https://orcid.org/0000-0003-3788-2265
Figures
Abstract
Emerging evidence implicates retinal microglia and inflammation as important components impacting the outcome of retinal regeneration, which is spontaneously achieved in zebrafish retina following acute damage but is limited or blocked in mammals. Here we describe the regenerative response in the larval zebrafish retina following ablation of cone photoreceptors. To investigate the role of microglia in the regenerative response, we used irf8st95 heterozygote (microglia-sufficient) and irf8st95 homozygous mutant (microglia-deficient) zebrafish. We compared multiple aspects of the regenerative response in irf8 + /- and irf8-/- larval retinas, including entry of the Müller glia (MG) into the cell cycle, the amplification of MG-derived progenitor cell (MGPC) proliferation, inflammatory and glial reactivity-associated gene expression, and the regeneration of cones. We found only modest impacts to early and late stages of MGPC proliferation and to inflammatory gene expression in irf8 mutants, with no obvious impacts to the regeneration of cones. Notably, we detected a population of immune cells in irf8 mutants that emerged following cone ablation, which expanded in number then were reduced over time, following a trajectory similar to microglia-sufficient siblings but at markedly reduced abundance. The immune cells detected in irf8 mutants included a subset with L-plastin/4C4 antibody staining patterns different than those in microglia-sufficient siblings, suggesting distinct origins and/or phenotype compared to resident microglia in controls. The presence of immune cells in irf8 mutants following cone ablation limited our ability to make strong conclusions about the role of microglia in regeneration of cones. However, our results are consistent with several reports that indicate a role for microglia in regulating MGPC proliferation in the regenerating retina. Collectively considered with other reports, our results further indicate that compensatory responses, which may include different immune cells and/or signaling from other retinal cell types such as Müller glia, may be elicited in microglia-deficient retinas upon neuronal damage.
Citation: Rumford JE, Farre AA, Mai J, Weimar HV, Shelton CD, Morales M, et al. (2026) Cone photoreceptor ablation in microglia-deficient larval zebrafish retina elicits a regenerative response alongside a compensatory immune cell response. PLoS One 21(7): e0344737. https://doi.org/10.1371/journal.pone.0344737
Editor: Yuk Fai Leung, Purdue University, UNITED STATES OF AMERICA
Received: February 19, 2026; Accepted: June 29, 2026; Published: July 23, 2026
Copyright: © 2026 Rumford et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the manuscript and its Supporting Information files.
Funding: This work was funded by NIH/NEI R01EY030467 awarded to Diana M Mitchell, a Bright Focus Foundation Postdoctoral Research award (M2025012F) to Ashley A Farre, a University of Idaho Office of Undergraduate Research Summer Research Fellowship to Claire D Shelton, and a NSF-Graduate Research Fellowship (Grant No. DGE-2235197) to Justin Mai. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the funding agencies named here.
Competing interests: The authors have declared that no competing interests exist.
Introduction
The zebrafish retina has a remarkable and well-appreciated intrinsic regenerative capacity. Because the cell types involved are also present in mammalian retinas, work in zebrafish has generated important translatable knowledge about the cells and genetic pathways that result in successful regeneration of retinal neurons. In zebrafish retina, Müller glia (MG) respond to acute damage of neurons by dividing asymmetrically [1] to produce a neuronal progenitor cell pool that then amplifies in abundance via proliferation [2–6]. The MG-derived progenitor cells (MGPCs) subsequently differentiate into neurons that integrate into retinal circuits to restore visual function [7–12]. Temporally alongside (or even prior to) the Müller glial response, microglia also respond to acute retinal damage [13–21]. These features of glial responses have been demonstrated in numerous experimental systems of neuronal damage to the zebrafish retina [1,3,6,7,14,18,21–24]. Several studies indicate that inflammatory signals, to which microglia likely contribute, are important for Müller glial responses and MGPC proliferation [14,15,21,25,26] and may also impact later outcomes such as the survival of regenerated neurons [21,27,28]. Given their activation in multiple retinal degenerative diseases, acute retinal injury [13,17,27,28], and intercellular interactions with the Müller glia [29,30], there is strong interest in identifying the function of microglia in retinal regeneration as this will lead to a better understanding of supportive versus detrimental activities of these cells which could be modulated therapeutically. Microglia could provide inflammatory signals and cues that play into a broad signaling axis [31,32]. Alternatively, or in addition, their role as phagocytes for clearance of debris and damaged/dying cells [13,29,33] could shape regenerative responses that follow neuronal death.
The first published studies to probe the role of microglia in retinal regeneration in zebrafish used distinct tools to modulate microglial presence in the retina upon acute damage. These included a transgenic system in which the mpeg1 promoter (active in zebrafish microglia and macrophages [34,35]) drives bacterial nitroreductase (NTR) conferring sensitivity to the pro-drug metronidazole [14], a pharmacological approach for CSF1R inhibition using the drug PLX3397 [20,36], and a genetic mutant with deficient microglial development [36]. Further, the immunosuppressive drug dexamethasone has been used to probe immune signaling in retinal regenerative responses [14,15,21,22,26,36], and though such experiments suggest inflammatory signaling may be involved, they do not directly implicate microglia because this drug has broad effects on multiple cell types including the Müller glia [37,38]. A more recent paper used nanoparticle technology to target dexamethasone to microglia, which increased kinetics of rod photoreceptor regeneration [39]. The generation of a zebrafish line with a loss-of-function mutation in the gene irf8 provided a novel genetic model for microglia deficiency [40] and is an attractive tool to further probe the role of microglia in retinal regeneration. To date this zebrafish irf8 mutant has been utilized in several studies focusing on the retina [15,18,36,41]. Other zebrafish genetic mutants have since been generated which are deficient in microglia; these include csf1r mutants though the two zebrafish paralogs encoding csf1ra and csf1rb have differential impact on microglial development [42]. To sufficiently deplete microglia, csf1ra/b double mutants may be required for experiments.
Currently, the literature does not provide a clear consensus about the role of microglia in retinal regeneration. While some studies in zebrafish suggest retinal regeneration is hindered in the absence of microglial functions [15,20,31,36,43] other work was less supportive of such a role [18]. Some of the underlying factors that may have resulted in different conclusions between these reports could potentially be the differences in microglia abundance or phenotype in the different mutant lines (irf8 vs csf1ra/b), use of different drug-based approaches, differences in the retinal damage systems used, or the age of the fish used in the experiments (larval versus adult). Timing of microglial inhibition may also be a factor [14,39]. Further, in mouse retina with forced neuronal production after damage, microglia were found to inhibit retinal regeneration [28], suggesting microglia could behave differently in different species.
Given our previous work demonstrating effects on Müller glia (MG) in microglia-deficient irf8st95 mutant zebrafish [41], an interest to further probe the role of microglia in retinal regeneration, and the ability to achieve microglial deficiency with a single gene mutation, we examined MG-mediated regenerative responses and the regeneration of cone photoreceptors in the larval zebrafish retina in the irf8st95 mutant. We used metronidazole (Mtz)-mediated ablation of cone photoreceptors in transgenic gnat2:nfsb-mCherry [23,44] fish at larval ages. Our results are consistent with several components of multiple reports investigating the role of microglia in retinal regeneration in zebrafish [14,15,18,36]. Homozygous irf8 mutants showed only modest decreases in MGPC proliferation and inflammatory gene expression following cone ablation, while regeneration of cones at 7- and 14-days post injury was not significantly impacted in irf8 mutants. While irf8 mutants were highly deficient in microglia in larval retina at baseline, upon induced retinal damage (i.e., ablation of cones), leukocytes consistent with microglia/macrophage features were detected in irf8 mutants causing difficulties in conclusive connections about the role of microglia in regeneration of cones. Expression patterns of selected microglia markers commonly used in zebrafish (4C4 and L-plastin antibody labeling), as well as leukocyte localization in the damaged and regenerating retina, were different in irf8 mutants compared to heterozygous siblings. Collectively, our results suggest that compensatory innate immune responses, which may include immune cell populations phenotypically distinct from resident microglia and/or signaling from other retinal cell types such as the Müller glia, are elicited in microglia-deficient retinas upon photoreceptor damage.
Results
Ablation of cones in larval zebrafish retina activates microglial responses
Using gnat2:nfsb-mCherry transgenic zebrafish larvae [23], in which the Nfsb/NTR enzyme tagged with mCherry fluorescent protein is expressed in cone photoreceptors, we administered DMSO (vehicle) or the pro-drug metronidazole (Mtz) at ~4.5 days post fertilization (dpf) via immersion to induce death of cones. After 24–48 hours of treatment, larvae were collected for analysis of cone death and glial responses (Fig 1). Although ablation of a central patch of cones has been reported in gnat2:nfsb-mCherry larvae with shorter Mtz treatment duration [23], we aimed to ablate all cones in our experiments and found that this required longer duration of Mtz immersion. We did not detect reliable, consistent ablation of cones in larval retinas with 24-hour Mtz exposure. However, by 48 hours of Mtz treatment, mCherry+ cones showed morphological alterations and disorganization indicating collapse compared to DMSO treatment (Fig 1A,B). Retinal cryosections further recapitulated this result at 48 hours post-treatment with Mtz (48 hpt) and additionally demonstrated abundant TUNEL staining in the outer nuclear layer (ONL) that co-localized with collapsed mCherry (cone reporter) signal, indicating that cones were ablated with this treatment regime (Fig 1C,D). Microglia were visualized by staining with the 4C4 antibody, thought to recognize LGALS3BP [45] and commonly used to detect zebrafish microglia. 4C4 + microglia were rarely detected within the ONL of DMSO-treated retinas (Fig 1C) but were observed within the damaged ONL of Mtztreated retinas at 48 hpt (Fig 1D). The microglia were positioned around and within dying cone signal and often observed to encircle TUNEL+ signal (Fig 1D,D’,D’’), consistent with microglial engulfment of dying cones.
A, B. Whole larval eyes imaged on intact gnat2:nfsb-mCherry zebrafish larvae, using confocal microscopy, at 48 hours post treatment (hpt) with DMSO (A) or metronidazole (Mtz, B) visualizing nuclear (DAPI) stain and mCherry fluorescence in cones. Annotation v, d indicate ventral and dorsal, respectively, with ventral located at the bottom of the image stack and dorsal at top. C, D. Retinal cryosections collected at 48 hpt following DMSO (C) or Mtz (D) treatment, and stained for microglia (4C4 antibody), TUNEL, and DAPI; mCherry fluorescence from cones. Microglia localized to collapsed, TUNEL+ cones in Mtz treated retinas (arrowheads, D). D’, D” is an enlarged region (dotted box in D) showing microglia (4C4, cyan) engulfing TUNEL+ cones (yellow) in Mtz treated retinas. E, F. Whole eyes from larval gnat2:nfsb-mCherry;TP1:mTurquoise double transgenic fish, stained for microglia (4C4), and imaged by confocal microscopy at 24 hpt (DMSO or Mtz treatment). Selected optical sections are projected showing retina radially around the lens. Arrows indicate microglia localized at basal face of ONL/OPL in DMSO treated samples (E) and their transition to the apical region of the ONL in Mtz treated samples (F). Asterisks (*) in F show Müller glia cell bodies that have shifted apically towards the ONL.
Examining whole eyes/retinas 24 hours earlier (at 24 hpt), we observed that microglia showed differential localization by 24 hpt with Mtz compared to DMSO (Fig 1E,F and Fig 2). Microglia were identified at the outer plexiform layer (OPL)/interface of the ONL in DMSO samples (Fig 1E) but in Mtz treated retinas, microglia were localized to the apical regions of the ONL (Fig 1F). We used the TP1:mTurquoise reporter to visualize Müller glia (MG) [29], which revealed that MG also showed morphological changes in Mtz treated samples, such as apical positioning of MG cell bodies (Fig 1E,F). Collectively, these images indicate that 48 hours of Mtz treatment reliably induces the death of cones in gnat2:nfsb-mCherry larval fish. Further, changes observed for microglia and Müller glia indicate that these cell types sense changes to cones by 24 hpt.
