Final Report Summary - TRHE (Translational research in human epilepsies) We have made significant progress towards understanding the genetic basis of the common human developmental epilepsy known as rolandic epilepsy. Over the course of the fellowship, we have demonstrated that there are several distinct genetic influences, some of which are are specific to certain component phenotypes, whereas others are shared between phenotypes and endophenotypes. Some of these influences may be structural, consisting of variations in genomic copy number.We initially showed that the EEG trait of centrotemporal spikes maps to chromosome 11p13 and is associated with variants at the ELP4 gene; we have now shown that the speech sound disorder phenotype that commonly aggregates in probands and relatives also maps to the 11p13 locus - for this study we developed a novel statistical test of pleiotropy. Furthermore, we demonstrated using acoustic analysis that the speech sound disorder results from a dyspraxic mechanism affecting parameters such as the timing of voice onset and vowel duration. Rolandic epilepsy is the strongest known risk factor for reading disability, with an odds in probands of 5.8 (CI:2.9-11.7) and an affected proband increasing the risk to relatives 2.9 fold (CI:1.4-5.8). We have now investigated the genetics of reading disability; the phenotype; and its risk factors. Genomewide linkage analysis showed no evidence of linkage to known dyslexia loci, suggesting genetic heterogeneity. We found evidence for novel loci at 1q42 (multipoint LOD 4.70) and at 7q21 in the French-Canadian population (MP LOD 3.03). We identified a neurocognitive endophenotype in rolandic epilepsy families: 31% of probands and 9% of siblings were reading impaired; 54% of probands and 36% of siblings were language impaired; 67% of probands and 70% of siblings were attention impaired; both probands and siblings displayed a similar deficit profile. Looking at clinical and demographic risk factors, male sex was strongly associated (OR 3.7 CI:1.1-12.9) as were a history of speech sound disorder (OR 9.8 CI: 2.4-40.5) and ADHD symptoms (OR 14.0 CI:2.6-75.9); interestingly seizure or treatment variables were not associated, and reading disability was not independently associated with the presence of interictal EEG discharges or proband status. A separate neuropsychological study showed that attention and phonological processing impairments, but not auditory processing deficits, predicted reading disability. We have also discovered major new susceptibility genes in the spectrum of idiopathic focal epilepsies, a group that includes rolandic epilepsy, Landau-Kleffner syndrome and Continuous Spikes in Slow-Wave Sleep. Mutations in the NMDA receptor subunit NR2A, coded by the gene GRIN2A, are found in 7-20% of individuals with these phenotypes. NR2A is important for sleep-dependent memory consolidation and therefore associative learning. This may be interesting in relation to our finding that children with rolandic epilepsy have a 30 minute shorter sleep duration than their healthy peers. Novel copy number variants, both de novo and inherited, occur in excess in these individuals, suggesting the presence of further causative or susceptibility factors that demand deeper investigation. In rolandic epilepsy, we discovered 117 rare CNVs in 86/195 cases, of average length 208 Kb. Only one was a known recurrent CNV, at 16p13.11 but others covered genes enriched for glutamate receptor, presynaptic membrane, ion homeostasis and chemokine activity. These pathway analyses will be useful in prioritising future large-scale genomewide studies.Details of our project are accessible at www.rolandic.com. All research is either published or in review for peer-reviewed, open-access scientific journals.
