Human daylight vision depends on cone photoreceptors and their degeneration results in visual impairment and blindness as observed in several eye diseases including age-related macular degeneration, cone-rod dystrophies, or late stage retinitis pigmentosa, with no cure available. Preclinical cell replacement approaches in mouse retina have been focusing on rod dystrophies, due to the availability of sufficient donor material from the rod-dominated mouse retina, leaving the development of treatment options for cone degenerations not well studied. Thus, an abundant and traceable source for donor cone-like photoreceptors was generated by crossing neural retina leucine zipper-deficient (Nrl 2/2 ) mice with an ubiquitous green fluorescent protein (GFP) reporter line resulting in double transgenic tg(Nrl 2/2 ; aGFP) mice. In Nrl 2/2 retinas, all rods are converted into cone-like photoreceptors that express CD73 allowing their enrichment by CD73-based magnetic activated cell sorting prior transplantation into the subretinal space of adult wild-type, cone-only (Nrl 2/2 ), or cone photoreceptor function loss 1 (Cpfl1) mice. Donor cells correctly integrated into host retinas, acquired mature photoreceptor morphology, expressed cone-specific markers, and survived for up to 6 months, with significantly increased integration rates in the cone-only Nrl 2/2 retina. Individual retinal ganglion cell recordings demonstrated the restoration of photopic responses in cone degeneration mice following transplantation suggesting, for the first time, the feasibility of daylight vision repair by cell replacement in the adult mammalian retina. STEM CELLS 2015;33:79-90

Retinal degeneration (rd) leads to progressive photoreceptor cell death, resulting in vision loss. Stimulation of the inner-retinal neurons by neuroprosthetic implants is one of the clinically approved vision-restoration strategies, providing basic visual percepts to blind patients. However, little is understood as to what degree the degenerating retinal circuitry and the resulting aberrant hyperactivity may prevent the stimulation of physiological electrical activity. Therefore, we electrically stimulated ex vivo retinas from wild-type (wt; C57BL/6J) and blind (rd10 and rd1) mice using an implantable subretinal microchip and simultaneously recorded and analyzed the retinal ganglion cell (RGC) output with a flexible microelectrode array. We found that subretinal anodal stimulation of the rd10 retina and wt retina evoked similar spatiotemporal RGC-spiking patterns. In both retinas, electrically stimulated ON and a small percentage of OFF RGC responses were detected. The spatial selectivity of the retinal network to electrical stimuli reveals an intact underlying network with a median receptive-field center of 350 μm in both retinas. An antagonistic surround is activated by stimulation with large electrode fields. However, in rd10 and to a higher percentage, in rd1 retinas, rhythmic and spatially unconfined RGC patterns were evoked by anodal or by cathodal electrical stimuli. Our findings demonstrate that the surviving retinal circuitry in photoreceptor-degenerated retinas is preserved in a way allowing for the stimulation of temporally diverse and spatially confined RGC activity. Future vision restoration strategies can build on these results but need to avoid evoking the easily inducible rhythmic activity in some retinal circuits.