Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2015 May 1.
Published in final edited form as: Mol Cancer Ther. 2014 Feb 25;13(5):1334–1344. doi: 10.1158/1535-7163.MCT-13-0924

The role of gene body cytosine modifications in MGMT expression and sensitivity to temozolomide

Erika L Moen 1, Amy L Stark 2, Wei Zhang 3, M Eileen Dolan 1,2,4,*, Lucy A Godley 1,2,4,*
PMCID: PMC4013207  NIHMSID: NIHMS571143  PMID: 24568970

Abstract

The DNA repair protein O6-methylguanine-DNA methyltransferase (MGMT) is known to play a role in sensitivity to temozolomide. Promoter hypermethylation of MGMT is commonly used to predict low expression levels of MGMT in gliomas, despite observed discordance between promoter methylation and protein levels. Here, we investigated the functional role of gene body cytosine modification in regulating levels of MGMT gene expression and sensitivity to temozolomide. In 91 human glioblastoma samples, we observed significant variation in MGMT expression levels in patients with an unmethylated promoter, with higher levels of gene body cytosine modification correlating with higher gene expression levels. Furthermore, inducing hypomethylation across the MGMT gene body with decitabine corresponded with decreased levels of MGMT gene expression in lymphoblastoid and glioblastoma cell lines, indicating an important functional role for gene body cytosine modifications in maintaining gene expression. We reasoned that the decrease in MGMT expression induced by decitabine may render resistant glioblastoma cell lines more sensitive to temozolomide. Consistent with this reasoning, we found that the MGMT-expressing glioblastoma cell lines exhibiting an unmethylated MGMT promoter that were pre-treated with decitabine became significantly more sensitive to temozolomide. Overall, our results suggest a functional role for gene body cytosine modification in regulating gene expression of MGMT and indicate that pre-treating patients whose tumors have an unmethylated MGMT promoter with decitabine prior to temozolomide treatment may increase their response to therapy.

Keywords: MGMT, temozolomide, glioblastoma, decitabine, DNA methylation

Introduction

The O6-methylguanine-DNA methyltransferase (MGMT) gene encodes for a DNA repair protein that repairs O6-alkylguanine-DNA adducts (1). This type of DNA damage can occur after certain environmental exposures, such as tobacco-specific nitrosamines, and from DNA alkylating agents used for chemotherapy (2, 3). Temozolomide is an alkylating agent that was FDA-approved to treat brain tumors in 2005, including anaplastic astrocytoma and glioblastoma multiforme, the most aggressive form of brain cancer (4). Only 5–10% of the methylation adducts on DNA resulting from temozolomide treatment are O6-methylguanine, but if MGMT is not present to repair the damage, these adducts trigger DNA mismatch repair and are highly cytotoxic (5, 6). Upon repair, MGMT is irreversibly inactivated, thereby requiring new protein synthesis for additional repair (7). Consequently, epigenetic alterations that affect transcriptional activity of MGMT, such as promoter hypermethylation, can have a significant effect on the number of O6-guanine lesions that are repaired (8, 9). Glioma patients with an aberrantly hypermethylated MGMT promoter respond better to temozolomide treatment compared with patients with an unmethylated MGMT promoter, since they lack MGMT protein expression (10, 11). Although approximately 40–70% of glioma patients have a methylated MGMT promoter (9, 12–14), promoter methylation does not always show a strong correlation with MGMT protein levels (13, 15). For example, one study investigating the correlation between MGMT promoter methylation and protein levels showed that 7/40 gliomas had an unmethylated promoter yet no detectable MGMT protein, and 16/40 gliomas had aberrant methylation at the promoter yet still had detectable MGMT protein (13). Therefore, additional mechanisms of transcriptional and translational regulation are likely affecting expression of MGMT.

The role of gene body cytosine methylation and hydroxymethylation in regulating gene expression levels is not fully understood, but several studies have observed a positive correlation between gene body cytosine modifications and gene expression levels (16–18). For MGMT in particular, gene body methylation was first shown to correlate with gene expression levels in 1992 (19). Two years later, it was demonstrated that glioma cell lines with lower expression of MGMT tend to have a more highly methylated promoter and low levels of methylation along the gene body compared with glioma cell lines showing higher levels of MGMT expression (8, 19). However, these studies have only shown correlation, and the field has been focused primarily on methylation of the promoter region after it was shown to predict glioblastoma patient survival. It remains unclear as to whether modulation of gene body cytosine modifications would be sufficient to disrupt gene expression levels without any change in the status of cytosine modification in the promoter region.

We hypothesized that incorporating gene body cytosine modification levels in models of temozolomide response may lead to better prediction of MGMT expression levels and, more importantly, improved methods of prediction for glioma patient response to temozolomide. In this study, we used lymphoblastoid cell lines, glioblastoma cell lines, and human glioblastoma tissue samples from The Cancer Genome Atlas (TCGA) to investigate the role of gene body cytosine modification in regulating MGMT expression levels and sensitivity to temozolomide.

Materials and Methods

Cell lines and reagents

Lymphoblastoid cell lines were cultured in RPMI supplemented with 15% fetal bovine serum (FBS) and 1% L-Glutamine at 37°C. Confirmation of cell line identities for the lymphoblastoid cell lines is described previously (18). The glioblastoma cell line U118MG (HTB-15) was purchased from the American Type Culture Collection (ATCC) in March of 2013, and was authenticated by ATCC by evaluating the short tandem repeat (STR) profile. U-87MG and A-172 were obtained from Dr. Maciej Lesniak in February of 2013, and T98G was obtained from Dr. Bakhtiar Yamini in March of 2013 at the University of Chicago and no subsequent authentication was performed on these cell lines. SF-188 was obtained from Dr. Joseph Costello at the Neurosurgery Tissue Bank at University of California, San Francisco in August of 2013. These cells were authenticated by UCSF using the PowerPlex16 System (Promega Corp). Glioblastoma cells were grown in DMEM (ATCC Catalog No. 30-2002) supplemented with 10% FBS and 1% Penicillin/Streptomycin.

