Promoter methylation analysis of O6-methylguanine-DNA methyltransferase in glioblastoma: detection by locked nucleic acid based quantitative PCR using an imprinted gene (<it>SNURF</it>) as a reference

Abstract

Abstract Background Epigenetic silencing of the MGMT gene by promoter methylation is associated with loss of MGMT expression, diminished DNA-repair activity and longer overall survival in patients with glioblastoma who, in addition to radiotherapy, received alkylating chemotherapy with carmustine or temozolomide. We describe and validate a rapid methylation sensitive quantitative PCR assay (MS-qLNAPCR) using Locked Nucleic Acid (LNA) modified primers and an imprinted gene as a reference. Methods An analysis was made of a database of 159 GBM patients followed between April 2004 and October 2008. After bisulfite treatment, methylated and unmethylated CpGs were recognized by LNA primers and molecular beacon probes. The SNURF promoter of an imprinted gene mapped on 15q12, was used as a reference. This approach was used because imprinted genes have a balanced copy number of methylated and unmethylated alleles, and this feature allows an easy and a precise normalization. Results Concordance between already described nested MS-PCR and MS-qLNAPCR was found in 158 of 159 samples (99.4%). The MS-qLNAPCR assay showed a PCR efficiency of 102% and a sensitivity of 0.01% for LNA modified primers, while unmodified primers revealed lower efficiency (69%) and lower sensitivity (0.1%). MGMT promoter was found to be methylated using MS-qLNAPCR in 70 patients (44.02%), and completely unmethylated in 89 samples (55.97%). Median overall survival was of 24 months, being 20 months and 36 months, in patients with MGMT unmethylated and methylated, respectively. Considering MGMT methylation data provided by MS-qLNAPCR as a binary variable, overall survival was different between patients with GBM samples harboring MGMT promoter unmethylated and other patients with any percentage of MGMT methylation (p = 0.003). This difference was retained using other cut off values for MGMT methylation rate (i.e. 10% and 20% of methylated allele), while the difference was lost when 50% of MGMT methylated allele was used as cut-off. Conclusions We report and clinically validate an accurate, robust, and cost effective MS-qLNAPCR protocol for the detection and quantification of methylated MGMT alleles in GBM samples. Using MS-qLNAPCR we demonstrate that even low levels of MGMT promoter methylation have to be taken into account to predict response to temozolomide-chemotherapy.