Anti-cancer compound screening identifies Aurora Kinase A inhibition as a means to favor CRISPR/Cas9 gene correction over knock-out

Abstract

CRISPR gene therapy holds the potential to cure a variety of genetic diseases by targeting causative mutations and introducing double stranded DNA breaks, subsequently allowing the host DNA repair mechanisms to introduce mutations. One option to introduce precise gene corrections is via the homology-directed repair (HDR) pathway. HDR can introduce a range of desired mutations dictated by a DNA template which holds a corrected DNA sequence which is written into the targeted gene. The problem in utilizing this pathway is that CRISPR-induced double stranded DNA breaks are repaired more often through the non-homologous end joining (NHEJ) pathway, which does not use a designed template and introduces random DNA damage in the form of insertions and deletions at the cut site. Since HDR activation depends on many interconnected processes in the cell, we aimed to screen a small library of drug compounds in clinical use or clinical development for cancer, to steer the DNA repair process towards preferential HDR activation. We included compounds in our screen based on three relevant mechanisms in CRISPR gene editing: the cell cycle, DNA repair processing and chromosomal packing. We included forty compounds, based on these criteria, screened their toxicity and dosed them in sub-toxic concentrations in cells during genome editing. Of these forty compounds we identified nine potential hits to have an effect on preferential activation of the HDR pathway over NHEJ. Alisertib, rucaparib and belinostat revealed a significant and major effect on gene editing pathway selection in further validation. Alisertib, an Aurora kinase A inhibitor, showed a particularly strong effect towards improving HDR over NHEJ. We subsequently investigated this effect at the genetic level and in a murine hepatoma cell line, which corroborated the initial findings. Alisertib led to an over 4-fold increase in preferential gene correction over gene knock-out, at a dose of 0.3 micromolar. However, the observations that Aurora kinase A inhibitors show considerable cytotoxicity (<50% cell viability) and can induce morphological changes at this concentration pose a limitation for the direct use of these inhibitors as HDR enhancers. However these findings do implicate that the pathways mediated by Aurora kinase A strongly influence HDR outcomes, which warrants further investigation into the downstream pathways driving this effect.