IDH1/2 genotyping

All six glioblastoma cell lines were genotyped for the IDH1 and IDH2 mutations. Primers for IDH1 (R132) are as follows: forward-GGTGGCACGGTCTTCAGAG, reverse-ATGTGTTGAGATGGACGCCT. Primers for IDH2 (R140 and R172) are as follows: forward-TTCTGGTTGAAAGATGGCGG, reverse-GGATGGCTAGGCGAGGAG. Genomic DNA was PCR-amplified using Platinum HiFi Taq polymerase (Life technologies, Grand Island, NY) under the following conditions: initial denaturation at 94°C for 2 min, 94°C for 30 sec, 62°C for 30 sec, and 68°C for 1 min for 35 cycles, and a final extension at 68°C for 5 minutes.

Drug treatment

Cells in a T25 flask were treated with 200 nM decitabine or vehicle control (H20) every 24 hours for five days. At the end of the fifth day, cells were plated at 10,000 cells/well (lymphoblastoid cell lines) or 5,000 cells/well (glioblastoma cell lines) in a 96-well plate, and the remainder was spun down into two pellets and kept at −80°C until subsequent DNA and RNA isolation. Twenty-four hours after plating, cells were treated with a range of doses of temozolomide (Sigma Aldrich, St. Louis, MO): 0, 50, 100, 200, 500, 750, and 1000 μM for lymphoblastoid cell lines and 0, 10, 50, 100, 500, 1000, 2000 μM for the glioblastoma cell lines for a total of 72 hours. AlamarBlue was added 24 hours before absorbance reading at wavelengths of 570 nm and 600 nm (Synergy-HT multidetection plate reader; BioTek, Winooski, VT). The percentage of cells that survived was quantified relative to vehicle control wells, and at least two separate experiments for each glioblastoma cell line were performed in technical triplicates for each concentration. To determine the significance of the effect of decitabine pre-treatment on temozolomide cytotoxicity, we fit the following linear model: Survival = temozolomide dose x decitabine, where decitabine is 1 for decitabine pre-treatment and 0 for vehicle control. This model was compared with a baseline model without the interaction term using ANOVA to determine if the interaction between decitabine and temozolomide dose was significant. The linear models were fit using the lm function in R.

Bisulfite sequencing

DNA was isolated from whole cell lysates using phenol-chloroform, and genomic DNA was treated with sodium bisulfite (20). For bisulfite PCR amplification, primers (Supplementary Table 1) were designed to amplify ~250 base pair fragments within the MGMT promoter and gene body that included CpG sites interrogated by the Illumina Infinium HumanMethylation 450 Beadchip (450K array) as well as adjacent CpGs. Primers for the CpG island in the promoter region were designed and optimized in a previous publication (21). ZymoTaq polymerase (Zymo Research, Irvine, CA) was used for PCR amplification at the indicated temperatures, according to manufacturer’s instructions. For the glioblastoma cell lines, the PCR product was TA cloned using the TOPO Cloning kit (Invitrogen, Grand Island, NY), and at least ten clones per condition were sequenced. Lymphoblastoid cell lines were sequenced by bulk sequencing of the PCR product. DNA sequencing was performed at the University of Chicago Sequencing Core Facility.

Quantitative real-time PCR

Total RNA was extracted from 1 million cells using TRIzol (Life technologies, Grand Island, NY) following the manufacturer’s protocol. RNA quality assessment and quantification were conducted using the optical spectrometry 260/280 nm ratio. Subsequently, mRNA was reverse transcribed to cDNA using Applied Biosystems High Capacity Reverse Transcription kit (Applied Biosystems, Foster City, CA). Quantitative RT-PCR was performed for MGMT, and huB2M (beta-2-microglobulin) and GAPDH (glyceraldehyde 3-phosphate dehydrogenase) were included as endogenous controls for lymphoblastoid and glioblastoma cell lines, respectively, using TaqMan Gene Expression Assays on the Applied Biosystems Viia7 RT-PCR system.

Statistical analyses of 450K array data

Initial analyses of the 450K array cytosine modification data (GSE39672) in the CEU and YRI lymphoblastoid cell lines are explained in detail previously (18). Level 3 450K array, exon array, and survival data were downloaded from TCGA Data Matrix. All three datasets were available for 91 glioblastoma patients. All of our analyses of TCGA glioblastoma patient data were performed using the R statistical software.

K-means clustering was performed using the ‘kmeans’ function in R. K-means clustering aims to divide n observations into k clusters in which each observation belongs to the cluster with the nearest mean, so as to minimize the within-cluster sum of squares. A plot of the within cluster sum of squares by number of clusters helped us to determine that three groups was the appropriate number of clusters to minimize the within-cluster sum of squares for clustering our data. After three groups, the within-group sum of squares plateaus, so increasing the number of groups would not significantly decrease the within-group sum of squares (Supplementary Fig. 1).

Results

A temozolomide-associated SNP is an mQTL for MGMT gene body cytosine modification levels

Our group has generated 450K array data for 60 Caucasian (CEU) and 73 African (YRI) lymphoblastoid cell lines (18). This array requires bisulfite-modified DNA, and modified cytosines, both 5-methylcytosine (5-mC) and 5-hydroxymethylcytosine (5-hmC), are distinguished from unmodified cytosines using this technique (22). Our analysis of cytosine modification and gene expression levels revealed that the averaged levels of cytosine modification within the gene body of MGMT were strongly positively correlated at genome-wide significance with gene expression levels (p<5.3×10−12) (Fig. 1A). Thus, we aimed to investigate the biological function of gene body cytosine modification in regulating MGMT transcription and how it may affect sensitivity to temozolomide. Interestingly, a recent genome-wide association study of temozolomide sensitivity in a cohort of Caucasian lymphoblastoid cell lines identified several SNPs in high linkage disequilibrium within the gene body of MGMT to be significantly associated with cellular sensitivity to temozolomide (23). The most significant SNP they identified was an expression quantitative trait locus for MGMT expression, but it was not correlated with promoter methylation levels in their lymphoblastoid cell line cohort. We hypothesized that the most significant temozolomide-associated SNP and MGMT expression quantitative trait locus (rs477692) identified in that study would be associated with gene body cytosine modification levels in our cohort of Caucasian (CEU) lymphoblastoid cell lines.

Figure 1.

