Dynamic changes in microglial and macrophage characteristics during degeneration and regeneration of the zebrafish retina

Dynamic changes in microglial and macrophage characteristics during degeneration and regeneration... Background: In contrast to mammals, zebrafish have the capacity to regenerate retinal neurons following a variety of injuries. Two types of glial cells, Müller glia (MG) and microglia, are known to exist in the zebrafish retina. Recent work has shown that MG give rise to regenerated retinal neurons, but the role of resident microglia, and the innate immune system more generally, during retinal regeneration is not well defined. Specifically, characteristics of the immune system and microglia following substantial neuron death and a successful regenerative response have not been documented. Methods: The neurotoxin ouabain was used to induce a substantial retinal lesion of the inner retina in zebrafish. This lesion results in a regenerative response that largely restores retinal architecture, neuronal morphologies, and connectivities, as well as recovery of visual function. We analyzed cryosections from damaged eyes following immunofluorescence and H&E staining to characterize the initial immune response to the lesion. Whole retinas were analyzed by confocal microscopy to characterize microglia morphology and distribution. Statistical analysis was performed using a two-tailed Student’s t test comparing damaged to control samples. Results: We find evidence of early leukocyte infiltration to the retina in response to ouabain injection followed by a period of immune cell proliferation that likely includes both resident microglia and substantial numbers of proliferating, extra-retinally derived macrophages, leading to rapid accumulation upon retinal damage. Following immune cell proliferation, Müller glia re-enter the cell cycle. In retinas that have regenerated the layers lost to the initial injury (histologically regenerated), microglia retain morphological features of activation, suggesting ongoing functions that are likely essential to restoration of retinal function. Conclusions: Collectively, these results indicate that microglia and the immune system are dynamic during a successful regenerative response in the retina. This study provides an important framework to probe inflammation in the initiation of, and functional roles of microglia during retinal regeneration. Keywords: Retina, Regeneration, Microglia, Macrophages, Zebrafish, Müller glia * Correspondence: dmitchell@uidaho.edu Department of Biological Sciences, University of Idaho, 875 Perimeter Drive, MS 3051, Moscow, ID 83844-3051, USA © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 2 of 20 Background other tissues participate in the phagocytosis and clearance The mammalian retina is unable to regenerate damaged of degenerative debris before a regenerative response be- neurons due to acute trauma or degenerative disease. gins [23]. In zebrafish, recruited macrophages appear to be Rather than restoration of lost neurons, a gliotic re- essential to regeneration of peripheral nervous system tis- sponse ensues, which is often associated with continual sue [24, 25]. The role of the innate immune system and inflammation [1]. Inflammation is the body’s response to resident macrophages, microglia, in regeneration of the tissue injury and/or infection, and is initiated by the innate central nervous system in organisms which have such re- immune system. Inflammation includes activation of im- generative capacity, such as teleost fish, is not well defined. mune cells and the production of pro-inflammatory cyto- Interestingly, a recent publication using larval zebrafish kines and molecular mediators. Particular cell types supports a role for inflammation and resident microglia in activated and cytokines/molecular mediators produced regeneration of rod photoreceptors in the retina [26]. In- will vary depending on the initial insult or injury, the tis- flammation has also been shown to be important in regen- sue location, and the immune response necessary to re- eration of the adult zebrafish brain [27]. Yet, it is not clear solve the infection or injury. Within the mammalian if these findings translate to retinal regeneration in an retina, resident microglia rapidly sense changes in the adult animal. microenvironment and can initiate and participate in both Under normal conditions, retinal microglia are highly acute and chronic inflammatory responses [2]. Acute in- ramified and integrated among the highly organized ret- flammation is crucial to clearing dead cells, debris, and inal tissue. Microglial responses to retinal insult and de- pathogens, which must occur prior to subsequent initi- generation have been studied mainly in rodent model ation of tissue repair. However, when inflammation be- systems [28]. Generally, when responding to neuro- comes chronic, it can be extremely damaging and logical insult, microglia transform morphology from contribute to tissue pathology. While chronic inflam- ramified to an ameboid shape and migrate to sites of cell mation and immune activation in mammalian neurode- death [29, 30]. However, microglial responses to retinal generative disease is well appreciated, the role of the injury in zebrafish have not been well documented, and immune system in contexts of successful retinal regen- it remains unclear under which conditions of retinal in- eration has not been well explored. jury extra-retinal immune cells may participate in the In contrast to mammals, teleost fish have a remarkable initial response to neuronal degeneration. Studies to date capacity to regenerate damaged retinas following a in zebrafish have focused on how inflammation, specific variety of lesions that destroy neurons [3–5]. The source cytokines, and microglia/macrophages may affect the of regenerated retinal neurons in zebrafish are the proliferation of Müller glial cells, which give rise to re- Müller glia (reviewed in [6–8]), retinal glial cells also generated retinal neurons [22, 26, 31]. There has been a present in mammals and other vertebrates. Upon injury, surge of interest in the role of microglia in development Müller glia re-enter the cell cycle [9–13] and undergo an and maintenance of the central nervous system, as well asymmetric division [14], ultimately generating multi- as their involvement in, and possibly sources of, neuro- potent progenitors that replace lost retinal neurons degenerative disease [2, 32]. Yet, little is known about [11, 14–16]. Intriguingly, Müller glia in mammals and the role of microglia in initiating and/or shaping success- other vertebrates appear to have properties suggesting ful regeneration of retinal neurons, and if these functions regenerative potential, including the upregulation of are shared with or unique from those in development. progenitor markers and cell cycle re-entry (reviewed In the present study, we use a tissue-disrupting lesion, in [1, 4, 17–19], butthisinstead resultsingliosisdue intraocular injection of the neurotoxin ouabain, to des- to the inability to differentiate into neurons (exten- troy inner retinal neurons while sparing glia and photo- sively reviewed in [1, 17]). Further, immune-Müller receptors [9, 10, 14, 33]. This type of lesion is known to glial crosstalk may be important in shaping the Müller result in a regenerative response that largely restores glia response to retinal injury [20–22]. retinal architecture, neuronal morphologies and con- In wound healing outside of the central nervous sys- nectivities, and behavioral and electrophysiological tem, the role of the immune system is well appreciated. measures of retinal function [10, 33]. We characterize In general, a period of degeneration is followed by re- the initial response to this cytotoxic damage to deter- generation in which the lost cell types are replaced and mine how the innate immune system responds to retinal tissue structure is restored. Tissue resident macrophages degeneration that is subsequently followed by a successful are crucial to coordinating the clearance of damaged tissue regenerative response. We then investigate microglial char- as well as instructing differentiation of new cell types to acteristics, including distribution, morphology, selective replace damaged tissue [23]. Resident macrophages as well markers, and histological features, in regenerated retinas. as recruited phagocytes, such as neutrophils, monocyte- We find that in damaged zebrafish retinas, both resident derived macrophages, and macrophages that migrate from and infiltrating immune cells contribute to mount a robust Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 3 of 20 response to neuronal cell death that is immediately Tissue collection and processing followed by Müller glia proliferation. Upon histological At selected time points post-injection, tissue samples regeneration of retinal layers, resident microglia localize were collected for analysis. For whole retina collection, to regions of regenerated neurons and maintain morph- fish were dark adapted for approximately 12 h, anesthe- ologies indicative of ongoing functional activation. Our tized in tricaine solution, and then eyes were enucleated findings indicate that retinal regeneration in adult zebra- using fine forceps, and placed into phosphate-buffered fish provides an excellent system to discover functional (pH = 7.4) saline (PBS) for dissection. The cornea and roles for the immune system and microglia that may be lens were removed and retinas were peeled from the crucial to supporting successful regenerative responses. whole eyecup. The retinal pigmented epithelium (RPE) detached from the retina due to the dark adaption, and Methods any remaining RPE was brushed away using a soft paint- Animals brush. Retinas were then rinsed several times in fresh Procedures involving zebrafish were performed in com- PBS using a plastic transfer pipette, transferred into fixa- pliance with protocols approved by the University of tive consisting of 4% paraformaldehyde in PBS for ap- Idaho Animal Care and Use Committee (IACUC). Zebra- proximately 1 h at room temperature with constant fish (Danio rerio) were maintained on a 14:10 light/dark gentle agitation, then washed several times with PBS cycle in 28.5 °C recirculating, monitored system water, containing 0.01% TritonX-100 (PBST). housed, and propagated according to [34]. Zebrafish To prepare retinal cryosections, whole eyes were enu- transgenic lines used in this study include mpeg1:GFP cleated using fine forceps, transferred to PBS, and the (gl22 Tg, GFP expressed in microglia/macrophages [35], lens was removed. Eyes were then fixed in phosphate- available from Zebrafish International Resource Center, buffered, 4% paraformaldehyde containing 5% sucrose ZIRC); mpeg1:mCherry (gl23 Tg, mCherry expressed in for 1 h at room temperature, washed in phosphate- microglia/macrophages [35], available from ZIRC). The buffered (pH = 7.4) 5% sucrose, and then washed in a gl22 Tg and gl23 Tg transgenic lines co-label cells in retinal graded series ending in 20% sucrose. The following day, tissue (Additional file 1: Figure S1). The wild-type strain tissues were embedded in blocks of a 1:2 solution of OCT used, referred to as SciH, was originally obtained from embedding medium (Sakura Finetek) and phosphate- Scientific Hatcheries (now Aquatica Tropicals). buffered, 20% sucrose, and frozen in isobutane super- cooled with liquid N . After freezing solid, tissues were Retinal lesion sectioned at 5 μm thickness using a Leica CM3050 cryo- Chemical lesioning of zebrafish retinas (age 6–14 months, stat. After overnight desiccation, tissue sections on glass both sexes) was performed by intravitreal injection of oua- slides were stored at − 20 °C until use. bain in order to destroy inner retinal neurons and spare To evaluate co-label of gl22 Tg and gl23 Tg, whole em- photoreceptors and Müller glia [9, 10, 33]. A working bryos were fixed at 3 days post-fertilization and washed stock of 40 μM ouabain (ouabain octahydrate, Sigma- with PBST as described for whole retina collection Aldrich) was prepared in 0.65% sterile saline (NaCl) above. Whole eyes were removed and mounted in solution. Fish were anesthetized by immersion in tricaine VECTASHIELD (Vector Labs), covered with a coverslip, solution, and an incision was made across the cornea then imaged as described below. using a sapphire blade. A Hamilton syringe was inserted behind the lens and 0.4–0.6 μLof 40 μM ouabain solution Immunofluorescence was injected into the vitreal chamber. Volume injected For staining of whole retinas, fixed tissue was blocked in was based on diameter of the eye (measured with calipers) 20% goat serum for 1 h at room temperature, incubated and upon calculations based on geometry and volumes of with primary antibody for 1–3 days at 4 °C, washed the eye and lens, resulting in an estimated intraocular con- three times in PBST, then incubated with conjugated centration of 2 μM[9, 36]. Lesions were unilateral and secondary antibody and DAPI (4′,6-diamidino-2-pheny- only the right eye was injected. For control samples, right lindole, 4.25 μM) for 1–3 days at 4 °C, washed twice in eyes of a separate group of fish were injected with 0.65% PBST with a final wash in PBS. All antibody incubations sterile saline (NaCl) solution. The same sterile saline solu- and washes were performed with constant agitation. tion was used for saline injections and for preparation of Whole retinas were then flat mounted by cutting four ouabain solution for injection. During the procedure, fish slits around the perimeter of the retinal cup and were were continuously flushed with tricaine solution. Immedi- mounted flat onto a microscope slide in VECTASHIELD ately following the procedure, fish were returned to tanks (Vector Labs), covered with a coverslip, and sealed with with clean system water. For all experiments, 3 control clear nail polish. (saline injected) and 3–4 ouabain injected fish were used To stain cryosections, tissue sections (5 μm thickness) for subsequent analysis. were blocked in 20% goat serum for 30 min at room Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 4 of 20 temperature, incubated in primary antibody overnight, were acquired at × 20 magnification using the large washed in PBST for at least 30 min, incubated in sec- stitched images feature in Nikon Elements software and ondary antibody for 1 h, and washed in PBST for at stitched based on DAPI staining. Z stacks of flattened least 30 min. Slides were then mounted in Vectashield retinas were generally obtained at 1 μm intervals. + DAPI (Vector Labs), covered with a coverslip, and Imaging of H&E stained cryosections was performed sealed with clear nail polish. Primary antibodies and di- using a Leica DM2500 compound microscope with Leica lutions used rabbit polyclonal