature using a RS 2000 X-ray Biological Irradiator operated at 160 kV/50 mA (Rad Source Technologies, USA). The treatment groups were as follows: control group (PBS); 5-azaC group (0 to 5 mmol/l); ionizing radiation (IR) group (0 to 8 Gy irradiation); and combined treatment group (pretreatment with 5-azaC for 24 h followed by IR). The 5-azaC was dissolved in phosphate-buffered saline (PBS); control cells were treated with media containing the same volume of PBS.

Colony formation assay
Single cells were seeded and treated with 5-azaC (0 or 1 mmol/l) alone, IR (0, 2, 4, 6 or 8 Gy) alone, or 5-azaC and IR. After culture for 10-14 days, the cells were fixed in ice-cold methanol, stained with Giemsa solution and colonies containing .50 cells were counted. Cell survival curves were fitted using the linearquadratic (LQ) formula: surviving fraction (SF) = exp (-aD-bD 2 ) [19], wherein a and b are the radiobiological cell survival parameters within the treatment volume, and D is irradiation dose; the sensitizer enhancement ratio (SER) was determined from the survival curves of irradiation in the presence and absence of the tested compounds using the following the equation: SER = D 0 untreated cells/D 0 treated cells, where D 0 values represent the radiation dose that led to 37% cell survival [20].

In vivo tumor xenograft model
Twenty 4-week-old female BALB/c nu/nu nude mice were purchased from the Guangdong Experimental Animal Center. All protocols were approved by the Institutional Animal Care and Use Committee of Sun Yat-Sen University (IACUC SYSU, NO.10212100E). CNE2 cells (1610 6 ) were subcutaneously injected into the left hind flank region. After the xenograft tumors reached 0.5 cm diameter, the mice were randomly assigned to four groups (n = 5) and treated as follows: the control group received twice weekly intraperitoneal (i.p.) injections of 100 ml PBS; the 5-azaC group received twice weekly i.p. injections of 100 ml 5-azaC (4 mg/kg); the IR group received twice weekly i.p. injections of 100 ml PBS and xenograft irradiation (8 Gy, once) at second week; and the 5-azaC + IR group received twice weekly i.p. injection of 100 ml 5-azaC (4 mg/kg) and xenograft irradiation (8 Gy, once) at second week. Tumor sizes and body weights were measured weekly. Tumor volumes (TVs) were calculated using the formula TV = LD 2 /2 (where L was the longest diameter and D was the shortest diameter) [21].

Cell cycle and apoptosis analysis
Cells were seeded in 6-well plates, incubated for 24 h, and treated with PBS, 5-azaC (1 mmol/l), IR (6 Gy), or 5-azaC + IR as described above. Cell cycle progression and apoptosis were analyzed with the Cell Cycle and Apoptosis Kit (Keygentec, China) using a CytomicsTM FC500 flow cytometer and CXP analysis software (Beckman Coulter, USA) following the manufacturer's instructions. Cell cycle analysis was performed using CXP analysis software; apoptotic cells were considered to include cells stained Annexin V (+)/propidium iodide (PI) (2) (lower right quadrant, early apoptosis) and late cells stained Annexin V (+)/PI (+) (upper right quadrant, late apoptosis) [22].For each sample, at least 10,000 cells were analyzed.

Immunofluorescent staining
Immunofluorescent staining of phosphorylated histone H2AX (cH2AX) was performed as previously described [23]. Cells were cultured on a chamber slide and treated with IR (6 Gy), or 5-azaC + IR as previously described. Cells then were fixed in 4% paraformaldehyde, permeabilized in Triton X-100 for 15 min, incubated with anti-c-H2AX antibody (Cell Signaling Technology, USA) overnight at 4uC, then incubated with secondary antibodies (Cell Signaling Technology, USA) for 1 h at 37uC. Nuclei were counterstained using 49, 69-diamidino-2-phenylindole (DAPI). The number of cH2AX-positive foci was determined in at least 50 cells at 4006magnification using a fluorescent microscope (Olympus); cH2AX repair kinetics were determined at 0, 1, 6, 24 and 48 h after irradiation.

RNA isolation and quantitative real-time RT-PCR
Total RNA was isolated using TRIzol reagent (Life Technologies, USA); cDNA was synthesized from 1 mg total RNA using the RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific, USA). Real-time RT-PCR was performed on a Bio-Rad CFX96 sequence detection system (Bio-Rad Laboratories Inc., USA) using Platinum SYBR Green qPCR SuperMix-UDG reagents (Life Technologies, USA). All reactions were incubated at 95uC for 3 min, followed by 40 cycles at 95uC for 15 s, annealing at 60uC for 15 s, and elongation at 72uC for 7 min. Primer sequences were obtained from previously published data for RASSF1A, CDNK2A, RPRM, 14-3-3s, and glyceraldehyde-3phosphate dehydrogenase (GAPDH) [24,25,26,27,28]. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was amplified as an endogenous control, and relative expression levels were calculated using the 2 2DDCT method [29].