Figure 1

Open in a new tab

MGMT gene body cytosine modification is positively correlated with gene expression levels in lymphoblastoid cell lines. A, MGMT expression is positively correlated with the averaged levels of cytosine modification across the gene body. B, A temozolomide-associated SNP (rs477692) is an expression quantitative trait locus for MGMT gene expression levels in the CEU HapMap lymphoblastoid cell lines. C, A temozolomide associated SNP (rs477692) is an mQTL for MGMT gene body cytosine modification levels in CEU HapMap lymphoblastoid cell lines. D, Cytosine modifications across the gene body are positively correlated with each other and negatively correlated with the promoter region. Grey bars represent the −log10 p-value of the association of each CpG with expression. Starred bars indicate the CpGs that, together, explain 69% of the variance of MGMT expression in CEU lymphoblastoid cell lines.

First, we validated that the SNP was an expression quantitative trait locus for MGMT in our Caucasian lymphoblastoid cell lines (p=0.004) (Fig. 1B). We did not observe this SNP to be an expression quantitative trait locus for MGMT expression in the African (YRI) lymphoblastoid cell lines (p=0.6), suggesting different genetic predictors of MGMT expression between the two populations. Furthermore, we found that this SNP is a modified cytosine QTL (mQTL) for the averaged levels of MGMT gene body cytosine modification (Fig. 1C). We also analyzed the SNP against each CpG within MGMT individually (n=94) using a linear model in the Caucasian lymphoblastoid cell lines, and identified 5 CpGs that were significantly associated with expression at p<0.05 after correcting for multiple testing. We performed bisulfite sequencing around three of these CpGs to demonstrate technical validation of the association using a separate methodology (Supplementary Fig. 2). From these results, we hypothesized that MGMT gene body cytosine modification levels are functionally important in regulating MGMT expression, thereby effecting cellular sensitivity to temozolomide.

Cytosine modification levels are correlated across the gene body of MGMT

We were interested in determining whether a certain pattern or region of cytosine modification across the MGMT gene would be able to predict gene expression levels more precisely than promoter methylation levels alone. In the Caucasian lymphoblastoid cell lines, a correlation matrix including all CpGs across the promoter region and gene body showed that the gene body cytosine modifications were positively correlated with each other, and they were slightly negatively correlated with CpGs in the promoter region (Fig. 1D). Across the gene body, there are several regions that show higher association with expression than others. Using a backwards elimination regression model (24), we were able to identify four CpGs, which were all located within the second intron, that could explain 69% of the variance in expression levels in the Caucasian lymphoblastoid cell lines (Fig. 1D marked by asterisks and Table 1). When we performed a backwards elimination regression model using the African lymphoblastoid cell lines, we were able to identify three CpGs, one in the promoter and two in the second intron, that could explain 29% of the variance in expression levels (Table 1). Therefore, in Caucasian and African lymphoblastoid cell lines, the second intron, which spans over 171 kb, contains several CpGs whose cytosine modification levels are significantly associated with expression levels of MGMT.

Table 1.

Certain gene body CpGs explain a high proportion of variance of MGMT expression levels.

Population CpGs r2 p-value
CEU cg11309842 0.69 1.13×10−13
cg18502933
cg06179303
cg14485787
YRI cg14194875 0.29 9.73×10−5
cg18651291
cg19680672
TCGA glioblastoma patients with an unmodified promoter cg25394042 0.18 0.001
cg14485787
cg26976729
cg25063211
cg24755725
Open in a new tab

Backwards elimination regression identifies four CpGs that explain 69% of variation and three CpGs that explain 29% of variation of MGMT expression levels in the Caucasian (CEU) and African (YRI) populations, respectively. In TCGA glioblastoma patients with an unmodified promoter, five CpGs were identified that explain 18% of the variance in expression levels, calculated using Spearman’s correlation backwards elimination.

Gene body cytosine modification levels predict MGMT expression levels in glioblastoma patients with an unmodified MGMT promoter

We next investigated whether MGMT gene body cytosine modification levels were correlated with gene expression levels in human glioblastoma samples. We obtained the 450K array, expression, and survival data for 91 glioblastoma samples from TCGA. To determine whether a certain pattern of cytosine modification across the entire gene was predictive of MGMT expression or patient survival, we performed K-means clustering to group the patients into three groups based on similar cytosine modification patterns across the MGMT promoter and gene body (Fig. 2A). Our rationale for clustering the patients into three groups is described in Methods and in Supplementary Fig. 1. Group 1 patients (black circles) had the highest levels of cytosine modification across the gene body, with the mean β-value for all CpGs being 0.803 (80.3% modified). This group has the highest levels of gene expression compared with the other two groups (Fig. 2B). Group 2 (grey circles) had moderate levels of cytosine modification across the gene body, with the mean β-value for all CpGs being 0.702 (70.2% modified) (Fig. 2A). Group 3 (white circles) had the lowest levels of cytosine modification across the gene body, with the mean β-value for all CpGs being 0.460 (46.0% modified) (Fig. 2A). These groups were not predictive of survival, which may be a result of the significant number of the glioblastoma patients who were lost to follow-up (Supplementary Fig. 3).

Figure 2.

Figure 2

Open in a new tab

MGMT gene body cytosine modification levels predict expression levels in glioblastoma patients with an unmodified MGMT promoter. A, K-means clustering of cytosine modification patterns along MGMT in 91 TCGA glioblastoma patients from TCGA. Group 1 patients (black) have the highest levels of modification across the gene body, Group 2 patients (grey) have intermediate levels of modification across the gene body, and Group 3 patients (white) have the lowest levels of modification across the gene body. B, Group 1 patients have higher levels of gene expression compared with Groups 2 and 3. C, Cytosine modification within the promoter is negatively correlated with gene expression. Patients with an unmodified promoter (β-value between 0 and 0.2, denoted by dotted line) still show significant variation in expression levels. D, MGMT gene body cytosine modification levels predict expression in glioblastoma patients with an unmodified promoter but not patients with a modified promoter.

