Abstract / Summary
Global incidence and outcomes of glioma have been found to vary significantly by region, however research into the disease continues to lack diversity. Here we investigated epigenetic patterns in glioma subtypes from cohorts collected from China and the USA. We retrospectively analysed the Chinese Glioma Genome Atlas (CGGA) and The Cancer Genome Atlas (TCGA) datasets following reclassification of glioma subtypes based on the WHO 2021 central nervous system (CNS) tumour classification. We used DNA methylation and transcriptomics data to identify methylation-driven cancer genes in the CGGA cohort, assessed their prognostic value and compared against the non-Hispanic White cohort in the TCGA database to consider ethnic influence. Furthermore, we used machine learning classification and clustering techniques to identify methylation patterns in glioma subgroups. Here, we showed that DNA methylation profiles of CGGA glioblastomas have a methylation signature more similar to TCGA high-grade astrocytomas: 36.7% of CGGA glioblastomas (n=49) were identified as high-grade astrocytomas using classification modelling. Assessment of survival revealed that CGGA glioblastoma patients had a significantly better survival rate than non-Hispanic White glioblastoma patients (p = 0.037). Four key methylation-driven genes were identified in the CGGA glioblastoma samples: GLDN, PRKDC, S100A1 and NCAPH. Hypermethylation of GLDN significantly suppressed gene expression in all glioma subtypes in only the East Asian cohort; a gene that has not been previously described as a driver in gliomas. Together these data suggest alternative epigenetic mechanisms occurring in glioma subtypes of different ethnic populations, which is important for our understanding of glioma and strategies for personalized treatment.