Quantitative analysis of O6-alkylguanine-DNA alkyltransferase in malignant glioma.

Promoter hypermethylation of the DNA repair protein O(6)-alkylguanine-DNA alkyltransferase (AGT) has been associated with an enhanced response to chloroethylating and methylating agents in patients with malignant glioma. The purpose of this study was to compare three distinct yet related indices for measuring AGT to determine if these assays could be used interchangeably when AGT status is to be used to guide chemotherapeutic decisions. Real-time methylation-specific PCR (MSP), assessed as the ratio of methylated AGT copies to internal beta-actin control, was used to quantitate AGT hypermethylation in 32 glioma samples. Data were compared with AGT enzyme activity as well as immunohistochemical detection of AGT protein from the same samples. Hypermethylation of the AGT promoter was detected in 19 of 31 (61%) samples evaluable by MSP. Low-level AGT, defined as <20% nuclear AGT staining by immunohistochemistry, was found in 10 of 32 samples (31%), whereas 12 of 32 (38%) had low levels of AGT activity. Correlation of immunohistochemistry to AGT activity was statistically significant (P = 0.014) as was the correlation of immunohistochemistry to MSP (P = 0.043), whereas MSP compared with AGT activity (P = 0.246) was not significant. Cross-tabulation of immunohistochemistry and MSP data based on prognostic groups, where good prognosis was represented by an immunohistochemistry of <20% and an MSP ratio >12, showed no significant relationship (P = 0.214), suggesting that one assay cannot be used interchangeably for another. The observed discordance between respective measures of AGT based on prognosis supports further standardization of AGT assays designed to guide therapeutic practice. The data also suggest that consideration be given to the large population of AGT-expressing cells within samples when therapeutic strategies based on tumor methylation are used.