Enhancement of In vitro and In vivo Tumor Cell Radiosensitivity by the DNA Methylation Inhibitor Zebularine

DNA methylation is a critical epigenetic process involved in regulating gene expression. It is well established that aberrant DNA hypermethylation is a frequent event in tumor cells, often resulting in the silencing of tumor suppressor genes as well as other cancer-related genes, including those involved in cell cycle regulation, apoptosis, and DNA repair (1–3). Given that abnormal DNA hypermethylation preferentially occurs in tumor cells compared with normal cells, use of demethylating agents has been suggested as a strategy for cancer therapy. Studies of demethylation as a cancer treatment approach have primarily focused on the two nucleoside analogues, 5-azacytidine and 5-aza-2′-deoxycytidine. Each is incorporated into DNA through substitution for cytosine during replication, ultimately leading to a loss of DNA methyltransferase activity. These compounds have been used extensively in studying the fundamental aspects of DNA methylation and have progressed to clinical trials (4–8). However, 5-azacytidine and 5-aza-2′-deoxycytidine are unstable in aqueous solution (9, 10), which has complicated their application in clinical settings. Moreover, each compound has been associated with considerable toxicity in experimental and in patient treatment situations (10).

In search of a more stable and less toxic DNA methylation inhibitor, Cheng et al. (11) have recently identified zebularine, which is also a cytidine analogue. While a potent inhibitor of DNA methylation, zebularine is stable in aqueous solution and induces minimal toxicity in animal studies. Daily oral or i.p. administration of zebularine to mice bearing human tumor xenografts was shown to inhibit tumor growth, which correlated with demethylation and reactivation of the p16 gene (11). Moreover, the tumor growth inhibition was not associated with any significant toxicity over the 18-day zebularine treatment period. Recently, zebularine was reported to preferentially reduce DNA methyltransferase activity and induce gene expression in tumor cells compared with normal fibroblasts (12). Thus, in vitro and in vivo data suggest that the zebularine-mediated inhibition of DNA methylation may have potential as a cancer treatment strategy.

However, because targeting DNA methylation seems to primarily affect genes involved in tumor cell proliferation (2, 13, 14), the antitumor effects of zebularine are likely to be limited to cytostasis. Furthermore, cessation of zebularine administration results in a remethylation of the demethylated genes (15), which would be associated with a resumption of tumor growth. Effective tumor control is thus likely to require chronic zebularine treatment. Inhibition of DNA methylation affects chromatin structure and gene expression (16–19), two parameters that have long been implicated in the regulation of cellular radioresponse. Moreover, with respect to the regulation of chromatin structure and gene expression, there seems to be an intimate relationship between DNA demethylation and histone hyperacetylation. We have recently shown that inhibitors of histone deacetylase activity enhance the radiosensitivity of a variety of human tumor cell lines (20). To determine whether the inhibition of DNA methylation has a similar effect, we have investigated the influence of zebularine on the radiosensitivity of three human tumor cell lines of different histologic origins. The data presented indicate that zebularine enhances tumor cell radiosensitivity, which correlates to changes in methylation status. Furthermore, the zebularine-mediated radiosensitization was associated with a prolonged expression of γH2AX foci, suggesting a decrease in the repair of radiation-induced DNA double-strand breaks.

Cell lines and treatment. Three human tumor cell lines originating from three different histologies were evaluated: MiaPaca, a pancreatic carcinoma; U251, a glioblastoma; and DU145, a prostate carcinoma. Each of these cell lines was obtained from the American Type Culture Collection (Gaithersburg, MD). The cell lines were grown in RPMI 1640 (Life Technologies, Rockville, MD) containing 5 mmol/L glutamate and 5% heat-inactivated fetal bovine serum and maintained at 37°C in an atmosphere of 5% CO₂ and 95% room air. Zebularine (NSC 309132), provided by the Drug Synthesis and Chemistry Branch, Developmental Therapeutics Program of the National Cancer Institute, was dissolved in PBS to a stock concentration of 10 mmol/L and stored at −20°C. Cultures were irradiated using a Pantak (Solon, OH) X-ray source at a dose rate of 1.55 Gy/min.

