A genome-wide association analysis of temozolomide response using lymphoblastoid cell lines reveals a clinically relevant association with MGMT

Recently, lymphoblastoid cell lines (LCLs) have emerged as an innovative model system for mapping gene variants that predict dose response to chemotherapy drugs. In the current study, this strategy was expanded to the in vitro genome-wide association approach, using 516 LCLs derived from a Caucasian cohort to assess cytotoxic response to temozolomide. Genome-wide association analysis using approximately 2.1 million quality controlled single-nucleotide polymorphisms (SNPs) identified a statistically significant association (p < 10 −8 ) with SNPs in the O 6 -methylguanine–DNA methyltransferase ( MGMT ) gene. We also demonstrate that the primary SNP in this region is significantly associated with differential gene expression of MGMT (p< 10 −26 ) in LCLs, and differential methylation in glioblastoma samples from The Cancer Genome Atlas. The previously documented clinical and functional relationships between MGMT and temozolomide response highlight the potential of well-powered GWAS of the LCL model system to identify meaningful genetic associations.