Epigenetic and Related Transcriptional Alterations Affecting Chromatin Remodelling and Synaptic Genes in Autism Spectrum Disorders

Background:  Autism Spectrum Disorders (ASD) are highly heritable and genetically complex conditions with likely environmental contribution. Although multiple candidate genes have been identified, they only account for the aetiology of 15-20% of cases so far. Additional evidence in support of an epigenetic contribution to ASD include the epigenetic deregulation in single-gene disorders associated with autism (i.e. Fragile X and Rett syndromes) and alterations in several chromosomal regions subjected to imprinting (chr7q, 15q). High throughput methylome studies have recently revealed a possible contribution of epigenetics to the molecular basis of ASD.

Objectives:  To explore the role of epigenetics in ASD by defining genome-wide methylation alterations along with their correlation with genetic and/or expression abnormalities in a cohort of idiopathic ASD. 

Methods:  We quantified global 5-methylcytosine (5mC) content by HPLC-MS and studied genomic methylation patterns by using the Illumina Infinium HumanMethylation 450K array in peripheral blood DNA from ASD patients (n=34, males, 2-15 year old, idiopathic aetiology) and controls (n=101). We also studied transcriptome expression in whole blood by RNAseq (n=19), and exclude possible genetic alterations (aCGH and exome sequencing). Functional annotation by enriched gene ontology-based analysis was performed with CPDB.

Results: We found that the genomic content of 5mC was significantly reduced in ASD with respect to age-matched controls (2.62+/-0.4 vs 4.06+/-0.16). The methylation high-throughput profile did not show remarkable global epigenetic alterations. After extensive filtering (discarding genetic alterations) and statistical analyses, we observed 427 genes containing differentially methylated CpGs (DMCpG) compared to controls (p-value<0.01). Most DMCpG showed relative hypomethylation in ASD (90%), in agreement with 5mC quantification. Among these 427 loci, 41 (10%) had been previously associated with ASD by genetic or genomic rearrangements, further reinforcing their putative role in the disease. Functional annotation revealed enrichment of genes involved in the glutamate neurotransmitter release cycle, mTOR signalling, synaptic plasticity, neuron projection, synapse organization and axon guidance.

Analysis of the putative transcriptional effects in cis, 50Kb up or downstream the target CpG, revealed expression changes in 10% of surrounding genes, including additional candidates coding for chromatin remodelling complexes (HR and SETD1A), synaptic transmission (STX1B), or linked to ASD by genetic studies (AGRP).

The analysis of the most relevant CpG sites by scalable techniques in a larger cohort of patients and controls is in process. 

Conclusions: Our data provide support for the hypotheses that epigenetic abnormalities may be pathogenically related with ASD in a significant proportion of cases, either as primary or secondary (genetic) events. Genes acting at the synapse and chromatin remodelling factors are differentially methylated in autism and show abnormal expression.