A major approach for the discovery of potential drug targets for tumorigenic disease is based around the identification of synthetic lethality. Synthetic lethality is defined as a genetic interaction where the combined disruption of two genes is lethal but either disruption alone has no effect. Genes that have synthetic lethal interactions with tumor suppressors are attractive as drug targets for cancers because treatments will specifically kill tumor cells but will have little effect on the surrounding wildtype tissue. However, these interactions have proved difficult to identify using existing screening technologies and those that are identified are often not reproducible between screens.
We combined CRISPR and RNAi in a novel combinatorial screening method in Drosophila cells. Using this approach we identified almost 200 robust candidate drug targets for Tuberous sclerosis complex (TSC), a tumorigenic disease caused by mutation of either the TSC1 or TSC2 tumor suppressors. Based on this work, a promising drug is now in development for this disease. However, despite this success, overlap of hits between screens is limited, indicating that many interactions are still missed. We hypothesized that many synthetic lethal effects occur only with incomplete gene knockdown. These interactions are therefore missed in many cases because weak knockdown has no effect and strong knockdown is lethal to both mutant and wildtype genetic backgrounds. We therefore developed an approach called variable dose analysis (VDA), which uses cytometry-based analysis to profile viability effects over a range of knockdown efficiencies in a single population of cells treated with RNAi. Using VDA to compare viability profiles between wildtype and TSC mutant cells, we found that this method has reduced noise compared to previous screening approaches. We were able to detect 83% of known synthetic lethal interactions with TSC compared to only 22% using established screening technologies. This is therefore a robust method to detect synthetic lethality that will likely be applicable to the identification of drug targets for many tumorigenic diseases.