A-G. Images of retinal cryosections from gnat2:nfsb-mCherry fish collected at the indicated timepoints (hours post treatment, hpt) following DMSO (A) or Mtz treatment (B-G) and stained for 4C4 (microglia) and DAPI. Cones visualized by mCherry fluorescence. Asterisks in E, F, and G indicate cones that likely emerged from CMZ progenitors after Mtz washout as a result of retinal growth in zebrafish larvae. H. Quantification of the total number of 4C4 + microglia per retinal cryosection at the indicated timepoints following DMSO or Mtz treatment. I. Quantification of the number of 4C4 + microglia localized to the outer retina/ONL per retinal cryosection for the indicated timepoints/treatments. Microglia counted in this region were in the ONL or apical to the ONL in regions consistent with photoreceptor outer segments and retinal pigment epithelium (RPE). Two-way ANOVA indicated an effect of treatment and timepoint (p < 0.05); Tukey’s post-hoc comparisons are shown between DMSO and Mtz groups for each timepoint, *p < 0.05, **p < 0.01, ***p < 0.001.
We next examined microglia over multiple timepoints following cone ablation, by staining retinal cryosections from gnat2:nfsb-mCherry fish collected at 24-hour intervals from 48–144 hours post-treatment (Fig 2). Again, at 48 hpt we observed responding 4C4 + microglia interacting with mCherry+ cone signal in the damaged ONL upon Mtz treatment (Fig 2C). Following Mtz washout, mCherry+ cones were largely absent from central retina (72–144 hpt, Fig 2D-G). Some mCherry+ cones were observed at the periphery and likely emerged from the ciliary marginal zone (CMZ), a region of neurogenesis contributing to growth of the fish retina [46,47], following Mtz removal (asterisks, Fig 2E-G). However, central retinal regions showed absence of mCherry+ cones and loss of nuclear organization after Mtz treatment (Fig 2C-G) consistent with cone ablation, where regenerative responses are required for repopulation of cones. We observed microglia localized to the damaged ONL at 48 and 72 hpt following Mtz treatment (Fig 2C, D). Quantification of 4C4 + cell numbers revealed that microglial numbers (total numbers per retinal cryosection and numbers localized to the outer retina/ONL) were increased at 48 and 72 hpt following Mtz treatment (Fig 2H, I) and total numbers per cryosection were elevated in Mtz samples compared to DMSO samples through 144 hpt (Fig 2H). Collectively, the increased microglial numbers and localization to the outer retina/ONL is consistent with their response to cone death. The increased total numbers of microglia through 144 hpt in retinal cryosections suggest that microglia could perform functions connected to regeneration.
Entry of Müller glia into the first asymmetric division is not altered in irf8 mutants
Previous work has shown that following retinal damage in zebrafish, the responding Müller glia enter an initial asymmetric division to produce a pool of proliferating progenitor cells [1]. To begin to dissect the role of microglia in regeneration of cones in our system, we first examined the timing of the initial Müller glial asymmetric division following Mtz treatment of gnat2:nfsb-mCherry larvae. We collected samples at 24 and 48 hpt and stained retinal sections for glutamine synthetase (GS) and PCNA to label Müller glia that had entered the cell cycle (Fig 3A-C). At 24 hours following DMSO or Mtz treatment, GS+PCNA+ cells were not significantly detected in the inner nuclear layer (INL) (Fig 3B, 3D). At 48 hpt, numerous GS+PCNA+ (Müller glia in S-phase) were detected in the INL of Mtz treated samples compared to DMSO treated samples (Fig 3C, 3D), indicating that Müller glia enter the cell cycle between 24 and 48 hours following Mtz treatment. We further confirmed the timing of MG cell cycle entry using EdU immersion. In these experiments, DMSO/Mtz treatment followed the same schedule, with EdU added from 24–48 hpt. Retinal sections from samples collected at 48 hpt were stained for GS, PCNA, and EdU (Fig 3E-E’’). Quantification of EdU incorporation revealed that by 48 hpt, a subset of PCNA+ Müller glia also incorporated EdU to detectable levels (Fig 3F) while there were essentially no GS+ Müller glia cells that stained positive for EdU but negative for PCNA (Fig 3G, which would indicate MG that had completed a cell cycle). These results indicate that the first responding Müller glia enter S-phase within the window of 24–48 hpt following Mtz treatment and that this window labels Müller glia in the first asymmetric division.
Representative images of retinal cryosections from gnat2:nfsb-mCherry fish collected at the indicated timepoints (hours post treatment, hpt) following DMSO (A) or Mtz treatment (B,C), and stained for Glutamine Synthetase (GS), PCNA (Proliferating Cell Nuclear Antigen) and DAPI. Scale bar in B applies to A-C. C. Arrows indicate GS+PCNA+ Müller glia in the inner nuclear layer (INL). D. Quantification of the number of GS+PCNA+ cells per retinal section following DMSO or Mtz treatment at the indicated timepoints. **p < 1x10-5, 2-way ANOVA with Tukey’s post-hoc. E-E.” Representative images of retinal cryosections from gnat2:nfsb-mCherry fish collected at 48 hpt of Mtz treatment, following immersion in EdU from 24 to 48 hpt. Images show staining and detection of GS, PCNA, and EdU; arrows indicate GS+ Muller glia in the INL that are also PCNA+ and EdU + . Scale bar in E” applies to E, E’, E.” F. The fraction of GS+PCNA+ cells that were also EdU+ detected in retinal cryosections collected at 48 hpt following Mtz treatment and exposed to EdU from 24 to 48 hpt. G. The number of GS+ cells that were PCNA-negative but EdU+ in the cryosections collected at 48 hpt.
We and others have shown microglia are deficient in numbers in irf8-/- larval retinas at timepoints prior to and spanning the Mtz treatment window in our experiments [36,41]. To determine if irf8-/- retinas remain microglia-deficient after cone ablation, we crossed gnat2:nfsb-mCherry;irf8 + /- breeders to irf8-/- fish to produce offspring carrying the gnat2:nfsb-mCherry transgene that are either irf8 + /- or irf8-/- genotype. At 48 hpt following DMSO or Mtz immersion, larvae were collected for retinal cryosections. We performed microglia/leukocyte staining in a subset of samples at 48 hpt (~6.5 dpf age). This analysis was done because microglia/leukocyte numbers may change over development and/or in response to DMSO/Mtz treatment. Further, peripheral macrophages are reported to develop in irf8 mutants around 6 days of age [40] (which coincides with our 48 hpt timepoint) and could potentially respond to cone damage. We examined retinal sections after staining with both 4C4 antibody and an antibody to L-plastin (Lcp1), a pan-leukocyte marker in zebrafish [13,48] (Fig 4). In irf8 + /- retinas, cells that stained for both Lcp1 and 4C4 were readily identified in the damaged cone layer (Fig 4A-A’’). In irf8-/- retinas, the number of cells staining for these markers was significantly reduced though a limited number of Lcp1 + cells (sometimes but not always co-labeling with 4C4) were detected in the damaged ONL (Fig 4B-B’’”). Quantifications of Lcp1 + and 4C4 + cell numbers revealed significantly reduced numbers in irf8-/- compared to irf8 + /- siblings (Fig 4C,D); however, it is worth noting that some cells staining with these markers were indeed detected in irf8 mutant retinal sections. Therefore, the irf8 mutant remains deficient in microglia/leukocyte numbers in response to cone damage at 48 hpt. We noted that a lower percentage of the Lcp1 + cells detected in irf8 mutants also co-labeled with the 4C4 microglia marker compared to irf8 + /- (Fig 4E), suggesting that the leukocytes detected in irf8 mutants may be phenotypically different than the responding microglia/macrophages in irf8 + /-.
Retinal cryosections from gnat2:nfsb-mCherry;irf8 + /- or gnat2:nfsb-mCherry;irf8-/- larval zebrafish samples collected at 48 hours post Mtz treatment and stained for 4C4, Lcp1, and DAPI; cones are visualized by mCherry fluorescence. A-A.” Microglia/leukocytes that stain for Lcp1 and 4C4 are detected in the ONL in response to cone ablation at 48 hpt in irf8 + /- retinas. Scale bar in A” applies to A-A.” B-B.” Some, but reduced, Lcp1 and 4C4 + cells are detected in the ONL in response to cone ablation at 48 hpt in irf8-/- retinas. Scale bar in B” applies to B-B.” C. Quantification of the number of 4C4 + cells in retinal cryosections of irf8 + /- or irf8-/- fish at 48 hpt following Mtz treatment. D. Quantification of the number of Lcp1 + cells in retinal cryosections of irf8 + /- or irf8-/- fish at 48 hpt following Mtz treatment. E. The percent of Lcp1 + cells that also co-label with 4C4 in the two genotypes. ***p < 10−7, * p < 0.05; pairwise comparisons between irf8 + /- and irf8-/- (Welch’s test).
Given that the irf8-/- fish still represent a system of microglia deficiency at 48hpt, we used staining for GS, PCNA, and EdU to analyze the first MG asymmetric division in microglia-sufficient (irf8 + /-) and microglia-deficient (irf8-/-) retinal sections from gnat2:nfsb-mCherry larvae (Fig 5). Larvae were immersed in Mtz for 48 hours to ablate cones, with EdU addition from 24–48 hpt (Fig 5A). GS+PCNA+ MG were detected in the INL of both genotypes, with some PCNA+ MG also showing EdU incorporation (Fig 5B-C’’’). Comparing the two genotypes, we found no statistically significant differences in the number of MG that stained positive for PCNA at 48 hpt (Fig 5D) nor in the fraction of these PCNA+ MG that had also incorporated EdU at 48 hpt (Fig 5E), indicating that entry of MG into the cell cycle in response to cone ablation occurred similarly by this timepoint in both microglia-sufficient and microglia-deficient conditions. We also quantified the number of GS-negative/PCNA+EdU+ cells in the INL (excluding the ciliary marginal zone) at 48 hpt, which would detect MG derived progenitor cells (MGPCs) (Fig 5F). At this timepoint, cells with these characteristics were not yet abundantly detected in either genotype (Fig 5F) indicating that a pool of abundant proliferating progenitors had not yet accumulated at 48 hpt in both microglia-sufficient and -deficient retinas.
A. Timeline showing treatment of gnat2:nfsb-mCherry;irf8 + /- or gnat2:nfsb-mCherry;irf8-/- fish with Mtz and exposure to EdU (dpf = days post fertilization, hpt = hours post treatment). B-B.” Images of retinal cryosections from irf8 heterozygotes collected at 48 hpt, showing PCNA, EdU, and GS signal. Scale bar in B” applies to B-B.” C-C.” Images of retinal cryosections from irf8 mutants collected at 48 hpt, showing PCNA, EdU, and GS signal. Arrows in B-C” indicate GS+PCNA+EdU+ cells. Scale bar in C” applies to C-C.” D. Quantification of the number of GS+PCNA+ cells per retinal section in the two genotypes. E. The fraction of GS+PCNA+ cells that were also EdU+ in the two genotypes. F. The number of GS-negative, PCNA+EdU+ cells (considered cycling MG-derived progenitors) in the two genotypes. Differences were not statistically significant, Welch’s test.
Given the detection of responding leukocytes in irf8 mutants (Fig 4), we attempted to inhibit microglial responses with the drug PLX3397. PLX3397 is a CSF1R inhibitor first used to deplete microglia in mouse [49,50] and more recently has been used for this purpose in zebrafish [20,36]. Treatment of larvae with PLX3397 (or vehicle) began at 2 dpf and solutions were refreshed every 24 hours. DMSO or Mtz treatment for 48 hours began at 4.5 dpf and EdU was added from 24–48 hpt (S1A Fig). In retinal sections from gnat2:nfsb-mCherry larvae with no cone ablation (DMSOtreated samples), we detected a reduction in both 4C4 + and Lcp1 + cell numbers with PLX3397 (S1B,C Fig) indicating that PLX3397 (PLX) treatment at least partially depletes microglial numbers in undamaged retinas. However, upon cone ablation with Mtz treatment, both 4C4 + and Lcp1 + cell counts were significantly increased resulting in similar overall numbers in both non-PLX and PLX-treated samples (S1B,C Fig) indicating that PLX treatment, even continued through 48 hpt/6.5 dpf alongside cone ablation, did not maintain microglial deficiency and/or prevent leukocyte responses to cone damage. Analysis of the first MG cell division did not detect differences in PCNA staining or EdU incorporation by Müller glia in non-PLX and PLX-treated groups at 48 hpt Mtz treatment (S1D-G Fig). In these experiments, a small number of proliferating progenitors were detected at 48 hpt (S1H Fig) but these did not return statistically significant differences in PLX-treated samples compared to controls. Given the microglial and leukocyte responses induced upon cone damage in PLX treated samples, we could not reliably derive conclusions about microglial functions in these experiments. In addition, this system did not provide a better alternative in terms of microglia/macrophage abundance to that of the irf8 mutants for probing microglial function in cone regeneration. Therefore, we elected to use only the irf8 genetic system for further experiments in this paper.
