Dysregulation of Cortical Neuron DNA Methylation in Autism: Implication of Gabaergic and Immune System-Related Genes

Background: Epigenetics has been implicated as a primary molecular mechanism that may mediate the interaction between environmental factors and the development of ASD. Previous epigenetic studies in post-mortem brains have been severely limited by the cellular heterogeneity of the brain tissue, which impairs our ability to determine specific and robust epigenetic changes in the brain.

Objectives:  Our main objective is to determine changes in the epigenetic signature in cortical neurons of individuals diagnosed with autism, compared to matched controls. Further objectives include to understand the biological function of the dysregulated epigenetic signature and to determine if genomic regions that display epigenetic plasticity during brain development are particularly sensitive to dysregulation in ASD.

Methods: To investigate methylation changes specifically in neurons from the frontal cortex of autistic and control subjects, we employed Fluorescent Activated Cell Sorting (FACS) of neuronal nuclei from post-mortem brains, followed by hybridization of DNA on 450K BeadArray. Bumphunting analysis (using CHAMP tool) was performed to find differentially methylated regions and WGCNA analysis was performed to determine co-methylated modules of CpGs that correlate with autism status, followed by gene ontology and protein-protein interaction analysis. Targeted Bisulfite Sequencing was used to validate differentially methylated regions. Differentially methylated regions and correlated modules were compared to published datasets of CpGs that are differentially methylated during neurodevelopment.

Results: We identified 58 Differentially Methylated Regions (DMRs) at FDR <0.05 that included genomic loci associated to GABAergic system genes, in particular ABAT and GABBR1, and brain-specific MicroRNAs, Mir124-1 and Mir124-2. We verified a subset of these DMRs by targeted bisulfite sequencing. At system level, Weighted Co-Methylation Network Analysis (WGCNA) detected three major modules significantly correlated to ASD. Two modules (p=2e-04; p=0.008) were inversely correlated with ASD and were enriched for regions underlying neuronal genes (synaptic genes, as well as GABAergic genes), whereas one module (p=0.001) showed direct correlation and was enriched for regions underlying immune genes. Protein-protein interaction networks determined a subset of neurotransmitter-related genes that were enriched in several of these modules. In a comparison between our data and published data on neurodevelopment-associated DMRS, A remarkable overlap of the 58 autism-related DMRs with age-specific (from embryonic to late-fetal neurodevelopment) and cell-type specific (neuron; glia) DMRs was observed. Finally, we established the specificity of these three modules to ASD by assessing their enrichment for GWAS databases related to other psychiatric and non-psychiatric disorders.

Conclusions: We have identified multiple genomic regions that display dysregulated DNA methylation in cortical neurons from individuals with ASD. These regions are highly relevant to biological mechanisms that have been implicated in autism. Our study suggests that robust analysis of specific cells in the brain is an important goal in understanding the epigenetics and molecular biology of the development of ASD. Our study identifies alterations of DNA methylation in cortical neurons as a plausible environment-mediated component to ASD aetiology.