Clonogenic assay. Cultures exposed to zebularine or vehicle control were trypsinized to generate a single cell suspension and a specified number of cells were seeded into each well of six-well tissue culture plates. After allowing cells time to attach (6 hours), the plates were irradiated at the specified doses. Twelve to fourteen days after seeding, colonies were stained with crystal violet; the number of colonies containing at least 50 cells was determined, and the surviving fractions were calculated. Survival curves were generated after normalizing for the cytotoxicity induced by zebularine alone. Data presented are the mean ± SE from at least three independent experiments.

Bisulfite modification and methylation-specific PCR. The methylation status of designated DNA promoters was determined by methylation-specific PCR. DNA was extracted from cells using DNeasy kit (Qiagen Sciences, Germantown, MD), and bisulfite modification of genomic DNA was done as reported by Herman et al. (21). Briefly, 1 μg of genomic DNA was denatured with NaOH (final concentration, 0.2 mol/L), and 10 mmol/L hydroquinone (Sigma, St. Louis, MO) and 3 mol/L sodium bisulfite (Sigma) were added and incubated at 50°C for 16 hours, which converts all unmethylated cytosines to uracils, whereas methylated cytosines remain unchanged (22). The modified DNA was purified using a Wizard DNA cleanup system (Promega, Madison, WI). The bisulfite-modified DNA was then mixed with 10× PCR buffer, 150 μmol/L of deoxynucleotide triphosphates, 0.4 μmol/L of primers, and 1 unit of HotStarTaq (Qiagen Sciences). The primers were designed according to the literature (23–25). The primer sequences for both the methylated and unmethylated templates of all the genes tested, annealing temperatures used, and expected PCR product size are summarized in Table 1. Amplification was carried out using a PTC-200 Peltier thermal cycler (MJ Research, Watertown, MA). The PCR products were resolved by electrophoresis in a 2% agarose gel containing ethidium bromide. Water blanks were included in each assay.

Reverse transcription-PCR. Total cellular RNA was isolated from cultured cells with Trizol (Invitrogen Life Technologies, Inc., Carlsbad, CA) and purified with RNeasy minikit (Qiagen Sciences) according to the manufacturer's instructions and then treated with 2 units/μL of DNase I (Qiagen Sciences) for 15 minutes at room temperature. Reverse transcription was done for 2 μg of total RNA with the SuperScript II First-Strand Synthesis using the Oligo (dT) primer system (Invitrogen Life Technologies) in 20 μL reaction mixture. The primers used in the PCR reactions for the genes analyzed (23, 26, 27) are summarized in Table 1. The PCR reactions were run in an appropriate linear range (37 cycles for RASSF1A, 33 cycles for HIC-1, 30 cycles for 14-3-3σ). The amplified products were subjected to electrophoresis on a 2% agarose gel containing ethidium bromide. Negative controls for PCR were generated under the same conditions without RNA or reverse transcriptase. Expression of the β-actin was used to examine the integrity of RNA in each sample and to standardize the amount of cDNA added to each of the PCR reactions.

Cell cycle phase analysis. Evaluation of cell cycle phase distribution was done using flow cytometry. The treatment protocols were essentially the same as in the clonogenic survival experiments, except that the cells were initially seeded into 10 cm dishes. All cultures were subconfluent at the time of collection. Cultures were collected for fixation, stained with propidium iodide, and analyzed using flow cytometry as previously described by the Clinical Services Program at National Cancer Institute-Frederick (28). To evaluate the activation of the G₂ cell cycle checkpoint, mitotic cells were distinguished from G₂ cells and the mitotic index was determined according to the expression of phosphorylated histone H3 (Upstate Biotechnology, Charlottesville, VA) as detected in the 4N DNA content population using the flow cytometric method of Xu et al. (29). In this assay, loss of mitotic cells (reduced mitotic index) reflects the onset of G₂ arrest.