Inflammatory gene expression during first 72 hours of response to cone ablation in irf8 mutants shows minor differences compared to irf8 heterozygotes
Since inflammation and microglia are implicated in MG-derived regenerative responses [14,17,20,21,26–28] and given the leukocyte/microglia responses seen in both irf8 heterozygote and mutant (Fig 4), we examined expression of selected inflammatory genes at 24, 48, and 72 hours post treatment of gnat2:nfsb-mCherry larvae using RT-qPCR (Fig 6). For this analysis, we isolated RNA and prepared cDNA from pairs of eyes dissected from individual larvae treated with either DMSO or Mtz. We selected il1b, il6, tnfa, tnfb, and il10 for the qPCR analysis because changes in these transcripts have previously been shown to be associated with retinal regeneration in zebrafish [16,21,51]. Although TNFα has been implicated in retinal regeneration in zebrafish [16,52], we could not reliably detect tnfa with the primer pairs used in our study. Although we detected il10, the Ct readings returned were highly variable and in high ranges (il10 Cts ranged 30–36, total n = 60 per genotype from both treatments at each timepoint), indicating that detection of il10 transcripts in larval eye homogenates was not reliable nor comparable between samples. We provide the full results of statistical analysis in S1 File and describe the prominent features of expression changes for each gene below. Graphs showing Log2FC over time are shown in Fig 6.
RT-qPCR was performed using RNA collected at the indicated timepoints from pairs of whole larval eyes from gnat2:nfsb-mCherry larvae of either irf8 + /- or irf8-/- genotypes, following DMSO or Mtz treatment. Gene expression for the indicated genes was analyzed compared to undamaged (DMSO treated) irf8 + /- samples at 24 hpt, using the 2^ddCt method and graphing as Log2(FC). The graphs show the mean + /- SEM for each group, calculated from n = 4–8 samples per group. Complete results from statistical analysis are shown for this data in S1 File and are discussed in the main text. For clarity, only one statistically significant result is annotated in this figure (***p = 0.0002, 3-way ANOVA with Tukey’s posthoc) indicating statistically significant difference between irf8 + /- and irf8-/- for il6 expression following Mtz treatment.
In general, il1b increased in irf8 + /- over time following Mtz treatment, but levels of il1b also increased in DMSO-treated irf8 + /- samples by 72 hpt, indicating developmental changes in il1b mRNA also occur over this timeframe. Trends in il1b expression were largely similar for irf8-/- samples and there were no statistically significant differences in il1b between irf8 + /- and irf8-/- at any timepoints examined. In irf8 + /-, il6 increased by 24 hours of Mtz treatment compared to if8 + /- treated with DMSO, remained elevated at 48 hpt, then decreased through 72 hpt. In irf8-/- samples, il6 transcripts did not increase and instead remained similar to DMSO groups of both genotypes. Differences in il6 for irf8-/- treated with Mtz compared to irf8 + /- treated with Mtz at 24 hpt were statistically significant (Fig 6). In general, tnfb levels showed a trend of increasing abundance in irf8 + /- samples treated with Mtz over time, though differences were not statistically significant. There were no significant changes in tnfb between irf8 + /- and irf8-/- (Fig 6).
We also examined transcripts associated with Müller glia reactivity and damage responses by comparing expression of gfap [53,54] and mmp9 [21] in the two treatments and genotypes (Fig 6). Levels of gfap were initially increased in irf8-/- (DMSO) compared to irf8 + /- (DMSO); this is consistent with our previous report [41]. In irf8 + /-, gfap increased with Mtz treatment and remained elevated through 72 hpt. Overall, there were no statistically significant differences in gfap expression between irf8 + /- and irf8-/- at specific timepoints. The trajectory of mmp9 expression in both irf8 + /- and irf8-/- showed increases at 48 hpt and remained elevated at 72 hpt, and this was seen in both the Mtz treated groups as well as the DMSO-treated groups compared to earlier timepoints (Fig 6), with no statistically significant differences between genotypes at any specific timepoints. The results for mmp9 indicated that increased expression in larval eyes over time is largely due to developmental changes, not an effect of Mtz treatment (Fig 6). Collectively, the RT-qPCR results did not find major differences in gene expression for the selected transcripts between irf8 genotypes following cone ablation, except for a reduced abundance of il6 mRNA in irf8 mutants at 24 hpt following Mtz treatment. The RT-qPCR results may be limited due to use of whole larval eyes for RNA extractions, sensitivity to detect transcripts from specific cell populations, and/or expression of the selected transcripts by multiple cell types.
Modest reduction in early and late progenitor cell proliferation in irf8 mutants following cone ablation
We next analyzed the amplification of MG-derived progenitor cell (MGPC) proliferation in irf8 + /- and irf8-/- genotypes using PCNA staining and quantification in gnat2:nfsb-mCherry larval retina cryosections at several timepoints following cone ablation (Fig 7). Both genotypes showed increasing numbers of PCNA+ cells in the INL and ONL (consistent with MGPC proliferation and apical migration) through 120 hpt/5 days post treatment (dpt), followed by a reduction in numbers of PCNA+ cells at 144 hpt/6 dpt (Fig 7A-C). The peak of total PCNA+ cell numbers detected in retina cryosections was at 5 dpt for both irf8 genotypes (Fig 7C). At 72 hpt/3 dpt, there were reduced numbers of PCNA+ cells in irf8-/- retinas compared to irf8 + /- (Fig 7C), indicating a delay in the initial amplification of MGPC proliferation in microglia deficient conditions. Numbers of PCNA+ cells were not statistically different for irf8-/- retinas at 96 hpt (4 dpt). Total numbers were similar between irf8 genotypes at the peak detection of PCNA+ cell numbers at 120 hpt/5 dpt (Fig 7C), then irf8-/- retinas again showed reduced numbers of PCNA+ cells at 144 hpt/6 dpt (Fig 7C). When the numbers of PCNA+ cells were scored by distribution amongst retinal layers over the timepoints (Fig 7D), we found that the majority of PCNA+ cells were localized to the INL from 72–96hpt (3–4 dpt) then at 120 hpt/5 dpt showed changes in distribution where PCNA+ cells were detected across both the INL and ONL (Fig 7D), consistent with the images in Fig 7A and with interkinetic nuclear migration described for MG-derived progenitor cells in regenerating zebrafish retinas [1]. The trends of PCNA+ cell distributions were similar between both irf8 + /- and irf8-/- genotypes, but at 72 hpt/3 dpt there were fewer PCNA+ cells in the INL of irf8-/- retinas compared to irf8 + /- (Fig 7D), again consistent with a delay in amplification of MGPC proliferation at this timepoint. Further, at 144 hpt/6 dpt, the irf8-/- retinas had fewer PCNA+ cells in the ONL than detected in irf8 + / (Fig 7D), consistent with reduced MGPC proliferation in later stages of the regenerative response.
A, B. Retinal cryosections from gnat2:nfsb-mCherry;irf8 + /- or gnat2:nfsb-mCherry;irf8-/- larval zebrafish samples collected at the indicated timepoints post Mtz treatment and stained for PCNA and DAPI. Representative images are shown, scale bars in A and B apply to images in the respective row. C. Quantification of the total number of PCNA+ cells per retinal section in the two genotypes at the indicated timepoints. (The peripheral CMZ was excluded from this analysis). 2-way ANOVA (p < 0.05) with post-hoc; annotation of all post hoc pairwise statistical comparisons on the graph were not included to prevent visual clutter. Pairwise comparisons indicated statistically significant increases in PCNA counts from 48 hpt (2 dpt) to 120 hpt (5 dpt) over time for both irf8 + /- and irf8-/- with p < 0.05 for comparisons across timepoints within the same genotypes, and for decreases in PCNA counts from 120 hpt (5 dpt) to 144 hpt (6 dpt) for both irf8 + /- and irf8-/-. Annotated **p = 0.0096, irf8 + /- vs irf8-/- at 72 hpt (3 dpt) and *p = 0.0189, irf8 + /- vs irf8-/- at 144 hpt (6 dpt). D. Distribution of PCNA+ cells in the two genotypes by retinal layer at the indicated timepoints (ONL = outer nuclear layer, INL = inner nuclear layer, OPL = outer plexiform layer, IPL = inner plexiform layer). 3-way ANOVA (p < 0.05) with posthoc, *p = 0.0225 for INL counts irf8 + /- vs irf8-/- at 72 hpt, **p = 0.0028 for ONL counts irf8 + /- vs irf8-/- at 144 hpt. E. Timeline showing treatment of gnat2:nfsb-mCherry;irf8 + /- or gnat2:nfsb-mCherry;irf8-/- fish with Mtz and exposure to EdU; dpf = days post fertilization, hpt = hours post treatment. The arrow indicates the timepoint of sample collection (72 hpt). Quantifications of (F) the number of EdU+ cells (*p = 0.0112, Mann Whitney U-test) and (G) the number of PCNA+EdU+ cells (**p = 0.0076, Mann Whitney U-test) at 72 hpt in the two genotypes. H-J. Regions of retinal sections collected at 72 hpt showing signal for DAPI, PCNA, and EdU relevant to quantifications in F and G. Scale bars apply to all images in the respective rows (I, J).
To further examine early MGPC proliferation at 72 hpt, we repeated previous experiments with EdU incorporation but extended EdU immersion exposure through 72 hpt/3 dpt (Fig 7E). We found reduced numbers of total EdU+ cells and PCNA+EdU + co-labeled cells in irf8-/- retinas compared to irf8 + /- retinas at 72 hpt/3 dpt (Fig 7F-J). These results, along with those for EdU incorporation at 48 hpt (Fig 5) and PCNA+ counts (Fig 7C), suggest that although entry into the cell cycle by responding MG is similar, MGPCs in irf8-/- may take longer to complete cell cycle progression than those in irf8 + /- retinas during the early response to cone ablation. However, by 96 hpt, this delay has likely been mitigated and both genotypes show a similar peak of PCNA amplification at 120 hpt (Fig 7C). The reduction of PCNA+ numbers at 144 hpt/6 dpt indicates that MG-derived progenitor proliferation is trending down at this timepoint in both genotypes, however this may happen more rapidly in irf8-/- retinas and further reduce the proliferating MGPCs in the ONL (Fig 7C,D). Consistent with a reduction in MGPC proliferation between 120 and 144 hpt, our previous images suggested that regenerated cones may be present by these timepoints (Fig 2D, mCherry+ cells visible in the ONL). We therefore examined 120 hpt/5 dpt retinal cryosections for EdU + mCherry+ cones from experiments with EdU washout at 72 hpt. By this timepoint, EdU signal was considerably reduced, consistent with rapid proliferation of MGPCs from after EdU washout at 72 hpt, yet we were able to detect a few (on average, fewer than 3) EdU + mCherry+ cones in the central ONL of both genotypes at 120 hpt/5 dpt (S2 Fig). Collectively, these results indicate that microglia may support cell cycle progression of MGPCs though the deficiency in irf8 mutants results in only modest reductions in numbers of dividing progenitors at early and late stages of MGPC amplification. In addition, our results indicate that regenerated cones can be detected by 120 hpt/5 dpt in this experimental system.
We attempted to measure changes in re-expression of ascl1a by neuronal progenitor cells generated upon cone damage, using RT-qPCR (S3 Fig, statistical results shown in S1 File). There was a modest increase in ascl1a in irf8 + /- detected at 24 hpt upon Mtz treatment, while increase in ascl1a was instead delayed to 48 hpt in irf8-/- upon Mtz treatment (S3 Fig). The active differentiation of neurons during development of the larval zebrafish retina may preclude the ability to detect significant changes in ascl1a and may be the reason we did not detect more robust changes in this gene upon Mtz treatment. However, these results are consistent with a delay in cell cycle progression in irf8 mutants from 48 to 72 hpt (Fig 7C), in which the re-expression of ascl1a in MG-derived progenitor cells could also be temporally delayed.
