Abstracts

Rationale: Efforts to find susceptibility genes in common epilepsies and febrile seizures by means of association studies have not yet led to convincing results. The etiopathogenesis that underlies Temporal Lobe Epilepsy (TLE) and its subgroups, as MTLE with Hippocampal Sclerosis (MTLE-HS) and TLE with Febrile Seizures (TLE-FS), is poorly understood. The human water channel aquaporin-4 (AQP4) and the potassium channels Kir 4.1 and GIRK3 are closely interrelated with epilepsy. Functional alterations of one or more of these channels will markedly influence brain excitability. Recent studies have indicated a probable link of these channels to MTLE-HS and FS. The current association study was designed to determine the contribution of AQP4 and the inwardly rectifying potassium channels Kir 4.1 and GIRK3 to the etiology of MTLE-HS and TLE-FS. Methods: 218 Norwegian patients with TLE were included. 56 of these patients with MTLE-HS were compared with 162 without MTLE-HS and 181 healthy controls; 102 TLE patients with FS were compared with 105 without FS and 181 healthy controls. SNPs were selected from HapMap in combination with SNPs identified by PCR resequencing of the AQP4, KCNJ10 and KCNJ9 genes using standard methods. Results: Analysis of the data using the HaploView 4.0 software package revealed a number of single SNPs in both KCNJ9 and KCNJ10 that were associated with disease in the TLE with FS group. The majority of these SNPs grouped into one haplotype block and one frequent haplotype within the block showed to be highly overrepresented in cases versus controls. Multivariate analysis revealed significant effect in a multivariate combination of particular alleles from 6 different SNPs from AQP4, 6 different SNPs from KCNJ10, 6 different SNPs from KCNJ9 and 2 different SNPs from the region between KCNJ9 and KCNJ10. Diagnostic classification modeling gave 70 % correct classification. Significance (p-value) of the obtained results was estimated to < 0.001 based on repeated analysis of 1000 random permutations of TLE with FS versus TLE without FS in the dataset. We found no significant associations in MTLE-HS. Conclusions: Our study supports an association between the potassium channel genes KCNJ9/KCNJ10 and TLE-FS, and also indicates a possible association with AQP4. (supported by GlaxoSmithKline)