Immunofluorescent staining for γH2AX. Cells were grown and treated in tissue culture chamber slides (Nalge Nunc International, Naperville, IL). At specified times, medium was aspirated and cells were fixed in 4% paraformaldehyde for 10 minutes at room temperature. Paraformaldehyde was aspirated, and the cells were treated with a 0.2% NP40/PBS solution for 15 minutes. Cells then were washed in PBS twice, and the anti-γH2AX antibody (Upstate Biotechnology, Lake Placid, NY) was added at a dilution of 1:200 in 1% bovine serum albumin and incubated overnight at 4°C. Cells again were washed twice in PBS before incubating in the dark with an FITC-labeled secondary antibody at a dilution of 1:50 in 1% bovine serum albumin for 1 hour. The secondary antibody solution then was aspirated and the cells were washed twice in PBS. Cells were then incubated in the dark with 4′,6-diamidino-2-phenylindole (1 μg/mL) in PBS for 30 minutes and washed twice, and coverslips were mounted with an antifade solution (DAKO Corp., Carpinteria, CA). Slides were examined on a Leica DMRXA fluorescent microscope (Leica, Wetzlar, Germany). Images were captured by a Photometrics Sensys CCD camera (Roper Scientific, Tucson, AZ) and imported into IP Labs image analysis software package (Scanalytics, Inc., Fairfax, VA) running on a Macintosh G3 computer (Apple, Cupertino, CA). For each treatment condition, the number of γH2AX foci was counted in at least 50 cells.

In vivo tumor growth delay assay. Male 6-week-old athymic nude mice (NCr nu/nu) were used in these studies. Mice, housed in filter-topped cages, were provided autoclaved feed and hyperchlorinated water ad libitum. U251 tumor cells were implanted s.c. on the lateral aspect of the rear leg. When tumors grew to a mean volume of 175 mm³ (7 × 7 mm), mice were randomized into four groups: vehicle, zebularine, radiation (4 Gy), and zebularine plus radiation (4 Gy). At the time of randomization, zebularine treatment, which consisted of i.p. injections of 350 mg/kg administered thrice per day at 8-hour interval for 3 days (nine injections), was initiated. For the zebularine plus radiation group, 8 hours after the last injection, 4 Gy was delivered to animals restrained in a custom lead jig, which allowed for the localized irradiation of the implanted tumors. Three days of zebularine treatment resulted in a slight decrease in tumor growth rate compared with control mice: the tumor volume of zebularine-treated mice was 248 ± 6.8 versus 306 ± 9.6 mm³ in saline-treated mice. Tumor volume is a critical parameter in determining radiation-induced growth delay with smaller tumors appearing more radiosensitive. To ensure that the zebularine-induced growth delay did not bias the results of the combination treatment (zebularine plus 4 Gy), it was important that the two irradiated groups (4 Gy and zebularine plus 4 Gy) received radiation when tumors were approximately the same size. Therefore, radiation was delivered to the 4 Gy–only group when tumors were 251 ± 4.7 mm³, which corresponded to day 2 after randomization compared with day 4 for the combination group. To obtain growth curves, perpendicular diameter measurements of each tumor were collected every 2 to 3 days with calipers, and the volumes were calculated using the formula (L × W × W) / 2. Tumors were followed until the tumors of the group were >1,400 mm³, which corresponds to three doublings in tumor volume. Absolute tumor growth delay was calculated as the number of days for the treated tumors to grow to 1,400 mm³ minus the number of days for the control group to reach the same size. Each experimental group contained 10 mice and the untreated group contained 20 mice. All animal studies were conducted in accordance with the principles and procedures outlined in the U.S. Public Health Service Guide for the Care and Use of Laboratory Animals in an Association for Assessment and Accreditation of Laboratory Animal Care–approved facility under an approved animal protocol.