anti-zebrafish L-plastin DFC700T color camera using × 20, × 40 (dry), or × 100 [37, 38] (1:10,000, a kind gift of Dr. Michael Redd), (Oil) objectives. Image processing and analysis was rabbit anti-zebrafish mpx (1:200, Genetex), mouse ZPR1 performed using FIJI (ImageJ). antibody (labeling cone-arrestin 3a [39], 1:200, ZIRC), mouse anti-PCNA (PC10, 1:200, Santa Cruz Biotech), rat Quantification of immune cells in retinal cryosections anti-PCNA (16D10, 1:200, Chromotek), rabbit anti- (24–72 h post-injection) phosphorylated histone 3 (PH3) (1:500, Cell Signaling), Using FIJI (ImageJ), a region of interest was drawn in rabbit polyclonal anti-PKCα (Santa Cruz Biotech), mouse each image obtained from cryosections corresponding to ZRF1 antibody (labeling GFAP, 1:200, ZIRC), and mouse the region of the retinal lesion in the inner retina: from anti-Glutamine Synthetase (1:1000, BD Transduction the basal border of the outer nuclear layer to the basal Laboratories). Secondary antibodies conjugated to Cy3, border of the ganglion cell layer. For 48 and 72 h post- FITC, or Alexa-Fluor647 (Jackson ImmunoResearch) were injury (hpi) time points, the basal layer of this border used at 1:200 dilution. was defined based on the limit of dense DAPI staining. DAPI+ nuclei surrounded by L-plastin+ cell bodies were H&E staining counted within the defined region of interest and nor- Slides containing cryosections were air dried at room malized per 1000 square microns. Note that the analyzed temperature for 30 min, immersed in 0.1% Hematoxylin region corresponds to the lesioned inner retina where (MHS-16, Sigma-Aldrich) 5 min, then washed in run- the x direction is parallel to retinal layers and the y dir- ning tap water for 5 min. Slides were then immersed in ection is perpendicular to retinal layers. Since the inner 0.5% eosin (Sigma-Aldrich) for 1 min, rinsed in deion- nuclear layer is damaged due to the ouabain-induced le- ized water until rinse was clear, then immersed in a sion and becomes thinner, the number of L-plastin+ series of ethanol solutions (50, 75, 95, and 100%) for cells is also reported per 100 μm of curvi-linear distance 30 s-1 min each, dipped five times in Xylenes (Fisher of retina using a line that follows parallel to the basal Scientific) and immediately mounted with Permount border of the ONL. Three to four non-adjacent cryosec- (Fisher Scientific) and sealed with a coverslip. tions were analyzed per fish. TUNEL staining Quantification of immune cells in whole, flat mounted TUNEL (Terminal deoxynucleotidyl transferase dUTP retinas (histologically regenerated retinas) nick end labeling) staining of 5 μm cryosections was Using FIJI (Image J), individual microglia (DAPI+ nuclei performed following the manufacturer’sinstructions surrounded by L-plastin+ signal) in flattened whole (Roche). Briefly, cryosections were washed for 20 min retinas were counted in individual image stacks contain- in × 1 PBS at room temperature then placed in ice-cold ing all z planes obtained at × 40 magnification, from the permeabilization solution for 2 min. Slides were washed vitreal face of the ganglion cell layer to the apical face of twice in × 1 PBS, 10 min per wash, incubated in the outer nuclear layer. Counts were normalized per TUNEL reaction mixture at 37 °C in the dark for 3 h, 1000 square microns of flattened retina where the x and then rinsed three times in PBS. Immediately following, y direction are parallel to retinal layers. Counts were cryosections were blocked in 20% goat serum for normalized to area rather than volume because the re- 30 min at room temperature and subsequent immuno- generated retinas are thinner than controls [33]. Six to fluorescence was performed as described above. eight images covering approximately 190,000 μm² per image were analyzed from each whole retina. Microscopy and image acquisition Imaging of whole flat mounted retinas (z series) or fluo- Analysis of microglia morphology rescently stained retinal sections (single plane) was per- We analyzed morphology of microglia in the ganglion formed with a Nikon Andor spinning disk confocal cell layer of histologically regenerated and control retinas. microscope equipped with a Zyla sCMOS camera run- Confocal z stacks obtained from whole, flat mounted ret- ning Nikon Elements software. Imaging was performed inas using a × 40 Oil objective corresponding to the region using × 20 (dry) or × 40 (oil immersion) objectives. For of the ganglion cell layer were flattened using FIJI (ImageJ). stitched images of entire retinal cryosections, images We used the marker L-plastin, which is a leukocyte-specific Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 5 of 20 form of the actin bundling protein plastin expressed by Statistical analysis zebrafish macrophages [37], to analyze microglia in A two-tailed Student’s t test was used to obtain p values control and histologically regenerated retinas because comparing saline injected to ouabain injected samples. (1) all retinal L-plastin+ cells co-label with GFP P values less than 0.05 are reported in the figures. expression in mpeg1:GFP transgenic fish at these time points and (2) L-plastin staining labels the entire cell Results body and cellular processes continuously, whereas the Immune cell response to a tissue-disrupting retinal lesion mpeg1 promoter-driven transgenic reporter labels To investigate the immune response to substantial neur- portions of the cell (Fig. 9A,A’,B,B’). Outlines of indi- onal damage in the retina, we used intravitreal injection vidual L-plastin+ cells were manually traced and mea- of the neurotoxin ouabain to generate this damage. In- surements of morphological features were analyzed jection of ouabain to a final concentration of 2 μM using the “measure” tool in FIJI. Approximately 35 causes death of inner retinal neurons, as demonstrated microglia, taken from 3 to 4 retinas per condition, by the loss of the PKCα+ bipolar neuron population at were analyzed for each condition and histograms of 3 days post-injection (dpi) (Fig. 1) and appearance of the distributions were created. Descriptions of the pa- TUNEL+ nuclei within the inner retina at 24 and 48 h rameters used to analyze microglia morphology are as post-injection (hpi) (Fig. 2). Photoreceptors and glia are follows. Area: area of an individual microglia in square spared [9, 10, 14, 33], as demonstrated by the continued microns; circularity: circularity of an individual micro- presence of zpr1+ cones and zrf1+ Müller glia at 3 dpi glial cell body. A value of 1.0 indicates a perfect circle, (Fig. 1). We visualized microglia/immune cells in control while values closer to 0 indicate elongated shapes; per- and ouabain-lesioned retinas using an antibody to zebra- imeter: the total length of a line tracing the outside fish L-plastin, which is used to mark leukocytes, includ- boundary of an individual microglial cell in microns; ing macrophages, in zebrafish [37, 38]. Similar to reports Feret diameter, also known as maximum caliper, the in embryonic and larval zebrafish [40], microglia in in- longest distance between any two points along the per- tact, undamaged adult zebrafish retina reside in the imeter of an individually traced microglial cell. inner nuclear layer, ganglion cell layer, and nerve fiber Fig. 1 Ouabain-induced retinal degeneration results in a robust accumulation of responding immune cells. Images show cryosections from undamaged (control, A, C) and ouabain-damaged retinas at 3 days post-injection (3 dpi, B, D). A ZPR1 (green), PKC-α (Red), and DAPI (blue) were used to label cones of the outer retina, bipolar neurons of the inner retina, and all nuclei, respectively, in undamaged retinas. B Images of damaged retinas sampled at 3 dpi; inner retinal neurons have been destroyed (note the absence of PKC-α staining and absence of DAPI+ layer corresponding to the ganglion cell layer; GCL), but photoreceptors are spared (ZPR1, green). C Ramified microglia, labeled by L-plastin (magenta), are present in undamaged retinas, along with radially patterned Müller glia (ZRF1, green). D Müller glia (green) are spared from the ouabain-induced lesion. A large number of immune cells (magenta) responding to the lesion are present in the damaged retina, primarily localized to the region of neuronal cell death (inner retina), although some appear to be recruited from regions apical to the retina (arrows). DAPI+ nuclei in the inner retina visible in B can be identified as immune cells and Müller glia. Scale bar in A, applies to all images, = 20 μm. ONL = outer nuclear layer, INL = inner nuclear layer, GCL = ganglion cell layer Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 6 of 20 Fig. 2 Progression of ouabain-induced retinal cell death and accumulation of immune cells in damaged retinas. Cryosections (5 μm thick) from retinas at 24 and 48 h post-intravitreal injection (24 and 48 hpi) of saline (A and B)or2 μM ouabain (C and D) were stained using the TUNEL cell death detection process (green), the immune cell marker L-plastin (magenta), and DAPI (blue). A–B Saline injection did not induce cell death (absence of TUNEL staining), and immune cells remain ramified (arrows). C–D TUNEL+ nuclei and debris (green) are present in the ganglion cell layer (GCL) and inner nuclear layer (INL) following intravitreal injection of 2 μM ouabain, and immune cells assume ameboid morphologies (arrows). As the lesion progresses over time, immune cells accumulate in regions of cell death, and the clearance of TUNEL+ nuclei and debris corresponds to this immune cell accumulation. Few immune cells contained TUNEL+, DAPI+ nuclei (quantification in the “Results” section), suggesting that immune cell death was not significantly induced by ouabain. E–G High magnification images of immune cells revealed immune cells phagocytosing TUNEL+ nuclei (E), immune cells with cytoplasmic TUNEL+ material (F, arrow), and immune cells with multiple nuclei, some of which are TUNEL+ (G,asterisks), suggesting that the accumulating immune cells are highly phagocytic and important for clearing dead cells and debris from the lesioned retina. Scale bar in A (applies to A–D)=20 μm. Scale bars in E, F,and G =5 μm. ONL = outer nuclear layer, INL = inner nuclear layer, GCL = ganglion cell layer layer (Fig. 1C). In addition, microglia flank the apical side To monitor the accumulation of immune cells over of the inner plexiform layer and the apical and basal the course of ouabain-induced cell death and to deter- sides of the outer plexiform layers (Fig. 1C). This is dif- mine how ouabain and the intravitreal injection process ferent than reports from mammals, which show that affect immune cells in the retina, we examined L-plastin microglia localize within plexiform layers [2]. After in- and TUNEL labeled cryosections from ouabain injected travitreal injection of ouabain, we found a dramatic ac- retinas at 24 and 48 hpi, compared to saline injected cumulation of immune cells in the region of the lesion controls (Fig. 2). In ouabain lesioned retinas, TUNEL+ by 3dpi (72hpi)(Fig. 1D). At 3 dpi, nearly all inner ret- nuclei and TUNEL+ debris within the inner retina ap- inal neurons have been destroyed, Müller glia are react- pear by 24 hpi, concomitant with immune cell accumu- ive, and some Müller glia have re-entered the cell cycle lation and activation as indicated by morphological ([9, 10, 14] and Fig. 1D). Many DAPI+ nuclei are present changes of L-plastin+ cells from ramified to ameboid cell in regions of neuronal death and degeneration of the shape (Fig. 2A, C). Immune cells in retinas from saline inner retina (Fig. 1B, D), and essentially all of these DAPI injected controls remained ramified, and in their typical + nuclei at 72 hpi correspond to immune cells and re- locations, and TUNEL signal was not detected (Fig. 2A, B). active Müller glia (Fig. 1D). By 48 hpi, immune cells with ameboid morphology Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 7 of 20 continue to accumulate in locations corresponding to 12 hpi following ouabain injection, ameboid immune ouabain-induced neuronal death and degeneration (INL, cells can be seen in central retina in a gradient originat- GCL, Fig. 2D). This accumulation correlates with a de- ing from the region of the optic nerve head (Fig. 3B, D, crease in TUNEL+ nuclei and debris from 24 to 48 hpi optic nerve head indicated) and vitreal to the ganglion (Fig. 2C, D), indicating that immune cells have cleared deb- cell layer in peripheral retina (arrows, Fig. 3B, C). We did ris resulting from neuronal death. Consistent with this, we not detect substantial PCNA signal in immune cells at observed immune cells engulfing TUNEL+ nuclei (Fig. 2E) this time point (Fig. 3C, D), indicating that the resident as well as TUNEL+ puncta within L-plastin+ immune cell microglia did not immediately enter the cell cycle. Cells bodies (Fig. 2F). However, only a small fraction of L- of the macrophage lineage (including microglia) in zeb- plastin+ cells contained TUNEL+, DAPI+, co-labeled rafish have been identified using the transgenic expres- nuclei (8.4 ± 5.7% at 24 hpi and 1.9 ± 1.9% at 48 hpi), indi- sion of mpeg1-driven transgenes [35], while neutrophils cating that ouabain only induced minor levels of immune are identified by expression of mpx (also called mpo) cell death. Further, some of the DAPI and TUNEL co- [41]. Nearly all of the L-plastin+ cells co-labeled with labeling within immune cells could potentially be ex- mpeg1:mCherry, indicating that the majority of these plained as phagocytosed TUNEL+ nuclei from dead or leukocytes are macrophages (Fig. 3E, F). Few mpx + neu- dying neurons. Consistent with this, in some cryosections, trophils were detected within the peripheral leukocyte we observed multiple nuclei, including TUNEL-nuclei, population vitreal to the ganglion cell layer at this time surrounded by L-plastin+ cytoplasm (Fig. 2G). point; however, mpx + neutrophils were not detected in other retinal regions, including regions surrounding the Origins and distributions of immune cells in damaged optic nerve head (Additional file 1: Figure S3). The pres- retina ence of few neutrophils supports that leukocytes in the To better understand the origins of the accumulation of blood have access to retinal tissue following ouabain in- immune cells in ouabain-lesioned retinas, we imaged jection. Previous reports indicate that apoptotic cells la- retinal cryosections at 12 hpi using the marker L-plastin beled by TUNEL are present as early as 3 hpi ouabain and the S-phase marker PCNA (Fig. 3). Saline injection injection, although the number of apoptotic cells peaks did not result in immune cell accumulation or PCNA at approximately 24 hpi [9]. The presence of ameboid im- expression (Fig. 3A and Additional file 1: Figure S2). At mune cells apparently infiltrating the retina, and absence Fig. 3 Evidence of early infiltration