5-AzaC Enhances Radiosensitivity in NPC
PLOS ONE | www.plosone.org Genomic DNA was isolated from cells treated with 1 mmol/l 5-azaC for 3 days using the EZ1 DNA tissue Kit (Qiagen, Germany). Bisulfite modification was performed using the EpiTect Plus DNA Bisulfite Kit (Qiagen) with 1 mg genomic DNA. Methylation of RASSF1A, CDKN2A, RPRM, and 14-3-3s in the bisulfite-treated DNA samples was quantitatively analyzed by bisulfite pyrosequencing as previously described [30], Pyrosequencing was carried out with primers designed by the PyroMark Assay Design Software 2.0 (Qiagen). Primer sequences and PCR conditions for bisulfite pyrosequencing are outlined in the Table 1. The sequencing reaction and quantitation of methylation was carried out using a PyroMark Q24 instrumentand software (Qiagen). Percentage methylation was calculated by averaging across all CpG sites interrogated.

Statistical analysis
Each in vitro experiment was repeated independently three times. Data is presented as the mean 6 SD values. Statistical analysis was performed using SPSS version 13.0 (SPSS, Chicago, IL, USA). Differences between groups were compared with the Student's t-test and one-way ANOVA was used for multiple comparisons. Two-tailed p values,0.05 were considered significant.

Cytotoxicity of 5-azaC in NPC cells in vitro
To investigate the cytotoxic effects of 5-azaC in NPC cells, CNE2 and SUNE1 cells were cultured with 0, 50, 100, 500, 1000, 3000, or 5000 nmol/l of 5-azaC. Cell proliferation was measured by the MTT assay after 24, 48, and 72 h of exposure to 5-azaC. Compared with the control group, no significant differences were observed in the survival rates of CNE2 and SUNE1 cells treated with 50 nmol/L to 1 mmol/L 5-azaC for 72 h (p.0.05). However, at 72 h, 3 mmol/L 5-azaC resulted in the cytotoxicity in CNE2 and SUNE1 cells (survival rate 6063.86% and 6864.94%, respectively; p,0.01), and major growth suppression was observed at concentrations of 5 mmol/L (survival rate 416 4.24% and 4462.68%, respectively; p,0.01) or higher (Fig. 1). These results demonstrate that concentrations of # 1 mmol/L 5-azaC do not supress the survival or proliferation of NPC cells in vitro.

5-AzaC increases the radiosensitivity of NPC cells in vitro
Based on the results above, we treated cells with 1 mmol/L 5-azaC to evaluate whether this demethylating agent increased the radiosensitivity measured in clonogenic survival assays. The survival rates of CNE2 and SUNE1 cells in the 5-azaC + IR group decreased significantly as the dose of radiation increased, compared to the IR alone groups (p,0.05) (Fig. 2).
The a and b parameters obtained using the L-Q model are displayed in Table 2. An increase of a or b has been suggested to represent an enhanced contribution due to the interaction of potentially lethal damage (PLD) or sublethal damage (SLD) [31], respectively. Increasing a and/or decreasing b values indicate higher radiosensitivity. When the 5-azaC treated and untreated groups were compared, the a value increased (CNE2 0. Gy; a higher SF2 indicates increased radioresistance. The SF2 values of two cells in the IR only groups were significantly higher than combined groups (CNE2 44%60.02% vs. 31%60.02%, SUNE1 54%60.02% vs. 45%60.02%, respectively; both p,0.01). The sensitivity-enhancement ratio (SER) of SF2 in CNE2 and SUNE1 cells was 1.4 and 1.2, respectively. Clonogenic assays showed that the a values and SER increased, while SF2 values decreased in the group pretreated with 5-AzaC before irradiation. These results show that pretreatment with 5-azaC increases the radiosensitivity of CNE2 and SUNE1 cells.