Regeneration of cones in larval retina is similar between irf8 heterozygotes and mutants
We next analyzed regeneration of cones in both gnat2:nfsb-mCherry;irf8 + /- and gnat2:nfsb-mCherry;irf8-/- larval retinas. After cone ablation, recovery of viable and healthy larvae later than 7 dpt (~11 dpf age) was challenging with standard larval fish rearing conditions, which we attribute to the cone-dominant function of the larval zebrafish retina through approximately 15 dpf [55–57]. The loss of cones from Mtz-mediated ablation likely impacts larval feeding, and therefore survival; standard rearing of the DMSO-treated larvae did not impact viability. Since we detected a small number of EdU+ regenerated cones at 5 dpt as described previously (S3 Fig), we selected 7 dpt as the first timepoint to analyze regenerated cones in the two irf8 genotypes. To increase the likelihood of encountering food following cone ablation, larvae were maintained in glass beakers and provided ample powder food daily with a window of time to allow feeding before water/solution changes. All unhealthy or nonviable larvae were removed prior to powder food addition. In these experiments, we included EdU exposure by immersion of larvae in water including EdU from 96 hpt/4 dpt to 144 hpt/6 dpt (Fig 8A), which corresponded to timeframes covering the peak of PCNA detection (Fig 7C).
A. Diagram of timing for cone ablation (Mtz treatment) and EdU immersion of gnat2:nfsb-mCherry;irf8 + /- or gnat2:nfsb-mCherry;irf8-/- larval zebrafish; dpf = days post fertilization (age of larvae), dpt = days post treatment (start of Mtz immersion). Mtz washout was at 2 dpt (6.5 dpf), EdU exposure was from 4–6 dpt, and retinal cryosections were collected at 7- or 14-days post Mtz treatment (7 dpt and 14 dpt). After EdU retrieval, cryosections were stained for PKCα and DAPI, mCherry signal was used to visualize cones. Selected images and individual or merged channels are shown for retinal sections from the two irf8 genotypes collected at 7 dpt (B, C) and at 14 dpt (D, E). Arrows annotate EdU + mCherry+ cones, * indicates regions where PKCα+ bipolar (BP) cell bodies are displaced into the ONL. Insets in B,D provide larger views of indicated regions to visualize EdU+ cones. C. Images show PKCα + BP cells with dendrites extended and contacting cones at locations adjacent to the radial direction of the retina in samples from both irf8 genotypes at 7 dpt. E. Images are higher magnification views of EdU + mCherry+ cones in retinal cryosections collected at 14 dpt; several EdU + mCherry+ cones are detected in the central retina of both irf8 genotypes as well as other EdU+ cells in the inner retina. F. EdU + mCherry+ cells (cones) were quantified per retinal section in the two genotypes at both timepoints (the peripheral CMZ was excluded from this analysis and counting was performed inside the “stripes” of EdU seen in the 14 dpt samples). Differences between genotypes and between timepoints were not statistically significant (Mann-Whitney U-test for pairwise comparisons between irf8 + /- and irf8-/- at 7 dpt and at 14 dpt, and pairwise comparisons within the same genotype between 7 dpt and 14 dpt). G. Quantification of the number of mCherry+ cones in the ONL in the two genotypes at 7 dpt and at 14 dpt (counting was performed in central retina, between the EdU+ “stripes” representing the boundary of the CMZ at the time of EdU immersion). *p < 0.05, **p < 0.01 for the indicated groups at the different timepoints. H. Quantification of the size of gaps between regions of mCherry+ signal in the ONL at 7 dpt and at 14 dpt. **p < 0.05 for the indicated groups at the different timepoints. Differences were not statistically significant between irf8 genotypes at each timepoint (two-way ANOVA).
At 7 dpt following Mtz treatment, mCherry+ cones were detectable in the ONL of both genotypes; several of these cones were EdU+ and centrally localized (and therefore not produced by growth at the CMZ) confirming that they were regenerated from MGPC (Fig 8B). Regions of the ONL in retinas from both genotypes still lacked cones and many of the mCherry+ cells within the central retina did not yet display typical cone morphology or patterns (Fig 8B), indicating that cone regeneration, differentiation, and/or maturation is ongoing at 7 dpt. We also included a stain for post-synaptic bipolar (BP) cell partners by staining for PKCα, which labels a subset of ON-BP types in zebrafish retina [58,59] (Fig 8B, C). Patterns of PKCα+ signal showed BPs contacting cones in the ONL of both genotypes and revealed that locations in which cones were absent had displaced BP cell bodies that protruded apically into the ONL region (Fig 8C). Further, some BP dendrites located basally to these “gaps” in cone signal appeared to “reach” for nearby cones tangential to the radial direction of the retina (Fig 8C); this was observed in both irf8 genotypes. These results indicate that in the absence of pre-synaptic partners, BP cell bodies and dendrites remodel and reposition at least temporarily, during regeneration [44,60,61].
We also performed experiments with similar timing of EdU exposure by immersion (Fig 8A), with collection at 14 dpt, to examine regeneration at a later timepoint. For these experiments, larvae were again exposed to EdU in glass beakers with feeding of ample powder food. Due to challenges with survival to this experimental endpoint and to ensure we could collect enough larvae, some experiments with collection at 14 dpt used multiple breeder pairs and clutches of larvae with known irf8 genotypes. We did not find differences in eye size measurements between irf8 + /- and irf8-/- samples at 14 dpt (S4 Fig). After EdU washout, larvae were transferred to a specialized housing apparatus in fish tanks and provided close feeding with powder and brine shrimp until tissue collection (details described in Methods). At 14 dpt, EdU+ cones were detected in regions of the ONL consistent with cone regeneration in samples from both genotypes: these cells were located interior of “stripes” of EdU visible in retinal sections, representing a boundary of growth from the CMZ formed during the EdU exposure window (Fig 8D,E). Quantification of EdU + mCherry+ cones at 7 dpt and at 14 dpt did not find statistically significant differences between irf8 + /- and irf8-/- samples (Fig 8F). We also noted that the number of EdU + mCherry+ cones did not increase in retinas collected at 14 dpt compared to those at 7 dpt (Fig 8F); however, the ONL of both irf8 genotypes was more populated with mCherry+ cones at 14 dpt compared to 7 dpt (Fig 8G). These observations indicate that regeneration is likely ongoing from 7 to 14 dpt, and not all regenerated cones are retained over time and/or EdU label is further diluted due to ongoing progenitor cell divisions and/or EdU immersion delivery is less effective at these later larval ages. We measured the curvilinear distance of the “gaps” in mCherry+ cone signal in the ONL and did not find differences in gap sizes between irf8 heterozygotes and irf8 mutants at 7 dpt or at 14 dpt (Fig 8H), and gap sizes decreased from 7 to 14 dpt for both genotypes (Fig 8H), again indicating that regeneration was occurring similarly between the two genotypes.
Some EdU+ cells were also detected in the inner retina at both 7 and 14 dpt that were not mCherry+ cones (Fig 8B,D,E), indicating that our EdU immersion schedule labeled other cell types. These could represent MG that were in the cell cycle at the time of EdU exposure and/or other neuron types produced from MGPC during this timeframe, as production of neuronal cell types different than the ones targeted has been described for zebrafish retinal regeneration [62–64]. The number of EdU+ cells that were scored as mCherry-negative “non-cones” were not statistically different between irf8 genotypes (S4 Fig).
To examine potential synapse formation of regenerated cones with BPs, we visualized OPL synaptic regions with higher magnification, and qualitatively compared features visualized in undamaged retinal cryosections (Fig 9A,C) to samples collected at 7- and 14-days post Mtz treatment, from both irf8 genotypes (Fig 9B,D). Staining for SV-2 (Synaptic Vesicle 2) was included in these images to visualize pre-synaptic vesicle+ regions in mCherry+ cones in relation to cone-BP contacts. SV-2 signal was detectable in the OPL, showing localization consistent with that of cone pedicles, of both irf8 genotypes at 7 and 14 dpt following cone ablation, though it was less uniform and appeared “patchier” in samples that had experienced cone ablation compared to undamaged retina (Fig 9B,D). In samples with regenerated cones, cone morphology did not fully recapitulate that of undamaged cones (upright in the radial direction of the retinal section, with clear axons and pedicles). Instead, cones in retinas following ablation were less elongated with axons harder to discern, often obliquely oriented relative to the radial axis of the tissue section, appeared disorganized, and this was the case for both irf8 genotypes (Fig 9B,D). As noted above, BP cell bodies were seen protruding into the ONL where gaps in cone signal were present (Fig 9B,D). However, regions with features consistent with re-established synapses (mCherry+ cone basal pedicle contacting PKCα + BP dendrites also co-labeling with SV-2) were notable in samples collected at 7 and 14 dpt following Mtz treatment (Fig 9B,D,E,F). Visualizing these features in the ONL selectively for regions with EdU + mCherry+ cones, confirmed that SV-2 + synapses were restored between regenerated cones and BPs and this was evident for both irf8 heterozygotes and mutants (Fig 9E,F).
Retinal cryosections from gnat2:nfsb-mCherry;irf8 + /- or gnat2:nfsb-mCherry;irf8-/- larval zebrafish were collected at 7 or 14 days post DMSO or Mtz treatment (dpt) and included EdU exposure as described in the text. Cryosections were stained for PKCα, SV-2, and DAPI; mCherry signal was used to visualize cones. A. Shows intact and typical outer retinal morphology in cryosections from undamaged samples (DMSO-treated) collected at 7 dpt, from each irf8 genotype. Outer plexiform layer (OPL) is visualized by SV-2 signal (white arrowhead in left panel) where cones and PKCα+ bipolar cells have synapses (black arrowhead, right). B. Shows images from selected samples collected at 7 dpt following Mtz treatment from each irf8 genotype. SV-2 signal is dim and patchy (annotated by *) though cones and BPs are again showing contact in the regenerating OPL. C. Shows intact and typical outer retinal morphology in cryosections from undamaged samples (DMSO-treated) collected at 14 dpt, from each irf8 genotype. Outer plexiform layer (OPL) is visualized by SV-2 signal (black arrowhead, right) where cones and PKCα+ bipolar cells have synapses. D. Shows images from selected samples collected at 14 dpt following Mtz treatment from each irf8 genotype. SV-2 signal is more apparent and stronger but still patchy (annotated by *) though cones and BPs are again showing contact in the regenerating OPL. Histological abnormalities (as described for Fig 8) are also visible here where BPs protrude into the OPL. E, F. Shows enlarged regions of images acquired from samples with cone regeneration where EdU+ (regenerated) mCherry+ cones have re-established contacts with PKCα+ bipolar cells. Samples were collected at 7 dpt (E) and at 14 dpt (F). The white arrows point out SV-2 + localization where EdU+ regenerated cones contact PKCα+ bipolar cells in the OPL.