of immune cells to the retina at 12 h post-ouabain injection. Images of retinal cryosections at 12 h post-saline (A), or ouabain (B–F) injection (12 hpi). A, B Show stitched images of entire cryosections stained for L-plastin (magenta) and DAPI (blue) obtained at × 20 magnification. Images in C and E show peripheral regions of retina. D, f Show central regions of retina adjacent to the optic nerve head (onh). Following ouabain injection, ameboid immune cells are seen in a gradient originating from the region of the optic nerve head (B, D, F), as well as vitreal to the ganglion cell layer in peripheral regions (C (arrows) and E). Cryosections were also stained for PCNA (green, C and D). PCNA expression was only rarely detected at this time point. E In peripheral retina, few L-plastin+ cells (green) co-label with mpeg1:mCherry transgene expression (red, co-labeled cells yellow). F In central retina, most L-plastin+ cells (green) co-label with mpeg1:mCherry transgene expression (red, co-labeled cells yellow). A–D were obtained at × 20 magnification, E and F were obtained at × 40 magnification. Scale bar in (A)=100 μm (applies to A and B); scale bar in C (applies to C and D)=20 μm; scale bars in E and F =20 μm Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 8 of 20 of PCNA expression, suggests an early infiltration of im- continued accumulation of immune cells through 72 hpi mune cells following ouabain injection, prior to the peak (Fig. 5A, B). These cell cycle markers indicate that infil- of retinal cell death. trating immune cells prepare for division upon their ar- We next imaged cryosections at 24, 48, and 72 hpi rival to damaged retinal tissue, and resident microglia using the markers L-plastin and PCNA (Fig. 4), and also respond to retinal damage by entering the cell cycle. quantified numbers of cells stained by these markers To our knowledge, only one study has used a histo- (Fig. 5). Injection of saline did not result in accumulation logical method (methylene blue/azure II staining of plas- of immune cells or significant expression of PCNA within tic sections) to document ouabain-induced retinal theretinathrough 72 hpi (Figs. 4A–C, G–I,and 5A–C). By degeneration in zebrafish [9], and this method does not 24 h post-ouabain injection, immune cells appear abun- readily reveal immune cell characteristics. Therefore, to dant in the region vitreal to the degenerating ganglion cell better observe immune cell accumulation concomitant layer in both peripheral and central retina (Fig. 4D), sug- with progression of the ouabain-induced inner retinal le- gesting that extra-retinal immune cells continue to invade sion and to simultaneously visualize features of respond- the retina in response to neuronal degeneration. Immune ing immune cells, H&E staining of cryosections was cells with ameboid shape begin to accumulate in regions used (Fig. 6). Saline-injected controls did not show of inner retina corresponding to the ouabain-induced le- changes in retinal structure or immune cell accumula- sion at 24 hpi and continue to accumulate through 72 hpi tion (Fig. 6A–C). However, in damaged retinas, immune (Figs. 1D, 4D–F and J–L, 5A, B). Further, evidence of im- cells and the progression of the inner retinal lesion were mune cell infiltration to the damaged retina remains at 48– readily visualized simultaneously using H&E stains 72 hpi, but this later migration appears to originate from (Fig. 6D–I). At 24 hpi, pyknotic nuclei corresponding regions and structures apical to the neural retina, which to dying neurons are visible in the ganglion cell and could include the RPE or sclera and associated blood ves- inner nuclear layer of the retina (Fig. 6D), while im- sels (Fig. 4K, L). We consider this to be the case, because mune cells are visible in the vitreal space as well as immune cells that appear directionally oriented to migrate within the retinal tissue (Fig. 6D, G). By 48 hpi, im- into the retina (i.e., radially oriented) are more frequently mune cells are mainly present within regions of degen- present in the outer retina and RPE (Figs. 1D, 4K, L). erating retinal tissue, following the progression of the A significant fraction of immune cells responding to lesion (Fig. 6E, H). Consistent with this, the presence the ouabain-induced lesion at 24 and 48 hpi stain posi- of extracellular matrix, debris, and pyknotic nuclei de- tive for the S-phase marker PCNA (Fig. 4, arrows, and creases from 24 to 72 hpi (Fig. 6D–F). Immune cells Fig. 5C). PCNA+ immune cells are located in the vitreal continue to accumulate in regions of retinal degener- space at 24 hpi ouabain injection (Fig. 4D), apical to the ationfrom48to72hpi (Fig. 6E, F), which is likely due retina at 48 hpi (Fig. 4E, K), and within retinal tissues at to accumulation of dividing immune cells as well as both 24 and 48 hpi (Fig. 4D, E, J, K), suggesting that by continued immune cell migration from regions apical 24 hpi, both resident and infiltrating immune cells have to the lesion (denoted by white arrows in Fig. 6E, F), entered the cell cycle in response to the lesion. Import- indicating that immune cell infiltration into the inner antly, the number of immune cells located within retinal retina continues at least through 72 hpi. Through tissues more than doubles between 24 and 48 hpi 72 hpi, evidence of intact vasculature structures at the (Fig. 5A, B) while fewer than 50% of these cells are vitreal face of ouabain injected eyes remains (Fig. 6H, PCNA+ at 24 hpi (Fig. 5C). This rate of PCNA ex- inset, and I, black arrows), suggesting that blood ves- pression is not sufficient to account for all of the accumu- sels are not immediately and/or fully destroyed by lated immune cells, supporting the interpretation that at 2 μM ouabain. These vessels could therefore possibly least a portion of them infiltrated the retina. Numbers of provide infiltrating immune cells access to damaged PCNA+ immune cells decrease at 72 hpi and instead retinal tissue. PCNA positive signal is present in non-immune cells (Fig. 4F, L asterisks, Fig. 5). Identity of S-phase cells at 72 h post-ouabain lesion Using the M-phase specific marker PH3, we did not Since Müller glia respond to retinal injury in zebrafish identify significant numbers of PH3+ cells at any of by re-entering the cell cycle [8, 11, 12, 42], and in order these time points (Additional file 1: Figure S4 and not to determine the identity of non-leukocyte PCNA+ cells shown). The absence of PH3+ cells could possibly be at 72 h post-ouabain injection, we co-stained immune due to rapid mitosis, making detection of mitotic cells in cells in retinal cryosections obtained at this time point fixed retinal tissue difficult. An alternative explanation, with Glutamine Synthetase (GS), which stains Müller that immune cells die after S phase and prior to M glia (Fig. 7). At 72 hpi ouabain injection, nearly all phase, is not supported by the observations of very few PCNA+ nuclei in central and central-peripheral retina L-plastin+, TUNEL+ cells (discussed above) and of the are associated with Glutamine Synthetase+ (GS, Fig. 7B’ Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 9 of 20 Fig. 4 (See legend on next page.) Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 10 of 20 (See figure on previous page.) Fig. 4 Distribution and proliferation markers in immune cells during the response to retinal damage. Images show staining of L-plastin (magenta), PCNA (green), and DAPI (blue) in cryosections of retinas injected with saline (A–C and G–I) or ouabain (E–F and J–L) at 24, 48, and 72 hpi. Images are from regions of peripheral retina (rows 1 and 2, A–F) or central retina (rows 3 and 4, G–L). PCNA+, L-plastin+ cells are indicated with white arrows, while PCNA+ non-immune cells are indicated with asterisks. In images of peripheral retina, regions corresponding to the ciliary marginal zone (CMZ) are within dotted ellipses. A–C and G–I Saline injection does not result in significant changes in L-plastin+ cell accumulation, location, or morphology and does not result in significant PCNA+ signal in L-plastin+ cells (A–C and G–I). D–F and J–L In ouabain-lesioned retinas, significant numbers of ameboid-shaped L-plastin+ cells are located in the region vitreal to the ganglion cell layer at 24 hpi and many are PCNA+ (D, J, arrows). At 48 hpi ouabain, L-plastin+ cells remain ameboid in shape, appear to accumulate at the region of the ouabain-induced lesion, and many are PCNA+ (E and K, arrows). By 72 hpi, PCNA+ signal is mainly localized to non-immune cells (asterisks*), although ameboid-shaped L-plastin+ cells remain in regions corresponding to the lesion (F and L). K, L Arrowheads indicate immune cells invading from regions posterior to the retina. G–L Regions denoted with dotted white lines indicate regions (corresponding to the lesion) used for quantification of L-plastin+ cells shown in Fig. 4.Scale bar in A (applies to all images) = 20 μm. ONL = outer nuclear layer, INL = inner nuclear layer, GCL = ganglion cell layer and C’, yellow arrows) or L-plastin+ cytoplasm (Fig. 7A’ reporter (Fig. 8D–F), supporting significant immune cell and C’, asterisk(*)), although we did occasionally detect infiltration to the damaged retina. Despite the absence PCNA+ nuclei in basal retina that were not associated of mpeg1 reporter expression, morphological features of with GS or L-plastin signal (pink arrow, Fig. 7C’). This the L-plastin+ immune cells in damaged retinal tissue indicates that by 72 hpi, Mülller glia have re-entered the are consistent with a macrophage identity, including ir- cell cycle in order to replace neurons lost to the ouabain regularly cell-shaped cytoplasm and the presence of nu- lesion. merous vacuoles (Fig. 8D–F, arrowheads). Using H&E staining, we confirmed that these immune cells are mac- Identities of immune cells in damaged retinas rophages (Fig. 8G–I and G’–I’). Histological features in- To determine the identities of the immune cells initially clude elongated shape, irregular borders, numerous responding to ouabain-induced neuronal cell death, we vacuoles (asterisks, Fig. 8G–I,G’–I’), and granular cyto- used cell lineage-specific markers of phagocytic immune plasmic inclusions with intensity and color similar to cells in cryosections. Mpeg1 expressing macrophages/ that of extracellular debris, evidence of phagocytosis, microglia are visible at all time points (Fig. 8D–F), but and lysosomal degradation (Fig. 8G–I and G’–I’,[43]). only a fraction of L-plastin+ immune cells present at any H&E staining also revealed numerous vacuoles within of these time points also express the mpeg1 transgenic macrophages that had phagocytosed pyknotic nuclei Fig. 5 Quantification of immune cells in cryosections of damaged retina. A Quantification of L-plastin+ immune cells per 1000 μm² of inner retina (5-μm-thick sections) from saline or ouabain injected eyes at 24, 48, and 72 hpi, representative images in Fig. 3G–L). For details, see the “Quantification of Immune Cells in Retinal Cryosections (24–72 h post-injection)” section. Regions depicted by dotted white lines in Fig. 3G–L are representative regions used for quantification. DAPI+ nuclei surrounded by L-plastin+ signal were counted within the region of interest and normalized per 1000 μm². p values comparing densities between 24 and 48 hpi and 48–72 hpi ouabain injection are shown on the graph; asterisks indicate the following − 12 − 12 p values comparing ouabain to saline injected at indicated time points: *p = 0.012, **p =7.84 × 10 , ***p =1.21×10 . B Since thickness of the INL changes due to ouabain-induced lesion, the number of L-plastin+ cells is also reported per 100 μm of curvi-linear distance of retina using a line that follows parallel to the basal border of the ONL. p values comparing densities between 24 and 48 hpi and 48–72 hpi ouabain injection are shown on the graph; asterisks indicate the following p values comparing ouabain injected to saline injected at indicated time − 6 − 9 − 10 points: *p =3.84× 10 ,**p =8.48×10 ,***p =5.5 ×10 . C PCNA+, DAPI+ nuclei surrounded by L-plastin+ signal quantified in retinal cryosections from saline or ouabain injected eyes at 24, 48, and 72 h hpi. Graph shows the fraction of L-plastin+ cell bodies with nuclear co-label of DAPI and PCNA. Only images from central and central-peripheral retina were quantified, and only L-plastin+ cell bodies within − 12 retinal tissue boundaries were counted. Error bars indicate standard deviation. *p = 0.0005, **p =1.01×10 , comparing ouabain injected to saline injected at indicated time points Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 11 of 20 Fig. 6 Progression of retinal degeneration and immune cell response visualized by Hematoxylin & Eosin (H&E). H&E staining of cryosections to visualize progression of the ouabain-induced lesion and immune cell accumulation. A–C Saline-injected controls did not show significant changes in retinal structure; ramified microglia are not readily visualized due to their thin processes with little cytoplasm and integration into retinal layers. D–F By 24 h post-ouabain injection (24 hpi), the inner retina has begun to swell, pyknotic nuclei are present (asterisks), and ameboid immune cells can be detected within the GCL and INL (D, white arrows and G, black arrows). Immune cells are visible in the vitreal space at 24 hpi (black arrows, G), consistent with invasion of extra-retinal immune cells from the retinal vitreal face. Immune cells accumulate in degenerating retinal tissue, which contains many pyknotic nuclei (asterisks), through 48 and 72 hpi ouabain injection (E and F). However, immune cells are no longer observed in the vitreal space at 48 and 72 hpi ouabain (E, F, H, I). Immune cells present in the degenerating retinal tissue are highly phagocytic, demonstrated by cytoplasmic color, regions of space immediately surrounding them, and the progressive disappearance of extracellular matrix and debris over time (D, E, F). Immune cells appear to invade from structures apical to the retina at 48 and 72 hpi ouabain (white arrows, E and F). G–I Evidence of surviving blood vessels at all time points following ouabain injection: at 24 hpi, blood vessel structures at the anterior of the eye adjacent to the peripheral retina are present (dashed box, G). Sections of intact retinal blood vessels at the vitreal face of the retina are present at 48 and 72 hpi ouabain injection (dashed box and inset, H, and black arrows, I). Scale bar in A (applies to A–F)=20 μm. Scale bar in G (applies to G–I)=40 μm (arrow, Fig. 8H’). Very few mpx + neutrophils were de- revealed very few immune cells bearing features of neu- tected in cryosections at 24, 48, or 72 h post-ouabain in- trophils (banded or segmented nuclei and pink, granular jection (Additional file 1: Figure S5), and H&E staining cytoplasm [43]), indicating that the few