Abstract

Glaucoma and optic neuropathies cause progressive and irreversible degeneration of retinal ganglion cells (RGCs) and the optic nerve and are currently without any effective treatment. Previous research into cell replacement therapy of these neurodegenerative diseases has been stalled due to the limited capability for grafted RGCs to integrate into the retina and project properly along the long visual pathway to reach their brain targets. In vivo RGC regeneration would be a promising alternative approach but mammalian retinas lack regenerative capacity even though cold-blood vertebrates such as zebrafish have the full capacity to regenerate a damaged retina using Muller glia (MG) as retinal stem cells. Nevertheless, mammalian MG undergo limited neurogenesis when stimulated by retinal injury. Therefore, a fundamental question that remains to be answered is whether MG can be induced to efficiently regenerate functional RGCs for vision restoration in mammals. Here we show that without stimulating proliferation, the transcription factor (TF) Math5 together with a Brn3 TF family member are able to reprogram mature mouse MG into RGCs with exceedingly high efficiency while either alone has no or limited capacity. The reprogrammed RGCs extend long axons that make appropriate intra-retinal and extra-retinal projections through the entire visual pathway including the optic nerve, optic chiasm and optic tract to innervate both image-forming and non-image-forming brain targets. They exhibit typical neuronal electrophysiological properties and improve visual responses in two glaucoma mouse models: Brn3b null mutant mice and mice with the optic nerve crushed (ONC). Together, our data provide evidence that mammalian MG can be reprogrammed by defined TFs to achieve robust in vivo regeneration of functional RGCs as well as a promising new therapeutic approach to restore vision to patients with glaucoma and other optic neuropathies.