Compensatory immune cell populations are present in irf8 mutants during regeneration of cones
Given the overall lack of significant differences observed in regenerative responses in irf8 mutants compared to heterozygotes and given the detection of some leukocytes/microglia at 48 hpt/2 dpt in irf8 mutant retinas (Fig 4), we performed further analysis of microglia/leukocytes at additional, selected timepoints following ablation of cones in gnat2:nfsb-mCherry larvae of both irf8 genotypes (Fig 10). For this analysis, we again stained retinal cryosections with the 4C4 antibody [45] and the pan-leukocyte marker Lcp1 [48] (Fig 10). We performed quantifications at timepoints corresponding to ablation of cones (48 hpt, Fig 1D, Fig 4), surrounding the peak of PCNA+ MGPC numbers (5–7 dpt, Fig 7C), and timepoints for which we examined cone regeneration (7 and 14 dpt, Figs 8 and 9). Comparing total numbers of cells staining with these markers (cells that stained positive for either or both 4C4 and Lcp1 markers), we found that in retinas of irf8 heterozygotes, immune cell populations expanded at 5 dpt compared to 48 hpt/2 dpt, then declined from 6 dpt through 14 dpt (Fig 10D). In comparison, irf8 mutant retinas had reduced numbers of immune cells at all timepoints compared to irf8 heterozygotes, though they did have detectable immune cell populations (Fig 10D). The immune cell populations in irf8 mutant retinas also showed expansion at 5 dpt and contraction by 14 dpt (Fig 10D). At 14 dpt, on average less than 2 immune cells were detected per retinal cryosection in irf8 mutants (Fig 10D). The populations detected in irf8 + /- retinas at each of these timepoints nearly all stained positive for Lcp1, with the majority also co-labeling with 4C4 (Fig 10E, F). Lcp1 + cells were also the most abundant immune cells detected in irf8 mutants (Fig 10A-A’, E, F). Interestingly, within the populations of immune cells detected in irf8 mutants from 5 to 7 dpt was a subpopulation of cells that stained positive only for the 4C4 marker, and not Lcp1 (Fig 10B-C’,C,C’,G). This “4C4-only” population localized primarily to the ONL of irf8 mutant retinas from 5–7 dpt (Fig 10B-B’, C, C’). Given that the neutrophil niche is expanded in irf8 mutant larvae after 6 dpf [40], we stained retinal cryosections with an antibody to Mpx (myeloperoxidase, as used previously in [13]) to determine if any of the immune cells detected could be neutrophils (S5 Fig). Though an occasional, isolated mpx+ cell was seen in retinal tissue, we did not detect significant numbers of neutrophils in retinas of either irf8 genotype (S5 Fig), indicating that neutrophils are not a substantial component of the immune response to cone ablation in larval retina and do not account for the immune cell populations detected in irf8 mutants. Overall, our staining for immune cell markers revealed that although immune populations are reduced in numbers in irf8 mutants compared to irf8 heterozygotes, compensatory immune cell populations are present during the regeneration of cones. The differential labeling with 4C4 antibody suggests that these compensatory populations in irf8 mutants are phenotypically different than those in irf8 heterozygotes.
A-A,” B-B.” Selected images shown here are from 7 dpt samples from the indicated irf8 genotypes. Images of retinal cryosections stained for DAPI, 4C4, and Lcp1; cones are visualized by mCherry fluorescence. A-A.” Images from irf8 + /- retinal cryosections showing selected channels and merged image. B-B.” Images from irf8-/- retinal cryosections showing selected channels and merged image. Arrowhead shows a cell staining positive for 4C4 but negative for Lcp1. C, C’. Enlarged region indicated by the dotted box in B” and signal from 4C4 (C) or from Lcp1 (C’) with mCherry and DAPI to better visualize the cell annotated by the arrowhead. D. Quantification of the total number of cells that stained positive for either or both 4C4 and L-plastin (Lcp1). Cells counted here included those that stained positive for one or both markers in retinal cryosections of gnat2:nfsb-mCherry;irf8 + /- or gnat2:nfsb-mCherry;irf8-/- larval zebrafish collected at the indicated timepoints post Mtz treatment. E. Quantification of the number of cells that stained positive for both 4C4 and Lcp1 (4C4+Lcp1 + cells) in the two genotypes. F. Quantification of the number of cells that were negative for 4C4 but stained positive for Lcp1(4C4-negative, Lcp1+) in the two genotypes. G. Quantification of the number of cells that were positive for 4C4 but were negative for Lcp1 (4C4 + , Lcp1-negative) in the two genotypes. Cells with such astaining pattern were not detected in irf8 + /- retinal cryosections at any timepoint, but a small number with this pattern were detected in irf8-/- retinal cryosections at 5–7 dpt, with statistically significant differences at 6 dpt. 2-way ANOVA (p < 0.05) with Tukey’s posthoc was used. P-values are annotated only (for simplicity) for pairwise comparisons between irf8 + /- and irf8-/- (**p < 0.02, ***p < 0.01, ****p < 0.00001). For the graphs in D and E, all p-values between counts at each timepoint compared to the others were <0.05 for both genotypes.
Discussion
We examined the regenerative response in the larval zebrafish retina following ablation of Nfsb-expressing cone photoreceptors achieved by exposure to metronidazole. Experiments were performed in microglia-sufficient larvae and in microglia-deficient irf8 mutants, using siblings wherever possible in our experimental design. We performed analysis over a time course spanning 14 days post-cone ablation and examined the ablation of cones, early glial cell responses, the first asymmetric division of the MG, amplification of MGPC proliferation, expression of selected inflammation-associated genes, and the regeneration of cones and their synapses with BPs. We also examined microglia and other immune cells (by staining with 4C4 and Lcp1 antibodies) at multiple timepoints in both irf8 heterozygotes and irf8 mutants. We found that overall, regeneration proceeds sufficiently in irf8 mutant larval zebrafish with moderate reduction in early and late MGPC proliferation. Notably, though irf8 mutant zebrafish retinas are highly deficient in microglia/immune cells prior to induction of damage, we identified a population of immune cells that emerge in irf8 mutant retinas following cone ablation. This compensatory immune cell response precludes the ability to make strong conclusions about the role of microglia in retinal regeneration in zebrafish based on our experiments presented here. Nonetheless, our results are consistent with multiple reports [14,15,31,41] which, when considered together with our work presented here, collectively indicate that (i) microglia likely support the proliferation of MGPC in the zebrafish retina following damage and (ii) that compensatory mechanisms and cell types provide function in microglia-deficient retinas.
Other studies utilizing the irf8st95 [40] mutant zebrafish line to manipulate microglial abundance during retinal regeneration have been published previously [15,18,36] and two of these reports involved photoreceptor regeneration (MG-mediated) in adult fish [15,18]. Bludau et al. [15] found a modest decrease in proliferating progenitors in the damaged irf8 mutant retina following light damage to photoreceptors, while Song et al. [18] found no difference in proliferating cells in the retina in irf8 mutants following light damage. These reports appear conflicting; however, it is worth noting that only one timepoint was examined in analysis of progenitor proliferation in these two studies, and the selected timepoint was a day apart. The other study using the irf8 mutant, by Leach et al. [36], examined regeneration of the retinal pigment epithelium (RPE) in larvae and is not known to involve MG-mediated progenitors towards the regeneration of RPE cells. Here, we performed analysis of MGPC proliferation in irf8 mutants and heterozygous siblings following cone photoreceptor damage over a comprehensive time course, as well as the initial entry of responding MG into the cell cycle. We found that in irf8 mutants, MG entered the cell cycle with similar timing (by 48 hours post drug exposure); however, early (72 hpt) and late (144 hpt) MGPC proliferation were reduced compared to sibling controls. Our results are consistent with findings for MGPC proliferation during retinal regeneration in larval zebrafish following rod ablation which manipulated microglia using a different strategy [14]. Further, multiple reports using dexamethasone to suppress microglia and inflammatory signals found impacts on MGPC proliferation [14,15,21,22]. Collectively, the emerging theme is that inflammatory signals promote MGPC proliferation in the damaged zebrafish retina and microglia are a likely source of such signals. Consistent with this, we found moderate changes in inflammatory gene expression in irf8 mutants following cone ablation, particularly for the cytokine il-6. A study using Dexamethasone and irf8 mutant zebrafish in light damaged retinas [32] supports the idea that microglia promote some inflammatory component of the response to retinal damage, which acutely may be important for the MGPC response but that could be detrimental in contexts of chronic photoreceptor degeneration. The molecular nature of such signals remains to be fully identified, and it is likely that multiple cell types can provide such signals. Moreover, microglia may function in an inter-cellular communication axis involving not only the MG but other cell types such as vascular endothelial cells [31].
Our analysis of regenerated cones did not detect differences in irf8 mutants compared to their siblings, at least through 14 dpt. Regenerated cones were detected at similar abundance in both genotypes, which we observed to make synaptic connections with BPs in the regenerating OPL. This is in contrast to previous studies that found delayed photoreceptor regeneration in laser damaged retinas when microglia were depleted with PLX3397 treatment [20], and that found delayed/reduced replacement of rods in larval retina when microglia were depleted using an Nfsb/Mtz strategy [14]. In addition, the regeneration of RPE in larval zebrafish eyes was reduced/abnormal when microglia were depleted using genetic and pharmacological tools [36,43]. The different outcome for photoreceptor regeneration in our studies here (which was overall normal) could be due to the age of fish used in the experiments, the presence of compensatory immune cell populations in irf8 mutants (discussed below), the methods used to quantify regenerated photoreceptors/RPE cells, the timepoints selected for analysis (later or earlier timepoints may reveal differences in cone regeneration, survival, or function), stem cell niches that do not involve MG and are available for regenerated RPE cells [65,66] and for rods [2], or other factors. Further, though we qualitatively described detection of synaptic connections of regenerated cones with BPs in the OPL in retinal cryosections, we did not deeply or quantitatively analyze regenerated synapses. It remains unknown whether microglia modulate synaptic rewiring or restoration of visual function in regenerated retinas [7,11,44]. Any experiments aimed to elucidate such microglial functions must consider impacts of microglial deficiency over development and/or the presence of compensatory immune cell populations.
As in previous reports using irf8 mutant zebrafish [15,18,36], we detected immune cells in irf8-/- retinas following neuronal damage. In our studies described in this paper, these other leukocytes expanded upon retinal damage then were reduced over time, with a trajectory similar to their microglia-sufficient siblings but at markedly reduced abundance. Neutrophils were not significantly detected with Mpx antibody staining in either genotype. Interestingly, the leukocyte populations detected in irf8 mutants in our experiments stained positive for markers L-plastin (Lcp1) and 4C4 but did not always express both markers. Notably, we detected a subset in microglia-deficient irf8 mutants that stained positive with 4C4 antibody but negative for Lcp1 and showed increased abundance compared to the microglia-sufficient siblings. This could indicate different phenotypes and/or ontogeny of the immune cells evoked in irf8 mutant retinas compared to microglia-sufficient retinas. The compensatory immune cell response strongly suggests an essential role for the immune system in retinal damage and possibly for regeneration. The ability of non-hematopoietic cell types, such as the neuroglia, to execute compensatory functions has been described [41,67,68]. Such responses may be limited and/or scale with the extent and nature of damage to retinal neurons.
The presence of such compensatory immune cell populations, however, makes it difficult to draw conclusive interpretations from our experiments in this paper. This has been emphasized previously [18] and more recently, others have instead used zebrafish with loss of function mutations in both csfr1a and csfr1b [31,43] to manipulate microglial abundance. Further, drugs such as PLX3397 which blocks CSFR1 signaling have been used to manipulate microglia in zebrafish [20,36]. However, in our results PLX3397 treated retinas had even more leukocytes than irf8 mutants, therefore we elected to forgo this approach. Though the use of irf8 mutant zebrafish may not be ideal for some experiments, such as those involving retinal damage that elicit compensatory immune cell responses as we and others found, this microglia-deficient irf8 mutant still has a place for developmental studies probing microglial functions. Indeed, we report here that undamaged larval irf8 mutant retinas are strongly deficient in microglia, consistent with our previous work [41]. The irf8 mutant offers advantages over the csfr1a/b double mutants (single homozygous mutation instead of two mutations) and compared to prolonged drug exposures aimed to deplete microglia (which could carry caveats as we encountered for maintaining depletion upon neuronal damage, or that may require extended immersions). Nonetheless, caution should be used in interpreting experiments with any microglia-deficient system, as compensatory responses from not only other immune cells but also neuroglia may be elicited in the absence of microglia [41,67,68].
In addition, the use of larval ages in our experiments may be a factor in the outcomes. While larvae offer many advantages (such as simpler and controlled co-housing and co-treatments of clutches of mixed genotypes, often larger sample sizes, repeatable experiments using few adult breeders to continuously produce clutches), the simultaneous progression of development with regeneration may confound experimental interpretation or not fully represent regeneration from a differentiated state as in adults. Also noteworthy, in Xenopus, regeneration and inflammation had differential outcomes following rod photoreceptor death in young vs older tadpoles [26]. However, without reliable inducible methods of microglial manipulations, adults that mature with genetic mutations or absence of microglia over a lifetime may also present confounding factors and impacts [18,41,69]. We also faced challenges with reliable Mtz-mediated ablation of cones in gnat2:nfsb-mCherry adult fish [70], and therefore selected larvae to pursue the experiments in this paper.
