Efficient enrichment of seamlessly gene-edited HSPCs by targeted integration into haploinsufficient genes.

Preclinical data in mouse models and occurrence of adverse events in some clinical trials have shown that random integration of strong promoters by viral vectors can be unpredictably genotoxic. We hypothesized that strong promoters may safely be exploited for clinical purposes when integrated into a single genomic locus, where genomic perturbations can be characterized beforehand. While gene editing (GE) by homology-direct repair (HDR) allows for target integration of a therapeutic cassette into a validated locus of interest in hematopoietic stem and progenitor cells (HSPCs), it is presently constrained by low efficiency and risk of heterogeneous genetic outcomes. We thus developed an innovative strategy that intrinsically enriches HDR-edited HSPCs, while purging out genotoxic byproducts. This strategy is based on nuclease-based disruption of haploinsufficient loci in HSPCs, and reconstitution of the same loci by HDR. As the HDR cassette also includes the gene of interest (GOI), cells that survive the nuclease also harbor the GOI in the desired site. Here we screened putative haploinsufficient loci and identified three candidate genes (UBA1, RPS19 and OGT) whose knock-out significantly impairs the clonogenic output of edited HSPCs; upon transplantation into immunodeficient mice, the edited-HSPCs bearing indels are spontaneously counter-selected over time. In UBA1 and in RPS19 loci, we have obtained a proof of concept of selection of edited cells by HDR mediated integration of a “rescue” cassette also carrying a reporter GOI. As compared to the KO only counterparts, the HDR-edited HSPCs showed partial restoration of clonogenic output, and 80-90% of the colonies were reporter positive. We then pursued in vivo validation for the target locus UBA1. While UBA1-KO HSPCs transplanted into NSGW mice resulted in very low engraftment in the bone marrow (<5% of live cells) as compared to a mock KO condition (55%), the HDR-edited cells in UBA1 had an engraftment of nearly 40%, and nearly 100% of human cells were GFP positive. This can be compared to the commonly used AAVS1 safe-harbor target, where at comparable levels of HDR efficiencies (60% of cells) in vitro (input at transplant), only 40% of cells remained GFP+ at the experiment endpoint. Characterization of genomic and transcriptional perturbations at single-cell resolution driven by different strong promoters integrated in our candidate loci is ongoing. We believe our strategy offers an appealing one-size-fits-all approach for HSPC gene therapy for metabolic diseases and beyond, allowing for targeted integration of a strongly expressed gene of interest into a well-known and characterized locus and spontaneous purge out of genotoxic editing byproducts.