O-methylguanine-DNA methyltransferase (MGMT) mRNA expression predicts outcome in malignant glioma independent of MGMT promoter methylation

Abstract

Background: We analyzed prospectively whether MGMT (O(6)-methylguanine-DNA methyltransferase) mRNA expression gains prognostic/predictive impact independent of MGMT promoter methylation in malignant glioma patients undergoing radiotherapy with concomitant and adjuvant temozolomide or temozolomide alone. As DNA-methyltransferases (DNMTs) are the enzymes responsible for setting up and maintaining DNA methylation patterns in eukaryotic cells, we analyzed further, whether MGMT promoter methylation is associated with upregulation of DNMT expression.

Methodology/principal findings: ADULT PATIENTS WITH A HISTOLOGICALLY PROVEN MALIGNANT ASTROCYTOMA (GLIOBLASTOMA: N = 53, anaplastic astrocytoma: N = 10) were included. MGMT promoter methylation was determined by methylation-specific PCR (MSP) and sequencing analysis. Expression of MGMT and DNMTs mRNA were analysed by real-time qPCR. Prognostic factors were obtained from proportional hazards models. Correlation between MGMT mRNA expression and MGMT methylation status was validated using data from the Cancer Genome Atlas (TCGA) database (N = 229 glioblastomas). Low MGMT mRNA expression was strongly predictive for prolonged time to progression, treatment response, and length of survival in univariate and multivariate models (p<0.0001); the degree of MGMT mRNA expression was highly correlated with the MGMT promoter methylation status (p<0.0001); however, discordant findings were seen in 12 glioblastoma patients: Patients with methylated tumors with high MGMT mRNA expression (N = 6) did significantly worse than those with low transcriptional activity (p<0.01). Conversely, unmethylated tumors with low MGMT mRNA expression (N = 6) did better than their counterparts. A nearly identical frequency of concordant and discordant findings was obtained by analyzing the TCGA database (p<0.0001). Expression of DNMT1 and DNMT3b was strongly upregulated in tumor tissue, but not correlated with MGMT promoter methylation and MGMT mRNA expression.

Conclusions/significance: MGMT mRNA expression plays a direct role for mediating tumor sensitivity to alkylating agents. Discordant findings indicate methylation-independent pathways of MGMT expression regulation. DNMT1 and DNMT3b are likely to be involved in CGI methylation. However, their exact role yet has to be defined.

Figure 1. MGMT mRNA expression and MGMT promoter methylation status in malignant glioma.

Horizontal bars indicate medians. The dotted line indicates the cut-off value distinguishing low from high MGMT expression values. A: Expression of MGMT determined with quantitative real-time PCR in non-cancerous brain tissue specimen (1, N = 10), in malignant glioma (2, N = 63), and in the glioma group stratified by MGMT promoter methylation status (methylated, 3, N = 32, and unmethylated, 4, N = 31). cDNA was synthesized from amplified RNA purified from tumor tissue obtained by stereotactic biopsy or open surgery and relative expression of MGMT with respect to expression of the reference genes SDHA and TBP was determined using real-time PCR. B: Validation set obtained from the TCGA database. All data were derived from array analyses and expression levels of methylated (3, N = 105) and unmethylated (4, N = 104) GBM tissue samples are shown.

Figure 2. Expression of DNMTs in non-cancerous brain tissue specimen and in malignant glioma.

cDNA was synthesized from amplified RNA purified from tumor tissue obtained by stereotactic biopsy or open surgery and expression of DNMTs was determined using real-time PCR. A: Expression pattern of DNMTs in normal brain tissue (1, N = 10) and in high grade glioma (2, N = 63). DNMT mRNA expression is calculated relative to the reference genes SDHA and TBP (*, p>0.01). B: Expression of DNMTs in non-cancerous brain tissue specimen (1, N = 10), in high grade glioma (2, N = 63), and in the glioma group stratified by MGMT promoter methylation status (methylated, 3, N = 32; unmethylated, 4, N = 31). All data were normalized to the reference genes SDHA and TBP, and the fold change of every DNMT relative to the median expression in the normal brain samples (arbitrarily set to 1) was calculated. Horizontal bars indicate medians.

Figure 3. Kaplan-Meier estimates of 63 patients with malignant glioma.

Tumor tissue obtained either by stereotactic biopsy or by open surgery. A: Progression free survival and overall survival of the whole study population, B: Overall survival stratified by the MGMT promoter methylation status.

Figure 4. Kaplan-Meier estimates of patients with malignant glioma stratified by MGMT mRNA expression.

Tumor tissue obtained either by stereotactic biopsy or by open surgery A: Progression free survival (N = 63), B: Overall survival (N = 63), C: Progression free survival of patients with methylated GBMs (N = 24) D: Survival of patients with methylated GBMs (N = 24).