5-AzaC enhances the radiosensitivity of NPC cells in vivo
To confirm that the enhancement of radiation sensitization observed in vitro could be translated into an in vivo tumor model, a tumor growth delay assay was performed in mice using CNE2 cells grown subcutaneously (sc).We first compared the tolerability by examining the relative body weights and tumor volumes of xenografts throughout the study (Fig. S1). When the dose of 5-azaC # 4 mg/kg was given twice weekly, body weights rose steadily (p.0.05; Fig. S1A-C) and tumor growth was not significantly suppressed (p.0.05; Fig. S1D, S1E) in each mouse group throughout the study.
Accordingly, we examined the radiosensitizing effect of the twice weekly administration of 4 mg/kg 5-azaC in nude mice bearing xenograft tumors. During the observation period,, no significant differences were observed in the average weight of the  These results indicate that 5-azaC enhances in vivo radiosensitivity in the CNE2 xenograft model.

5-AzaC enhances radiation-induced apoptosis, but does not affect cell cycle progression and DNA double-strand breaks (DSBs) repair
To assess the possible mechanism underlying the enhanced radiosensitivity of cells treated with 5-azaC, we evaluated the apoptosis, the cell cycle progression and the expression and activation of regulators for radiation-induced DSBs in CNE2 and SUNE1 cells. Flow cytometry analysis indicated that there was no significant difference in the rate of apoptosis (Annexin V (+)/PI (2) plus Annexin V (+)/PI (+)) between the control and 5-azaC groups of either CNE2 or SUNE1 cells that did not receive IR (both p.0.05; Fig. 4A). IR alone resulted in 10.83%61.92% of CNE2 cells and 8.97%61.10% of SUNE1 cells undergoing apoptosis  Flow cytometry analysis showed that, when treated with and without 5-azaC, there was no statistical difference in the percentage of cells in the G2/M phase (CNE2 15 (Fig. 4B).
Immunofluorescent staining indicated that the number of cH2AX foci per cell that received combined 5-azaC/IR treatment was not significantly different over time compared to the corresponding IR only groups (p.0.05; Fig. 5A, 5B), which indicated that DSBs repair was not hindered by treatment with 5-azaC. Subsequently, western blot analysis demonstrated that 5-azaC pretreatment did not inhibit the expression of the nonhomologous end-joining (NHEJ) regulators (DNA-PK, Phospho-DNA-PK (Ser2056), XRCC4, and XLF) and homologous recombination (HR) regulators (RPA2 and RAD51) over time in CNE2 and SUNE1 cells after IR (Fig. 5C). Taken together, these data indicate that 5-azaC enhances radiation-induced apoptosis; however, this was not associated with cell cycle progression and the modulation of the expression of DSBs repair regulators in CNE2 and SUNE1 cells.

