The link between methylation age acceleration and incidence of brain glioblastoma: Two-sample bidirectional and multivariable Mendelian randomization analyses.

173

Background: Glioblastoma is a highly aggressive primary brain tumor strongly associated with increasing chronological age; however, the role of biological aging in its etiology remains unclear. Epigenetic age acceleration clocks and telomere length reflect complementary dimensions of biological aging and have been linked to cancer risk. Here, we aim to evaluate potential causal relationships between methylation age acceleration, telomere length, and glioblastoma risk. Methods: We performed bidirectional two-sample Mendelian Randomization (MR) to assess causal effects between epigenetic age acceleration measures: Horvath Intrinsic Age (24 single nucleotide polymorphisms [SNPs]; n = 34461), HannumAge (9 SNPs; n = 34449), PhenoAge (11 SNPs; n = 34463), and GrimAge (4 SNPs; n = 34467); telomere length (150 SNPs; n = 472174); and glioblastoma. Genetic instruments were obtained from IEU OpenGWAS, and glioblastoma summary statistics were derived from FinnGen finn-b-C3_GBM (150 SNPs; 91 cases, 174006 controls). Independent genome-wide significant SNPs were selected (P < 5 × 10 ^-8 ) and clumped for linkage disequilibrium (r² < 0.001; 10,000 kb window). Causal estimates were calculated using the inverse-variance weighted (IVW) method. Multivariable MR (MVMR) was conducted to estimate independent effects while adjusting telomere length for each of the epigenetic age acceleration measures. All datasets used included data exclusively from European populations. Results: No evidence of a causal association was observed between methylation age acceleration and glioblastoma risk: Intrinsic Age (odds ratio [OR] = 1.198 [0.916–1.567]), HannumAge (OR = 1.095 [0.667–1.799]), PhenoAge (OR = 1.237 [0.885–1.729]), and GrimAge (OR = 0.541 [0.245–1.195]). Reverse MR analyses showed no causal effect of glioblastoma on epigenetic age acceleration: Intrinsic Age (OR = 1.113 [0.945–1.312]), HannumAge (OR = 1.090 [0.969–1.226]), PhenoAge (OR = 1.101 [0.676–1.793]), and GrimAge (OR = 1.104 [0.971–1.254]). Telomere length demonstrated a significant association with increased glioblastoma risk in univariable MR (OR = 3.555 [1.056–11.]). However, after adjustment for Horvath Intrinsic Age, HannumAge, PhenoAge, and GrimAge in MVMR, this association was no longer statistically significant (p > 0.05). Conclusions: This study does not suggest a causal role for methylation age acceleration in glioblastoma development. Although longer telomere length was associated with increased glioblastoma risk in univariable analysis, this effect was not independent of epigenetic aging measures. However, these findings should be interpreted cautiously given potential limitations in instrument strength and outcome GWAS power.