Author’s Name(s): Lia Huo, Daniela Isaacs-Bernal, Ellie Monroe, Carter Teal, Thierry Léveillard, Shira Albeck, Tamar Unger, Valerie Wallace, Robert Fluhr, Anubhav Garg, Molly Shoichet

Author’s Disclosure Block: Lia Huo, none; Daniela Isaacs-Bernal, none; Ellie Monroe, none; Carter Teal, none; Thierry Léveillard, none; Shira Albeck, none; Tamar Unger, none; Valerie Wallace, none; Robert Fluhr, none; Anubhav Garg, none; Molly Shoichet, none

Abstract Body
Purpose:Rod-derived cone viability factor (RdCVF) is a promising therapeutic protein that could rescue dying cones in retinitis pigmentosa (RP); however, its rapid clearance, limited bioavailability, and the need for repeated injections hinder translation to the clinic.Therefore, there is a need to develop a delivery system that can sustain the release of RdCVF locally and reduce the number of ocular injections required. We previously developed a biocompatible hyaluronan (HA)-oxime hydrogel that displays mechanical and optical properties like that of the vitreous.Here, we designed an affinity-based strategy for the controlled release of RdCVF from HA-oxime, leveraging the interactions between a Src homology (Sh3) domain that we engineered onto RdCVF (Sh3-RdCVF) and Sh3 binding peptides (SBP) immobilized onto HA-oxime.
Study Design/ Methods: To determine the tunability of release, HA-oxime was modified with varying molar excesses of SBP to Sh3-RdCVF andin vitroprotein release was quantified using an ELISA over 14 days. To test our drug delivery system in anin vivomodel of RP, rd1 mice were injected intravitreally at postnatal day 21 with 1 μL of either Sh3-RdCVF encapsulated in HA-oxime (N=10), Sh3-RdCVF alone (N=10), HA-oxime alone (N=8) or PBS (N=8). Cone function was assessed using electroretinogram (ERG) weekly until 4 weeks post-injection, followed by quantification of cone survival and outer segment health by histology. To validate our delivery system in larger rodents and to demonstrate its gene-agnostic properties, P23H rats were injected intravitreally at postnatal day 150 with 3 μL of either Sh3-RdCVF encapsulated in HA-oxime (N=4), Sh3-RdCVF alone (N=4), HA-oxime alone (N=3) or PBS (N=3) and followed up weekly with the same histological assessments as in the rd1 mice. Results: Ourin vitroprotein release showed that as the molar excess of SBP in our hydrogel increases with respect to the Sh3-RdCVF protein, the slower the release. Subsequentin vivoinjection in rd1 mice demonstrated that our slow-release Sh3-RdCVF increases photopic b-wave amplitudes on ERG compared to the three control conditions at 4 weeks post-injection. Whole retinal cone counts and outer segment length were significantly increased with our slow-release RdCVF formulation compared to mice injected with HA-oxime alone, protein alone, or PBS. At 1 and 3 weeks post-injection, Sh3-RdCVF was only detected by Western blot in retinas injected with our slow-release RdCVF, which indicates that the cone rescue seenin vivois indeed due to our hydrogel’s ability to retain Sh3-RdCVF at the retina for at least three weeks while a single protein injection is cleared within a week. In P23H rats, cone survival and outer segment length is similarly increased at 4 weeks post-injection, indicating therapeutic efficacy in another model of RP with a gentler rate of degeneration.