Though our findings presented in this paper are indeed limited and not fully conclusive, other literature emphasizes the impact of microglia on the outcome of retinal regeneration [27,28,31,39,43]. This is an important area of scientific pursuit given the involvement of microglia in numerous neurodegenerative diseases, many of which may involve compensatory cellular and molecular mechanisms and pathways. Continued interrogation of microglial functions will require focused, targeted approaches with advanced experimental design.
Materials and methods
Animals
All experiments used zebrafish (Danio rerio) and procedures were approved by the University of Idaho Animal Care and Use Committee (IACUC). Adult zebrafish were housed in an aquatic system with filtered, recirculating water maintained at 28oC and on a 14:10 light:dark cycle as described in Westerfield, 2007 [71]. Light onset was considered 0 hours post-fertilization (0 hpf). Upon fertilization, embryos were collected into glass beakers with system water and either reared in the fish facility or transferred to an incubator with 14:10 light:dark cycle at 28oC until treatments commenced. Water for embryos/larvae was refreshed daily. All experiments used transgenic gnat2:nfsb-mCherry fish [23] in which cone photoreceptors express bacterially derived nitroreductase tagged with mCherry under the cone-specific promoter gnat2. In some experiments, the gnat2:nfsb-mCherry line was crossed to the TP1:mTurquoiseuoi2523Tg transgenic line [29]. The gnat2:nfsb-mCherry fish were crossed with irf8^st95 [40] homozygous mutant partners, to generate gnat2:nfsb-mCherry;irf8 + /- breeders. The gnat2:nfsb-mCherry;irf8 + /- breeders were then crossed to irf8 mutants (irf8-/-) to generate gnat2:nfsb-mCherry+ microglia-sufficient (irf8 + /-) and -deficient (irf8-/-) offspring for experiments. Whenever possible, experiments were performed blind to genotype at the treatment and collection stages, using clutches of offspring with unknown irf8 genotype. Genotyping for the irf8st95 mutation was performed after sample collection; gDNA extraction and the PCR/AvaI restriction digest protocol as previously described [40,41]. Feeding of larvae began at 6 dpf with powder food provided once daily. At 10 dpf, larvae were transferred to feeding compartments in tanks and received brine shrimp in the morning and powder food in the afternoon. Zebrafish cannot be sexed before reproductive maturity. Since larval ages were used for experiments and analyses, sex as a biological variable is not a component of the experiments in this paper.
Treatments
At 4 dpf, embryos were screened using a fluorescent sorting microscope for the presence of mCherry signal in eyes (also visible in pineal gland), indicative of transgene expression in cone photoreceptors. Solutions were prepared by dissolving dimethyl sulfoxide (DMSO, vehicle) to 0.1% final concentration with or without metronidazole (Mtz, Thermo Scientific 210341000) at 10 mM final concentration in fish water. Solutions of DMSO/Mtz were protected from light and mixed for a minimum of 4 hours on an orbital shaker before use. Embryos positive for mCherry were split into two groups (genotypes unknown) and immersed in either 0.1% DMSO or 10 mM Mtz for 48 hours total (approximately 4.5–6.5 dpf), with refreshment of treatment solutions at 24 hrs. Because Mtz is light sensitive, these treatments were done in petri dishes covered loosely with foil. After treatments, larvae were removed from the dishes and transferred to beakers with fresh water for washout and recovery. In some experiments, larvae were immersed in 175 μM 5-ethynyl 2’-deoxyuridine (EdU, Invitrogen) for 24–48 hours total, with EdU solutions refreshed at 24-hour intervals. Feeding of powder food was administered prior to the time of EdU solution refreshment.
For regeneration timepoints (collection at 7- or 14-days post treatment, dpt): at 10 dpf (6 dpt = 144 hpt), the larvae were transferred to a feeding compartment within a 2.8L tank. The feeding compartment used a 0.5L plastic container with a fine mesh screen bottom with edges that extend and rest on the lip of the tank, allowing the feeding compartment to sit under the feeding hole and water feed of the tank lid while on the aquatic housing system. The larvae were placed in the 0.5L feeding compartment, permitting food to be temporarily restricted to the location of the larvae, increasing chances of successful feeding. The mesh bottom allowed for constant water flow that slowly removed the excess brine shrimp and powder food from the compartment, maintaining water quality. Feeding larval zebrafish in this housing situation consisted of morning brine shrimp (pre-chilled at 4oC to reduce locomotion of the shrimp) and afternoon powder feedings of ~0.05g/feeding. Buildup of excess food was removed daily by gentle aspiration with a plastic transfer pipette to maintain water quality.
PLX3397 treatments
Solutions were prepared by dissolving DMSO to 0.1% final concentration with or without PLX3397 (Sellick Chemicals) at 1µM final concentration in fish water. Solutions were protected from light during mixing and treatment. At 2 dpf, embryos were dechorionated and split into treatment groups. Embryos were treated with either DMSO or PLX3397/DMSO with solution refreshes occurring every 24 hours until 4.5 dpf, when they were sorted for expression of the gnat2:nfsb-mCherry transgene. Embryos positive for mCherry signal were split into 4 groups based on prior PLX3397 treatments: DMSO, PLX3397/DMSO, Mtz, and PLX3397/Mtz. Mtz solutions were prepared as previously described, either with or without the addition of 1µM PLX3397. Solutions were refreshed every 24 hours until collection at 6.5 dpf.
Cryopreservation and sectioning of retinal tissue
On the day of sample collection, larvae were euthanized with 0.4% v/w tricaine/MS-222 and then immediately fixed in 4% paraformaldehyde in 5% sucrose phosphate overnight at 4oC with constant rocking. The following day, larvae were washed in a graded series of phosphate buffered solution (pH 7.4) of 5% to 20% sucrose. Tails were clipped from fixed larvae and transferred to hot NaOH gDNA extraction buffer [72] then genotyped for irf8st95 mutation as described previously. Larval samples were transferred to a 1:2 solution of OCT embedding medium (Sakura Finetek): 20% sucrose phosphate and mixed for 30 minutes at room temperature. Larvae were embedded in 1:2 OCT:20% sucrose phosphate and frozen by immersion in 2-methylbutane supercooled with liquid nitrogen. After freezing, blocks containing larvae were stored at -20oC for at least 24 hours. Sectioning was performed using a Leica CM3050 cryostat, with 10 μM thick sections placed onto glass slides (FisherBrand Superfrost Plus Microscope Slides). Sections on slides were desiccated overnight then stored at -20oC until use.
Immunostaining, TUNEL labeling, and EdU detection
Retinal cryosections on slides were thawed in a humidified chamber at room temperature. For PCNA staining, citrate buffer antigen retrieval was performed prior to the blocking step as previously described [13,73]. TUNEL staining was performed following the manufacturer’s instructions (Roche), and slides were rinsed in PBS, prior to blocking and antibody staining. EdU detection, where appropriate, was completed using an EdU Click-iT kit (Invitrogen) or EdU Assay/EdU Staining Proliferation Kit (iFlour 647) (Abcam) following the manufacturer’s instructions. EdU detection was completed prior to immunostaining, then slides were transferred to antibody blocking buffer and stained as described below.
Blocking solution (20% normal donkey serum, 0.1% sodium azide, 0.1% Triton X-100, prepared in phosphate buffered saline (PBS)) was added to slides and incubated for 30 minutes at room temperature. Primary antibodies were prepared in antibody dilution buffer (1% normal donkey serum, 0.1% sodium azide, 0.1% Triton X-100, prepared in PBS) then added to slides and incubated overnight at 4oC. The following day, slides were washed in PBS + 0.1% Triton-X-100 (PBST) at room temperature for 30 minutes. Secondary antibodies and DAPI were prepared in antibody dilution buffer. Slides were incubated in secondary antibody solution at room temperature for at least 1 hour. Slides were washed for 30 minutes in PBST, rinsed in PBS, then mounted with coverslips using Vectashield Vibrance.
Primary antibodies and dilutions were as follows: mouse 4C4 antibody (1:100, used as hybridoma supernatant from clone 7.4.C4 sourced from Sigma-Aldrich 92092321), rabbit anti-Lplastin (Lcp1, 1:10,000, a kind gift of Dr. Michael Redd), rabbit anti-mpx (1:200, Genetex GTX128379), rabbit anti-proliferating cell nuclear antigen (PCNA; 1:200, Proteintech 10205), mouse anti-Glutamine Synthetase (GS; 1:200, BD Biosciences 610517), mouse anti-SV2 (1:2000, Developmental Studies Hybridoma Bank), Rabbit anti-PKCa (1:500, Santa Cruz Biotechnology). Secondary antibodies were from Jackson Immunoresearch and used at 1:200 dilution (donkey anti-mouse and donkey anti-rabbit antibodies conjugated to Alexa-Fluor488, Cy3, or Alexa-Fluor647) or from AAT Bioquest (goat anti-mouse Cy7, used at 1:200 dilution).
Confocal microscopy, image processing, and analysis
Imaging was performed on a Nikon CrestOptics X-Light spinning disc confocal microscope running Nikon Elements software, using 20X dry (Plan Apo λ 20X Air 0.75 NA DIC), 40X water immersion (Apo LWD λS DIC N2 1.15 NA), or 60X oil immersion (Plan Apo λ 40X oil 1.40 NA WD = 130µm) objectives. Z stacks (optical sections) were collected at 0.3–3 micron step sizes. Images were viewed and analyzed using ImageJ (FIJI, RRID:SCR_003070) or Nikon Elements software. For retinal cryosections, 3–6 sections per eye were imaged (representing technical replicates) from which the average of technical replicates was determined for each sample that was quantified/counted. Sections containing central retina were selected for analysis. Counting of microglia, PCNA/EdU+ cells, and/or eye measurements were performed in entire retinal cryosections, using cell specific markers (from immunostaining or fluorescent reporters) and signal from the nuclear stain DAPI to identify retinal layers. Regenerated cone photoreceptors were quantified from confocal z-stack images using ImageJ/Fiji. The region of interest was defined by EdU labeling of proliferating cells originating from CMZ, serving as an anatomical reference point across regenerative samples. Cells positive for mCherry and DAPI located in the ONL between the EdU+ boundaries were scored manually using the cell counter plugin. Counting was performed in 3–6 sections per sample as technical replicates, then averaged and reported as number of cones per sample for the two genotypes.
RNA extraction and cDNA synthesis
On day of collection, larvae were euthanized with 0.4% w/v tricaine, water was removed, and the larvae were flash frozen on liquid nitrogen. Flash frozen larvae were stored in 100% methanol at -80oC until RNA extraction. Pairs of eyes were dissected from larval heads and transferred to RNA lysis buffer then immediately homogenized with a handheld homogenizer (Bio-Gen Series PRO200, Pro Scientific). RNA was extracted using Zymogen Quick-RNA microprep kit following the manufacturer’s instructions. After eye removal, bodies were transferred to gDNA extraction buffer followed by irf8 genotyping. RNA was quantified and checked using a NanoDropTM One Microvolume UV-Vis Spectrometer (Thermo Scientific). Following RNA extraction, cDNA was synthesized using SuperScript IV First-Strand cDNA Synthesis Kit (Invitrogen) using random hexamers. The cDNA samples were stored at -20oC until qPCR.
Quantitative PCR (qPCR)
Gene-specific primers for qPCR and their sources are shown in Table 1. Quantitative PCR (qPCR) was performed using PowerTrack SYBR-Green Master Mix (Applied Biosystems), using 2 ng of cDNA template per reaction, with reactions done in technical duplicates. Reactions were run on a QuantStudioTM 3 Real-Time PCR machine (ThermoFisher). qPCR reactions underwent 40 amplification cycles, and non-template controls were included for all primer pairs. After reactions, amplification and melt-curves were inspected for clean, specific peaks. Fold change was calculated using the delta delta Ct method using elf1a Ct readings as the calibrator gene.