neutrophils Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 12 of 20 Fig. 7 PCNA expression in immune cells and Müller glia at 72 h post-ouabain injection. Representative images of retinal cryosections at 72 h post-ouabain injection (72 hpi) stained for L-plastin to mark immune cells (green), glutamine synthetase to mark Müller glia (GS, red), PCNA (white), and DAPI (blue). A L-plastin and DAPI. B GS and DAPI. C L-plastin, GS, and DAPI. A’ L-plastin, PCNA, and DAPI. B’ GS, PCNA, and DAPI. C’ Four color merge of all stains. Nearly all PCNA+ nuclei can be attributed to immune cells or Müller glia at this time point. Asterisk (*) in A’ and C’ denotes L-plastin+, PCNA+ cell. Yellow arrows in B’ and C’ are provided to emphasize selected GS+ and PCNA+ cells. Occasionally, we observed PCNA+ nuclei at the vitreal surface at 3 dpi that could not be attributed to L-plastin+ or GS+ cells (C’, pink arrow). Scale bar in A (applies to all images) = 20 μm present in peripheral regions at 12 hpi did not signifi- Microglia morphologies and distributions in histologically cantly increase in number, and that neutrophils do not regenerated retinas comprise a significant portion of responding leuko- We next investigated microglia morphology and retinal cytes. However, it is possible that the few neutrophils distribution at 14–21 days following ouabain injection. present could represent some of the TUNEL+ immune Although there has yet to be a demonstration of func- cells at 24 or 48 hpi (Fig. 2, discussed above), as neu- tional restoration at these times, we refer to 14 and trophils undergo apoptosis shortly after completing 21 dpi retinal tissues as “histologically regenerated,” due their phagocytic functions [43, 44]. We conclude that to the presence of new neurons and the emergence of the immune cells infiltrating the retina in response to plexiform layers [9, 10]. These histologically regenerated ouabain-induced neuron death may be derived from retinas, however, do show lamination defects. Typical microglia and/or macrophages in nearby tissues and lamination defects are the presence of neuronal nuclei monocytes from circulation that rapidly differentiate within plexiform layers, particularly the inner plexiform to macrophages to perform crucial phagocytic func- layer; such defects have been referred to as “laminar fu- tions. The absence of detectable mpeg1:mCherry ex- sions” [10, 45, 46], some of which resolve over further pression could be explained by delays in expression of recovery time [10]. We reasoned that microglia are ex- the transgenic reporter, or alternatively, that mpeg1 cellent candidates for involvement in resolution of histo- expression is downregulated in this particular activa- logical errors and clearing non-functional or apoptotic tion state. neurons that may arise during retinal regeneration, and Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 13 of 20 Fig. 8 (See legend on next page.) Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 14 of 20 (See figure on previous page.) Fig. 8 Immune cells responding to ouabain-induced retinal degeneration identify as macrophages. A–F. Retinal cryosections from transgenic mpeg1:mCherry fish following intravitreal injection of saline (A–C) or ouabain (D–F) were immunolabeled with L-plastin (green) and counterstained with DAPI (blue). A–C. In saline-injected controls, all L-plastin+ immune cells (green) co-label with the macrophage marker mpeg1:mCherry (red); co-label is yellow (A–C), indicating ramified microglia. D–F. Following ouabain injection, ameboid, L-plastin+ immune cells accumulate (green, D–F), but only a subset of these also express the mpeg1:mCherry reporter (yellow, D–F). Arrowheads indicate vacuoles in selected cells. G–I High magnification images of retinal cryosections from ouabain damaged retinas stained with H&E reveal classical characteristics of phagocytic macrophages in accumulating immune cells (G–I and G’–I’), indicating that these accumulating immune cells are in fact macrophages. These features include irregular shapes of cell bodies and nuclei, cytoplasm similar in color to the environment, space immediately surrounding the cell borders, and the presence of vacuoles (asterisks). H’ Pyknotic nucleus within the cytoplasm of a macrophage (arrow), indicating phagocytosis of apoptotic cells. The red signal in the ONL in images A–F is due to autofluorescence from photoreceptors. Scale bar in A (applies to A–F)= 20 μm. Scale bar in G (applies to G–I and G’–I”)=10 μm so selected these time points for further analysis. By each parameter begin to shift towards those in control 14 dpi, all immune cells in the retina express the mpeg1: retinas (Fig. 10), indicating dynamic changes are occurring GFP reporter (Fig. 9B–B”). Densities of microglia in his- within the microglial population, possibly a trend towards tologically regenerated and saline-injected retinas were ramification. Collectively, the morphological features of calculated from whole, flat mounted retinas considering microglia in newly histologically regenerated retinas indi- all retinal layers (described in the “Methods” section). cate that rather then returning to a ramified state, micro- Densities of microglia in saline and histologically regen- glia remain dynamic as histological features of regenerated erated retinas were calculated from image stacks ob- retinas continue to improve. tained from whole, flattened retinas (representative images in Fig. 9A and B). Densities of microglia were Discussion slightly elevated in histologically regenerated retinas In this study, we examined the immune response to compared to saline controls, but the differences were ouabain-induced degeneration of inner retinal neurons not statistically significant (Fig. 9E). However, microglia in zebrafish, as well as morphology of microglia in histo- have not returned to their normal retina distribution logically regenerated retinas. Key findings include (1) patterns and instead remain localized to regions of histo- immune cells accumulate rapidly and in large numbers logically regenerated retina, particularly the ganglion cell following substantial retinal damage; (2) in addition to layer (Fig. 9D, F), suggesting that they are indeed per- local microglia, extra-retinally derived macrophages ap- forming functional activities in regions containing new pear to comprise a significant portion of responding im- neurons. mune cells; and (3) Microglia in histologically regenerated Further examination of microglia morphologies within retinas retain morphological features of activation. Col- the ganglion cell layer at 14 and 21 dpi reveal that lectively, these results indicate that microglia and the im- microglia retain morphological features of activation ra- mune system are dynamic during retinal degeneration and ther than ramification (Fig. 10), even though retinal layers regeneration and could play important roles in a success- have been histologically regenerated by 14 dpi [9, 10, 46]. ful regenerative response in the retina. In control retinas (saline injected), microglia display long, complex processes and appear ramified (Fig. 10A, B), as in- Response of immune cells to damaged retinal tissue dicated by long perimeters and increased Feret diameter We find that immune cells accumulate rapidly in re- (Fig. 10J and K). In histologically regenerated retinas, sponse to ouabain-induced retinal lesion. We conclude microglia remain ameboid in shape, especially at 14 dpi that these immune cells are composed of responding ouabain (Fig. 10C, D compared to Fig. 10A, B), as indicated resident microglia as well as extra-retinally derived mac- by more circular morphology with reduced perimeter and rophages, providing a clear identity to the abundant nuclei reduced Feret diameter (Fig. 10I, K). Some microglia are of unidentified cell type in the inner retina following oua- associated with multiple nuclei (Fig. 10C asterisk, and E–G), bain damage documented in previous studies [9, 10, 33]. possibly due to phagocytosis of improperly localized, non- Several publications to date indicate that the innate im- functional, or apoptotic neurons. In addition, chains of mune response, particularly macrophages, are crucial to re- rod-shaped microglia were observed in regenerated retinas generation of a variety of tissues in zebrafish [24–27, 48]. (Fig. 10D). Rod-shaped microglia have been observed fol- However, it remains unclear to what extent and in which lowing optic nerve transection, which may phagocytose context(s) recruited/peripheral immune cells are involved. retinal ganglion cell debris and appear to be highly prolifer- In contexts of peripheral nervous system degeneration and ative [47]. In histologically regenerated retinas, between 14 regeneration, two studies found a requirement for recruited and 21 dpi, distributions of microglia measurements for innate immune cells [24, 25]. In the zebrafish retina, Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 15 of 20 Fig. 9 Microglia distribute to regions containing histologically regenerated retinal neurons. A–B. Images (A and B) show z projections from saline injected (A–A”) or regenerated (B–B”) mpeg1:GFP (green) flat-mounted whole retinas labeled with L-plastin (magenta). By 14 dpi, all immune cells in the histologically regenerated retinas express the mpeg1:GFP reporter (B–B”), and the marker L-plastin can be used to label these mpeg1:GFP+ immune cells. C–D Resliced images from whole flat-mounted retinas labeled with L-plastin (magenta) are shown to demonstrate distribution of microglia in control (C) and histologically regenerated retinas at 14 dpi ouabain (D). E Densities of microglia in whole, flat mounted saline injected (n = 3) and histologically regenerated retinas (14 (n = 4) or 21 dpi (n = 3) ouabain). Densities were calculated by counting microglia in individual z stacks obtained from flat mounted retinas, which included retinal layers from the ganglion cell layer to the outer nuclear layer (representing ap- proximately 50–80 μm of depth; Fig. 7A, B in their entirety show representative regions that were quantified). Counts were normalized to 1000 μm² (area) rather than volume because regenerated retinas are thinner than controls. This area represents x and y directions parallel to the layers of the flattened retina. F Fraction of microglia in the ganglion cell layer of saline injected (n = 3) and histologically regenerated retinas (14 (n = 4) or 21 dpi (n = 3) ouabain). **p = 0.048 (two-tailed Student’s t test comparing regenerated to saline injected at indicated time point), effect size 1.56. Error bars indicate standard deviation. Scale bar in A” (applies to A–A” and B–B”)= 20 μm. Vertical scale bar in C and D =20 μm White et al. [26] found that resident microglia were mpeg1:GFP+ cells were sequenced. Further, the selective unique in their response to selective rod cell death; re- neuronal ablation used in the Oosterhof et al. [49]and cruited phagocytes did not enter the retina. However, White et al. [26] studies initiated neuronal cell death these studies were performed in a developmental context through cell-specific drug sensitivity, rather than a more (larvae), making it difficult to determine if findings may be wide-spread and tissue-disrupting chemical insult. In our tied to a developmental microenvironment or could be system, which results in neuron death throughout the en- shared with degeneration and regeneration in adult ani- tire inner retina of adult fish due to chemical insult, re- mals. Using adult zebrafish following a cell-selective abla- cruited/invading macrophages comprise a significant tion of brain neurons, Oosterhof et al. [49] concluded that portion of responding immune cells. Likewise, following peripheral macrophages did not infiltrate the injured cryolesion of the peripheral adult tench retina, infiltrating brain. This conclusion, however, was based on sequencing immune cells were documented [50]. Likely, the executed results which yielded mainly markers of microglia rather cell death pathway(s) and resulting molecules released than macrophages (based on mouse orthologues), but only into the tissue environment by dying neurons, as well as Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 16 of 20 Fig. 10 Morphological features of microglia in ganglion cell layer of histologically regenerated retinas. A–D Images show L-plastin+ microglia (magenta) in the ganglion cell layer (GCL) of saline injected (A and B) or histologically regenerated retinas at 14 or 21 dpi ouabain (C and D). Microglia in saline injected retinas appear ramified, displaying long complex processes with multiple tips (A and B). Microglia in regenerated retinas at 14 dpi appear rounded (ameboid) in shape and have few processes (C). In regenerated retinas at 21 dpi, microglia remain more ameboid than those seen in saline injected controls; however, some microglia begin to display more cellular processes (D). Microglia in regenerated retinas are occasionally associated with multiple nuclei (asterisk, C, resliced projections E–G). Chains of rod-shaped microglia are observed in regenerated retinas (D, arrowheads). E–G Individual channels in resliced projections of the microglia denoted by a asterisk in (C). Arrows indicate DAPI+ nuclei. H–K Violin plots show distributions of area (H), circularity (I), perimeter (J), and Feret diameter (K) of individually traced microglia located in the GCL at 14 and 21 dpi following injection of saline (red plots) or ouabain (regenerated, teal plots). Violin shapes show distribution of all measurements; circles within violin plots represent individual microglia. Area is reported in square microns; circularity as a value from 0 to 1.0 (1.0 indicating perfect circle); perimeter and Feret’s diameter are reported in microns. More information on morphological parameters/measurements is located in the “Methods” section. Indicated p values (two-tailed Student’s t test) in black compare measurements from regenerated to saline injected at the same time point; p values in purple (bottom of graphs) compare measurements from 21 to 14 dpi ouabain. Scale bar in A (applies to A–D)=20 μm; scale bar in E (applies to E–G)= 5 μm the extent of tissue degeneration, are important factors in PCNA+ immune cells detected at 24 h post-ouabain in- determining if peripheral immune cells are recruited to jection present at locations consistent with ramified ret- the retina. inal microglia (Fig. 4J), as well as those residing at the We also find that a significant portion of immune cells vitreal face of the ganglion cell layer (Fig. 4D, J), suggest enter the cell cycle upon ouabain-induced retinal injury, that both infiltrating immune cells and resident micro- similar to findings following acute injury in the zebrafish glia enter the cell cycle in response to ouabain-induced brain [49] and selective cell death in the retina [26]. retinal lesion. As cell division in responding macrophages Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 17 of 20 weans by approximately 3 days post-injury, PCNA signal genes were among the top hits yielded by microarray is instead detected in non-immune cells, which we have analysis in a context of ongoing rod photoreceptor death identified to be almost entirely Müller glia (Fig. 7)that are and regeneration [60]. Further work is needed to identify likely entering the cell cycle to replace the damaged neu- key inflammatory signals that are detrimental to regener- rons. Collectively, this suggests that a relatively large and ation versus those that are supportive. Likely, spatio- robust response of phagocytic macrophages and microglia temporal control of both pro- and anti-inflammatory is required to clear dead cell bodies and debris before a re- mediators is important, and these signals could differ generative response begins. depending on the particular disease or damage system. Importantly, our results support that the immune envir- Infiltration of immune cells to degenerating retina onment in mammals could potentially be modulated to suggests systemic inflammatory signals promote retinal regeneration. The act of phagocytosis by macrophages does not neces- sarily initiate an inflammatory response [51]. However, because our data supports substantial levels of immune Morphologically, microglia in histologically regenerated cell infiltration into the damaged retina following a oua- retinas appear functionally active bain lesion, this indicates that at least some systemic in- To date, most studies of inflammation and/or microglia flammatory signals are received by extra-retinal immune in retinal regeneration have focused on effects on Müller cells, even before the peak of retinal cell death. Acute in- glia proliferation [22, 26]. Further, recent studies that de- flammation appears to be a crucial signal initiating CNS pleted microglia/macrophages prior to or during the ini- regeneration and specific inflammatory mediators alone tial immune response to neuronal degeneration altered may be able to trigger neuronal progenitor proliferation the dynamics and timing of regeneration [24, 26]. How- in zebrafish [27]. In our system of retinal damage in zeb- ever, regeneration of neurons, although delayed, was ul- rafish, as well as others [26, 52], responding microglia/ timately recovered. We therefore characterized microglia macrophages localize almost exclusively to regions of ret- in histologically regenerated retinas, after the peak of inal degeneration, suggesting that inflammatory signals Müller glia proliferation [9], at the time that plexiform are spatially controlled. In particular, the localization of layers become evident, but prior to documented restor- responding macrophages in our system moves in a wave ation of visual function [10]. We found that microglia from the central region originating from the optic nerve localize to regions of regenerated neurons, rather than head and the vitreal surface of the retina at the periphery, returning to their typical distributions. In addition, to the inner nuclear layer, which correlates temporally microglia retain morphology that indicates activation for with the progression of the ouabain-induced retinal lesion. several weeks following the initial lesion, suggesting that Further, the switch in proliferative activity from microglia are active and performing functional roles. It responding immune cells to Müller glia (discussed remains to be determined when, or if, microglia return above) suggests that inflammation is indeed acute and to their original distribution patterns and ramified state, temporally controlled. and whether this coincides with restoration of retinal Hallmarks of neurodegenerative disease include function. Although visually mediated behaviors and chronic inflammation and infiltration of immune cells; ERGs are restored by 60–80 dpi [10, 33], it is not known however, it is now recognized that these features may if ERGs are restored prior to this time point. It has been have both pathological and beneficial effects [53–56]. In documented that histological organization of the regen- the context of damaged zebrafish retina, however, in- erated retina improves over time [9, 10], and it is pos- flammation appears to be beneficial rather than detri- sible that this may correlate temporally with recovery of mental, which is in contrast to what has been described retinal function. Microglia/macrophages are excellent in mammals. It is possible that zebrafish have the ability candidates to facilitate improvement of histological to rapidly switch any harmful inflammatory signals to organization as they are mobile, highly responsive, and those that are supportive of regeneration. Alternatively, highly phagocytic making them readily able to locate the molecular aspects of inflammation in zebrafish may and eliminate apoptotic or non-functional neurons, ex- differ from that of mammals [57]. Interestingly, aspects cess neurons shown to be generated in zebrafish retinal of reactive gliosis still exist during retinal regenerative regeneration [10, 61], and/or prune regenerated synaptic responses in zebrafish [58], suggesting that aspects of processes that arise during retinal regeneration. There- harmful inflammation are not completely absent, but fore, functional roles of microglia/macrophages during supportive signals likely dominate. Consistent with this retinal regeneration in zebrafish should be explored. idea, both pro- and anti-inflammatory cytokines have These functions may include known roles that are essen- been found to be important to retinal regeneration in tial to central nervous system development and normal zebrafish following light damage [31, 59] and immune processes [62–69], yet it is also possible that currently Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 18 of 20 unknown or novel functional roles may be involved and of Health under Grant #P20GM103408 (DMM), start-up funds from the University of Idaho (DMM), and NIH R21EY026814 (DLS). NIH Grant S10 are yet to be discovered. OD0108044 (DLS) funded purchase of the Nikon Andor spinning disk confocal microscope and camera. We are grateful to Ann Norton (Director, University of Idaho IBEST Optical Imaging Core), Dr. Michael Redd (University Conclusions of Utah) for the L-plastin antibody, the Zebrafish International Resource Center Interactions of the immune and central nervous system (ZIRC) for transgenic zebrafish lines and antibodies, Dr. Onesmo Balemba have seen a surge of interest, supporting the idea of im- (University of Idaho) for the helpful discussion of H&E, and members of the Stenkamp laboratory for the review of figures. mune system influence on regenerative potential [70]. With goals to regenerate damaged retinal tissue in Funding humans, further work in investigating neuro-immune in- Faculty Seed Grant from the University of Idaho (DMM). Start-up Funds, University of Idaho (DMM). teractions during successful retinal regeneration, which Institutional Development Award (IDeA) from the National Institute of occurs in a variety of vertebrate organisms including General Medical Sciences of the National Institutes of Health under Grant teleost fish, is crucial. The present work provides a #P20GM103408 (DMM). NIH R21EY026814 (DLS). framework for future studies and confirms that the zeb- NIH S10 OD0108044 (DLS). rafish retina is an excellent model to study immune- The funders had no role in the study design, data collection/analysis, or neuron interactions during neuronal regeneration. This writing of the manuscript. will allow us to uncover key inflammatory signals and Availability of data and materials the function of the innate immune system in the initi- The datasets used and/or analyzed during the current study are available ation and successful execution of retinal regeneration. from the corresponding author upon reasonable request. Further, this will allow us to probe the functional roles Authors’ contributions of microglia in the retina, both in normal and regenera- DMM designed and performed experiments, collected/analyzed the data, tive contexts. and wrote the manuscript. AGL performed the experiments and collected/ analyzed the data. DLS designed the experiments, analyzed the data, and wrote the manuscript. All authors read and approved the final manuscript. Additional file Ethics approval Additional file 1: Figure S1.A–A”. To assess co-label of mpeg1:GFP and All procedures using zebrafish were performed in compliance with protocols mpeg1:mCherry transgenes in retina, double transgenic embryos from a approved by the University of Idaho Animal Care and Use Committee cross of gl22 Tg x gl23 Tg fish at 3 dpf were fixed, washed, and eyes re- (IACUC). No human subjects were used in this work. moved then mounted for imaging. A z series (5 μm step size) was ob- tained. Images show selected z-projections from whole eyes; white lines Competing interests indicate the eye boundary. A. mpeg1:GFP signal. A’ mpeg1:mCherry sig- The authors declare that they have no competing interests. nal. A” Merge to show colabel. Scale bar in A” =20 μm. Whole retinas from adult mpeg1:GFP and mpeg1:mCherry fish were stained for L- Publisher’sNote plastin. Images show expression of individual transgenes with L-plastin Springer Nature remains neutral with regard to jurisdictional claims in (magenta, B’, and C’). Essentially all transgene signal coincides with L- published maps and institutional affiliations. plastin (B” and C”). Scale bars in B” and C= 100 μm. Figure S2. Retinal cryosections corresponding to peripheral (A) or central regions (B), adja- Received: 7 November 2017 Accepted: 30 April 2018 cent to the optic nerve head (onh, denoted by **) at 12 h post-injection (12 hpi) saline. Cryosections were stained for PCNA (green), L-plastin (ma- genta), and DAPI (blue). Scale bar in B = 20 μm, applies to both images. References Figure S3. Retinal cryosections corresponding to peripheral (A) or central 1. Bringmann A, Iandiev I, Pannicke T, Wurm A, Hollborn M, Wiedemann P, regions (B, adjacent to onh, denoted by **) at 12 hpi ouabain from Osborne NN, Reichenbach A. Cellular signaling and factors involved in mpeg1:mCherry transgenic fish stained for mpx (green) and DAPI (blue). Müller cell gliosis: neuroprotective and detrimental effects. Prog Retin Eye Scale bar in B = 20 μm, applies to both images. Figure S4. Images show Res. 2009;28:423–51. retinal cryosections from mpeg1:mCherry fish following intravitreal saline 2. Karlstetter M, Scholz R, Rutar M, Wong WT, Provis JM, Langmann T. Retinal (A) or ouabain (B) injection at 72 hpi. Cryosections were labeled with microglia: just bystander or target for therapy? Prog Retin Eye Res. 2015;45: anti-phosphorylatedhistone 3 (PH3, green) and DAPI (blue). Arrows indi- 30–57. cate PH3+ nuclei. Red signal in the outer retina is autofluorescence from 3. Wan J, Goldman D. Retina regeneration in zebrafish. Curr Opin Genet Dev. photoreceptors. Scale bar in A = 20 μm, applies to A and B. Figure S5. 2016;40:41–7. Images show retinal cryosections following intravitreal injection of oua- 4. Hamon A, Roger JE, Yang X-J, Perron M. Müller glial cell-dependent bain at 24, 48, and 72 hpi. Cryosections were stained for mpx (red) and regeneration of the neural retina: an overview across vertebrate model DAPI (blue). Scale bar = 20 μm, applies to all images. (PDF 1222 kb) systems. Dev Dyn. 2016;245:727–38. 5. Stenkamp DL. Neurogenesis in the fish retina. Int Rev Cytol. 2007;259:173–224. Abbreviations 6. Lenkowski JR, Raymond PA. Müller glia: stem cells for generation and CMZ: Ciliary marginal zone; DAPI: 4′,6-diamidino-2-phenylindole; dpi: Days regeneration of retinal neurons in teleost fish. Prog Retin Eye Res. 2014;40: post-injection; GCL: Ganglion cell layer; H&E: Hematoxylin and Eosin; 94–123. hpi: Hours post-injection; INL: Inner nuclear layer; MG: Müller glia; ONL: Outer 7. Goldman D. Müller glial cell reprogramming and retina regeneration. Nat nuclear layer; PCNA: Proliferating cell nuclear antigen; PKC: Protein kinase c; Rev Neurosci. 2014;15:431–42. TUNEL: Terminal deoxynucleotidyl transferase dUTP nick end labeling 8. Gorsuch RA, Hyde DR. Regulation of Muller glial dependent neuronal regeneration in the damaged adult zebrafish retina. Exp Eye Res. 2014; Acknowledgements 123(c):131–40. Funding for this work was provided by a Faculty Seed Grant from the 9. Fimbel SM, Montgomery JE, Burket CT, Hyde DR. Regeneration of inner University of Idaho (DMM), an Institutional Development Award (IDeA) from retinal neurons after intravitreal injection of ouabain in zebrafish. J Neurosci. the National Institute of General Medical Sciences of the National Institutes 2007;27:1712–24. Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 19 of 20 10. Sherpa T, Lankford T, McGinn TE, Hunter SS, Frey RA, Sun C, Ryan M, 35. Ellett F, Pase L, Hayman JW, Andrianopoulos A, Lieschke GJ. mpeg1 Robison BD, Stenkamp DL. Retinal regeneration is facilitated by the promoter transgenes direct macrophage-lineage expression in zebrafish. presence of surviving neurons. Dev Neurobiol. 2014;74:851–76. Blood. 2011;117:e49–56. 11. Bernardos RL, Barthel LK, Meyers JR, Raymond PA. Late-stage neuronal 36. Raymond PA, Reifler MJ, Rivlin PK. Regeneration of goldfish retina: rod progenitors in the retina are radial Müller glia that function as retinal stem precursors are a likely source of regenerated cells. J Neurobiol. 1988;19:431–63. cells. J Neurosci. 2007;27:7028–40. 37. Herbomel P, Thisse B, Thisse C. Ontogeny and behaviour of early 12. Fausett BV, Goldman D. A role for alpha1 tubulin-expressing Müller glia in macrophages in the zebrafish embryo. Development. 1999;126:3735–45. regeneration of the injured zebrafish retina. J Neurosci. 2006;26:6303–13. 38. Le Guyader D, Redd MJ, Colucci-Guyon E, Murayama E, Kissa K, Briolat V, 13. Thummel R, Kassen SC, Montgomery JE, Enright JM, Hyde DR. Inhibition of Mordelet E, Zapata A, Shinomiya H, Herbomel P. Origins and unconventional Müller glial cell division blocks regeneration of the light-damaged zebrafish behavior of neutrophils in developing zebrafish. Blood. 2008;111:132–41. retina. Dev Neurobiol. 2008;68:392–408. 39. Renninger SL, Gesemann M, Neuhauss SCF. Cone arrestin confers cone vision 14. Nagashima M, Barthel LK, Raymond PA. A self-renewing division of zebrafish of high temporal resolution in zebrafish larvae. Eur J Neurosci. 2011;33:658–67. Müller glial cells generates neuronal progenitors that require N-cadherin to 40. Herbomel P, Thisse B, Thisse C. Zebrafish early macrophages colonize regenerate retinal neurons. Development. 