Abstracts

Rationale: Autism and epilepsy are common complex disorders which independently result in significant behavioral and developmental problems. Their co-occurrence is conservatively estimated at 30%. Biologic mechanisms that account for this co-occurrence have eluded discovery. Several conceptual models have proposed a common brain pathology in which autism and epilepsy are independent consequences of the same underlying disorder. Given the overlap in these two disorders we proposed that epilepsy risk genes could be etiologically relevant to autism. The objective of this study was to test the hypothesis that epilepsy related candidate genes may confer risk to autism. Methods: Using existing genome wide association study (GWAS) data, we examined 33 candidate genes, selected on the basis of previous reports of association or biological relevance to epilepsy or epilepsy and co-occurring autism. The discovery dataset consisted of 438 autism families from the Hussman Institute for Human Genomics (HIHG) autism program genotyped on the Illumina 1M chip. The validation dataset consisted of 457 autism families from the Autism Genetics Resource Exchange (AGRE) genotyped on the Illumina 550K. The 1M Beadchip is redundant to the 550K Beadchip with the addition of approximately 500,000 more SNPs. Results: We examined single nucleotide polymorphisms (SNPs) in our autism GWAS dataset for each of candidate genes. All SNPs were tested for association to autism using a family based test for association. A finding was declared significant if a SNP was nominally significant in the HIHG and AGRE datasets and showed greater significance in the joint analysis. Two SNPs in CACNA1G (rs11079919 and rs9898731) were significant in the HIHG, AGRE, and joint analyses. Both are in CACNA1G but are in intronic regions and are in high linkage disequilibrium (LD). A third SNP, (rs2240119), while not significant in the HIHG dataset, is significant in the AGRE dataset (p=0.002) and highly significant (9.55E-04) in the joint analysis. Additional analysis of an ASD-epilepsy only dataset (N=71) revealed SNPs of interest in GABRG2, GABRB3, and SCN2A Conclusions: Testing genes with biological relevance to epilepsy yielded a significant association to SNPs in CACNA1G, a calcium channel gene which has recently been implicated in autism as well as idiopathic generalized epilepsy. The role of ion channel genes in autism risk is supported by evidence showing that calcium channel dysfunction is tied to both non-syndromic and syndromic autism (e.g., Timothy Syndrome, a multisystem disorder characterized by cardiac, immune, and cognitive abnormalities along with a clearly defined autism phenotype). Calcium dependent defects that perturb neural development lead to changes common to those found in autism (e.g., cell-packing density, decreases in neuron size and arborization, and alterations in connectivity. Further, calcium channel variants in autism (e.g.CACNA1G) are tied to increased calcium signaling suggesting a role for calcium dependent activation in this disorder. The role of sodium channel and GABA-ergic variants are also of great interest.