Graphing and statistical analysis
Graphing and statistical analysis were performed in the R coding environment or using GraphPad Prism. For pairwise comparisons of quantitative data, either parametric (t-test for normally distributed data),non-parametric Mann-Whitney U-tests, or for data with unequal variances Welch’s test, were performed. For proportion data, Fisher Exact test was used. For data with multiple factors, 2-way ANOVA, Kruskal-Wallis, or 3-way ANOVA test was used, followed by post-hoc pairwise comparisons when warranted, as indicated in the figure legends. A p-value <0.05 was considered statistically significant. Statistically significant differences are annotated in the figures and stated in the figure legends.
Supporting information
S1 Fig. Experiments using PLX3397 treatment.
A. Timeline showing treatment of gnat2:nfsb-mCherry;irf8 + /- or gnat2:nfsb-mCherry;irf8-/- fish with Mtz and exposure to PLX3397 (PLX) and EdU. B. Quantification of the number of 4C4 + cells in retinal cryosections of fish at 48 hpt following Mtz treatment, and with (+PLX) or without (No PLX) PLX3397 exposure. C. Quantification of the number of Lcp1 + cells in retinal cryosections of fish at 48 hpt following Mtz treatment, and with (+PLX) or without (No PLX) PLX3397 exposure. D-D.” Images of retinal cryosections from non-PLX treated fish collected at 48 hpt following Mtz treatment, showing PCNA, EdU, and GS signal. E-E.” Images of retinal cryosections from PLX3397 treated fish collected at 48 hpt following Mtz treatment, showing PCNA, EdU, and GS signal. Arrows show GS+PCNA+EdU+ cells in the INL. F. Quantification of the number of GS+PCNA+ cells per retinal section for samples without (No PLX) or with PLX3397 (+PLX) treatment. G. The fraction of GS+PCNA+ cells that were also EdU+ for samples without (No PLX) or with PLX3397 (+PLX) treatment. H. The number of GS-negative, PCNA+EdU+ cells (considered cycling MG-derived progenitors) for samples without (No PLX) or with PLX3397 (+PLX) treatment. ** indicates p < 0.05 (Welch’s test) for comparisons shown in B and C. Differences were not statistically significant for comparisons shown in F, G, and H (Welch’s test).
https://doi.org/10.1371/journal.pone.0344737.s001
(TIF)
S2 Fig. Detection of a small number of EdU+ cones at 120 hpt/5 dpt.
gnat2:nfsb-mCherry;irf8 + /- or gnat2:nfsb-mCherry;irf8-/- larvae were treated with Mtz for 48 hours with EdU immersion following the regime shown in Figure 7E, with EdU washout at 72 hours post treatment (hpt). Retinal cryosections were collected at 120 hpt (5 days post treatment, dpt) and stained for EdU; mCherry was used to identify cones. The graph shows the number of EdU + mCherry+ cells (cones) in retinal cryosections from each genotype detected at 120 hpt (5 dpt).
https://doi.org/10.1371/journal.pone.0344737.s002
(TIF)
S3 Fig. RT-qPCR for ascl1a.
RT-qPCR was performed using RNA collected at the indicated timepoints from pairs of whole larval eyes from gnat2:nfsb-mCherry larvae of either irf8 + /- or irf8-/- genotypes, following DMSO or Mtz treatment. Gene expression for ascl1a was analyzed compared to undamaged (DMSO treated), irf8 + /- samples at 24 hpt, using the 2^ddCt method and graphing as Log2(FC). Complete results from statistical analysis are shown for this data in S1 File. **p = 0.0125 irf8-/- + Mtz 48 hpt vs irf8-/- + DMSO 48 hpt, *p = 0.043; irf8-/- + Mtz 48 hpt vs irf8-/- + Mtz 72 hpt.
https://doi.org/10.1371/journal.pone.0344737.s003
(TIF)
S4 Fig. Additional measurements from retinal cryosections at 7 and 14 dpt.
A, B. Eye size measurements between gnat2:nfsb-mCherry;irf8 + /- or gnat2:nfsb-mCherry;irf8-/- larvae at 14 dpt following Mtz treatment for (A) dorsal-ventral (D-V) axis direction diameter length and (B) distance from the lens to the back of the eye boundary (front-back). Measurements were not different between genotypes (Mann-Whitney U-test). C. Counts of EdU+ cells that were mCherry-negative in retinal cryosections for the indicated timepoints and genotypes. Differences between genotypes and between timepoints were not statistically significant (Mann-Whitney U-test for pairwise comparisons between irf8 + /- and irf8-/- at 7 dpt and at 14 dpt, and pairwise comparisons within the same genotype between 7 dpt and 14 dpt).
https://doi.org/10.1371/journal.pone.0344737.s004
(TIF)
S5 Fig. Staining for Mpx.
A. We performed staining of retinal cryosections with antibody to Mpx (myeloperoxidase) and DAPI to determine if neutrophils were detected in the responding immune cell populations following cone ablation in gnat2:nfsb-mCherry;irf8 + /- or gnat2:nfsb-mCherry;irf8-/- larvae. Images shown are from samples collected at 7 dpt and at 14 dpt, from both irf8 genotypes as indicated. B. Quantification of the number of Mpx+ cells detected in retinal tissue from the two irf8 genotypes at the indicated timepoints following cone ablation. Mpx+ cells were not abundantly detected in retinas of either genotype, though staining was confirmed by the presence of Mpx+ cells (green) in other tissues outside of retina. 2-way ANOVA (no statistical significance was found).
https://doi.org/10.1371/journal.pone.0344737.s005
(TIF)
S1 File. Statistical Analysis for RT-qPCR data.
P-values for statistical analysis of RT-qPCR data pertaining to Fig 6.
https://doi.org/10.1371/journal.pone.0344737.s006
(XLSX)
Acknowledgments
We thank Dr. Mikiko Nagashima (University of Michigan) for the kind gift of the gnat2:nfsb-mCherry transgenic line made in Rachel Wong’s laboratory (University of Washington). We thank the University of Idaho Imaging and Data Acquisition Core (IDAC, managed by the Institute of Modeling Complex Interactions/IMCI) and IIDS for support of our research. We thank Mae Dawson (undergraduate research student) for technical assistance.
References
- 1. Nagashima M, Barthel LK, Raymond PA. A self-renewing division of zebrafish Müller glial cells generates neuronal progenitors that require N-cadherin to regenerate retinal neurons. Development. 2013;140(22):4510–21. pmid:24154521
- 2. Bernardos RL, Barthel LK, Meyers JR, Raymond PA. Late-stage neuronal progenitors in the retina are radial Müller glia that function as retinal stem cells. J Neurosci. 2007;27(26):7028–40. pmid:17596452
- 3. Thummel R, Kassen SC, Enright JM, Nelson CM, Montgomery JE, Hyde DR. Characterization of Müller glia and neuronal progenitors during adult zebrafish retinal regeneration. Exp Eye Res. 2008;87(5):433–44. pmid:18718467
- 4. Fausett BV, Goldman D. A role for alpha1 tubulin-expressing Müller glia in regeneration of the injured zebrafish retina. J Neurosci. 2006;26(23):6303–13. pmid:16763038
- 5. Fimbel SM, Montgomery JE, Burket CT, Hyde DR. Regeneration of inner retinal neurons after intravitreal injection of ouabain in zebrafish. J Neurosci. 2007;27(7):1712–24. pmid:17301179
- 6. Thummel R, Kassen SC, Montgomery JE, Enright JM, Hyde DR. Inhibition of Müller glial cell division blocks regeneration of the light-damaged zebrafish retina. Dev Neurobiol. 2008;68(3):392–408. pmid:18161852
- 7. McGinn TE, Mitchell DM, Meighan PC, Partington N, Leoni DC, Jenkins CE, et al. Restoration of Dendritic Complexity, Functional Connectivity, and Diversity of Regenerated Retinal Bipolar Neurons in Adult Zebrafish. J Neurosci. 2018;38(1):120–36. pmid:29133431
- 8. McGinn TE, Galicia CA, Leoni DC, Partington N, Mitchell DM, Stenkamp DL. Rewiring the Regenerated Zebrafish Retina: Reemergence of Bipolar Neurons and Cone-Bipolar Circuitry Following an Inner Retinal Lesion. Front Cell Dev Biol. 2019;7:95. pmid:31245369
- 9. Sherpa T, Lankford T, McGinn TE, Hunter SS, Frey RA, Sun C, et al. Retinal regeneration is facilitated by the presence of surviving neurons. Dev Neurobiol. 2014;74(9):851–76. pmid:24488694
- 10. Barrett LM, Mitchell DM, Meighan PC, Varnum MD, Stenkamp DL. Dynamic functional and structural remodeling during retinal regeneration in zebrafish. Front Mol Neurosci. 2022;15:1070509. pmid:36533135
- 11. Abraham E, Hartmann H, Yoshimatsu T, Baden T, Brand M. Restoration of cone-circuit functionality in the regenerating adult zebrafish retina. Dev Cell. 2024;59(16):2158–2170.e6. pmid:39096897
- 12. Hammer J, Röppenack P, Yousuf S, Schnabel C, Weber A, Zöller D, et al. Visual Function is Gradually Restored During Retina Regeneration in Adult Zebrafish. Front Cell Dev Biol. 2022;9:831322. pmid:35178408
- 13. Mitchell DM, Lovel AG, Stenkamp DL. Dynamic changes in microglial and macrophage characteristics during degeneration and regeneration of the zebrafish retina. J Neuroinflammation. 2018;15(1):163. pmid:29804544
- 14. White DT, Sengupta S, Saxena MT, Xu Q, Hanes J, Ding D, et al. Immunomodulation-accelerated neuronal regeneration following selective rod photoreceptor cell ablation in the zebrafish retina. Proc Natl Acad Sci U S A. 2017;114(18):E3719–28. pmid:28416692
- 15. Bludau O, Weber A, Bosak V, Kuscha V, Dietrich K, Hans S, et al. Inflammation is a critical factor for successful regeneration of the adult zebrafish retina in response to diffuse light lesion. Front Cell Dev Biol. 2024;12:1332347. pmid:39071801
- 16. Iribarne M, Hyde DR. Different inflammation responses modulate Müller glia proliferation in the acute or chronically damaged zebrafish retina. Front Cell Dev Biol. 2022;10:892271. pmid:36120571
- 17. Fischer AJ, Zelinka C, Gallina D, Scott MA, Todd L. Reactive microglia and macrophage facilitate the formation of Müller glia-derived retinal progenitors. Glia. 2014;62(10):1608–28. pmid:24916856
- 18. Song P, Parsana D, Singh R, Pollock LM, Anand-Apte B, Perkins BD. Photoreceptor regeneration occurs normally in microglia-deficient irf8 mutant zebrafish following acute retinal damage. Sci Rep. 2024;14(1):20146. pmid:39209978
- 19. Xu H, Cao L, Chen Y, Zhou C, Xu J, Zhang Z, et al. Single-cell RNA sequencing reveals the heterogeneity and interactions of immune cells and Müller glia during zebrafish retina regeneration. Neural Regen Res. 2025;20(12):3635–48. pmid:38934409
- 20. Conedera FM, Pousa AMQ, Mercader N, Tschopp M, Enzmann V. Retinal microglia signaling affects Müller cell behavior in the zebrafish following laser injury induction. Glia. 2019;67(6):1150–66. pmid:30794326
- 21. Silva NJ, Nagashima M, Li J, Kakuk-Atkins L, Ashrafzadeh M, Hyde DR, et al. Inflammation and matrix metalloproteinase 9 (Mmp-9) regulate photoreceptor regeneration in adult zebrafish. Glia. 2020;68(7):1445–65. pmid:32034934
- 22. Fogerty J, Song P, Boyd P, Grabinski SE, Hoang T, Reich A, et al. Notch Inhibition Promotes Regeneration and Immunosuppression Supports Cone Survival in a Zebrafish Model of Inherited Retinal Dystrophy. J Neurosci. 2022;42(26):5144–58. pmid:35672150
- 23. D’Orazi FD, Suzuki SC, Darling N, Wong RO, Yoshimatsu T. Conditional and biased regeneration of cone photoreceptor types in the zebrafish retina. J Comp Neurol. 2020;528(17):2816–30. pmid:32342988
- 24. Thomas JL, Nelson CM, Luo X, Hyde DR, Thummel R. Characterization of multiple light damage paradigms reveals regional differences in photoreceptor loss. Exp Eye Res. 2012;97(1):105–16. pmid:22425727
- 25. Zhang Z, Hou H, Yu S, Zhou C, Zhang X, Li N, et al. Inflammation-induced mammalian target of rapamycin signaling is essential for retina regeneration. Glia. 2020;68(1):111–27. pmid:31444939
- 26. García-García D, Vidal-Gil L, Parain K, Lun J, Audic Y, Chesneau A, et al. Neuroinflammation as a cause of differential Müller cell regenerative responses to retinal injury. Sci Adv. 2024;10(40):eadp7916. pmid:39356769
- 27. Rumford JE, Grieshaber A, Lewiston S, Reed JL, Long SS, Mitchell DM. Forced MyD88 signaling in microglia impacts the production and survival of regenerated retinal neurons. Front Cell Dev Biol. 2024;12:1495586. pmid:39633708
- 28. Todd L, Finkbeiner C, Wong CK, et al. Microglia Suppress Ascl1-Induced Retinal Regeneration in Mice. Cell Reports 2020; 33:108507.
