Preclinical evaluation of a homing endonuclease gene therapy for adRP in models of P23H retinopathy.

Autosomal dominant Retinitis pigmentosa (adRP) is a blinding eye disease that results in loss of rods and eventually cone photoreceptors. A large fraction of adRP is caused by mutations in the rhodopsin (Rho) gene, and the most common mutation in North America causes a P23H mutation in rhodopsin. Patients typically present with mid- to late-stage disease when many rod photoreceptors already have degenerated. We will evaluate genome editing using a meganuclease genome editing tool (Rho1-2) that is specific for the sequence causing the P23H mutation in several preclinical animal models of P23H adRP. We will deliver AAV5- GRK1-Rho1-2 via subretinal injection to infect photoreceptors. We will measure efficacy, the ability of Rho1-2 to alter rod degeneration and functional decline. We will quantitatively determine Rho1-2 efficiency, the proportion of P23H alleles with insertions/delections (indels), specificity the proportion of P23H to WT alleles with indels, and its safety, the absence of indels in other genes throughout the human genome (off target editing). In Pig P23H adRP models we will evaluate Rho1-2 editing efficacy because of the similarities between pigs and humans in eye size and retinal structure and because the pig has a cone rich visual streak. We will use the same non-invasive tools used in the clinic to measure retinal function (full field electroretinograms) and structure, (spectral domain OCT). We will use two existing transgenic pig lines that harbor the entire P23H hRHO gene, but differ in their time course, resulting in a fast or slow loss of rods and rod function. We will match treatment time with the natural history of rod structural and functional decline, compare Rho1-2 editing efficacy in early-, mid- and late-stage P23H adRP and define its temporal treatment window. In Humanized adRP mouse models we will evaluate Rho1-2 editing efficiency and specificity because they contain WT and P23H human alleles in the appropriate genomic context. In human retinal organoids we will evaluate Rho1-2 editing safety (off-target cutting throughout the genome) because they contain human photoreceptors as well as the rest of the human genome in a relevant context.. Taken together the results of our experiments will determine if Rho1-2 should move forward for eventual use in human clinical trials and will define endpoints for those trials.