Clcf1/Crlf1a-mediated signaling is neuroprotective and required for Müller glia proliferation in the light-damaged zebrafish retina

Zebrafish possess the innate ability to fully regenerate any neurons lost following a retinal injury. This response is mediated by Müller glia that reprogram and divide asymmetrically to produce neuronal precursor cells that differentiate into the lost neurons. However, little is understood about the early signals that induce this response. Ciliary neurotrophic factor (CNTF) was previously shown to be both neuroprotective and pro-proliferative within the zebrafish retina, however CNTF is not expressed following injury. Here we demonstrate that alternative ligands of the Ciliary neurotrophic factor receptor (CNTFR), such as Cardiotrophin-like cytokine factor 1 (Clcf1) and Cytokine receptor-like factor 1a (Crlf1a), are expressed within Müller glia of the light-damaged retina. We found that CNTFR, Clcf1, and Crlf1a are required for Müller glia proliferation in the light-damaged retina. Furthermore, intravitreal injection of CLCF1/CRLF1 protected against rod photoreceptor cell death in the light-damaged retina and induced proliferation of rod precursor cells in the undamaged retina, but not Müller glia. While rod precursor cell proliferation was previously shown to be Insulin-like growth factor 1 receptor (IGF-1R)-dependent, co-injection of IGF-1 with CLCF1/CRLF1 failed to induce further proliferation of either Müller glia or rod precursor cells. Together, these findings demonstrate that CNTFR ligands have a neuroprotective effect and are required for induction of Müller glia proliferation in the light-damaged zebrafish retina.

One pathway of particular interest is the ciliary neurotrophic factor receptor (CNTFR) mediated signalling pathway. CNTFR is a tripartite receptor that binds a ligand, which recruits both Il6st (Gp130) and M17 (LIFR), resulting in CNTFR phosphorylation (samuel et al., 1993;Stahl & Yancopoulos, 1994). This allows CNTFR to signal via the Stat3, MAPK, and PI3K pathways (Rezende et al., 2009;Leibinger et al., 2013;Askvig & Watt, 2015;Hu et al., 2020), all of which were shown to regulate the Müller glia response to injury (Kassen et al., 2009;Nelson et al., 2012;Wan et al., 2012;Wan et al., 2014). CNTFR was previously found to be expressed within zebrafish Müller glia following injury . Several studies also demonstrated that CNTFR and its ligand, ciliary neurotrophic factor (CNTF), have a wide array of effects within the zebrafish retina. Knockdown of gp130 in the damaged zebrafish retina decreased Müller glia proliferation , suggesting that Gp130-mediated signalling is required for retinal regeneration. Further, intravitreal injection of CNTF was sufficient to induce zebrafish Müller glia proliferation in a Stat3-dependent manner and was neuroprotective in a MAPK pathway-dependent manner (Kassen et al., 2009;. CNTF treatment also resulted in reduced photoreceptor cell death in both rodents (Rhee et al., 2013;Lipinksi et al., 2015) and humans (Sieving et al., 2006;Zhang et al., 2011). Additionally, several different cell types, including Müller glia, rods, and cones, were shown to be CNTF-sensitive (Leibinger et al., 2009;Rhee et al., 2013;Li et al., 2018), with CNTF exposure stimulating significant transcriptional changes . Interestingly, CNTF also induced regeneration of cones in a rat model of retinal degeneration (Li et al., 2010). Finally, CNTF was also shown to confer neuroprotective effects in other areas of the central nervous system (Beurrier et al., 2010;Jeong et al., 2015).
Despite the effects CNTF exert in vivo, cntf was not shown to be expressed within the regenerating zebrafish retina, however, the alternative CNTFR ligands Cardiotrophin-like cytokine factor 1 (Clcf1) and Cytokine receptor-like factor 1a (Crlf1a), which form a ligand complex, were expressed within the retina following injury . Additionally, knockdown of either clcf1 or crlf1a inhibited regeneration following optic nerve crush in zebrafish (Elsaeidi et al., 2014). Despite the presence of these ligands within the retina, no in-depth studies have examined their role in Müller gliadependent regeneration of retinal neurons. Understanding the roles of the different CNTFR ligands during zebrafish retinal damage and regeneration may guide the development of cell-based retinal therapeutics in humans.

