The impact of germline variants in DNA repair pathways on survival and temozolomide toxicity in adults with glioma

Abstract

Gliomas are highly fatal malignant brain tumors with prognostically relevant molecular subtypes. Genomic instability is a hallmark of cancer, including glioma, and DNA repair mechanisms are crucial to maintaining cell integrity. Temozolomide (TMZ), which utilizes specific DNA repair pathways in its mechanism of action, is one of the few standard-of-care drugs with a meaningful impact on IDH wildtype glioblastoma survival. However, its utility in non-glioblastomas is complicated by temozolomide-induced hypermutation. This study aims to identify if germline polymorphisms in DNA repair genes are associated with overall survival within glioma subtypes and if the polymorphisms modify survival in subjects treated with temozolomide. We utilized genotype data of 2078 adults with glioma collected through the UCSF Adult Glioma Study, Mayo Clinic, and TCGA, to study 1393 SNPs within 22 specific DNA repair genes on overall survival for all cases and for those with/without temozolomide exposure during first course of treatment. All models were fit separately for individuals grouped by major molecular subtypes. We identified four germline SNPs (within MLH1, MSH4, MSH3, and MUTYH) with a significant (p<0.00019) polygenic effect on survival, where the presence of at least one SNP decreased overall survival among grade 2 and 3 glioma cases with IDH mutated tumors (Hazard ratio per unit increase in number of present polymorphic sites, HR-PS=1.52, 95% confidence interval, CI: 1.26-1.83, p=1.3x10-5). In IDH mutant gliomas without somatic 1p/19q chromosomal arm codeletion, we identified a group of SNPs (in MSH4, ERCC2, and ERCC1) which showed a stepwise significant improvement on overall survival for each present germline polymorphism (HR-PS=0.67, 95% CI: 0.56-0.80, p=8.9x10-6). Within IDH mutant glioma cases, we also identified two mutations (rs1540354-T in MLH1, rs71636247-G in MSH3) which were significantly associated with decreased overall survival only amongst those with known temozolomide usage (HR-PS=2.23, 95% CI: 1.55-3.20, p=1.56x10-5), with no survival effect observed in the non-temozolomide group (HR=0.90, 95% CI: 0.63-1.28, p=0.56). One SNP in MSH2 was suggestively associated (p<0.005) with altered IDH wildtype glioblastoma response to temozolomide (rs149630102-T, HR=1.74, 95% CI: 1.27-2.39, p=0.00052). Suggestive associations for other glioma subtypes are also reported. We found evidence that germline polygenic alterations within DNA repair pathways may alter prognosis and potentially modulate temozolomide toxicity in adults with glioma. Further validation and functional work are warranted to assess if these markers can assist clinicians in individually tailoring therapeutic treatments for adults with glioma tumors.