As expected, cytosine modification within the gene promoter was negatively associated with gene expression (Fig. 2C). However, we observed a significant amount of variation in expression levels among glioblastoma patients with an unmodified promoter (Fig. 2C). When analyzing only patients with an unmodified promoter (averaged β-values between 0 and 0.2), we found that levels of gene body cytosine modification were linearly correlated with expression levels of MGMT (Fig. 2D, right). Hence, among the patients with an unmodified promoter, patients who had the lowest levels of gene body modification (Group 3) expressed significantly lower levels of MGMT compared with patients with an unmodified promoter and high levels of gene body cytosine modification (Group 1). The patients whose cytosine modification levels within the promoter averaged to be at least 20% modified (average β-value >0.2) did not show a correlation between gene body cytosine modification levels and expression, and, as expected, all groups had very low levels of expression (Fig. 2D, left). This suggests that disruption of cytosine modification levels across the gene body may introduce variability in expression levels in patients who have an unmodified promoter. In order to identify which CpGs along the gene body contributed most to variation in expression within patients with an unmodified promoter, we performed backwards elimination to identify five CpGs, one in the first intron and four in the second intron, that explained 18% of the variation in expression (Table 1).

Decitabine decreases MGMT expression in cell lines with an unmethylated MGMT promoter

We next investigated whether hypomethylating the MGMT gene body correlated with reduced gene expression levels. To do so, we treated lymphoblastoid cell lines that have an unmodified promoter and high levels of MGMT expression with decitabine, a hypomethylating agent. After a five-day treatment with a low dose of decitabine (200 nM), we examined changes in gene expression and gene body cytosine modification levels. We included six lymphoblastoid cell lines that all had relatively equal baseline levels of MGMT expression and gene body cytosine modification: GM7000, GM7357, GM12005, GM12043, GM12810, and GM12812. After decitabine treatment, three lymphoblastoid cell lines (GM7357, GM12005, and GM12810) showed a decrease in MGMT expression, which corresponded to a decrease in cytosine modification levels in the examined regions along the gene body, measured by bisulfite sequencing (Fig. 3A–C, black bars, left). The other three lymphoblastoid cell lines (GM7000, GM12043, and GM12812) showed no change in expression and experienced less hypomethylation across the gene (Fig. 3A–C, white bars, right) after decitabine treatment. These phenotypes were reproducible over two biological replicates, indicating that decitabine was consistently more effective in the same three cell lines. We noted that the growth rates of the three lymphoblastoid cell lines that showed hypomethylation of the MGMT gene body were faster than the other three lymphoblastoid cell lines. It is thought that decitabine is dependent on cell division (25), and the cells that divide more frequently will incorporate more of the drug into the nascent DNA strand. Thus, the difference in MGMT gene body hypomethylation and expression observed between these two groups is likely due to the differences in cellular growth rate. The unmethylated MGMT promoter status remained completely unchanged among the lymphoblastoid cell lines that showed a decrease in MGMT expression, confirming that the change in gene expression was not due to changes within the promoter (Supplementary Fig. 4).

Figure 3.

Figure 3

Open in a new tab

Lymphoblastoid cell lines that show a decrease of MGMT expression after decitabine treatment are more hypomethylated across the gene body. A, qRT-PCR of six lymphoblastoid cell lines after decitabine treatment, normalized to vehicle treated controls. Three show a decrease in MGMT expression after decitabine (DAC) treatment (black bars, left) and three show no change in MGMT expression (white bars, right). B and C, Bisulfite sequencing to measure hypomethylation across a region at the beginning of intron 2 (B) and a region at the end of intron 2 (C). Each vertical bar represents a CpG captured by the bisulfite PCR. The degree of cytosine hypomethylation after decitabine treatment is represented on the y-axis. P-values were calculated using the Student’s t-test.

We also tested this observation in five glioblastoma cell lines. We obtained two glioblastoma cell lines that had high expression levels of MGMT (T98G and SF188), one glioblastoma cell line with low expression of MGMT (U118MG), and two glioblastoma cell lines with no MGMT expression due to promoter hypermethylation (U87MG and A172). It has been reported that IDH1/2 mutation status is associated with distinct methylation and hydroxymethylation patterns (26–28), so we genotyped all six cell lines for IDH1 and IDH2 mutations and found that they were all wild-type for both. Therefore, IDH1/2 mutations could not explain changes in cytosine modifications between these cell lines.

We observed that a five-day treatment of low dose decitabine (200 nM) resulted in a significant decrease in MGMT expression in T98G, SF188, and U118MG (Fig. 4A). This correlated with loss of modified cytosines across regions of the gene body of MGMT, measured by bisulfite sequencing (Fig. 4B). Conversely, U87MG and A172 have a methylated promoter, and decitabine treatment could therefore hypomethylate the promoter and increase MGMT expression. In U87 and A172, we observed some hypomethylation of the promoter region after decitabine treatment (p<0.05) (Supplementary Fig. 5). U87 experienced a very small increase in MGMT expression, yet A172 experienced no increase in MGMT expression (Fig. 4C). This suggests that increased hypomethylation and/or other factors may be required for reactivation of MGMT expression in these cell lines.

Figure 4.

Figure 4

Open in a new tab

Glioblastoma cell lines that express MGMT show a decrease in expression and gene body cytosine modification levels following decitabine treatment. A and B, qRT-PCR of MGMT expression levels in MGMT-expressing glioblastoma cell lines (A) and MGMT-silenced glioblastoma cell lines (B) treated with either 200 nM decitabine (DAC) for 5 days or vehicle controls. Each bar represents the mean of three independent experiments, and error bars denote the standard deviation of the mean. P-values were calculated using the Student’s t-test. C, Bisulfite sequencing of two regions within the gene body of MGMT in MGMT-expression glioblastoma cell lines treated with decitabine or vehicle control. Each pie chart represents the proportion of modified cytosines at a particular CpG, calculated from at least 10 clones. P-values were calculated using a χ2 statistical test. N.S.= not significant.