Zebrafish husbandry
Adult albino and transgenic albino; Tg[gfap:EGFP] nt11 zebrafish were maintained in the Center for Zebrafish Research at the University of Notre Dame Freimann Life Sciences Center as previously described (Recasens et al., 2021). Fish were maintained on a 14h/10h light-dark cycle for standard conditions. Approximately equal numbers of male and female fish were used in the study. All fish used in these experiments were 6-12 months of age and 2-3 cm in length. All animal care protocols were approved by the University of Notre Dame Animal Care and Use Committee and are compliant with the National Institutes of Health guide for the care and use of laboratory animals.

Drug and protein treatments
Adult albino and albino; Tg[gfap:EGFP] nt11 zebrafish were anesthetized in 2-phenoxyethanol (2-PE, 1:1000) until unresponsive to tail pinch and intravitreally injected every 12 h with 0.5 µl of either PBS (vehicle), 1 mg/ml recombinant mouse CLCF1/CRLF1 (R&D systems; Minneapolis, MN), 1 mg/ml recombinant human IGF-1 (R&D systems; Minneapolis, MN), or a combination of both, using a 2.5 µl syringe with 33-gauge rounded needle (Hamilton; Reno, NV) and then maintained at 32°C. Eyes were collected 72 h after the initial injection (hpi). For light treatments, dark-adapted albino zebrafish received intravitreal injections every 12 h for 48 h, at which point the fish were exposed to constant light and received an additional injection at 12 h of light treatment. These eyes were collected at 24 h of light treatment. For EdU treatment, fish received intraperitoneal injections of 1 mg/ml EdU using a 30-gauge needle while simultaneously receiving intravitreal injections of CLCF1/CRLF1 every 12 h and eyes were collected at 72 h hpi.

NVP-ADW742 (NVP)
albino; Tg[gfap:EGFP] nt11 zebrafish were intraperitoneally injected with 50 µl of either 5% DMSO (vehicle) or 500 µM NVP (Selleck Chemical; Houston, TX) every 12 h throughout the course of either constant light treatment or intravitreal injection of CLCF1/CRLF1, using a 30-gauge needle. For light treatment experiments, the initial intraperitoneal injection of NVP corresponded with the commencement of light treatment and the final injection occurred at 24 h of light treatment, with the eyes collected at 36 h of light treatment. During CLCF1/CRLF1 treatment, intraperitoneal injections of NVP were performed simultaneously with intravitreal CLCF1/CRLF1 injections every 12 h and the eyes were collected at 72 hpi.
For EdU labelling, slides were washed and rehydrated in PBS for 20 min, washed in 3% bovine serum albumin (BSA) in PBS for 10 min, then permeabilized by washing in 0.5% Triton X-100 (Fisher Scientific; Pittsburgh, PA) in PBS for 20 min. Slides were then labelled using the Click-iT EdU labeling kit (Life Technologies; Carlsbad, CA), per the manufacturer's instructions, and washed in 3% BSA for 5 min before blocking and continuing with immunohistochemistry as described above.

