­­­­Quiescent Müller glia heterogeneity influences regenerative response following photoreceptor ablation in the zebrafish retina

Loss of neurons in the neural retina is a leading cause of vision loss. Retinal regeneration in zebrafish is attributed to Müller glia, which are activated, adopt a stem cell-like state, proliferate, and generate new neurons. Despite the same endogenous Müller glia being present in all vertebrates including humans, Müller glia activation in mammals leads to glial scarring instead of neurogenesis. Therefore, understanding Müller glia cellular states and molecular processes during zebrafish regeneration may allow us to improve mammalian regeneration. Since only a subset of Müller glia activate following neuron loss, we asked whether a specialised subset of Müller glia might be genetically primed for activation. Single-cell RNA sequencing (sc-RNAseq) from integrated zebrafish Müller glia samples reveals heterogeneity of gene expression across the quiescent Müller glia pool with 4 main clusters emerging, two of which are genetically similar. Labelling for key differentially expressed markers identified that these main clusters correlated with glia spatially distributed across dorsal, central and ventral retina. Using a genetically driven, chemically induced nitroreductase approach, three distinct photoreceptor types were ablated: the long (Lws2), short wavelength-sensitive (Sws2) and rod (Xops) photoreceptors, which also differ in their abundance. As expected, histological analysis of the three injuries and sc-RNAseq data for the two cone photoreceptor ablations identified common genetic program involved in the transition of Müller glia quiescence to activation, and differential responses to injury related to either injury extent or subtype of photoreceptor targeted. Interestingly, despite differences in the distributed across the retina, biased activation of Müller glia was observed in dorsal and central regions. Gene ontology analysis revealed that these injury-responsive dorsal and central Müller glia express genes related to dorsal/ventral pattern formation, growth factor activity, and regulation of developmental process. Müller glia upregulated genes related to homeostasis as well as certain AP-1 and injury-responsive transcription factors, followed by expression of genes involved in cell cycle, chromatin remodeling, and microtubule organisation. Prior to cell cycle entry, Müller glia show a transient upregulation of genes involved in gliogenesis, and Notch signalling. These findings enhance our understanding of heterogeneous states of quiescent Müller glia, and how these differences relate to activation and regeneration potential following neural ablation. A comparison of the distinct quiescent Müller glia with glia states in mammals will reveal key molecular pathways that could be targeted for improved mammalian neural regeneration. Zebrafish lines Tg(her4.1:dRFP/gfap:GFP) for the uninjured control dataset,  Tg(her4.1:dRFP/gfap:GFP/lws2:nfsb-mCherry) for red cone ablation, and Tg(her4.1:dRFP/gfap:GFP/sws2:nfsb-mCherry) for blue cone ablation were used in this study. Zebrafish larvae at 6 days post-fertilisation (dpf) were swum in a 10 mM solution of metronidazole for 48 hours, to specifically eliminate the relevant subtype of cone photoreceptor. At 8 dpf, fish were rinsed in fresh system water under standard housing conditions. At 9 dpf (3 days post-injury; dpi), fish were humanely killed. The single-cell suspensions of 9 dpf zebrafish were prepared by following a published protocol (Lopez-Ramirez et al., 2016). Retinae were dissected and digested in 350 ml papain solution at 37°C for 15 minutes. The papain solution was prepared as follows: 100 ml papain (Worthington, LS003126), 100ul of 1% DNAse (Sigma, DN25), and 200 ml of 12mg/ml L-cysteine (Sigma, C6852) were added to a 5 ml DMEM/F12 (Invitrogen, 11330032). During digestion, retinal tissue was mixed by pipetting up and down 4 x 10 times. Following digestion, 1400 ml of washing buffer was added, containing 65 ml of 45% glucose (Invitrogen, 04196545 SB), 50ul of 1M HEPES (Sigma, H4034), and 500 ml FBS (Gibco, 10270106) in 9.385ml of 1x DPBS (Invitrogen, 14190-144). All solutions were filtered through a 0.22 mm filter (Millipore) and stored at 4°C prior to use.