Effect of zebularine on DNA methylation and gene expression. Three human tumor cell lines of different histologic origins, which typically receive radiotherapy as part of their treatment protocol, were used in this study: MiaPaCa (pancreatic carcinoma), U251 (glioblastoma), and DU145 (prostate carcinoma). In addition, multiple cell lines were evaluated to reduce the potential for a cell line–specific effect of zebularine on radiosensitivity. These studies were based on the hypothesis that zebularine-induced DNA demethylation modifies tumor cell radiosensitivity. Therefore, as an indicator of demethylation, initial experiments used methylation-specific PCR analysis to determine the effects of zebularine on the DNA methylation status of three genes RASSF1A, HIC-1, and 14-3-3σ, which have been frequently reported to be hypermethylated in tumor cells (23–27, 30–34). As shown in Fig. 1A, the promoter regions of RASSF1A, HIC-1, and 14-3-3σ were heavily methylated in control (untreated) MiaPaCa and U251 cells, whereas in control DU145 cells, RASSF1A was heavily methylated, 14-3-3σ was unmethylated, and HIC-1 was partially methylated. In each of the cell lines, zebularine treatment resulted in an increase in the level of unmethylated RASSF1 and HIC-1 promoters, whereas demethylation of 14-3-3σ was only detected in MiaPaCa2 and U251 cells. These data indicate that although there is some cell type specificity, zebularine inhibits DNA methylation in each of the tumor cell lines. Consistent with the changes in methylation, zebularine exposure increased the expression of RASSF1 and HIC-1 in each cell line and increased the expression of 14-3-3σ in the MiaPaCa and U251 cell lines (Fig. 1B). To determine the kinetics of demethylation, cells were exposed for 24 to 72 hours to zebularine and a similar set of analyses were done for the RASSF1A gene, which was heavily methylated in each of the three cell lines. The results for the MiaPaCa cells are shown in Fig. 1C. The unmethylated band specific for the promoter region of RASSF1A was detected after 24 hours reaching a maximum by ∼48 hours of exposure to zebularine. Consistent with the increase in the unmethylated promoter, the expression of RASSF1A gene was detectable by 24 hours of zebularine exposure, reaching a maximum from 48 to 72 hours. Similar results were obtained for the other two cell lines (data not shown). These results indicate that zebularine induces demethylation and reactivation of silent genes in each of the tumor cell lines evaluated.

Zebularine and tumor cell radiosensitivity. To determine whether zebularine enhances cellular sensitivity to ionizing radiation, cell lines were exposed to zebularine for 24 or 48 hours before irradiation and performance of the colony-forming efficiency assay. Radiation survival curves (Fig. 2) were generated for each cell line after normalization for the level of cell killing induced by zebularine alone. Exposure to zebularine for 24 hours before irradiation resulted in an increase in radiosensitivity of each cell line with dose enhancement factors (at a surviving fraction of 0.1) of 1.25, 1.15, and 1.24 for MiaPaCa, DU145, and U251, respectively. Zebularine treatment alone resulted in surviving fractions of 0.93 ± 0.095, 0.66 ± 0.093, and 0.65 ± 0.068 for MiaPaCa, DU145, and U251, respectively. When the cell lines were exposed to zebularine for 48 hours before irradiation, there was a greater degree of enhancement of radiation-induced cell killing with dose enhancement factors of 1.51, 1.67, and 1.59 for MiaPaCa, DU145, and U251, respectively. Using this protocol, the surviving fractions for zebularine exposure alone were 0.35 ± 0.066, 0.22 ± 0.038, and 0.42 ± 0.041 for MiaPaCa, DU145, and U251, respectively. These data indicate that zebularine enhances the radiosensitivity of the three tumor cell lines with the maximum effect attained by 48 hours, which follows a time course consistent with decreases in DNA methylation.