2.6
In situ hybridization RNA in situ hybridization was performed using the RNAscope Multiplex Fluorescent v2 Assay (Advanced Cell Diagnostics; Newark, CA) according to the protocol for fixed-frozen tissue sample preparation with some modifications . Tissue sections were prepared as described for immunohistochemistry. Frozen sections were washed in PBS for 5 min followed by baking for 1 h at 55°C in an oven. Tissue sections were post-fixed with 4% paraformaldehyde (PFA; Sigma Aldrich; St. Louis, MO) at room temperature for 1 h. Slides were dehydrated for 5 min each in 50%, 70%, and twice in 100% ethanol and then baked for 1 h at 55°C. Sections were treated with hydrogen peroxide solution (Advanced Cell Diagnostics; Newark, CA) for 10 min at room temperature, followed by a distilled water wash and placed in boiling Target Retrieval Buffer (Advanced Cell Diagnostics; Newark, CA) for 15 min. The slides were immediately washed with distilled water, dehydrated with 100% ethanol, and dried. Slides were baked for 1 h at 55°C, during which time a hydrophobic barrier was applied to the slide (ImmEdge Hydrophobic Pen; Vector Laboratories; Burlingame, CA) and dried overnight at room temperature. Sections were treated with the Protease III solution (Advanced Cell Diagnostics; Newark, CA) for 30 min at 40°C and washed with distilled water prior to probe incubation at 40°C for 2 h. The following probes (Advanced Cell Diagnostics; Newark, CA) were used: Dr-il6st-C1 (cat no. 1108591), Dr-clcf1-C2 (cat no. 1102851), Dr-cntfr-C2 (cat no. 1102711), and Dr-crlf1a-C3 (cat no. 1102861). The 3-plex negative control probe mixture was used on a separate slide. Probe amplification with AMP1, AMP2, and AMP3 proceeded according to manufacturer instructions. Development of the HRP signal also proceeded according to manufacturer instructions with the Opal 520, Opal 570, and Opal 690 dyes (1:1000; Akoya Biosciences; Menlo Park, CA). After signal development for each probe, slides were either incubated for 5 min at room temperature with DAPI solution (Advanced Cell Diagnostics), mounted with Prolong Gold Antifade Mountant (Thermo Fisher Scientific; Waltham, MA), and coverslipped or washed in PBS-T for 5 min before proceeding with immunohistochemistry. The primary antibodies used were rabbit anti-GFP (1:1000; Abcam; Cambridge, United Kingdom; AB6556) and mouse anti-PCNA (1:1000; MilliporeSigma; Burlington, MA; P8825), followed by secondary antibodies Alexa Fluor 488conjugated goat anti-rabbit (1:500; Thermo Fisher Scientific; Waltham, MA; A11034) and Alexa Fluor 647-conjugated goat antimouse (1:500; Thermo Fisher Scientific; Waltham, MA; A21245).

Terminal deoxynucleotidyl transferase dUTP nick end labelling (TUNEL)
Slides were washed and rehydrated in PBS and fixed in 4% PFA for 5 min at room temperature. Slides were washed 3 times for 5 min in PBS and permeabilized for 15 min in PBS-T (5% Tween-20) at room temperature. Slides were washed 3 times for 5 min in PBS and incubated for 30 min with 1:150 proteinase K (Takara Bio; Kusatsu, Shiga, Japan) in PBS at room temperature. Slides were then washed 3 times for 5 min in PBS, before the TUNEL protocol was performed as described previously (Lahne et al., 2021).

Image acquisition and analysis
A Nikon A1 confocal microscope, equipped with a ×40 plan-fluor oil immersion objective, was used to acquire 6.5 μm z-stacks in 0.8 μm steps, of 1024 × 1024 images of the central-dorsal region of the retina. Counts were performed manually using Fiji software throughout the 6.5 μm z-stack and normalized to a 300 μm length of the retina as previously described (Lahne et al., 2015). Statistical analyses were performed using GraphPad Prism 9 (San Diego, CA). Statistical tests used included either a Student's t-test or a one-way ANOVA with a Tukey or Dunnett post hoc test. A p-value of less than 0.05 was considered significant. The mean ± SEM for each experiment is stated in the text of the appropriate Results section and the sample size (n) is stated in the appropriate figure legend.

Single cell RNA-seq (scRNA-seq) analysis
We used scRNA-seq data from whole light-damaged retinas that was previously published (Hoang et al., 2020). scRNA-seq analysis was Frontiers in Cell and Developmental Biology frontiersin.org performed using Seurat (Hao et al., 2021). Established clusters from the previous publication were maintained and differential expression analysis was performed comparing resting and activated Müller glia.

