RHO1-2 meganuclease gene editing targets human P23H rhodopsin-induced retinitis pigmentosa to rejuvenate rods and maintain cones.

Autosomal dominant retinitis pigmentosa (adRP) is an inherited retinal dystrophy characterized by progressive vision loss and eventual blindness. The P23H mutation (proline to histidine substitution at codon 23) in the rhodopsin (RHO) gene represents the most common form of adRP in North Americans. Currently, there is no cure for P23H adRP. Genome editing targeting the mutant RHO allele, leaving a functional wildtype (WT) allele, is an attractive approach for P23H adRP, as only one copy of RHO is needed for normal retinal function. We re-engineered an I-Cre meganuclease, called RHO1-2, to target a 22bp recognition sequence encompassing the mutation responsible for the p.P23H RHO mutation. In vitro, RHO1-2, cuts human P23H RHO but not WT RHO. In vivo, we delivered scAAV5:GRK1:RHO1-2 via subretinal injection in early-stage degeneration using the only large animal model of human p.P23H RHO adRP (TgP23H pigs). We tested RHO1-2 efficacy and durability, on retinal function using full-field electroretinograms and on retinal structure using spectral domain optical coherence tomography and immunohistochemistry. We observe that RHO1-2 treatment: arrests rod photoreceptor degeneration, resurrects rod-driven retinal function that does not exist in untreated TgP23H pigs, restores mislocalized rhodopsin expression and rebuilds rod inner and outer segments (IS/OS). Rod rescue maintains cones. A year after RHO1-2 treatment, we show that TgP23H pigs use rod-driven vision to navigate a maze. Our results demonstrate that genome editing via RHO1-2 meganuclease is a viable treatment to cure human p.P23H RHO adRP. They also suggest that meganuclease-based editors can be effective for other IRDs.