Abstract
In light with the view that KEAP1 loss of function may impact tumour behavior and modify response to chemotherapeutical agents, we sought to determine whether KEAP1 gene is epigenetically regulated in malignant gliomas. We developed a Quantitative Methylation Specific PCR (QMSP) assay to analyze 86 malignant gliomas and 20 normal brain tissues. The discriminatory power of the assay was assessed by Receiving Operating Characteristics (ROC) curve analysis. The AUC value of the curve was 0.823 (95%CI: 0.764-0.883) with an optimal cut off value of 0.133 yielding a 74% sensitivity (95%CI: 63%-82%) and an 85% specificity (95%CI: 64%-95%). Bisulfite sequencing analysis confirmed QMSP results and demonstrated a direct correlation between percentage of methylated CpGs and methylation levels (Spearman's Rho 0.929, P=0.003). Remarkably, a strong inverse correlation was observed between methylation levels and KEAP1 mRNA transcript in tumour tissue (Spearman's Rho -0.656 P=0.0001) and in a cell line before and after treatment with 5-azacytidine (P=0.003). RECPAM multivariate statistical analysis studying the interaction between MGMT and KEAP1 methylation in subjects treated with radiotherapy and temozolomide (n=70), identified three prognostic classes of glioma patients at different risk to progress. While simultaneous methylation of MGMT and KEAP1 promoters was associated with the lowest risk to progress, patients showing only MGMT methylation were the subgroup at the higher risk (HR 5.54, 95% CI 1.35-22.74). Our results further suggest that KEAP1 expression is epigenetically regulated. In addition we demonstrated that KEAP1 is frequently methylated in malignant gliomas and a predictor of patient's outcome.