CNTFR ligands Clcf1 and Crlf1a are expressed in the activated Müller glia in lightdamaged retinas
To investigate the spatial expression of CNTFR ligands in the light-damaged retina, we analyzed previously published scRNA-seq datasets (Hoang et al., 2020). We did not detect expression of cntf in any of the retinal cell types (Figure 1). In contrast, we observed expression of both clcf1 and crlf1a primarily within activated Müller glia (Figure 1), suggesting that Müller glia express and secrete the Clcf1 and Crlf1a ligands in response to retinal damage. These expression patterns were similar to other key pro-proliferative factors within the zebrafish retina, such as hbegfa and lepb (Wan et al., 2012;. In contrast, the cytokine LIF (encoded by the m17 gene) was expressed predominantly in microglia and not in Müller glia (Figure 1). To begin to understand how these ligands interact with cells within the damaged retina, we determined what cell types expressed any of the corresponding receptor components. We observed only low levels of cntfr expression in the scRNA-seq datasets, suggesting its expression was not induced during the first 36 h of retinal regeneration. Additionally, of the two LIF receptor genes encoded in zebrafish (Ogai et al., 2014), lifra was not significantly expressed in the retina and lifrb expression was upregulated in the activated Müller glia relative to the resting Müller glia ( Figure 1).
Finally, expression of il6st (encoding Gp130) was primarily limited to activated Müller glia, vascular endothelial, retinal pigmented epithelium, and pericytes, suggesting these cells are particularly sensitive to Gp130-mediated signalling. Together, these data suggest that Clcf1 and Crlf1a are released from activated Müller glia following injury and potentially interact with a variety of different cell types within the retina.
We further explored the expression of clcf1 and crlf1a, and several associated genes, in both resting and activated Müller glia, as well as across the damage time course within the activated Müller glia population. Both clcf1 (p < 1.00 × 10 −200 ) and crlf1a (p < 1.00 × 10 −200 ) exhibited a significant increase in expression within the activated population of Müller glia (Figures 2A,B, respectively). In addition, clcf1 and crlf1a expression were upregulated by 10 h and 4 h of light damage (Figures 2A,B, respectively). This suggests that Müller glia upregulates the expression of both genes prior to photoreceptor cell death and they both may be required for either neuroprotection or proliferation. Little expression of cntf was observed in either resting or activated Müller glia ( Figure 2C), indicating that Clcf1 and Crlf1a are the primary CNTFR ligands expressed within the retina. In contrast, clcf1 expression was significantly increased (p = 7.40 × 10 −183 ) in activated Müller glia in the light-treated mouse retina (Supplementary Figure S1A), while crlf1 expression was significantly decreased (p = 0.03) in activated Müller glia relative to resting Müller glia (Supplementary Figure S1B). Similar to zebrafish, we did not detect expression of cntf in either mouse Müller glia population following light treatment (Supplementary Figure S1C).
In the zebrafish retina, expression of the receptor components cntfr ( Figure 2D) and lifra (Figure 1) was relatively low in both Müller glia populations. In contrast, lifrb expression ( Figure 2E) was significantly higher in activated Müller glia (p = 6.00 × 10 −153 ) and  Figure 2F). This rapid induction and sustained expression of il6st suggests that Gp130-mediated signalling may have both neuroprotective and pro-proliferative effects. The expression differences seen in the receptor genes may be due to Gp130 forming a receptor complex in both a CNTFR-and LIFRindependent manner, and therefore is likely required for additional processes (White & Stevens, 2011). Interestingly, Cntfr expression was observed in resting mouse Müller glia, however Cntfr was significantly downregulated in activated Müller glia (Supplementary Figure S1D; p = 7.11 × 10 −24 ). No change was observed in Lifr expression (Supplementary Figure S1E), but Il6st expression was significantly increased within activated Müller glia (Supplementary Figure S1F; p = 3.80 × 10 −11 ). To confirm that clcf1 and crlf1a were expressed in Müller glia, we performed in situ hybridization using the RNAscope system. Dark-adapted albino; Tg[gfap:EGFP] nt11 fish were placed in intense light treatment for either 0 (undamaged), 10, 20, or 36 h before eyes were collected and retinal sections were labeled for either clcf1 or crlf1a, GFP, and PCNA. A broad and low level of clcf1 expression was detected at 0 h ( Figures 3A-A''). However, clcf1 expression increased and colocalized with GFP-expressing Müller glia (Tg(gfap:EGFP)) from 10 through 20 h ( Figures 3B-C''). By 36 h of light treatment, the clcf1 signal appeared weaker and more dispersed throughout the retina ( Figures 3D-D''). However, clcf1 puncta were still observed within proliferating Müller glia ( Figures 3D, D''), suggesting that clcf1 was expressed in Müller glia that would enter the cell cycle. Little expression of crlf1a was observed at 0 h ( Figures  3E-E''). However, by 10 and 20 h of light treatment, crlf1a was observed within Müller glia ( Figures 3F-G'') and persisted within proliferating Müller glia through 36 h of light treatment ( Figures  3H-H''). Furthermore, clcf1 and crlf1a were observed within the same Müller glia at 36 h of light treatment (Supplementary Figures  S2A, S2B). Additionally, PCNA-negative Müller glia also did not Frontiers in Cell and Developmental Biology frontiersin.org express either clcf1 or crlf1a ( Figures 3D,H), which likely represent resting Müller glia. These expression patterns closely resemble the expression time course identified by scRNA-seq and revealed that clcf1 and crlf1a are highly upregulated within Müller glia early in the damage paradigm, before photoreceptor cell death. We also examined the expression patterns of the cntfr and il6st receptors. Expression of cntfr was observed broadly throughout the retina from 0 to 36 h of light treatment (Supplementary Figures  S3A-S3D'') with no obvious changes in expression level or pattern, which is consistent with the scRNA-seq data (Figure 1). A similar expression pattern was observed for il6st (Supplementary Figures  S3E-S3H''). This data suggests that cells within each nuclear layer may respond to the release of Clcf1 and Crlf1a from Müller glia following injury.