Pre-treatment with decitabine sensitizes MGMT-expressing glioblastoma cell lines to temozolomide

A significant clinical problem is that glioblastoma patients whose tumors express MGMT, typically with an unmethylated MGMT promoter, respond poorly to temozolomide, which is the drug of choice in this disease. To determine whether the decrease in MGMT expression observed in the decitabine-treated glioblastoma cell lines is sufficient to sensitize the cells to temozolomide, we pre-treated the glioblastoma cell lines with decitabine to determine the effect on cellular sensitivity to temozolomide. The cells were treated with 200 nM of decitabine every 24 hours for five days and then plated for a temozolomide dose-curve cytotoxicity assay. T98G, SF188, and U118MG, each of which experienced a decrease in MGMT expression following the decitabine pre-treatment, were significantly more sensitive to temozolomide compared with vehicle-treated controls (Fig. 5A). Interestingly, decitabine did not affect the sensitivity to temozolomide in either of cell lines that had a methylated MGMT promoter (Fig. 5B). This suggests that the effectiveness of decitabine pre-treatment is more pronounced in glioblastoma cell lines that have an unmethylated MGMT promoter and are initially resistant to temozolomide.

Figure 5.

Figure 5

Open in a new tab

Glioblastoma cell lines that express MGMT are sensitized to temozolomide following decitabine treatment. A, Temozolomide cytotoxicity assay of T98G, SF-188, and U118MG with or without a 5-day pre-treatment with 200 nM decitabine (DAC). Each graph represents at least 2 biological replicates performed in technical triplicates. B, Temozolomide cytotoxicity assay of U87MG and A172 with or without a 5-day pre-treatment with 200 nM decitabine. Each graph represents at least 2 biological replicates performed in technical triplicates.

We also tested the effect of decitabine pre-treatment to cellular sensitivity to temozolomide in lymphoblastoid cell lines that showed both no change (GM7000 and GM12812) and a decrease in MGMT expression (GM7357) following decitabine treatment. We observed that none of the lymphoblastoid cell lines tested were sensitized to temozolomide (Supplementary Fig. 6). Lymphoblastoid cell lines are non-cancerous cells with high levels of MGMT, which makes them highly resistant to temozolomide compared with cancer cell lines, even when the levels of MGMT in lymphoblastoid cell lines are reduced. Thus, decitabine pre-treatment may increase sensitivity of cancer cells to temozolomide with no increase in sensitivity in normal cells, which would potentially reduce unwanted toxicities during combination treatment.

Discussion

In this study, we investigated the role of MGMT gene body cytosine modification in regulating MGMT expression levels. Using primary human glioblastoma data from TCGA, we demonstrated that levels of cytosine modification across the MGMT gene body are associated with expression levels specifically in patients with an unmodified promoter, and those patients with the lowest levels of gene body cytosine modification have similar levels of MGMT expression as those patients with a methylated promoter. Our results may shed light on previous studies that show discordance between MGMT promoter methylation and MGMT protein levels (13, 15), at least in the cases where the promoter is unmethylated yet there is little or no detectable MGMT protein. An unmethylated promoter leaves the MGMT gene open for transcription, but additional transcription factors and chromatin states, including histone marks and DNA methylation and hydroxymethylation, are likely required for transcriptional initiation and elongation. This interplay between epigenetic modifications and transcription factors may contribute to the variation in expression levels seen in glioblastoma patients with an unmethylated promoter.

The importance of gene body cytosine modifications in regulating gene transcription is not completely understood. In fact, the interaction between gene body cytosine modification and expression may be bi-directional. It has been shown that transcriptional elongation may promote DNA methylation along gene bodies (29, 30), while the interaction of 5-mC with histone marks, such as H3K36me3, may be important for transcriptional elongation (29). It is also thought that gene body cytosine modifications may be important in defining exon-intron boundaries and alternative splicing (31). Furthermore, levels 5-mC and 5-hmC located within gene bodies may have distinct effects on gene expression. A recent study demonstrated that, in the brain, the ratio of 5-hmC:5-mC within the gene body is a better predictor of gene expression than either epigenetic mark on their own (32). We propose that gene body cytosine modifications may be one of the factors influencing expression levels of MGMT, such that when the gene body becomes hypomethylated, the cell may not be able to successfully transcribe the gene even when the transcriptional machinery has access to the promoter. This implies that, at least in the case of MGMT, maintaining high levels of cytosine modification along the gene body may be required for the cell to continue to transcribe high levels of the gene. Gaining a better understanding of the functional importance of cytosine modifications along the gene body will help us to predict gene expression patterns better, which could have implications on downstream phenotypes such as response to chemotherapeutic agents.

In addition, previous work by our group investigating population-specific gene expression patterns has shown that MGMT is differentially expressed at baseline between Caucasian and African lymphoblastoid cell lines, with Caucasians expressing lower levels of MGMT compared with Africans (33). However, the overall poor prognosis for all patients has made it difficult to discern if health disparities exist between different ethnic populations (34). Overall, our results indicate that measuring cytosine modification levels in the gene body in addition to those in the promoter will more accurately predict MGMT expression. Importantly, inclusion of gene body cytosine modification levels may allow for better prediction of glioblastoma patient response to temozolomide treatment and long-term survival, because a study of long-term survival of glioblastoma patients showed that 8/36 of glioblastoma long-term survivors had an unmethylated MGMT promoter (35). Our data would suggest that these patients would have low MGMT gene body cytosine modification levels.

The primary model for the mechanism of action of hypomethylating agents decitabine and azacitidine is that they incorporate into DNA to induce global hypomethylation by covalently trapping the maintenance methyltransferase, DNMT1, to the DNA strand (36). Thus, over the course of cell divisions, the genome becomes hypomethylated. Several studies have investigated decitabine-induced hypomethylation of promoter regions and have shown subsequent upregulation of gene expression (37–39). However, decitabine has been shown to induce both upregulation and downregulation of genes (38, 40). One study investigated genome-wide changes in methylation induced by decitabine and azacitidine and showed that hypomethylation had a preference for CpGs that were not in CpG islands (41). Gene body CpGs may be frequently effected by decitabine, but the effects on gene expression may be more subtle than changes in the promoter region, which acts more like an on/off switch. After treatment with decitabine, DNMT1 may not be able to keep levels of methylation high across the MGMT gene body after several replication events, leading to a decrease in gene expression. Although this gene may be more dependent on gene body cytosine modification levels compared with other genes, this phenomenon could occur in other genes as well, and may explain why certain genes become downregulated following decitabine treatment. However, because decitabine causes global hypomethylation, we cannot rule out the possibility that treatment with decitabine is affecting the expression levels of transcription factors involved in MGMT expression, leading to changes in MGMT expression that are independent of the observed changes in gene body methylation levels.

