Monash ‘Neuroscience in a Flash’ Competition 2020

Title of presentation: Using DNA methylation patterns to predict disease severity in multiple sclerosis
Department: CCS, Alfred

Abstract: Multiple sclerosis (MS) is a chronic autoimmune and neurodegenerative disease. Currently, disease management is a challenge for clinicians as they cannot predict an individual’s likely outcome at diagnosis. Published genetic biomarkers of disease severity explain up to 33% of variation in severity, suggesting that it is heavily influenced by environmental factors. Epigenetics is the biological interface of the interaction between genetic and environmental factors. DNA methylation is a dynamic and reversible epigenetic mechanism that regulates gene transcription, and is heavily influenced by environmental factors. Therefore, we hypothesised that DNA methylation plays a role in regulating disease severity in MS. In a cohort of Australian females with relapse-onset MS (n=235), we tested the association between DNA methylation patterns in whole-blood, and immune cell subsets, with disease severity. Association testing was conducted with the Chip Analysis Methylation Pipeline (ChAMP) and a false-discovery rate (FDR) of 0.05 was used to assess statistical significance. We further investigated the ability of DNA methylation to predict disease severity. We identified 745 differentially methylated positions (DMPs), enriched in the PDGFR-beta signalling pathway, and 15 differentially methylated regions (DMRs) in whole-blood. One DMR, CXCR4, has previously been associated with MS pathology. We identified 95 cell-specific DMPs in eosinophils. Using machine learning, we showed that methylation can be used to predict disease severity with an AUC of 0.91. Our results demonstrate that DNA methylation patterns are associated with, and can be used to predict, disease severity in Australian females with relapse-onset MS. Using DNA methylation patterns to predict disease severity in multiple sclerosis.