sentence: This study identifies gene regulatory networks controlling proliferative and neurogenic competence in retinal Müller glia.
AbstractInjury induces retinal Müller glia of cold-blooded, but not mammalian, vertebrates to regenerate neurons. To identify gene regulatory networks that control neuronal reprogramming in retinal glia, we comprehensively profiled injury-dependent changes in gene expression and chromatin accessibility in Müller glia from zebrafish, chick and mice using bulk RNA-Seq and ATAC-Seq, as well as single-cell RNA-Seq. Crossspecies integrative analysis of these data, together with functional validation, identified evolutionarily conserved and species-specific gene networks controlling glial quiescence, gliosis and neurogenesis. In zebrafish and chick, transition from the resting state to gliosis is essential for initiation of retinal regeneration, while in mice a dedicated network suppresses neurogenic competence and restores quiescence. Selective disruption of NFI family transcription factors, which maintain and restore quiescence, enables Müller glia to proliferate and generate neurons in adult mice following retinal injury. These findings may aid in the design of cell-based therapies aimed at restoring retinal neurons lost to degenerative disease. previously shown to induce MG reprogramming independent of injury (20). In the mouse, we also tested FGF2 and insulin treatment, which induces limited MG proliferation, but not neurogenic competence (21).For bulk RNA-Seq and ATAC-Seq, in zebrafish, we used the reporter line Tg[gfap:EGFP]nt11 (22) and cell sorting to enrich MG cells, purifying 9.8% of total input retinal cells (Fig. S1A). In mouse, we performed intraperitoneal injection of tamoxifen at postnatal day (P) 21 to induce MG-specific Sun1-GFP expression in GlastCreERT2;CAG-lsl-Sun1-GFP mice (23), purifying 3.5% of total input cells ( Fig. S1A). Based on RT-qPCR, we had an overall enrichment of 30-fold of canonical MG marker genes such as rlbp1a/Rlbp1 and Glul in the GFP-positive (GFP+) fraction in both species (Fig. S1B). Bulk RNA-Seq samples were generated from both the GFP+ MG and GFP-neuronal fractions in zebrafish and mouse, while ATAC-Seq samples were generated from the GFP+ cells. Each sample had a minimum of two biological replicates. In both species, we analyzed retinas with a time course following either inner retinal injury induced by NMDA excitotoxicity or outer retinal injury induced by light damage. In total, we generated 100 bulk RNA-Seq libraries and 40 bulk ATAC-Seq libraries ( Fig. 1A, Table S1, Supplementary Data 1).In parallel, we conducted scRNA-Seq analysis of whole retinas at the same time points as used for the bulk RNA-Seq. We profiled retinas following either NMDA or light damage in zebrafish and mouse, as well as following NMDA damage in P10 chick. To distinguish injury-responsive genes from injury-independent reprogramming genes, we profiled zebrafish retinas at multiple time points following T+R treatment, and a single time point in chick and mouse retinas...