2013;140:4510–21. cephalic mesenchyme and developing brain, retina, and epidermis through 15. Raymond PA, Barthel LK, Bernardos RL, Perkowski JJ. Molecular a M-CSF receptor-dependent invasive process. Dev Biol. 2001;238:274–88. characterization of retinal stem cells and their niches in adult zebrafish. BMC 41. Mathias JR, Dodd ME, Walters KB, Yoo SK, Ranheim EA, Huttenlocher A. Dev Biol. 2006;6:36. Characterization of zebrafish larval inflammatory macrophages. Dev Comp 16. Thummel R, Kassen SC, Enright JM, Nelson CM, Montgomery JE, Hyde DR. Immunol. 2009;33:1212–7. Characterization of Müller glia and neuronal progenitors during adult 42. Lenkowski JR, Raymond PA. Muller glia: stem cells for generation and zebrafish retinal regeneration. Exp Eye Res. 2008;87:433–44. regeneration of retinal neurons in teleost fish. Prog Retin Eye Res. 2014; 17. Bringmann A, Pannicke T, Grosche J, Francke M, Wiedemann P, Skatchkov 40(C):94–123. SN, Osborne NN, Reichenbach A. Müller cells in the healthy and diseased 43. Kumar V, Abbas AK, Aster JC. Robbins Basic Pathology. Ninth edition. retina. Prog Retin Eye Res. 2006;25:397–424. Philadelphia, PA: Elsevier/Saunders; 2012. 18. Jadhav AP, Roesch K, Cepko CL. Development and neurogenic potential of 44. Kolaczkowska E, Kubes P. Neutrophil recruitment and function in health and Müller glial cells in the vertebrate retina. Prog Retin Eye Res. 2009;28:249–62. inflammation. Nature Publishing Group. Natur Rev Immunol. 2013;13:159– 19. Fischer AJ, Bongini R. Turning Müller glia into neural progenitors in the 75. retina. Mol Neurobiol. 2010;42:199–209. 45. Hitchcock PF, Lindsey Myhr KJ, Easter SS, Mangione-Smith R, Jones DD. Local 20. Wang M, Ma W, Zhao L, Fariss RN, Wong WT. Adaptive Müller cell responses regeneration in the retina of the goldfish. J Neurobiol. 1992;23:187–203. to microglial activation mediate neuroprotection and coordinate 46. Sun C, Galicia C, Stenkamp DL. Transcripts within rod photoreceptors of the inflammation in the retina. J Neuroinflammation. 2011;8:173. zebrafish retina. BMC Genomics. 2018;19:127. 21. Wang M, Wang X, Zhao L, Ma W, Rodriguez IR, Fariss RN, Wong WT. 47. Yuan T-F, Liang Y-X, Peng B, Lin B, So K-F. Local proliferation is the main Macroglia-microglia interactions via TSPO signaling regulates microglial source of rod microglia after optic nerve transection. Sci Rep. 2015;5:10788. activation in the mouse retina. J Neurosci. 2014;34:3793–806. 48. Petrie TA, Strand NS, Tsung-Yang C, Rabinowitz JS, Moon RT. Macrophages 22. Fischer AJ, Zelinka C, Gallina D, Scott MA, Todd L. Reactive microglia and modulate adult zebrafish tail fin regeneration. Development. 2014;141:2581–91. macrophage facilitate the formation of Müller glia-derived retinal 49. Oosterhof N, Holtman IR, Kuil LE, van der Linde HC, Boddeke EWGM, Eggen progenitors. Glia. 2014;62:1608–28. BJL, van Ham TJ. Identification of a conserved and acute 23. Wynn TA, Vannella KM. Macrophages in tissue repair, regeneration, and neurodegeneration-specific microglial transcriptome in the zebrafish. Glia. fibrosis. Immunity. 2016;44:450–62. 2016;65:138–49. 24. Carrillo SA, Anguita-Salinas C, Peña OA, Morales RA, Muñoz-Sánchez S, 50. Lillo C, Velasco A, Jimeno D, Cid E, Aijón J, Lara JM. Non-neuronal cells Muñoz-Montecinos C, Paredes-Zúñiga S, Tapia K, Allende ML. Macrophage involved in the degeneration and regeneration of the fish retina. J recruitment contributes to regeneration of mechanosensory hair cells in the Neurocytol. 2001;30:475–91. zebrafish lateral line. J Cell Biochem. 2016;117:1880–9. 51. Hochreiter-Hufford A, Ravichandran KS. Clearing the dead: apoptotic cell 25. Villegas R, Martin SM, Donnell KCO, Carrillo SA, Sagasti A, Allende ML. sensing, recognition, engulfment, and digestion. Cold Spring Harb Perspect Dynamics of degeneration and regeneration in developing zebrafish Biol. 2013;5:a008748. peripheral axons reveals a requirement for extrinsic cell types. Neural Dev. 52. Craig SEL, Calinescu A-A, Hitchcock PF. Identification of the molecular 2012;7:1–1. signatures integral to regenerating photoreceptors in the retina of the 26. White DT, Sengupta S, Saxena MT, Xu Q, Hanes J, Ding D, Ji H, Mumm JS. zebra fish. J Ocul Biol Dis Inform. 2008;1:73–84. Immunomodulation-accelerated neuronal regeneration following selective 53. Glass CK, Saijo K, Winner B, Marchetto MC, Gage FH. Mechanisms underlying rod photoreceptor cell ablation in the zebrafish retina. Proc Natl Acad Sci U inflammation in neurodegeneration. Cell. 2010;140:918–34. S A. 2017;114:E3719–28. 54. Amor S, Puentes F, Baker D, van der Valk P. Inflammation in 27. Kyritsis N, Kizil C, Zocher S, Kroehne V, Kaslin J, Freudenreich D, Iltzsche A, neurodegenerative diseases. Immunology. 2010;129:154–69. Brand M. Acute inflammation initiates the regenerative response in the 55. Amor S, Woodroofe MN. Innate and adaptive immune responses in adult zebrafish brain. Science. 2012;338:1353–6. neurodegeneration and repair. Immunology. 2014;141:287–91. 28. Sieger D, Peri F. Animal models for studying microglia: the first, the popular, 56. Rezai-Zadeh K, Gate D, Town T. CNS infiltration of peripheral immune cells: and the new. Glia. 2012;61:3–9. D-day for neurodegenerative disease? J NeuroImmune Pharmacol. 29. Langmann T. Microglia activation in retinal degeneration. J Leukoc Biol. 2009;4:462–75. 2007;81:1345–51. 57. Kyritsis N, Kizil C, Brand M. Neuroinflammation and central nervous system 30. Kettenmann H, Hanisch U-K, Noda M, Verkhratsky A. Physiology of microglia. regeneration in vertebrates. Trends Cell Biol. 2014;24:128–35. Physiol Rev. 2011;91:461–553. 58. Thomas JL, Ranski AH, Morgan GW, Thummel R. Reactive gliosis in the adult 31. Nelson CM, Ackerman KM, O'Hayer P, Bailey TJ, Gorsuch RA, Hyde DR. zebrafish retina. Exp Eye Res. 2016;143(C):98–109. Tumor necrosis factor-alpha is produced by dying retinal neurons and is 59. Lenkowski JR, Qin Z, Sifuentes CJ, Thummel R, Soto CM, Moens CB, required for Muller glia proliferation during zebrafish retinal regeneration. Raymond PA. Retinal regeneration in adult zebrafish requires regulation of J Neurosci. 2013;33:6524–39. TGFβ signaling. Glia. 2013;61:1687–97. 32. Salter MW, Stevens B. Microglia emerge as central players in brain disease. 60. Morris AC, Forbes-Osborne MA, Pillai LS, Fadool JM. Microarray analysis of Nat Med. 2017;23:1018–27. XOPS-mCFP zebrafish retina identifies genes associated with rod 33. McGinn TE, Mitchell DM, Meighan PC, Partington N, Leoni DC, Jenkins CE, photoreceptor degeneration and regeneration. Invest Ophthalmol Vis Sci. Varnum MD, Stenkamp DL. Restoration of dendritic complexity, functional 2011;52:2255–66. connectivity, and diversity of regenerated retinal bipolar neurons in adult 61. Powell C, Cornblath E, Elsaeidi F, Wan J, Goldman D. Zebrafish Müller glia- zebrafish. J Neurosci. 2018;38:120–36. derived progenitors are multipotent, exhibit proliferative biases and 34. Westerfield M. The zebrafish book. A guide for the laboratory use of regenerate excess neurons: Nature Publishing Group. Scientific Reports. zebrafish (Danio rerio). 5th ed. Eugene: University of Oregon Press; 2007. 2016;6:24851. Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 20 of 20 62. Paolicelli RC, Bolasco G, Pagani F, Maggi L, Scianni M, Panzanelli P, Giustetto M, Ferreira TA, Guiducci E, Dumas L, Ragozzino D, Gross CT. Synaptic pruning by microglia is necessary for normal brain development. Science. 2011;333:1456–8. 63. Stevens B, Allen NJ, Vazquez LE, Howell GR, Christopherson KS, Nouri N, Micheva KD, Mehalow AK, Huberman AD, Stafford B, Sher A, Litke AM, Lambris JD, Smith SJ, John SWM, Barres BA. The classical complement cascade mediates CNS synapse elimination. Cell. 2007;131:1164–78. 64. Schafer DP, Lehrman EK, Kautzman AG, Koyama R, Mardinly AR, Yamasaki R, Ransohoff RM, Greenberg ME, Barres BA, Stevens B. Microglia sculpt postnatal neural circuits in an activity and complement-dependent manner. Neuron. 2012;74:691–705. 65. Sipe GO, Lowery RL, Tremblay M-È, Kelly EA, Lamantia CE, Majewska AK. Microglial P2Y12 is necessary for synaptic plasticity in mouse visual cortex. Nat Commun. 2016;7:10905. 66. Sierra A, Encinas JM, Deudero JJP, Chancey JH, Enikolopov G, Overstreet- Wadiche LS, Tsirka SE, Maletic-Savatic M. Microglia shape adult hippocampal neurogenesis through apoptosis-coupled phagocytosis. Cell Stem Cell. 2010; 7:483–95. 67. Huang T, Cui J, Li L, Hitchcock PF, Li Y. The role of microglia in the neurogenesis of zebrafish retina. Biochem Biophys Res Commun. 2012;421: 214–20. 68. Mazaheri F, Breus O, Durdu S, Haas P, Wittbrodt J, Gilmour D, Peri F. Distinct roles for BAI1 and TIM-4 in the engulfment of dying neurons by microglia. Nat Commun. 2014;5:1–11. 69. Peri F, Nüsslein-Volhard C. Live imaging of neuronal degradation by microglia reveals a role for v0-ATPase a1 in phagosomal fusion in vivo. Cell. 2008;133:916–27. 70. Kokaia Z, Martino G, Schwartz M, Lindvall O. Cross-talk between neural stem cells and immune cells: the key to better brain repair? Nat Neurosci. 2012; 15:1078–87. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Neuroinflammation Springer Journals

Dynamic changes in microglial and macrophage characteristics during degeneration and regeneration of the zebrafish retina

Free
20 pages

Loading next page...
 
/lp/springer_journal/dynamic-changes-in-microglial-and-macrophage-characteristics-during-F20QqLdZJi
Publisher
Springer Journals
Copyright
Copyright © 2018 by The Author(s).
Subject
Biomedicine; Neurosciences; Neurology; Neurobiology; Immunology
eISSN
1742-2094
D.O.I.
10.1186/s12974-018-1185-6
Publisher site
See Article on Publisher Site

Abstract

Background: In contrast to mammals, zebrafish have the capacity to regenerate retinal neurons following a variety of injuries. Two types of glial cells, Müller glia (MG) and microglia, are known to exist in the zebrafish retina. Recent work has shown that MG give rise to regenerated retinal neurons, but the role of resident microglia, and the innate immune system more generally, during retinal regeneration is not well defined. Specifically, characteristics of the immune system and microglia following substantial neuron death and a successful regenerative response have not been documented. Methods: The neurotoxin ouabain was used to induce a substantial retinal lesion of the inner retina in zebrafish. This lesion results in a regenerative response that largely restores retinal architecture, neuronal morphologies, and connectivities, as well as recovery of visual function. We analyzed cryosections from damaged eyes following immunofluorescence and H&E staining to characterize the initial immune response to the lesion. Whole retinas were analyzed by confocal microscopy to characterize microglia morphology and distribution. Statistical analysis was performed using a two-tailed Student’s t test comparing damaged to control samples. Results: We find evidence of early leukocyte infiltration to the retina in response to ouabain injection followed by a period of immune cell proliferation that likely includes both resident microglia and substantial numbers of proliferating, extra-retinally derived macrophages, leading to rapid accumulation upon retinal damage. Following immune cell proliferation, Müller glia re-enter the cell cycle. In retinas that have regenerated the layers lost to the initial injury (histologically regenerated), microglia retain morphological features of activation, suggesting ongoing functions that are likely essential to restoration of retinal function. Conclusions: Collectively, these results indicate that microglia and the immune system are dynamic during a successful regenerative response in the retina. This study provides an important framework to probe inflammation in the initiation of, and functional roles of microglia during retinal regeneration. Keywords: Retina, Regeneration, Microglia, Macrophages, Zebrafish, Müller glia * Correspondence: dmitchell@uidaho.edu Department of Biological Sciences, University of Idaho, 875 Perimeter Drive, MS 3051, Moscow, ID 83844-3051, USA © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 2 of 20 Background other tissues participate in the phagocytosis and clearance The mammalian retina is unable to regenerate damaged of degenerative debris before a regenerative response be- neurons due to acute trauma or degenerative disease. gins [23]. In zebrafish, recruited macrophages appear to be Rather than restoration of lost neurons, a gliotic re- essential to regeneration of peripheral nervous system tis- sponse ensues, which is often associated with continual sue [24, 25]. The role of the innate immune system and inflammation [1]. Inflammation is the body’s response to resident macrophages, microglia, in regeneration of the tissue injury and/or infection, and is initiated by the innate central nervous system in organisms which have such re- immune system. Inflammation includes activation of im- generative capacity, such as teleost fish, is not well defined. mune cells and the production of pro-inflammatory cyto- Interestingly, a recent publication using larval zebrafish kines and molecular mediators. Particular cell types supports a role for inflammation and resident microglia in activated and cytokines/molecular mediators produced regeneration of rod photoreceptors in the retina [26]. In- will vary depending on the initial insult or injury, the tis- flammation has also been shown to be important in regen- sue location, and the immune response necessary to re- eration of the adult zebrafish brain [27]. Yet, it is not clear solve the infection or injury. Within the mammalian if these findings translate to retinal regeneration in an retina, resident microglia rapidly sense changes in the adult animal. microenvironment and can initiate and participate in both Under normal conditions, retinal microglia are highly acute and chronic inflammatory responses [2]. Acute in- ramified and integrated among the highly organized ret- flammation is crucial to clearing dead cells, debris, and inal tissue. Microglial responses to retinal insult and de- pathogens, which must occur prior to subsequent initi- generation have been studied mainly in rodent model ation of tissue repair. However, when inflammation be- systems [28]. Generally, when responding to neuro- comes chronic, it can be extremely damaging and logical insult, microglia transform morphology from contribute to tissue pathology. While chronic inflam- ramified to an ameboid shape and migrate to sites of cell mation and immune activation in mammalian neurode- death [29, 30]. However, microglial responses to retinal generative disease is well appreciated, the role of the injury in zebrafish have not been well documented, and immune system in contexts of successful retinal regen- it remains unclear under which conditions of retinal in- eration has not been well explored. jury extra-retinal immune cells may participate in the In contrast to mammals, teleost fish have a remarkable initial response to neuronal degeneration. Studies to date capacity to