A recent avenue in current strategies for cancer treatment is the combination of hypomethylating agents with cytotoxic chemotherapy to increase response to the chemotherapeutic agent (42–45). The goal of precision medicine is to use patient-specific biomarkers to help determine their individualized treatment strategy. In this work, we have demonstrated that decitabine can be used to sensitize resistant glioblastoma cell lines with an unmethylated MGMT promoter status to temozolomide. Importantly, the effects of decitabine appear to be dependent on the changes to the MGMT expression levels. Cells that express MGMT become hypomethylated across the gene body after decitabine treatment, corresponding with a decrease in expression levels and an increase in temozolomide sensitivity. On the other hand, cells with a hypermethylated MGMT promoter are not sensitized further to temozolomide following decitabine pre-treatment. Although MGMT promoter methylation status is one of the most requested molecular assays in clinical neuro-oncology, there are few therapeutic alternatives to temozolomide for patients with an unmethylated MGMT promoter even though it has been shown that those patients are less likely to respond to treatment (46). In fact, many of the other agents used to treat brain tumors such as carmustine (BCNU), temozolomide, streptozotocin, and dacarbazine also form O6-alkylguanine DNA adducts that are repaired by MGMT (5). Therefore, our strategy of combining decitabine would likely be beneficial with these other agents. In addition, MGMT promoter methylation status and low MGMT expression both were identified as positive prognosticators in patients with newly diagnosed glioblastoma who underwent surgical resection and received Gliadel wafer implants followed by adjuvant radiotherapy and concomitant oral TMZ chemotherapy (47). Finally, we observed that lymphoblastoid cell lines were not sensitized to temozolomide by pre-treatment with decitabine, suggesting that this combination may not significantly increase toxicities in non-cancerous tissues.

Decitabine crosses the blood brain barrier, but the dose of decitabine needed to induce hypomethylation in the brain in human patients is not yet known (48). Decitabine is thought to require cell division to induce hypomethylation, and therefore may only be effective on tumors that are actively proliferating. Thus, more work is needed to determine the appropriate combination of decitabine and temozolomide in human patients. For example, to ensure the cancer cells are exposed to decitabine during cellular divisions, a longer duration of a lower dose may be required. In addition, this drug combination may be beneficial to the survival of patients with other malignancies. Diffuse large B-cell lymphoma is another disease in which the promoter of MGMT is hypermethylated and is predictive of survival among patients treated with alkylating chemotherapy (49). Thus, it will be worthwhile to study this drug combination in a lymphoma model as well. In conclusion, our work argues that pre-treating glioblastoma patients with an unmethylated MGMT promoter with decitabine prior to the start of temozolomide treatment may sensitize their tumors to the chemotherapy.

Supplementary Material

1

Acknowledgments

Financial support: This work was supported by grants from the NIH: R21HG006367 (to WZ, MED, and LAG) and U01GM061393 (to MED).

The authors would like to acknowledge Dr. Heather Wheeler for insightful conversations regarding statistical analyses. This work was supported by grants from the National Institute of Health R21HG006367 (to W.Z., L.A.G., M.E.D.) and U01GM061393 (to M.E.D.). E.L.M. is supported by an HHMI:Med-into-Grad Fellowship (56006772) and a Cancer Biology Training Grant (T32-CA009594).

Footnotes

The authors declare no conflicts of interests.