- 29. Morales M, Findley AP, Mitchell DM. Intercellular contact and cargo transfer between Müller glia and to microglia precede apoptotic cell clearance in the developing retina. Development. 2024;151(1):dev202407. pmid:38174987
- 30. Harada T, Harada C, Kohsaka S, Wada E, Yoshida K, Ohno S, et al. Microglia-Müller glia cell interactions control neurotrophic factor production during light-induced retinal degeneration. J Neurosci. 2002;22(21):9228–36. pmid:12417648
- 31. Mitra S, Devi S, Lee M-S, Jui J, Sahu A, Goldman D. Vegf signaling between Müller glia and vascular endothelial cells is regulated by immune cells and stimulates retina regeneration. Proc Natl Acad Sci U S A. 2022;119(50):e2211690119. pmid:36469778
- 32. Raghavan D, Jeakle O, Berry Y, Victor M, Thummel R. Microglia response and function in a chronic model of photoreceptor damage. Front Cell Dev Biol. 2025;13:1699271. pmid:41476611
- 33. Blume ZI, Lambert JM, Lovel AG, Mitchell DM. Microglia in the developing retina couple phagocytosis with the progression of apoptosis via P2RY12 signaling. Dev Dyn. 2020;249(6):723–40. pmid:32072708
- 34. Ellett F, Pase L, Hayman JW, Andrianopoulos A, Lieschke GJ. mpeg1 promoter transgenes direct macrophage-lineage expression in zebrafish. Blood. 2011;117(4):e49–56. pmid:21084707
- 35. Mitchell DM, Sun C, Hunter SS, New DD, Stenkamp DL. Regeneration associated transcriptional signature of retinal microglia and macrophages. Sci Rep. 2019;9(1):4768. pmid:30886241
- 36. Leach LL, Hanovice NJ, George SM, Gabriel AE, Gross JM. The immune response is a critical regulator of zebrafish retinal pigment epithelium regeneration. Proc Natl Acad Sci U S A. 2021;118(21):e2017198118. pmid:34006636
- 37. Gallina D, Zelinka C, Fischer AJ. Glucocorticoid receptors in the retina, Müller glia and the formation of Müller glia-derived progenitors. Development. 2014;141(17):3340–51. pmid:25085975
- 38. Gallina D, Zelinka CP, Cebulla CM, Fischer AJ. Activation of glucocorticoid receptors in Müller glia is protective to retinal neurons and suppresses microglial reactivity. Exp Neurol. 2015;273:114–25. pmid:26272753
- 39. Emmerich K, White DT, Kambhampati SP, Casado GL, Fu T-M, Chunawala Z, et al. Nanoparticle-based targeting of microglia improves the neural regeneration enhancing effects of immunosuppression in the zebrafish retina. Commun Biol. 2023;6(1):534. pmid:37202450
- 40. Shiau CE, Kaufman Z, Meireles AM, Talbot WS. Differential requirement for irf8 in formation of embryonic and adult macrophages in zebrafish. PLoS One. 2015;10(1):e0117513. pmid:25615614
- 41. Thiel WA, Blume ZI, Mitchell DM. Compensatory engulfment and Müller glia reactivity in the absence of microglia. Glia. 2022;70(7):1402–25. pmid:35451181
- 42. Oosterhof N, Kuil LE, Linde HC van der. Colony-Stimulating Factor 1 Receptor (CSF1R) Regulates Microglia Density and Distribution, but Not Microglia Differentiation In Vivo. Cell Reports. 2018;24:1203-1217.e6.
- 43. Leach LL, Gonzalez RG, Jayawardena SU, Gross JM. Macrophage/microglia-dependent mechanisms drive retinal pigment epithelium regeneration in zebrafish. Cell Rep. 2025;44(10):116325. pmid:40974576
- 44. D’Orazi FD, Zhao X-F, Wong RO, Yoshimatsu T. Mismatch of Synaptic Patterns between Neurons Produced in Regeneration and during Development of the Vertebrate Retina. Curr Biol. 2016;26(17):2268–79. pmid:27524481
- 45. Rovira M, Miserocchi M, Montanari A, Hammou L, Chomette L, Pozo J, et al. Zebrafish Galectin 3 binding protein is the target antigen of the microglial 4C4 monoclonal antibody. Dev Dyn. 2023;252(3):400–14. pmid:36285351
- 46. Wehman AM, Staub W, Meyers JR, Raymond PA, Baier H. Genetic dissection of the zebrafish retinal stem-cell compartment. Dev Biol. 2005;281(1):53–65. pmid:15848388
- 47. Fischer AJ, Bosse JL, El-Hodiri HM. The ciliary marginal zone (CMZ) in development and regeneration of the vertebrate eye. Exp Eye Res. 2013;116:199–204. pmid:24025744
- 48. Le Guyader D, Redd MJ, Colucci-Guyon E, Murayama E, Kissa K, Briolat V, et al. Origins and unconventional behavior of neutrophils in developing zebrafish. Blood. 2008;111(1):132–41. pmid:17875807
- 49. Anderson SR, Roberts JM, Ghena N, Irvin EA, Schwakopf J, Cooperstein IB, et al. Neuronal apoptosis drives remodeling states of microglia and shifts in survival pathway dependence. Elife. 2022;11:e76564. pmid:35481836
- 50. Elmore MRP, Najafi AR, Koike MA, Dagher NN, Spangenberg EE, Rice RA, et al. Colony-stimulating factor 1 receptor signaling is necessary for microglia viability, unmasking a microglia progenitor cell in the adult brain. Neuron. 2014;82(2):380–97. pmid:24742461
- 51. Lu C, Hyde DR. Cytokines IL-1β and IL-10 are required for Müller glia proliferation following light damage in the adult zebrafish retina. Front Cell Dev Biol. 2024;12:1406330. pmid:38938553
- 52. Nelson CM, Ackerman KM, O’Hayer P, Bailey TJ, Gorsuch RA, Hyde DR. Tumor necrosis factor-alpha is produced by dying retinal neurons and is required for Muller glia proliferation during zebrafish retinal regeneration. J Neurosci. 2013;33(15):6524–39. pmid:23575850
- 53. Bringmann A, Iandiev I, Pannicke T, Wurm A, Hollborn M, Wiedemann P, et al. Cellular signaling and factors involved in Müller cell gliosis: neuroprotective and detrimental effects. Prog Retin Eye Res. 2009;28(6):423–51. pmid:19660572
- 54. Thomas JL, Ranski AH, Morgan GW, Thummel R. Reactive gliosis in the adult zebrafish retina. Exp Eye Res. 2016;143:98–109. pmid:26492821
- 55.
Clark DT. Visual responses in developing zebrafish (Brachydanio rerio). University of Oregon.
- 56. Bilotta J, Saszik S, Sutherland SE. Rod contributions to the electroretinogram of the dark-adapted developing zebrafish. Dev Dyn. 2001;222(4):564–70. pmid:11748826
- 57. Branchek T, Bremiller R. The development of photoreceptors in the zebrafish, Brachydanio rerio. I. Structure. J Comp Neurol. 1984;224:107–15.
- 58. Haug MF, Berger M, Gesemann M, Neuhauss SCF. Differential expression of PKCα and -β in the zebrafish retina. Histochem Cell Biol. 2019;151(6):521–30. pmid:30604284
- 59. Suzuki S, Kaneko A. Identification of bipolar cell subtypes by protein kinase C-like immunoreactivity in the goldfish retina. Vis Neurosci. 1990;5(3):223–30. pmid:2134845
- 60. D’Orazi FD, Suzuki SC, Wong RO. Neuronal remodeling in retinal circuit assembly, disassembly, and reassembly. Trends Neurosci. 2014;37(10):594–603. pmid:25156327
- 61. Saade CJ, Alvarez-Delfin K, Fadool JM. Rod photoreceptors protect from cone degeneration-induced retinal remodeling and restore visual responses in zebrafish. J Neurosci. 2013;33(5):1804–14. pmid:23365220
- 62. Powell C, Cornblath E, Elsaeidi F, Wan J, Goldman D. Zebrafish Müller glia-derived progenitors are multipotent, exhibit proliferative biases and regenerate excess neurons. Sci Rep. 2016;6:24851. pmid:27094545
- 63. Ng Chi Kei J, Currie PD, Jusuf PR. Fate bias during neural regeneration adjusts dynamically without recapitulating developmental fate progression. Neural Dev. 2017;12(1):12. pmid:28705258
- 64. Lahne M, Brecker M, Jones SE, Hyde DR. The Regenerating Adult Zebrafish Retina Recapitulates Developmental Fate Specification Programs. Front Cell Dev Biol. 2021;8:617923. pmid:33598455
- 65. Hanovice NJ, Leach LL, Slater K, Gabriel AE, Romanovicz D, Shao E, et al. Regeneration of the zebrafish retinal pigment epithelium after widespread genetic ablation. PLoS Genet. 2019;15(1):e1007939. pmid:30695061
- 66. George SM, Lu F, Rao M, Leach LL, Gross JM. The retinal pigment epithelium: Development, injury responses, and regenerative potential in mammalian and non-mammalian systems. Prog Retin Eye Res. 2021;85:100969. pmid:33901682
- 67. Konishi H, Okamoto T, Hara Y, Komine O, Tamada H, Maeda M, et al. Astrocytic phagocytosis is a compensatory mechanism for microglial dysfunction. EMBO J. 2020;39(22):e104464. pmid:32959911
- 68. Beachum AN, Salazar G, Nachbar A. Glia preserve their own functions while compensating for neighboring glial cell dysfunction. Glia. 2025.
- 69. Wang X, Zhao L, Zhang J, Fariss RN, Ma W, Kretschmer F, et al. Requirement for Microglia for the Maintenance of Synaptic Function and Integrity in the Mature Retina. J Neurosci. 2016;36(9):2827–42. pmid:26937019
- 70. Weimar HV, Farre AA, Rumford JE, Mitchell DM. Müller glial and microglial responses coupled to recovery of cone photoreceptors following limited cone ablation in zebrafish retina. Sci Rep. 2026;16(1):9058. pmid:41691075
- 71.
Westerfield M. The zebrafish book. A guide for the laboratory use of zebrafish (Danio rerio). 5th ed. University of Oregon Press. 2007.
- 72. Meeker ND, Hutchinson SA, Ho L, Trede NS. Method for isolation of PCR-ready genomic DNA from zebrafish tissues. Biotechniques. 2007;43(5):610, 612, 614. pmid:18072590
- 73. Lovel AG, Mitchell DM. Axon regeneration, methods and protocols. Methods Mol Biol. 2023;2636:389–400.
- 74. Tsarouchas TM, Wehner D, Cavone L, Munir T, Keatinge M, Lambertus M, et al. Dynamic control of proinflammatory cytokines Il-1β and Tnf-α by macrophages in zebrafish spinal cord regeneration. Nat Commun. 2018;9(1):4670. pmid:30405119
- 75. Kirchberger S, Shoeb MR, Lazic D, Wenninger-Weinzierl A, Fischer K, Shaw LE, et al. Comparative transcriptomics coupled to developmental grading via transgenic zebrafish reporter strains identifies conserved features in neutrophil maturation. Nat Commun. 2024;15(1):1792. pmid:38413586