Mechanism of zebularine-induced radiosensitization. To gain insight into the potential mechanism responsible for zebularine-induced radiosensitization, we addressed a number of cellular processes that can serve as determinants of radiosensitivity. One such process is the activation of the G₂ checkpoint, which arrests cells in G₂ allowing for the repair of DNA damage before progression into mitosis (35). To evaluate the effects of radiosensitizing concentrations of zebularine on the activation of the G₂ checkpoint, the method of Xu et al. (29) was used, which distinguishes between G₂ and mitotic cells. This assay determines the percentage of mitotic cells in the 4N population according to the flow cytometric analysis of phosphorylated histone H3, which is specifically expressed in mitotic cells. Done as a function of time after irradiation, this analysis provides a measure of the progression of G₂ cells into M phase and, thus, the activation of the G₂ checkpoint. As shown in Fig. 3, radiation alone decreased the percentage of mitotic cells by 3 hours, consistent with the rapid onset of G₂ arrest and with previously published results (36). Zebularine exposure alone also reduced the mitotic index, but not to the same degree as radiation in MiaPaCa and U251 cells. The combination of zebularine followed by radiation also reduced mitotic index, although to varying degrees. In MiaPaCa, the mitotic index for cells exposed to the zebularine/radiation combination was increased compared with those irradiated alone, yet remained below untreated levels. In DU145 cells, the combination treatment seemed to result in an additive decrease in mitotic index and in U251 the combination was essentially the same as radiation alone. Although the specific effect of the zebularine/radiation combination varied between cell lines, in each case the mitotic index remained substantially below control levels, suggesting that zebularine does not abrogate radiation-induced G₂ arrest.

Because the susceptibility to apoptosis can be a determinant of radiosensitivity, the effects of zebularine on radiation-induced apoptosis were evaluated using flow cytometry and the determination of the sub-G₁ population. Each cell line was exposed to a radiosensitizing concentration of zebularine for 48 hours and irradiated with 6 Gy; cultures were then analyzed 24 and 48 hours later. Radiation alone did not induce a significant level of apoptosis, nor did zebularine. Moreover, the combination of radiation and zebularine did not result in a significant increase in the sub-G₁ population (apoptotic cells) at 24 or 48 hours (data not shown). These data indicate that the zebularine-mediated increase in radiosensitivity cannot be attributed to an enhanced susceptibility to apoptosis.

A critical determinant of radiation-induced lethality is the induction and repair of DNA damage, specifically double-strand breaks. To determine the effects of zebularine on DNA damage in irradiated cells, we evaluated foci of phosphorylated histone H2AX (γH2AX), which has been established as a sensitive indicator of DNA double-strand breaks (37–39). As shown by the representative micrographs of U251 cells in Fig. 4, γH2AX foci could be clearly distinguished after exposure to 2 Gy. The expression of γH2AX foci were quantitatively evaluated in each of the three tumor cell lines after exposure to 2 Gy with and without zebularine pretreatment (Fig. 5). Zebularine treatment alone had only a minor effect, if any, on γH2AX foci over the 24-hour time course of evaluation. Radiation (2 Gy) alone induced a significant increase in γH2AX foci at 1 hour in each of the cell lines, which was reduced by 6 hours and returning to background levels by 24 hours after irradiation. For cells exposed to the combination of zebularine and radiation, there was no consistent difference in γH2AX foci compared with radiation only at the 1-hour time point. However, at the 24-hour time point, γH2AX foci levels in cultures receiving the combined zebularine/radiation treatment were significantly greater compared with the radiation-only group. The maintenance of γH2AX foci levels suggests that the zebularine-mediated radiosensitization involves the inhibition of the repair of DNA damage (39).