regenerate damaged retinas following a in zebrafish have focused on how inflammation, specific variety of lesions that destroy neurons [3–5]. The source cytokines, and microglia/macrophages may affect the of regenerated retinal neurons in zebrafish are the proliferation of Müller glial cells, which give rise to re- Müller glia (reviewed in [6–8]), retinal glial cells also generated retinal neurons [22, 26, 31]. There has been a present in mammals and other vertebrates. Upon injury, surge of interest in the role of microglia in development Müller glia re-enter the cell cycle [9–13] and undergo an and maintenance of the central nervous system, as well asymmetric division [14], ultimately generating multi- as their involvement in, and possibly sources of, neuro- potent progenitors that replace lost retinal neurons degenerative disease [2, 32]. Yet, little is known about [11, 14–16]. Intriguingly, Müller glia in mammals and the role of microglia in initiating and/or shaping success- other vertebrates appear to have properties suggesting ful regeneration of retinal neurons, and if these functions regenerative potential, including the upregulation of are shared with or unique from those in development. progenitor markers and cell cycle re-entry (reviewed In the present study, we use a tissue-disrupting lesion, in [1, 4, 17–19], butthisinstead resultsingliosisdue intraocular injection of the neurotoxin ouabain, to des- to the inability to differentiate into neurons (exten- troy inner retinal neurons while sparing glia and photo- sively reviewed in [1, 17]). Further, immune-Müller receptors [9, 10, 14, 33]. This type of lesion is known to glial crosstalk may be important in shaping the Müller result in a regenerative response that largely restores glia response to retinal injury [20–22]. retinal architecture, neuronal morphologies and con- In wound healing outside of the central nervous sys- nectivities, and behavioral and electrophysiological tem, the role of the immune system is well appreciated. measures of retinal function [10, 33]. We characterize In general, a period of degeneration is followed by re- the initial response to this cytotoxic damage to deter- generation in which the lost cell types are replaced and mine how the innate immune system responds to retinal tissue structure is restored. Tissue resident macrophages degeneration that is subsequently followed by a successful are crucial to coordinating the clearance of damaged tissue regenerative response. We then investigate microglial char- as well as instructing differentiation of new cell types to acteristics, including distribution, morphology, selective replace damaged tissue [23]. Resident macrophages as well markers, and histological features, in regenerated retinas. as recruited phagocytes, such as neutrophils, monocyte- We find that in damaged zebrafish retinas, both resident derived macrophages, and macrophages that migrate from and infiltrating immune cells contribute to mount a robust Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 3 of 20 response to neuronal cell death that is immediately Tissue collection and processing followed by Müller glia proliferation. Upon histological At selected time points post-injection, tissue samples regeneration of retinal layers, resident microglia localize were collected for analysis. For whole retina collection, to regions of regenerated neurons and maintain morph- fish were dark adapted for approximately 12 h, anesthe- ologies indicative of ongoing functional activation. Our tized in tricaine solution, and then eyes were enucleated findings indicate that retinal regeneration in adult zebra- using fine forceps, and placed into phosphate-buffered fish provides an excellent system to discover functional (pH = 7.4) saline (PBS) for dissection. The cornea and roles for the immune system and microglia that may be lens were removed and retinas were peeled from the crucial to supporting successful regenerative responses. whole eyecup. The retinal pigmented epithelium (RPE) detached from the retina due to the dark adaption, and Methods any remaining RPE was brushed away using a soft paint- Animals brush. Retinas were then rinsed several times in fresh Procedures involving zebrafish were performed in com- PBS using a plastic transfer pipette, transferred into fixa- pliance with protocols approved by the University of tive consisting of 4% paraformaldehyde in PBS for ap- Idaho Animal Care and Use Committee (IACUC). Zebra- proximately 1 h at room temperature with constant fish (Danio rerio) were maintained on a 14:10 light/dark gentle agitation, then washed several times with PBS cycle in 28.5 °C recirculating, monitored system water, containing 0.01% TritonX-100 (PBST). housed, and propagated according to [34]. Zebrafish To prepare retinal cryosections, whole eyes were enu- transgenic lines used in this study include mpeg1:GFP cleated using fine forceps, transferred to PBS, and the (gl22 Tg, GFP expressed in microglia/macrophages [35], lens was removed. Eyes were then fixed in phosphate- available from Zebrafish International Resource Center, buffered, 4% paraformaldehyde containing 5% sucrose ZIRC); mpeg1:mCherry (gl23 Tg, mCherry expressed in for 1 h at room temperature, washed in phosphate- microglia/macrophages [35], available from ZIRC). The buffered (pH = 7.4) 5% sucrose, and then washed in a gl22 Tg and gl23 Tg transgenic lines co-label cells in retinal graded series ending in 20% sucrose. The following day, tissue (Additional file 1: Figure S1). The wild-type strain tissues were embedded in blocks of a 1:2 solution of OCT used, referred to as SciH, was originally obtained from embedding medium (Sakura Finetek) and phosphate- Scientific Hatcheries (now Aquatica Tropicals). buffered, 20% sucrose, and frozen in isobutane super- cooled with liquid N . After freezing solid, tissues were Retinal lesion sectioned at 5 μm thickness using a Leica CM3050 cryo- Chemical lesioning of zebrafish retinas (age 6–14 months, stat. After overnight desiccation, tissue sections on glass both sexes) was performed by intravitreal injection of oua- slides were stored at − 20 °C until use. bain in order to destroy inner retinal neurons and spare To evaluate co-label of gl22 Tg and gl23 Tg, whole em- photoreceptors and Müller glia [9, 10, 33]. A working bryos were fixed at 3 days post-fertilization and washed stock of 40 μM ouabain (ouabain octahydrate, Sigma- with PBST as described for whole retina collection Aldrich) was prepared in 0.65% sterile saline (NaCl) above. Whole eyes were removed and mounted in solution. Fish were anesthetized by immersion in tricaine VECTASHIELD (Vector Labs), covered with a coverslip, solution, and an incision was made across the cornea then imaged as described below. using a sapphire blade. A Hamilton syringe was inserted behind the lens and 0.4–0.6 μLof 40 μM ouabain solution Immunofluorescence was injected into the vitreal chamber. Volume injected For staining of whole retinas, fixed tissue was blocked in was based on diameter of the eye (measured with calipers) 20% goat serum for 1 h at room temperature, incubated and upon calculations based on geometry and volumes of with primary antibody for 1–3 days at 4 °C, washed the eye and lens, resulting in an estimated intraocular con- three times in PBST, then incubated with conjugated centration of 2 μM[9, 36]. Lesions were unilateral and secondary antibody and DAPI (4′,6-diamidino-2-pheny- only the right eye was injected. For control samples, right lindole, 4.25 μM) for 1–3 days at 4 °C, washed twice in eyes of a separate group of fish were injected with 0.65% PBST with a final wash in PBS. All antibody incubations sterile saline (NaCl) solution. The same sterile saline solu- and washes were performed with constant agitation. tion was used for saline injections and for preparation of Whole retinas were then flat mounted by cutting four ouabain solution for injection. During the procedure, fish slits around the perimeter of the retinal cup and were were continuously flushed with tricaine solution. Immedi- mounted flat onto a microscope slide in VECTASHIELD ately following the procedure, fish were returned to tanks (Vector Labs), covered with a coverslip, and sealed with with clean system water. For all experiments, 3 control clear nail polish. (saline injected) and 3–4 ouabain injected fish were used To stain cryosections, tissue sections (5 μm thickness) for subsequent analysis. were blocked in 20% goat serum for 30 min at room Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 4 of 20 temperature, incubated in primary antibody overnight, were acquired at × 20 magnification using the large washed in PBST for at least 30 min, incubated in sec- stitched images feature in Nikon Elements software and ondary antibody for 1 h, and washed in PBST for at stitched based on DAPI staining. Z stacks of flattened least 30 min. Slides were then mounted in Vectashield retinas were generally obtained at 1 μm intervals. + DAPI (Vector Labs), covered with a coverslip, and Imaging of H&E stained cryosections was performed sealed with clear nail polish. Primary antibodies and di- using a Leica DM2500 compound microscope with Leica lutions used rabbit polyclonal anti-zebrafish L-plastin DFC700T color camera using × 20, × 40 (dry), or × 100 [37, 38] (1:10,000, a kind gift of Dr. Michael Redd), (Oil) objectives. Image processing and analysis was rabbit anti-zebrafish mpx (1:200, Genetex), mouse ZPR1 performed using FIJI (ImageJ). antibody (labeling cone-arrestin 3a [39], 1:200, ZIRC), mouse anti-PCNA (PC10, 1:200, Santa Cruz Biotech), rat Quantification of immune cells in retinal cryosections anti-PCNA (16D10, 1:200, Chromotek), rabbit anti- (24–72 h post-injection) phosphorylated histone 3 (PH3) (1:500, Cell Signaling), Using FIJI (ImageJ), a region of interest was drawn in rabbit polyclonal anti-PKCα (Santa Cruz Biotech), mouse each image obtained from cryosections corresponding to ZRF1 antibody (labeling GFAP, 1:200, ZIRC), and mouse the region of the retinal lesion in the inner retina: from anti-Glutamine Synthetase (1:1000, BD Transduction the basal border of the outer nuclear layer to the basal Laboratories). Secondary antibodies conjugated to Cy3, border of the ganglion cell layer. For 48 and 72 h post- FITC, or Alexa-Fluor647 (Jackson ImmunoResearch) were injury (hpi) time points, the basal layer of this border used at 1:200 dilution. was defined based on the limit of dense DAPI staining. DAPI+ nuclei surrounded by L-plastin+ cell bodies were H&E staining counted within the defined region of interest and nor- Slides containing cryosections were air dried at room malized per 1000 square microns. Note that the analyzed temperature for 30 min, immersed in 0.1% Hematoxylin region corresponds to the lesioned inner retina where (MHS-16, Sigma-Aldrich) 5 min, then washed in run- the x direction is parallel to retinal layers and the y dir- ning tap water for 5 min. Slides were then immersed in ection is perpendicular to retinal layers. Since the inner 0.5% eosin (Sigma-Aldrich) for 1 min, rinsed in deion- nuclear layer is damaged due to the ouabain-induced le- ized water until rinse was clear, then immersed in a sion and becomes thinner, the number of L-plastin+ series of ethanol solutions (50, 75, 95, and 100%) for cells is also reported per 100 μm of curvi-linear distance 30 s-1 min each, dipped five times in Xylenes (Fisher of retina using a line that follows parallel to the basal Scientific) and immediately mounted with Permount border of the ONL. Three to four non-adjacent cryosec- (Fisher Scientific) and sealed with a coverslip. tions were analyzed per fish. TUNEL staining Quantification of immune cells in whole, flat mounted TUNEL (Terminal deoxynucleotidyl transferase dUTP retinas (histologically regenerated retinas) nick end labeling) staining of 5 μm cryosections was Using FIJI (Image J), individual microglia (DAPI+ nuclei performed following the manufacturer’sinstructions surrounded by L-plastin+ signal) in flattened whole (Roche). Briefly, cryosections were washed for 20 min retinas were counted in individual image stacks contain- in × 1 PBS at room temperature then placed in ice-cold ing all z planes obtained at × 40 magnification, from the permeabilization solution for 2 min. Slides were washed vitreal face of the ganglion cell layer to the apical face of twice in × 1 PBS, 10 min per wash, incubated in the outer nuclear layer. Counts were normalized per TUNEL reaction mixture at 37 °C in the dark for 3 h, 1000 square microns of flattened retina where the x and then rinsed three times in PBS. Immediately following, y direction are parallel to retinal layers. Counts were cryosections were blocked in 20% goat serum for normalized to area rather than volume because the re- 30 min at room temperature and subsequent immuno- generated retinas are thinner than controls [33]. Six to fluorescence was performed as described above. eight images covering approximately 190,000 μm² per image were analyzed from each whole retina. Microscopy and image acquisition Imaging of whole flat mounted retinas (z series) or fluo- Analysis of microglia morphology rescently stained retinal sections (single plane) was per- We analyzed morphology of microglia in the ganglion formed with a Nikon Andor spinning disk confocal cell layer of histologically regenerated and control retinas. microscope equipped with a Zyla sCMOS camera run- Confocal z stacks obtained from whole, flat mounted ret- ning Nikon Elements software. Imaging was performed inas using a × 40 Oil objective corresponding to the region using × 20 (dry) or × 40 (oil immersion) objectives. For of the ganglion cell layer were flattened using FIJI (ImageJ). stitched images of entire retinal cryosections, images We used the marker L-plastin, which is a leukocyte-specific Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 5 of 20 form of the actin bundling protein plastin expressed by Statistical analysis zebrafish macrophages [37], to analyze microglia in A two-tailed Student’s t test was used to obtain p values control and histologically regenerated retinas because comparing saline injected to ouabain injected samples. (1) all retinal L-plastin+ cells co-label with GFP P values less than 0.05 are reported in the figures. expression in mpeg1:GFP transgenic fish at these time points and (2) L-plastin staining labels the entire cell Results body and cellular processes continuously, whereas the Immune cell response to a tissue-disrupting retinal lesion mpeg1 promoter-driven transgenic reporter labels To investigate the immune response to substantial neur- portions of the cell (Fig. 9A,A’,B,B’). Outlines of indi- onal damage in the retina, we used intravitreal injection vidual L-plastin+ cells were manually traced and mea- of the neurotoxin ouabain to generate this damage. In- surements