References

  • 1.Pegg AE, Dolan ME, Moschel RC. Structure, function, and inhibition of O6-alkylguanine-DNA alkyltransferase. Prog Nucleic Acid Res Mol Biol. 1995;51:167–223. doi: 10.1016/s0079-6603(08)60879-x. [DOI] [PubMed] [Google Scholar]
  • 2.Weller M, Steinbach JP, Wick W. Temozolomide: a milestone in the pharmacotherapy of brain tumors. Future Oncol. 2005;1:747–54. doi: 10.2217/14796694.1.6.747. [DOI] [PubMed] [Google Scholar]
  • 3.Hecht SS. DNA adduct formation from tobacco-specific N-nitrosamines. Mutat Res. 1999;424:127–42. doi: 10.1016/s0027-5107(99)00014-7. [DOI] [PubMed] [Google Scholar]
  • 4.Cohen MH, Johnson JR, Pazdur R. Food and Drug Administration Drug approval summary: temozolomide plus radiation therapy for the treatment of newly diagnosed glioblastoma multiforme. Clin Cancer Res. 2005;11:6767–71. doi: 10.1158/1078-0432.CCR-05-0722. [DOI] [PubMed] [Google Scholar]
  • 5.Gerson SL. Clinical relevance of MGMT in the treatment of cancer. J Clin Oncol. 2002;20:2388–99. doi: 10.1200/JCO.2002.06.110. [DOI] [PubMed] [Google Scholar]
  • 6.Kaina B, Christmann M. DNA repair in resistance to alkylating anticancer drugs. Int J Clin Pharmacol Ther. 2002;40:354–67. doi: 10.5414/cpp40354. [DOI] [PubMed] [Google Scholar]
  • 7.Pegg AE. Mammalian O6-alkylguanine-DNA alkyltransferase: regulation and importance in response to alkylating carcinogenic and therapeutic agents. Cancer Res. 1990;50:6119–29. [PubMed] [Google Scholar]
  • 8.Costello JF, Futscher BW, Tano K, Graunke DM, Pieper RO. Graded methylation in the promoter and body of the O6-methylguanine DNA methyltransferase (MGMT) gene correlates with MGMT expression in human glioma cells. J Biol Chem. 1994;269:17228–37. [PubMed] [Google Scholar]
  • 9.Esteller M, Hamilton SR, Burger PC, Baylin SB, Herman JG. Inactivation of the DNA repair gene O6-methylguanine-DNA methyltransferase by promoter hypermethylation is a common event in primary human neoplasia. Cancer Res. 1999;59:793–7. [PubMed] [Google Scholar]
  • 10.Esteller M, Garcia-Foncillas J, Andion E, Goodman SN, Hidalgo OF, Vanaclocha V, et al. Inactivation of the DNA-repair gene MGMT and the clinical response of gliomas to alkylating agents. N Engl J Med. 2000;343:1350–4. doi: 10.1056/NEJM200011093431901. [DOI] [PubMed] [Google Scholar]
  • 11.Hegi ME, Liu L, Herman JG, Stupp R, Wick W, Weller M, et al. Correlation of O6-methylguanine methyltransferase (MGMT) promoter methylation with clinical outcomes in glioblastoma and clinical strategies to modulate MGMT activity. J Clin Oncol. 2008;26:4189–99. doi: 10.1200/JCO.2007.11.5964. [DOI] [PubMed] [Google Scholar]
  • 12.Blanc JL, Wager M, Guilhot J, Kusy S, Bataille B, Chantereau T, et al. Correlation of clinical features and methylation status of MGMT gene promoter in glioblastomas. J Neurooncol. 2004;68:275–83. doi: 10.1023/b:neon.0000033385.37098.85. [DOI] [PubMed] [Google Scholar]
  • 13.Brell M, Tortosa A, Verger E, Gil JM, Viñolas N, Villá S, et al. Prognostic significance of O6-methylguanine-DNA methyltransferase determined by promoter hypermethylation and immunohistochemical expression in anaplastic gliomas. Clin Cancer Res. 2005;11:5167–74. doi: 10.1158/1078-0432.CCR-05-0230. [DOI] [PubMed] [Google Scholar]
  • 14.Komine C, Watanabe T, Katayama Y, Yoshino A, Yokoyama T, Fukushima T. Promoter hypermethylation of the DNA repair gene O6-methylguanine-DNA methyltransferase is an independent predictor of shortened progression free survival in patients with low-grade diffuse astrocytomas. Brain Pathol. 2003;13:176–84. doi: 10.1111/j.1750-3639.2003.tb00017.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Christmann M, Nagel G, Horn S, Krahn U, Wiewrodt D, Sommer C, et al. MGMT activity, promoter methylation and immunohistochemistry of pretreatment and recurrent malignant gliomas: a comparative study on astrocytoma and glioblastoma. Int J Cancer. 2010;127:2106–18. doi: 10.1002/ijc.25229. [DOI] [PubMed] [Google Scholar]
  • 16.Ball MP, Li JB, Gao Y, Lee JH, LeProust EM, Park IH, et al. Targeted and genome-scale strategies reveal gene-body methylation signatures in human cells. Nat Biotechnol. 2009;27:361–8. doi: 10.1038/nbt.1533. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Bell JT, Pai AA, Pickrell JK, Gaffney DJ, Pique-Regi R, Degner JF, et al. DNA methylation patterns associate with genetic and gene expression variation in HapMap cell lines. Genome Biol. 2011;12:R10. doi: 10.1186/gb-2011-12-1-r10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Moen EL, Zhang X, Mu W, Delaney SM, Wing C, McQuade J, et al. Genome-wide variation of Cytosine modifications between European and african populations and the implications for complex traits. Genetics. 2013;194:987–96. doi: 10.1534/genetics.113.151381. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Fritz G, Kaina B. Genomic differences between O6-methylguanine-DNA methyltransferase proficient (Mex+) and deficient (Mex−) cell lines: possible role of genetic and epigenetic changes in conversion of Mex+ into Mex−. Biochem Biophys Res Commun. 1992;183:1184–90. doi: 10.1016/s0006-291x(05)80315-8. [DOI] [PubMed] [Google Scholar]
  • 20.Clark SJ, Harrison J, Paul CL, Frommer M. High sensitivity mapping of methylated cytosines. Nucleic Acids Res. 1994;22:2990–7. doi: 10.1093/nar/22.15.2990. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Mikeska T, Bock C, El-Maarri O, Hübner A, Ehrentraut D, Schramm J, et al. Optimization of quantitative MGMT promoter methylation analysis using pyrosequencing and combined bisulfite restriction analysis. J Mol Diagn. 2007;9:368–81. doi: 10.2353/jmoldx.2007.060167. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Huang Y, Pastor WA, Shen Y, Tahiliani M, Liu DR, Rao A. The behaviour of 5-hydroxymethylcytosine in bisulfite sequencing. PLoS One. 2010;5:e8888. doi: 10.1371/journal.pone.0008888. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Brown CC, Havener TM, Medina MW, Auman JT, Mangravite LM, Krauss RM, et al. A genome-wide association analysis of temozolomide response using lymphoblastoid cell lines shows a clinically relevant association with MGMT. Pharmacogenet Genomics. 2012;22:796–802. doi: 10.1097/FPC.0b013e3283589c50. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Eadon MT, Wheeler HE, Stark AL, Zhang X, Moen EL, Delaney SM, et al. Genetic and epigenetic variants contributing to clofarabine cytotoxicity. Hum Mol Genet. 2013;22:4007–20. doi: 10.1093/hmg/ddt240. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Stresemann C, Lyko F. Modes of action of the DNA methyltransferase inhibitors azacytidine and decitabine. Int J Cancer. 2008;123:8–13. doi: 10.1002/ijc.23607. [DOI] [PubMed] [Google Scholar]
