Abstracts

Rationale: Ion channels regulate function in all brain networks and form the targets for most antiepileptic drugs. Despite their central role in physiology and pharmocotherapy, the genetic variation in this candidate gene set, composed of over 400 members and comprising almost 1% of the human genome is substantially understudied. Because 17/20 confirmed monogenic Mendelian epilepsy syndromes arise in individuals with a mutations in a single channel subunit gene, we explored the personal genetic profiles of individuals with epilepsy of unknown cause (idiopathic epilepsy, IE) and compared them to neurologically healthy individuals to assess ion channel variation as a contributor to personal risk in epilepsy. We then utilized computational approaches to simulate individuals with combinations of multiple alleles to better understand the relative pathogenicity of known deleterious mutations. Methods: We performed parallel exomic Sanger sequencing of 237 candidate ion channel genes in a clinically characterized cohort of nearly 300 individuals, comparing the channotype variant profiles between those with sporadic idiopathic epilepsy (IE) and normal neurologically healthy controls. Computational (Traub) models of a single hippocampal pyramidal neuron were used to combinatorially evaluate differing variant patterns with equivalent mutational load demonstrating the differential effects on firing properties.Results: We discovered rich patterns of variation in all individuals regardless of disease status, including rare missense variation in many known Mendelian disease genes. Each channotype was unique, with complex overlapping variant portfolios comprised of both rare and common nonsynonymous SNPS. Some individuals carried mutations in as many as 7 known Mendelian epilepsy channel genes. In some cases, multiple variants were found in a single gene. When the severity of channels alleles is modeled by simulating their effects on membrane current, the combination of multiple SNPs in an individual neuron has the ability to mask or enhance the excitability defect, enabling a broad spectrum of possible network signaling defects. Conclusions: Our work has three central findings of immediate clinical significance for personalized prediction in epilepsy; 1) the architecture of ion channel variation in both groups is comprised of complex patterns of alleles 2) structural variants in known epilepsy genes appear in otherwise healthy individuals, and 3) individuals with epilepsy typically carry more than one mutation in known human epilepsy genes. Our findings suggest that genetic screens of ion channel genes are only the first step in evaluating personal disease risk. We propose that the complexity of a single individual s unique variant portfolio will require the development of in silico models of channel variation at the protein, cell and network level to accurately assess the functional implications and predict drug response in those with a broad range of excitability disorders.