Effect of zebularine on radiation-induced tumor growth delay. To determine whether the radiosensitizing effects of zebularine could be extended to an in vivo model, we used U251 cells grown as xenografts in nude mice. As in the in vitro studies, it was first necessary to determine the effects of zebularine on DNA methylation in U251 xenografts. Mice bearing xenografts received 350 mg/kg administered by i.p. injection at 8-hour intervals for up to 5 days. Preliminary studies showed that this 5-day zebularine treatment protocol induced a significant tumor growth delay in the absence of overt toxicity. Because of their susceptibility to zebularine-mediated reactivation in vitro (Fig. 1), the expression of RASSF1A, HIC-1, and 14-3-3σ were used as indicators of DNA demethylation in vivo. As shown in Fig. 6, by 3 days (9 doses) of zebularine treatment, the expression of each of the indicator genes reached a maximum expression level compared with vehicle-treated mice; no further increase was detected by 5 days (14 doses) of treatment. These gene expression data indicated that 3 days for drug treatment was sufficient to induce maximum DNA demethylation, which was then used to design an in vivo combination protocol aimed at determining whether zebularine enhances the radioresponse of U251 tumor xenografts.

Mice bearing s.c. U251 leg tumor xenografts were randomized into four treatment groups: vehicle (PBS), zebularine, radiation (4 Gy), or the combination of zebularine followed by radiation. Treatment with zebularine consisted of nine i.p. injections of 350 mg/kg delivered at 8-hour intervals over 3 days, which began the day of randomization. Radiation (4 Gy) was then delivered 8 hours after the last zebularine dose (time of maximum gene expression). The growth rates of U251 tumors in mice treated with vehicle (control), zebularine, 4 Gy, and zebularine/4 Gy combination are shown in Fig. 7. For each group, the time to grow from 175 mm³ (volume at the time of zebularine treatment initiation) to 1,400 mm³ was calculated using the tumor volumes from the individual mice in each group (mean ± SE). Zebularine treatment alone inhibited tumor growth (P < 0.01 versus vehicle): the time required for tumors to grow from 175 to 1,400 mm³ increased from 15.2 ± 0.9 days for vehicle-treated mice to 19.9 ± 0.8 days for zebularine-treated mice. Radiation treatment alone increased the time for tumors to reach 1,400 mm³ to 20.4 ± 1.1 days (P < 0.01 versus vehicle). However, in mice that received the zebularine/radiation combination, the time for tumors to grow to 1,400 mm³ increased to 30.1 ± 2.2 days (P < 0.01 versus vehicle), which is also significantly greater than the individual treatment groups (P < 0.01 versus zebularine and versus 4 Gy). The absolute growth delays (the time in days for tumors in treated mice to grow from 175 to 1,400 mm³ minus the time in days for tumors to reach the same size in vehicle-treated mice) were 4.7 for zebularine only, 5.2 days for 4 Gy only, and 14.9 for the combination. Thus, the growth delay after the combined treatment was more than the sum of the growth delays caused by zebularine alone and radiation alone. To obtain a dose enhancement factor comparing the tumor radioresponse in mice with and without zebularine treatment, the normalized tumor growth delay was determined, which accounts for the contribution of zebularine to tumor growth delay induced by the combination treatment. Normalized tumor growth delay is defined as the time in days for tumors to grow from 175 to 1,400 mm³ in mice treated with the combination of zebularine and radiation minus the time in days for tumors to grow from 175 to 1,400 mm³ in mice treated with zebularine only, which was 10.2 days (30.1 minus 19.9 days). The dose enhancement factor, obtained by dividing the normalized tumor growth delay in mice treated with the zebularine/radiation combination by the absolute growth delay in mice treated with radiation only (5.2 days), was 2.0. Thus, zebularine alone slows tumor growth and enhances the effect of radiation, which is similar to the results obtained in vitro. This occurs in the absence of body weight loss in any of the treatment groups (data not shown).