5-AzaC induces demethylation and the re-expression of epigenetically silenced genes
5-AzaC changes genome-wide methylation status. To further characterize the relevance between gene methylation changes and radiosensitivity, we evaluated the differential methylation and expression status of RASSF1A, RPRM, CDKN2A, and 14-3-3s both in vitro and in vivo in NPC by pyrosequencing and real-time RT-PCR. Pyrosequencing revealed that RASSF1A and RPRM exhibited hypermethylation at the promoter region, whereas CDKN2A and 14-3-3s showed hypomethylation at the promoter in CNE2 and SUNE1 cells and in the CNE2 xenografts (Fig. 6A, 6B) (Fig. 6A, 6B). Subsequently, the detection of mRNA expression demonstrated that 5-azaC treatment resulted in an obvious increase in the expression of RASSF1A, RPRM, CDKN2A, and 14-3-3s in both CNE2 and SUNE1 cells and in the CNE2 xenografts. In particular, the increases in the expression of RPRM (3-fold upregulation, p,0.01; Fig. 6C) and CDKN2A (2fold upregulation, p,0.01; Fig. 6C) observed in CNE2 cells and the CNE2 xenografts were significant. Taken together, these results from both in vitro and in vivo experiments demonstrate that 5-azaC changes methylation levels and restores the expression of mRNA in key genes involved in DNA repair in NPC cell lines.
Discussion 5-AzaC, as a nucleotide analog, is known to inhibit DNA methyltransferases (DNMTs), which results in DNA hypomethylation and the re-expression of epigenetically silenced genes [18]. In the present study, we evaluated the optimal treatment schedule for the cytotoxicity of 5-azaC in vitro and in vivo and used minimally toxic drug concentrations combined with radiotherapy for further investigations. Our findings agree with previous studies that showed that concentrations of 1 mmol/L, as used in our current study, did not induce substantial cytotoxicity in colorectal carcinoma or bladder cancer cell lines [6,32]. Our in vitro experiments showed that the clonogenic ability of CNE2 and SUNE1 cells was suppressed by IR alone, and further suppressed by the combination of 5-azaC and IR. In addition, 5-azaC also potentiated X-ray antitumor activity in vivo. Our observations confirmed prior reports that this combination results in greater degrees of tumor regression in human cancer cell lines [4,5,6].
Based on the potent radiosensitizing activity, further investigations of the mechanism by which 5-azaC mediates its radiosensitizing effect were performed in NPC cells. In general, the ability of the cell to undergo apoptosis is one of the most important mechanisms of radiosensitivity [33]. In this study, nontoxic concentrations of 5-azaC were unable to trigger apoptosis in nonirradiated CNE2 or SUNE1 cells. However, a dramatic increase in apoptosis was induced by 5-azaC during irradiation; consequently, a strong, synergistic, cytotoxic effect is achieved. The findings were in line with those reported in other studies. For example, Hong et al. showed that the radiosensitivity of gastric cancer cells to a demethylating agent was dependent on an increasing rate of apoptosis [4]. Juergen et al. found that apoptosis was strongly increased after combined 4 Gy/5-azaC treatment was administered to head and neck squamous carcinoma cells [34]. These findings suggest that the switching on and off of apoptosis may be the critical event during radiosensitivization.
Cell cycle arrest is one of most common causes of a radiosensitizing effect [5]. Pretreatment with 5-azaC did not affect the cell cycle in both cell lines in our experiment. This result is consistent with findings from colorectal carcinoma, lung cancer, and glioblastoma cell lines [6,9]. However, breast carcinoma and hepatic cancer cell lines have been shown to undergo G2/M phase arrest after treatment with demethylating agents [35,36]. This discrepancy may be due to the use of different cell lines or differing concentrations of demethylating agents.
NHEJ and HR are the major repair pathways for DNA DSBs in mammalian species [37]. cH2AX is important for sensing ionizing radiation-induced DSBs [38]. In this study, the number of irradiated cells with cH2AX-positive foci and the expression levels of DNA repair-related proteins (NHEJ or HR) were not significantly inhibited by 5-azaC. Hak et al. [9] found an inverse association between 5-azaC treatment and DSBs repair in lung cancer or glioblastoma cells, whereas Harlinde et al. [39], in accordance with our current findings, found no such association in head and neck squamous carcinoma cells. This discrepancy may result from the use of different cells lines, differences in the demethylating agents used and/or radiation treatment schedules. 5-AzaC can induce genome-wide DNA hypomethylation and alter the expression of a variety of genes [18]. According to methylation status analysis, RASSF1A and RPRM exhibited promoter hypermethylation, which was consistent with the findings of previous studies [11,12]. In our experiment, the hypomethylated promoters were detected in CDKN2A and 14-3-3s, while the promoters were reported hypermethylated in other studies [11,13]. The difference in measured CpG sites may be responsible for the discrepancy. It was noteworthy that noncytotoxic concentrations of 5-azaC significantly reduced methyl-ation levels and upregulated the mRNA expression of RPRM. A recent report indicated that RPRM is a tumor suppressor, and the downregulation of the RPRM transcript is associated with promoter methylation. RPRM re-expression can activate the downstream effector caspase 3 in apoptotic pathways [16]. In our study, we also observed the re-expression of the tumor suppressor CDKN2A without significant changes in methylation. The possible reasons may be the upstream components that are epigenetically silenced in CDKN2A signaling network are activated after 5-azaC administration, and their re-expression leads to the upregulation of CDKN2A expression either directly or indirectly, a study has found [40]. It has recently been shown that CDKN2A is  not only a cell cycle checkpoint protein, but is also a master regulator of gene expression in the p16INK4a-cyclinD/cdk4-pRb-E2F1 regulating cell proliferation pathway [41,42]. E2F1 is an important downstream effector that mediates apoptosis in cancer cells [43]. CDKN2A re-expression upregulates E2F1 expression and, subsequently, the induction of apoptotic cell death [44]. Therefore, a possible explanation for radiation-induced apoptosis is that apoptotic pathways in cancer cells are activated by ionizing radiation, and the re-expression of apoptosis-associated genes after 5-azaC treatment, which were previously silenced by epigenetic inactivation, further increases the ability of tumor cells to undergo apoptosis.
In summary, our experiments indicate that 5-azaC potentiate antitumor radiotherapy activity and resulted in greater levels of tumor regression than radiotherapy alone both in vitro and in vivo in NPC cells. Possible mechanisms of radiosensitization may increase the ability of tumor cells to undergo apoptosis after changes in DNA methylation status, which enables the re-expression of tumor suppressor genes. Our findings suggest that the clinical use of epigenetic modulators may be a promising approach to enhance radiosensitivity during the treatment of NPC. Figure S1 In vivo drug tolerability study. (A) Mice bearing CNE2 tumor xenografts were randomized into four groups; each group contained 4 mice. The mice were treated using the following schedule for three weeks: Group 1 received intraperi-