Intravitreal injection of recombinant CLCF1/CRLF1 protects against photoreceptor cell death during light treatment
Previous studies determined that CNTF is neuroprotective within both the zebrafish and mammalian retinas (Kassen et al., 2009;Aslam et al., 2013). We examined whether Clcf1 and Crlf1a, which are alternative ligands for the CNTF receptor, can provide similar neuroprotective benefits. We intravitreally injected albino zebrafish with either PBS (vehicle) or 1 mg/ml CLCF1/CRLF1 every 12 h until 48 h after the initial injection, at which point fish were placed in constant intense light. They received a further injection at 12 h of constant light and eyes were collected 12 h later (24 h light treatment). To quantify cell death, we performed the TUNEL assay on retinal sections ( Figures 4A-D). We observed a significant decrease in the number of TUNEL-positive ONL cells in Clcf1/Crlf1a-co-injected retinas relative to vehicle control ( Figure 4E; PBS: 86.68 ± 6.19, CLCF1/CRLF1: 47.34 ± 7.18, p = 0.0005). This demonstrates that clcf1/crlf1a co-expression by zebrafish Müller glia may be neuroprotective in the light-damaged retina.
We next examined whether the reduction in Müller glia proliferation was a result of reduced cell death in the morphants by electroporating albino zebrafish with either S.C. (Supplementary Figures  S4A, S4E), cntfr (Supplementary Figures S4B, S4F), clcf1/crlf1a (Supplementary Figures S4C, S4G), or all three morpholinos (Triple MO, Supplementary Figures S4D, S4H) and collecting the retinas at 24 h of light treatment. The TUNEL assay was performed to assess cell death ( Supplementary Figures S4A-S4H). No significant differences in the number of TUNEL-positive cells were observed between any treatment group relative to S.C. (Supplementary Figure S4I; S.C.: 109.9 ± 8.75 TUNEL-positive cells, cntfr: 112.5 ± 4.19 p = 0.98, clcf1/crlf1a: 114.7 ± 7.46 p = 0.92, Triple: 112.4 ± 5.12 p = 0.99), suggesting that the reduced number of proliferating Müller glia in either the cntfr, clcf1/ crlf1a or the Triple morphants was not due to decreased cell death, but rather their requirement for stimulating Müller glia proliferation, with Clf1/Crlf1a likely acting as ligands for CNTFR in the light-treated retina.

