Correcting CPS1 Deficiency in Primary Human Hepatocytes In Vivo Using AAV Delivered Cell-Cycle Independent Targeted Genome Editing

Abstract

Carbamoyl phosphate synthetase I (CPS1) deficiency is a rare genetic metabolic disorder that impairs nitrogen metabolism. Loss of CPS1 enzyme function can cause neonatal death or intellectual disability. The only treatment for severe CPS1 deficiency is liver transplantation, however the disease is an ideal candidate for gene therapy. An effective gene therapy for CPS1 deficiency must create permanent changes to the genome such that therapeutic effects are maintained during paediatric liver growth. Repair must also overcome a high therapeutic threshold that correlates to hepatic cell mass corrected. We aimed to develop a cell-cycle independent targeted genome editing approach to treat CPS1 deficiency in patient-derived primary human hepatocytes in vivo. A dual recombinant adeno-associated virus (rAAV) approach was designed to deliver CRISPR/SaCas9 guided editing machinery and GFP or CPS1 donor DNA for homology-independent genome insertion at the CPS1 locus. Both vectors were packaged in the highly human hepatotropic AAV capsid NP59. Editing efficacy was tested in primary human hepatocytes xenografted into the FRG (Fah-/-Rag2-/-Il2rg-/-) mouse liver. Editing events were characterised at a DNA and RNA level using ddPCR and next-generation sequencing (NGS), and the protein level using immunofluorescence and in situ activity staining. Insertion of donor DNA at the CPS1 locus was observed in all dual-treated human hepatocytes, as well as high levels of edited CPS1 transcript. Activity of CPS1 following dual-vector treatment was observed in CPS1-deficient hepatocytes via in situ activity staining. Investigation of unwanted on-target events revealed rearranged subgenomic fragments of both vector genomes inserted at the cleavage site in a dose-dependent manner. Off-target integration at unique sites across the genome was observed by NGS, consistent with AAV-mediated integration events, and no events were observed in in silico-predicted CRISPR/Cas9 off-target sites.