  • 26.Noushmehr H, Weisenberger DJ, Diefes K, Phillips HS, Pujara K, Berman BP, et al. Identification of a CpG island methylator phenotype that defines a distinct subgroup of glioma. Cancer Cell. 2010;17:510–22. doi: 10.1016/j.ccr.2010.03.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Christensen BC, Smith AA, Zheng S, Koestler DC, Houseman EA, Marsit CJ, et al. DNA methylation, isocitrate dehydrogenase mutation, and survival in glioma. J Natl Cancer Inst. 2011;103:143–53. doi: 10.1093/jnci/djq497. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Turcan S, Rohle D, Goenka A, Walsh LA, Fang F, Yilmaz E, et al. IDH1 mutation is sufficient to establish the glioma hypermethylator phenotype. Nature. 2012;483:479–83. doi: 10.1038/nature10866. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Hahn MA, Wu X, Li AX, Hahn T, Pfeifer GP. Relationship between gene body DNA methylation and intragenic H3K9me3 and H3K36me3 chromatin marks. PLoS One. 2011;6:e18844. doi: 10.1371/journal.pone.0018844. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Jones PA. Functions of DNA methylation: islands, start sites, gene bodies and beyond. Nat Rev Genet. 2012;13:484–92. doi: 10.1038/nrg3230. [DOI] [PubMed] [Google Scholar]
  • 31.Laurent L, Wong E, Li G, Huynh T, Tsirigos A, Ong CT, et al. Dynamic changes in the human methylome during differentiation. Genome Res. 2010;20:320–31. doi: 10.1101/gr.101907.109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Mellén M, Ayata P, Dewell S, Kriaucionis S, Heintz N. MeCP2 binds to 5hmC enriched within active genes and accessible chromatin in the nervous system. Cell. 2012;151:1417–30. doi: 10.1016/j.cell.2012.11.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Zhang W, Duan S, Kistner EO, Bleibel WK, Huang RS, Clark TA, et al. Evaluation of genetic variation contributing to differences in gene expression between populations. Am J Hum Genet. 2008;82:631–40. doi: 10.1016/j.ajhg.2007.12.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Curry WT, Barker FG. Racial, ethnic and socioeconomic disparities in the treatment of brain tumors. J Neurooncol. 2009;93:25–39. doi: 10.1007/s11060-009-9840-5. [DOI] [PubMed] [Google Scholar]
  • 35.Krex D, Klink B, Hartmann C, von Deimling A, Pietsch T, Simon M, et al. Long-term survival with glioblastoma multiforme. Brain. 2007;130:2596–606. doi: 10.1093/brain/awm204. [DOI] [PubMed] [Google Scholar]
  • 36.Kihslinger JE, Godley LA. The use of hypomethylating agents in the treatment of hematologic malignancies. Leuk Lymphoma. 2007;48:1676–95. doi: 10.1080/10428190701493910. [DOI] [PubMed] [Google Scholar]
  • 37.Negrotto S, Ng KP, Jankowska AM, Bodo J, Gopalan B, Guinta K, et al. CpG methylation patterns and decitabine treatment response in acute myeloid leukemia cells and normal hematopoietic precursors. Leukemia. 2012;26:244–54. doi: 10.1038/leu.2011.207. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Fabiani E, Leone G, Giachelia M, D’alo’ F, Greco M, Criscuolo M, et al. Analysis of genome-wide methylation and gene expression induced by 5-aza-2′-deoxycytidine identifies BCL2L10 as a frequent methylation target in acute myeloid leukemia. Leuk Lymphoma. 2010;51:2275–84. doi: 10.3109/10428194.2010.528093. [DOI] [PubMed] [Google Scholar]
  • 39.Daskalakis M, Nguyen TT, Nguyen C, Guldberg P, Köhler G, Wijermans P, et al. Demethylation of a hypermethylated P15/INK4B gene in patients with myelodysplastic syndrome by 5-Aza-2′-deoxycytidine (decitabine) treatment. Blood. 2002;100:2957–64. doi: 10.1182/blood.V100.8.2957. [DOI] [PubMed] [Google Scholar]
  • 40.Hollenbach PW, Nguyen AN, Brady H, Williams M, Ning Y, Richard N, et al. A comparison of azacitidine and decitabine activities in acute myeloid leukemia cell lines. PLoS One. 2010;5:e9001. doi: 10.1371/journal.pone.0009001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Hagemann S, Heil O, Lyko F, Brueckner B. Azacytidine and decitabine induce gene-specific and non-random DNA demethylation in human cancer cell lines. PLoS One. 2011;6:e17388. doi: 10.1371/journal.pone.0017388. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Plimack ER, Stewart DJ, Issa JP. Combining epigenetic and cytotoxic therapy in the treatment of solid tumors. J Clin Oncol. 2007;25:4519–21. doi: 10.1200/JCO.2007.12.6029. [DOI] [PubMed] [Google Scholar]
  • 43.Gravina GL, Festuccia C, Marampon F, Popov VM, Pestell RG, Zani BM, et al. Biological rationale for the use of DNA methyltransferase inhibitors as new strategy for modulation of tumor response to chemotherapy and radiation. Mol Cancer. 2010;9:305. doi: 10.1186/1476-4598-9-305. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Appleton K, Mackay HJ, Judson I, Plumb JA, McCormick C, Strathdee G, et al. Phase I and pharmacodynamic trial of the DNA methyltransferase inhibitor decitabine and carboplatin in solid tumors. J Clin Oncol. 2007;25:4603–9. doi: 10.1200/JCO.2007.10.8688. [DOI] [PubMed] [Google Scholar]
  • 45.Scandura JM, Roboz GJ, Moh M, Morawa E, Brenet F, Bose JR, et al. Phase 1 study of epigenetic priming with decitabine prior to standard induction chemotherapy for patients with AML. Blood. 2011;118:1472–80. doi: 10.1182/blood-2010-11-320093. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.von Deimling A, Korshunov A, Hartmann C. The next generation of glioma biomarkers: MGMT methylation, BRAF fusions and IDH1 mutations. Brain Pathol. 2011;21:74–87. doi: 10.1111/j.1750-3639.2010.00454.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Lechapt-Zalcman E, Levallet G, Dugué AE, Vital A, Diebold MD, Menei P, et al. O(6)-methylguanine-DNA methyltransferase (MGMT) promoter methylation and low MGMT-encoded protein expression as prognostic markers in glioblastoma patients treated with biodegradable carmustine wafer implants after initial surgery followed by radiotherapy with concomitant and adjuvant temozolomide. Cancer. 2012;118:4545–54. doi: 10.1002/cncr.27441. [DOI] [PubMed] [Google Scholar]
  • 48.Momparler RL. Pharmacology of 5-Aza-2′-deoxycytidine (decitabine) Semin Hematol. 2005;42:S9–16. doi: 10.1053/j.seminhematol.2005.05.002. [DOI] [PubMed] [Google Scholar]
  • 49.Esteller M, Gaidano G, Goodman SN, Zagonel V, Capello D, Botto B, et al. Hypermethylation of the DNA repair gene O(6)-methylguanine DNA methyltransferase and survival of patients with diffuse large B-cell lymphoma. J Natl Cancer Inst. 2002;94:26–32. doi: 10.1093/jnci/94.1.26. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS571143-supplement-1.pdf (538.6KB, pdf)

RESOURCES