Abstracts

Rationale: Only 35% of epilepsy cases have a clear acquired cause and the remaining 65% were historically termed “idiopathic” and of presumed genetic origin. While there have been many genetic discoveries in recent years, there remains a void in our understanding of the biological mechanisms underpinning these epilepsies and the role of epigenetic mechanisms warrants deeper reflection. Genetic factors are unlikely when studying discordant monozygotic (MZ) twin pairs, and this allows focus on mechanisms such as epigenetics, an environment-controlled ‘thermostat’ of gene activity. This work represents the first twin epigenetic study focusing on peripheral biomarkers for epilepsy. Epilepsy biomarkers do not necessarily need to resemble brain methylation patterns, but rather be an accessible and reproducible peripheral indicator of disease.  Methods: 15 epilepsy-discordant MZ twin pairs without a known acquired cause were recruited, and affected individuals classified into generalized and focal (benign rolandic, idiopathic photosensitive occipital and non-lesional temporal lobe) epilepsies based on clinical history, electroencephalogram and neuroimaging findings. DNA obtained from buccal samples was utilized for genome-wide methylation analysis using the Illumina Infinium EPIC array and the raw data was normalized and analysed to yield a list of differentially methylated probes (DMPs) ranked by false discovery rate (FDR) values. Differentially methylated regions (DMRs) were determined using the BumpHunter package, part of Bioconductor in R. Gene ontology analysis was performed using the gometh function in missmethyl package in R to identify molecular function and biological processes that were over-represented by the top differentially methylated sites. Pathway analysis was performed using Ingenuity Pathways Analysis (IPA) software to identify underlying biological pathways corresponding to DMPs and DMRs. Results: DNA methylation analysis of the epilepsy subtypes (generalized vs focal) of the affected cases demonstrated clear hierarchical clustering of differentially methylated CpG sites between the subtypes. Methylation analysis of discordant generalized epilepsy twin pairs (generalized vs unaffected) demonstrated two top DMR-associated genes: orthodenticle homeobox 1 (OTX1) and glial cell line-derived neurotrophic factor (GDNF), both having large and consistent effect sizes across all pairs (6.7-16.5% for OTX1; 5.2-15.5% for GDNF). DNA methylation analysis of twin pairs discordant for focal epilepsy (focal vs unaffected) found distal-less homeobox 5 (DLXR5) to be the top DMR-associated gene with an effect size of between 5% and 8.9%. Biological function of the top DMPs and DMRs and their study in animal models support these regions as plausible candidate biomarkers for epilepsy.  Conclusions: The distinctive methylation clustering demonstrated when comparing generalized and focal epilepsies supports divergent biological mechanisms underpinning these sub-types. A deeper understanding of the mechanisms driving epilepsy will enable focus on peripheral biomarkers for earlier diagnosis and treatment choice, therapeutic targets and ultimately disease-modifying treatments.  Funding: This research is funded by a Health Research Fellowship, The Royal Brisbane and Women’s Hospital Research Foundation and The University of Queensland Mayne Bequest Funds.