Search

The anatomical differences between the retinas of humans and most animal models pose a challenge for testing novel therapies. Non-human primate (NHP) retina is anatomically closest to the human retina with the presence of a high acuity region called the fovea. However, there is a lack of relevant NHP models for retinal degeneration that can be used for preclinical studies of vision restoration. To address this unmet need we aimed to generate inducible NHP models of photoreceptor degeneration. We generated three cynomolgus macaque models using distinct strategies. We used two genetically targeted strategies using optogenetics and Crispr-Cas9 to ablate specifically rods to mimic rod-cone dystrophy. Additionally, we created an acute model by physical separation of the photoreceptors and retinal pigment epithelium using a polymer patch. Retinal degeneration was evaluated in all three models by in-life exams such as fundus imaging, optical coherence tomography, adaptive optics and electroretinography. In the genetic models we observed punctuate areas of degeneration in the injected area marked by disorganization of outer segments, loss of rod photoreceptors and thinning of the outer nuclear layer. In the acute model, the degeneration was faster and involved both rods and cones. Among the three distinct NHP models, the Crispr-Cas9 based approach was the most advantageous model in view of recapitulating disease specific features and its ease of implementation. The acute model however resulted in the fastest degeneration making it the most relevant model for testing end-stage vision restoration therapies such as stem cell transplantation.

Many neural interfaces used for therapeutic applications are based on extracellular electrical stimulation to control cell polarization and thus functional activity. Amongst them, retinal implants have been designed to restore visual perception in blind patients affected by photoreceptor degeneration diseases, such as age-related macular degeneration (AMD) or retinitis pigmentosa (RP). While designing such a neural interface, several aspects must be taken into account, like the stimulation efficiency related to the current distribution within the tissue, the bio-interface optimization to improve resolution and tissue integration, and the material biocompatibility associated with long-term aging. In this study, we investigate the use of original microelectrode geometries for subretinal stimulation. The proposed structures combine the use of 3D wells with protuberant mushroom shaped electrode structures in the bottom, implemented on a flexible substrate that allows the in vivo implantation of the devices. These 3D microelectrode structures were first modeled using finite element analysis. Then, a specific microfabrication process compatible with flexible implants was developed to create the 3D microelectrode structures. These structures were tested in vivo to check the adaptation of the retinal tissue to them. Finally, preliminary in vivo stimulation experiments were performed.