Abstracts

Rationale: According to the multidrug transporter hypothesis, pharmacoresistance in epilepsy results from impaired drug penetration into the epileptic focus secondary to localized dysregulation of drug transporters. Solute carrier transporters (SLCs) are known to mediate chemoresistance in a number of diseases but have not been systematically studied in epilepsy. As SLCs are primarily influx transporters, our aim was to identify the SLCs most significantly downregulated in the pharmacoresistant epileptic hippocampus by the initial use of an in silico strategy and then to confirm this by ex vivo analysis of human brain tissue. Methods: We extracted data relating to SLCs from (1) our previously published integrative analysis of microarray studies on brain tissue from epilepsy surgery (Hum Mol Genet 2011;20:4381-94), and (2) a comprehensive review of published literature on epilepsy pharmacoresistance, and integrated all the data using ‘convergent functional genomics (CFG)'—a validated technique for prioritizing genes involved in complex diseases by collating evidence using a pre-defined scoring system. To identify SLCs which have not been studied in epilepsy pharmacoresistance but could potentially be involved, we employed a computational gene prioritization tool called Endeavour, using SLCs with the highest CFG-scores as the training genes. We validated this computationally prioritized gene list by (1) prioritizing the same candidate genes using a robust independent training set comprising the 10 most consistently downregulated genes from the aforementioned integrative analysis and, then, demonstrating significant rank order similarity with the original list, and (2) prioritizing the same candidate genes using a training set comprising 10 randomly chosen genes from the integrative analysis and demonstrating no significant rank order similarity with the original list (see figure). In parallel with our bioinformatics analysis, we used a custom Agilent oligonucletoide microarray containing exon probes for all known SLCs to analyse 24 hippocampal samples obtained from surgery for pharmacoresistant mesial temporal lobe epilepsy and 24 hippocampal samples from normal post-mortem controls. The whole-transcript amplification protocol of exon arrays allows more accurate measurement of gene expression than standard microarrays. Results: Our exon array identified 15 SLCs significantly (FDR <0.05) downregulated in the epileptic hippocampus by 1.5 fold or more (see table). There was a highly significant overlap between the genes identified by our in silico and ex vivo strategies: p <1.5x10⁻⁷ (hypergeometric distribution) for the overlap between the top 30 bioinformatically-identified genes and the 15 experimentally-identified genes. Conclusions: The most significantly downregulated SLCs in the pharmacoresistant epileptic human hippocampus were identified using a robust in silico methodology and confirmed ex vivo. Our analysis shows the power of stringently applied bioinformatic techniques. The role of these SLCs in the epileptic hippocampus will need to be defined through functional studies.