FIGURE 4
Intravitreal injection of CLCF1/CRLF1 protects against photoreceptor cell death during light treatment. (A-D) Single z-plane confocal images from dark-adapted albino zebrafish that were lighttreated for 24 h while being intravitreally injected with either PBS (A, C) or CLCF1/CRLF1 (B, D). TUNEL assay was performed to assess cell death, and nuclei were counterstained with DAPI. (E) Quantification showing significant decrease in the number of TUNEL-positive nuclei within the ONL of CLCF1/CRLF1-treated retinas. Student's t-test, p = 0.0005, n ≥ 11. Mean ± SEM, ***p < 0.001. ONL, outer nuclear layer, INL, inner nuclear layer, GCL, ganglion cell layer. Scale bar represents 20 µm.
Frontiers in Cell and Developmental Biology frontiersin.org ( Figures 7B,D). Fish received both injections every 12 h. At 72 h after the initial injection, retinas were isolated, sectioned, and labelled for GFP and PCNA to assess Müller glia proliferation. We observed no significant difference in the number of PCNA-positive Müller glia between the DMSO and NVP-treated groups ( Figure 7E; DMSO: 0.57 ± 0.38, NVP: 0.49 ± 0.33, p = 0.87). However, there was a significant decrease in the number of proliferating rod precursor cells in the NVP-treated group relative to vehicle control ( Figure 7F; DMSO: 25.17 ± 2.43, NVP: 11.23 ± 1.36, p < 0.0001). This suggests that CLCF1/CRFL1-induced rod precursor cell proliferation was IGF-1R dependent. To confirm that increased rod precursor cell proliferation was not a product of increased Müller glia proliferation, dark-adapted albino; Tg[gfap:EGFP] nt11 zebrafish were placed in constant intense light treatment, while also receiving intraperitoneal injections of either DMSO (Figures 7G,I) or NVP ( Figures 7H,J) beginning at the start of the light treatment and every 12 h thereafter. Eyes were collected after 36 h of constant light treatment and sections were labelled for GFP and PCNA to assess Müller glia proliferation. We did not detect a significant change in the number of PCNA-positive Müller glia between treatment groups ( Figure 7K; DMSO: 38.04 ± 2.63, NVP: 40.03 ± 1.88, p = 0.54), suggesting that inhibiting IGF-1R does not affect Müller glia proliferation. This is consistent with scRNA-seq data that revealed little expression of either igf1 or the associated receptor components (igf1ra and igf1rb) in Müller glia in the light-damaged retina ( Figure 1). Together these data suggest that CLCF1/CRFL1 induces rod precursor proliferation in an IGF-1R-dependent manner.

