Abstracts

This abstract is a recipient of the Young Investigator Award.

Rationale: Focal cortical dysplasia (FCD) is a common cause of focal epilepsy and often results from somatic, mosaic mutations. The sparse nature of mutated cells (1%-6% of cells in surgically resected tissue) presents many experimental barriers to FCD gene discovery, and only a small number of FCD genes have been identified based upon hypothesis-driven targeted sequencing. Although the mutations are difficult to detect in human tissues, phosphorylation of S6 ribosomal protein is a reliable biomarker of FCD. Leveraging this biomarker, we have developed an unbiased screening platform for identifying novel FCD genes in vitro using a genome-wide CRISPR interference library on inducible neurons generated from human iPSCs.

Methods: We engineered an iPSC line with doxycycline-inducible CRISPR interference (KRAB-dCAS) and iNeuron (NGN1,2) cassettes. Libraries of gRNA-containing lentiviruses were generated, and cells were transduced with the library at a low MOI. FACS for TagBFP was used to select for transduced cells. After 7 days of doxycycline treatment, iNeurons were dissociated, fixed, and immunolabelled for phospho-S6 (S235/236). FACS sorting was performed for pS6-high and -low cell populations using gates set by positive (DEPDC5 gRNA) and negative (non-targeting gRNA) controls. Next-gen sequencing was used to calculate a log fold change (LFC) in gRNA abundance between these two samples. Screening was performed with test (27 gRNAs), whole genome (>100,000 gRNAs), and candidate (130 gRNAs) libraries. After screening, a set of 24 gRNAs were individually evaluated for significant changes to S6 phosphorylation by In-Cell Western blotting.

Results: Three independent whole genome library screenings were performed. Gene otology analysis of the average rank order across the three experiments identified “negative regulation of TOR signaling,” “TSC1-TSC2 complex”, and “GATOR1 complex” as the most enriched terms (Table 1). We next performed candidate screening of top candidate genes and known FCD genes (130 gRNAs total). Linear regression of two replicate independent experiments shows correlation with an R² of 0.4242 and a slope of 0.9312 (Figure 1). All 9 known FCD gene gRNAs (TSC1, TSC2, DEPDC5, NPRL2, PTEN) had LFCs > 0.22 and were in the top 25% in the screen. Candidate genes with LFC > 0.22 underwent in-cell western analysis with 13/14 candidate gRNAs significantly increasing S6 phosphorylation. Four of 4 candidate gRNAs with LFC < 0.22 were non-significant for changes in pS6.