It is well established that methylation of CpG islands within the DNA of cancer cells is associated with transcriptional gene silencing. Because this epigenetic event often includes genes involved in tumor cell proliferation and survival, DNA methylation has been suggested as a target for cancer therapy. Zebularine has recently been identified as an inhibitor of DNA methylation, which, in contrast to previously studied inhibitors, is stable in aqueous solution and has considerably less toxicity (11). Moreover, zebularine has recently been shown to be preferentially incorporated into the DNA of tumor cells compared with normal cells in vitro (12). Given the effects of zebularine on the expression of genes associated with tumorigenesis, its selectivity for tumor cells, and its growth inhibitory effect in xenografts systems, this demethylating agent seems to be an appropriate candidate for molecularly targeted cancer therapy. However, in experimental systems, withdrawal of zebularine or the other inhibitors of DNA methylation results in remethylation of demethylated gene promoters and a resilencing of the expression of the corresponding gene (15). Return to the initial gene expression profile after cessation of treatment suggests that the use of demethylating agents may be limited as a single modality approach in the treatment of most solid tumors. However, the effects of demethylating hypermethylated DNA on chromatin structure and gene expression suggested that inhibitors of methylation may also enhance tumor cell radiosensitivity. The data presented here indicate that zebularine increases radiation-induced cell death as evaluated in vitro and in an in vivo xenograft model. Although not establishing a casual relationship, these experiments are suggestive of a correlation between zebularine-induced demethylation and radiosensitization.

Although the expression of a considerable number of genes are potentially repressed by DNA methylation in tumor cells (1, 2), in this study we focused on RASSF1A, HIC-1, and 14-3-3σ, which have been frequently reported to be epigenetically inactivated by hypermethylation of the CpG island in their promoters (23–27, 30–33). Whereas HIC-1 and 14-3-3σ were hypermethylated in two of the three cell lines investigated, the promoter region of RASSF1A was strongly methylated in each of the three cell lines. Consistent with its demethylation effect, zebularine restored the expression of HIC-1 and 14-3-3σ in MiaPaca and U251 cells and RASSF1A in each of the tumor cell lines. At present, there is no evidence suggesting a causal association between the reexpression of HIC-1 or 14-3-3σ with enhanced radiosensitivity. Although the specific functions of RASSF1A have not been fully defined, it has been reported to form heterodimers with the Ras effector NORE1 and to subsequently enhance apoptosis (40). However, such as the fact that zebularine-induced radiosensitization did not seem to involve apoptosis, it is also unclear how reactivation of RASSF1A contributes to the enhanced radiosensitivity. It should be emphasized that the gene promoters evaluated in this study were merely intended to serve as markers for demethylation. Whether the reactivation of a specific gene or set of genes is responsible for zebularine-induced radiosensitization remains to be determined. It is also possible that the process of gene reactivation involving the relaxation of chromatin structure may play a role. To our knowledge, the results presented here, although far from establishing a causal relationship, are the first to show a correlation between zebularine-induced demethylation and the enhancement of radiosensitivity.

Whereas defining the molecular event responsible for zebularine-induced radiosensitization requires further investigation, the effects of this demethylating agent on cellular processes known to influence cellular radiosensitivity have been investigated. Apoptosis, the G₂ checkpoint, and repair of DNA damage are each considered cellular processes that can affect intrinsic radiosensitivity. Apoptotic cell death generally occurs at lower radiation doses than death occurring through mitotic catastrophe. However, radiation-induced apoptosis is a relatively infrequent event for cells from most solid tumor; the ability to convert the mode of death from mitotic catastrophe to apoptosis can be considered a strategy for radiosensitization. According to sub-G₁ content, zebularine exposure of all three of the tumor cell lines investigated did not induce a significant increase in apoptotic death and did not modify the already low frequency of radiation-induced apoptosis. The activation of the G₂ cell cycle checkpoint, which rapidly arrests cells in G₂ allowing time for DNA repair before progression into mitosis, is also considered a determinant of radiosensitivity (41). Moreover, the G₂ checkpoint is considered a target for potential radiosensitizing agents. Because evaluation of this checkpoint requires distinguishing G₂ cells from mitotic cells in the 4N DNA content population, the procedure of Xu et al. (29) was used to identify mitotic cells based on the expression of phosphorylated histone H3. Whereas the effect of zebularine on radiation-induced G₂ arrest in each of the cell lines varied slightly, in none of the cell lines was arrest in G₂ abrogated. These data indicate that zebularine-induced radiosensitization is not the result of inhibiting the activation of the G₂ checkpoint.

