Abstract
Purpose :
An overlapping dual vector system for the treatment of Stargardt disease has been developed in which the entire ABCA4 expression cassette is delivered in two fragments, each of which is packaged separately into an AAV8 Y733F capsid. Although ABCA4 is expressed at therapeutic levels following dual vector recombination, the effect of non-recombined ABCA4 fragments potentially expressed by a single vector is unknown. Hence we sought to identify any toxic effects from either of the two single vectors, corresponding to the 5’ sand 3’ end of the ABCA4 transgene.
Methods :
In a blinded study, Abca4 KO mice received a subretinal injection in the superior retina of the right eye at 4-5 weeks of age (n=8-11 per group). Injected materials tested were: AAV diluent (PBS PF68 0.001%); GRK1.GFP.pA high dose (2E+10 genome copies); 5’-ABCA4 vector low dose (2E+9 genome copies); 5’-ABCA4 vector high dose (2E+10 genome copies); 3’-ABCA4 vector (1E+10 genome copies); dual vector low dose (2E+9 total genome copies); dual vector high dose (2E+10 total genome copies). All vectors were AAV8 Y733F. Optical coherence tomography (OCT) and ERG assessments were performed at 3 and 6 months post-injection.
Results :
As expected, mice in all cohorts revealed varying degrees of damage relating to the surgical procedure. The superior retina around the subretinal injection site was thinner than uninjected paired eyes at both 3 and 6 months post-injection (3 months two-way ANOVA: eye p<0.001, cohort p=0.7012, interaction p=0.6203; 6 months two-way ANOVA: eye p<0.001, cohort p=0.6858, interaction p=0.6230). The reduction in total retinal thickness was not however influenced by the injection material, as there was no additional loss observed in vector injected mice compared to those that received AAV diluent only. All cohorts exhibited reductions in ERG amplitude at 1cd.s/m2 between injected and uninjected eyes (3 months two-way ANOVA: eye p<0.0001, cohort p=0.0173, interaction p=0.5954; 6 months two-way ANOVA: eye p<0.0001, cohort p=0.0102, interaction p=0.4437). The ERG changes were similar between vector and sham injected eyes at both time points (two-way ANOVA: time point p=0.8507, cohort p=0.4014).
Conclusions :
Performing a subretinal injection in Abca4 KO mice led to focal retinal damage around the injection site. There was no evidence of any toxic effect from AAV8 expressing non-recombined ABCA4 fragments, either alone or in tandem.
This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.