Discussion
This is the first demonstration that the Ciliary neurotrophic factor receptor (CNTFR) ligands Cardiotrophin-like cytokine factor 1 Frontiers in Cell and Developmental Biology frontiersin.org (Clcf1) and Cytokine receptor-like factor 1a (Crlf1a) are required for retinal regeneration. We investigated expression of these ligands and their corresponding receptor components within the damaged retina and showed that Clcf1 and Crlf1a are neuroprotective, are required for Müller glia proliferation, and can induce rod precursor proliferation in an IGF-1R-dependent manner. Previously, it was demonstrated that both clcf1 and crlf1a are expressed in Müller glia-derived precursor cells at 96 h post injury , however no in-depth analyses have fully investigated the spatio-temporal expression or function of either ligand. Our scRNA-seq data demonstrate that both clcf1 and crlf1a are expressed rapidly following retinal damage with Müller glia expressing these ligands as early as 4 h post injury and maintaining expression until at least 36 h post injury. This expression pattern was further confirmed by in situ hybridization, which demonstrates coexpression of clcf1 and crfl1a in proliferating MG, indicating that these ligands are expressed specifically in activated MG. In contrast, we observed little expression of cntf, which was previously shown to be neuroprotective in both mammalian (Tao et al., 2002;Sieving et al., 2006;Azadi et al., 2007;Rhee et al., 2013) and zebrafish models of retinal damage (Kassen et al., 2009) and can induce Müller glia proliferation within zebrafish (Kassen et al., 2009;. This suggests that Clcf1 and Crlf1a are the primary CNTFR ligands expressed within the regenerating retina and may have similar effects on the retina. Similar to the effects of CNTF (Kassen et al., 2009), intravitreal injection of CLCF1/CRLF1 significantly reduces photoreceptor cell death during light treatment, indicating that Clcf1 and Crlf1a may play a neuroprotective role following damage, which is further supported by the expression of these ligands prior to the occurrence of cell death. Additionally, knockdown of clcf1 and crlf1a significantly reduced Müller glia proliferation following light treatment, however intravitreal injection of CLCF1/CRLF1 was insufficient to induce Müller glia proliferation in the undamaged retina. Therefore, Clcf1 and Crlf1a are necessary, but not sufficient, for the induction of Müller glia proliferation. Clcf1 and Crlf1a have also been shown to be required for optic nerve regeneration (Elsaeidi et al., 2014), suggesting a general role for these ligands in neuronal regeneration. The contrast in Müller glia proliferation between CNTF and CLCF1/ CRLF1-treated retinas may be due to differences in downstream actions of these ligands. Mutations in CRLF1 or CLCF1 have been linked to cold-induced sweating syndrome, however mutations in CNTF do not induce similar effects (Murakami et al., 2019), although both have been demonstrated to induce expression and phosphorylation of Stat3 (Kassen et al., 2009;Nahlé et al., 2019;Rezende et al., 2009;Savin et al., 2015 which has been shown to be essential in Müller glia for retinal regeneration (Nelson et al., 2012). Despite the inability to increase Müller glia proliferation, intravitreal injection of CLCF1/CRLF1 did induce a significant increase in rod precursor cell proliferation. A similar increase in rod precursor cell proliferation was observed following intravitreal CNTF injection (Kassen et al., 2009). It has previously been demonstrated that proliferation of rod precursor cells is IGF-1R1R-dependent (Lahne et al., 2019). We demonstrated that CLCF1/CRLF1-induced rod precursor cell proliferation is IGF-1R-dependent, although Müller glia proliferation following injury is not. It has previously been shown that IGF-1 can synergize with FGF2 to induced Müller glia proliferation, and that igfra is required for Müller glia proliferation following injury (Wan et al., 2014), however co-injection of IGF-1 with CLCF1/ CRLF1 failed to induce additional Müller glia or rod precursor cell proliferation. Therefore, the factors that may synergize with Clcf1 and Crlf1a to induce Müller glia proliferation following injury remain unclear.
To better understand how these ligands operate within the zebrafish retina, we also investigated expression of the receptor complex components. NPCs have previously been shown to express cntfr at 96 h post injury . Within scRNA-seq datasets we observed little expression of cntfr, however in situ hybridization demonstrated that cntfr is expressed throughout the retina at all timepoints studied. Knockdown of cntfr also demonstrated that CNTFR is required for Müller glia proliferation. We observed a significant increase in expression of lifrb and il6st within activated Müller glia, and in situ hybridization confirmed that il6st is expressed throughout the retina at all timepoints. These findings suggest that Müller glia, and likely most cell types of the retina, are sensitive to CNTFR ligands. This is further supported by transcriptional changes in multiple cell types following CNTF exposure in the mouse retina .
Together this data suggests a model in which Müller glia respond to injury by producing the CNTFR ligands Clcf1 and Crlf1a, which play a dual role within the regenerating retina showing both neuroprotective and pro-proliferative properties. Based upon expression of receptor components, Clcf1 and Crlf1a can signal to most cell types. It has previously been shown within the mouse retina that the neuroprotective effects of CNTF require GP130 within Müller glia (Rhee et al., 2013), therefore changes induced in Müller glia in response to CNTFR ligands may provide the observed neuroprotective effects, rather than direct CNTFRmediated signalling within damaged cell types. CNTF has also previously been shown to interact with microglia (Kahn et al., 1995;Baek et al., 2018), which are known to be required during retinal regeneration (Conedera et al., 2019;Mitchell et al., 2019). Therefore, the potential interaction of Clcf1 and Crlf1a with microglia within the damaged retina may be one way in which these two cell types interact.

Data availability statement
The datasets presented in this study can be found in online repositories. All scRNA-seq data and source codes are available at GitHub https://github.com/jiewwwang/Single-cell-retinalregeneration. The scRNA-seq data can be queried interactively at https:// proteinpaint.stjude.org/F/2019.retina.scRNA.html.

Ethics statement
The animal study was reviewed and approved by University of Notre Dame Animal Care and Use Committee.