Another process involved in determining cellular radiosensitivity is DNA repair, specifically the repair of DNA double-strand breaks. Typically, the repair of double-strand breaks has been evaluated using pulse-field gel electrophoresis (42). However, this technique is fairly insensitive, requiring radiation doses of 10 to 20 Gy, which brings into question its relevance to clinically applicable radiation doses. With the increased understanding of the fundamental DNA repair mechanisms, there has been the identification of molecular markers that can serve as sensitive indicators of double-strand break induction and repair. At sites of radiation-induced double-strand breaks, the histone H2AX becomes rapidly phosphorylated (γH2AX), forming readily visible nuclear foci (37, 38). Although the specific role of γH2AX in DNA repair has not been defined, recent reports indicate that the dispersal of γH2AX foci in irradiated cells correlates with the repair of double-strand breaks (39, 41, 43). Moreover, Macphail et al. (44), in their study of 10 cell lines, reported that the loss of γH2AX correlates with clonogenic survival after irradiation. Thus, the results presented here, in which the radiation-induced expression of γH2AX in cells treated with zebularine were similar to radiation exposure only at 1 hour but significantly greater at 24 hours, is suggestive of an inhibition of DNA repair.

The specific process resulting in a putative DNA repair defect in zebularine-treated cells remains to be determined. However, DNA methylation has been intimately associated with histone deacetylation with respect to the epigenetic regulation of gene expression (16). Both processes act to silence gene expression through modulation of transcription factor activity and by condensing local chromatin structure (45, 46). Moreover, DNA methylation acts to recruit histone deacetylases to gene promoter regions. We have recently shown that the histone deacetylase inhibitors MS-275 and valproic acid enhance tumor cell radiosensitivity (including that of DU145 and U251 cells), and that this radiosensitization is also associated with a prolongation of γH2AX expression (20, 47). It should be noted that Stoilov et al. (48) showed that the histone deacetylase inhibitor sodium butyrate reduced the repair of chromosome breaks as detected by the premature chromosome condensation technique, suggestive of an inhibition of DNA double-strand breaks. Given the physical and functional interactions between DNA methylation and histone acetylation (16, 18, 19, 45), it could be speculated that the radiosensitization induced by the DNA methylation inhibitor zebularine may involve a similar mechanism as that mediating the sensitization induced by histone deacetylase inhibitors. Alternatively, because the prolonged expression of radiation-induced γH2AX foci may reflect the end result of disparate processes and events leading to the maintenance of unrepaired double-strand breaks, distinctly different mechanism may be involved. Clearly, additional investigations are required to define the molecular processes responsible for zebularine-mediated radiosensitization. However, these results do suggest that enhanced radiosensitivity induced by zebularine involves an inhibition of the repair of DNA damage.

This is the initial investigation into the combination of radiation and the DNA demethylating agent zebularine. Results of this study indicate that zebularine enhanced the radiosensitivity of three human tumor cell lines, which corresponded to the loss of promoter-specific methylation and consequent gene reactivation. Furthermore, zebularine enhanced the radiation-induced tumor growth delay in a human tumor xenograft model, which also correlated with reactivation of a hypermethylated gene. These results suggest that combining zebularine and radiation may be of clinical potential as a cancer treatment strategy.