of morphological features were analyzed jection of ouabain to a final concentration of 2 μM using the “measure” tool in FIJI. Approximately 35 causes death of inner retinal neurons, as demonstrated microglia, taken from 3 to 4 retinas per condition, by the loss of the PKCα+ bipolar neuron population at were analyzed for each condition and histograms of 3 days post-injection (dpi) (Fig. 1) and appearance of the distributions were created. Descriptions of the pa- TUNEL+ nuclei within the inner retina at 24 and 48 h rameters used to analyze microglia morphology are as post-injection (hpi) (Fig. 2). Photoreceptors and glia are follows. Area: area of an individual microglia in square spared [9, 10, 14, 33], as demonstrated by the continued microns; circularity: circularity of an individual micro- presence of zpr1+ cones and zrf1+ Müller glia at 3 dpi glial cell body. A value of 1.0 indicates a perfect circle, (Fig. 1). We visualized microglia/immune cells in control while values closer to 0 indicate elongated shapes; per- and ouabain-lesioned retinas using an antibody to zebra- imeter: the total length of a line tracing the outside fish L-plastin, which is used to mark leukocytes, includ- boundary of an individual microglial cell in microns; ing macrophages, in zebrafish [37, 38]. Similar to reports Feret diameter, also known as maximum caliper, the in embryonic and larval zebrafish [40], microglia in in- longest distance between any two points along the per- tact, undamaged adult zebrafish retina reside in the imeter of an individually traced microglial cell. inner nuclear layer, ganglion cell layer, and nerve fiber Fig. 1 Ouabain-induced retinal degeneration results in a robust accumulation of responding immune cells. Images show cryosections from undamaged (control, A, C) and ouabain-damaged retinas at 3 days post-injection (3 dpi, B, D). A ZPR1 (green), PKC-α (Red), and DAPI (blue) were used to label cones of the outer retina, bipolar neurons of the inner retina, and all nuclei, respectively, in undamaged retinas. B Images of damaged retinas sampled at 3 dpi; inner retinal neurons have been destroyed (note the absence of PKC-α staining and absence of DAPI+ layer corresponding to the ganglion cell layer; GCL), but photoreceptors are spared (ZPR1, green). C Ramified microglia, labeled by L-plastin (magenta), are present in undamaged retinas, along with radially patterned Müller glia (ZRF1, green). D Müller glia (green) are spared from the ouabain-induced lesion. A large number of immune cells (magenta) responding to the lesion are present in the damaged retina, primarily localized to the region of neuronal cell death (inner retina), although some appear to be recruited from regions apical to the retina (arrows). DAPI+ nuclei in the inner retina visible in B can be identified as immune cells and Müller glia. Scale bar in A, applies to all images, = 20 μm. ONL = outer nuclear layer, INL = inner nuclear layer, GCL = ganglion cell layer Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 6 of 20 Fig. 2 Progression of ouabain-induced retinal cell death and accumulation of immune cells in damaged retinas. Cryosections (5 μm thick) from retinas at 24 and 48 h post-intravitreal injection (24 and 48 hpi) of saline (A and B)or2 μM ouabain (C and D) were stained using the TUNEL cell death detection process (green), the immune cell marker L-plastin (magenta), and DAPI (blue). A–B Saline injection did not induce cell death (absence of TUNEL staining), and immune cells remain ramified (arrows). C–D TUNEL+ nuclei and debris (green) are present in the ganglion cell layer (GCL) and inner nuclear layer (INL) following intravitreal injection of 2 μM ouabain, and immune cells assume ameboid morphologies (arrows). As the lesion progresses over time, immune cells accumulate in regions of cell death, and the clearance of TUNEL+ nuclei and debris corresponds to this immune cell accumulation. Few immune cells contained TUNEL+, DAPI+ nuclei (quantification in the “Results” section), suggesting that immune cell death was not significantly induced by ouabain. E–G High magnification images of immune cells revealed immune cells phagocytosing TUNEL+ nuclei (E), immune cells with cytoplasmic TUNEL+ material (F, arrow), and immune cells with multiple nuclei, some of which are TUNEL+ (G,asterisks), suggesting that the accumulating immune cells are highly phagocytic and important for clearing dead cells and debris from the lesioned retina. Scale bar in A (applies to A–D)=20 μm. Scale bars in E, F,and G =5 μm. ONL = outer nuclear layer, INL = inner nuclear layer, GCL = ganglion cell layer layer (Fig. 1C). In addition, microglia flank the apical side To monitor the accumulation of immune cells over of the inner plexiform layer and the apical and basal the course of ouabain-induced cell death and to deter- sides of the outer plexiform layers (Fig. 1C). This is dif- mine how ouabain and the intravitreal injection process ferent than reports from mammals, which show that affect immune cells in the retina, we examined L-plastin microglia localize within plexiform layers [2]. After in- and TUNEL labeled cryosections from ouabain injected travitreal injection of ouabain, we found a dramatic ac- retinas at 24 and 48 hpi, compared to saline injected cumulation of immune cells in the region of the lesion controls (Fig. 2). In ouabain lesioned retinas, TUNEL+ by 3dpi (72hpi)(Fig. 1D). At 3 dpi, nearly all inner ret- nuclei and TUNEL+ debris within the inner retina ap- inal neurons have been destroyed, Müller glia are react- pear by 24 hpi, concomitant with immune cell accumu- ive, and some Müller glia have re-entered the cell cycle lation and activation as indicated by morphological ([9, 10, 14] and Fig. 1D). Many DAPI+ nuclei are present changes of L-plastin+ cells from ramified to ameboid cell in regions of neuronal death and degeneration of the shape (Fig. 2A, C). Immune cells in retinas from saline inner retina (Fig. 1B, D), and essentially all of these DAPI injected controls remained ramified, and in their typical + nuclei at 72 hpi correspond to immune cells and re- locations, and TUNEL signal was not detected (Fig. 2A, B). active Müller glia (Fig. 1D). By 48 hpi, immune cells with ameboid morphology Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 7 of 20 continue to accumulate in locations corresponding to 12 hpi following ouabain injection, ameboid immune ouabain-induced neuronal death and degeneration (INL, cells can be seen in central retina in a gradient originat- GCL, Fig. 2D). This accumulation correlates with a de- ing from the region of the optic nerve head (Fig. 3B, D, crease in TUNEL+ nuclei and debris from 24 to 48 hpi optic nerve head indicated) and vitreal to the ganglion (Fig. 2C, D), indicating that immune cells have cleared deb- cell layer in peripheral retina (arrows, Fig. 3B, C). We did ris resulting from neuronal death. Consistent with this, we not detect substantial PCNA signal in immune cells at observed immune cells engulfing TUNEL+ nuclei (Fig. 2E) this time point (Fig. 3C, D), indicating that the resident as well as TUNEL+ puncta within L-plastin+ immune cell microglia did not immediately enter the cell cycle. Cells bodies (Fig. 2F). However, only a small fraction of L- of the macrophage lineage (including microglia) in zeb- plastin+ cells contained TUNEL+, DAPI+, co-labeled rafish have been identified using the transgenic expres- nuclei (8.4 ± 5.7% at 24 hpi and 1.9 ± 1.9% at 48 hpi), indi- sion of mpeg1-driven transgenes [35], while neutrophils cating that ouabain only induced minor levels of immune are identified by expression of mpx (also called mpo) cell death. Further, some of the DAPI and TUNEL co- [41]. Nearly all of the L-plastin+ cells co-labeled with labeling within immune cells could potentially be ex- mpeg1:mCherry, indicating that the majority of these plained as phagocytosed TUNEL+ nuclei from dead or leukocytes are macrophages (Fig. 3E, F). Few mpx + neu- dying neurons. Consistent with this, in some cryosections, trophils were detected within the peripheral leukocyte we observed multiple nuclei, including TUNEL-nuclei, population vitreal to the ganglion cell layer at this time surrounded by L-plastin+ cytoplasm (Fig. 2G). point; however, mpx + neutrophils were not detected in other retinal regions, including regions surrounding the Origins and distributions of immune cells in damaged optic nerve head (Additional file 1: Figure S3). The pres- retina ence of few neutrophils supports that leukocytes in the To better understand the origins of the accumulation of blood have access to retinal tissue following ouabain in- immune cells in ouabain-lesioned retinas, we imaged jection. Previous reports indicate that apoptotic cells la- retinal cryosections at 12 hpi using the marker L-plastin beled by TUNEL are present as early as 3 hpi ouabain and the S-phase marker PCNA (Fig. 3). Saline injection injection, although the number of apoptotic cells peaks did not result in immune cell accumulation or PCNA at approximately 24 hpi [9]. The presence of ameboid im- expression (Fig. 3A and Additional file 1: Figure S2). At mune cells apparently infiltrating the retina, and absence Fig. 3 Evidence of early infiltration of immune cells to the retina at 12 h post-ouabain injection. Images of retinal cryosections at 12 h post-saline (A), or ouabain (B–F) injection (12 hpi). A, B Show stitched images of entire cryosections stained for L-plastin (magenta) and DAPI (blue) obtained at × 20 magnification. Images in C and E show peripheral regions of retina. D, f Show central regions of retina adjacent to the optic nerve head (onh). Following ouabain injection, ameboid immune cells are seen in a gradient originating from the region of the optic nerve head (B, D, F), as well as vitreal to the ganglion cell layer in peripheral regions (C (arrows) and E). Cryosections were also stained for PCNA (green, C and D). PCNA expression was only rarely detected at this time point. E In peripheral retina, few L-plastin+ cells (green) co-label with mpeg1:mCherry transgene expression (red, co-labeled cells yellow). F In central retina, most L-plastin+ cells (green) co-label with mpeg1:mCherry transgene expression (red, co-labeled cells yellow). A–D were obtained at × 20 magnification, E and F were obtained at × 40 magnification. Scale bar in (A)=100 μm (applies to A and B); scale bar in C (applies to C and D)=20 μm; scale bars in E and F =20 μm Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 8 of 20 of PCNA expression, suggests an early infiltration of im- continued accumulation of immune cells through 72 hpi mune cells following ouabain injection, prior to the peak (Fig. 5A, B). These cell cycle markers indicate that infil- of retinal cell death. trating immune cells prepare for division upon their ar- We next imaged cryosections at 24, 48, and 72 hpi rival to damaged retinal tissue, and resident microglia using the markers L-plastin and PCNA (Fig. 4), and also respond to retinal damage by entering the cell cycle. quantified numbers of cells stained by these markers To our knowledge, only one study has used a histo- (Fig. 5). Injection of saline did not result in accumulation logical method (methylene blue/azure II staining of plas- of immune cells or significant expression of PCNA within tic sections) to document ouabain-induced retinal theretinathrough 72 hpi (Figs. 4A–C, G–I,and 5A–C). By degeneration in zebrafish [9], and this method does not 24 h post-ouabain injection, immune cells appear abun- readily reveal immune cell characteristics. Therefore, to dant in the region vitreal to the degenerating ganglion cell better observe immune cell accumulation concomitant layer in both peripheral and central retina (Fig. 4D), sug- with progression of the ouabain-induced inner retinal le- gesting that extra-retinal immune cells continue to invade sion and to simultaneously visualize features of respond- the retina in response to neuronal degeneration. Immune ing immune cells, H&E staining of cryosections was cells with ameboid shape begin to accumulate in regions used (Fig. 6). Saline-injected controls did not show of inner retina corresponding to the ouabain-induced le- changes in retinal structure or immune cell accumula- sion at 24 hpi and continue to accumulate through 72 hpi tion (Fig. 6A–C). However, in damaged retinas, immune (Figs. 1D, 4D–F and J–L, 5A, B). Further, evidence of im- cells and the progression of the inner retinal lesion were mune cell infiltration to the damaged retina remains at 48– readily visualized simultaneously using H&E stains 72 hpi, but this later migration appears to originate from (Fig. 6D–I). At 24 hpi, pyknotic nuclei corresponding regions and structures apical to the neural retina, which to dying neurons are visible in the ganglion cell and could include the RPE or sclera and associated blood ves- inner nuclear layer of the retina (Fig. 6D), while im- sels (Fig. 4K, L). We consider this to be the case, because mune cells are visible in the vitreal space as well as immune cells that appear directionally oriented to migrate within the retinal tissue (Fig. 6D, G). By 48 hpi, im- into the retina (i.e., radially oriented) are more frequently mune cells are mainly present within regions of degen- present in the outer retina and RPE (Figs. 1D, 4K, L). erating retinal tissue, following the progression of the A significant fraction of immune cells responding to lesion (Fig. 6E, H). Consistent with this, the presence the ouabain-induced lesion at 24 and 48 hpi stain posi- of extracellular matrix, debris, and pyknotic nuclei de- tive for the S-phase marker PCNA (Fig. 4, arrows, and creases from 24 to 72 hpi (Fig. 6D–F). Immune cells Fig. 5C). PCNA+ immune cells are located in the vitreal continue to accumulate in regions of retinal degener- space at 24 hpi ouabain injection (Fig. 4D), apical to the ationfrom48to72hpi (Fig. 6E, F), which is likely due retina at 48 hpi (Fig. 4E, K), and within retinal tissues at to accumulation of dividing immune cells as well as both 24 and 48 hpi (Fig. 4D, E, J, K), suggesting that by continued immune cell migration from regions apical 24 hpi, both resident and infiltrating immune cells have to the lesion (denoted by white arrows in Fig. 6E, F), entered the cell cycle in response to the lesion. Import- indicating that immune cell infiltration into the inner antly, the number of immune cells located within retinal retina continues at least through 72 hpi. Through tissues more than doubles between 24 and 48 hpi 72 hpi, evidence of intact vasculature structures at the (Fig. 5A, B) while fewer than 50% of these cells are vitreal face of ouabain injected eyes remains (Fig. 6H, PCNA+ at 24 hpi (Fig. 5C). This rate of PCNA ex- inset, and I, black arrows), suggesting that blood ves- pression is not sufficient to account for all of the accumu- sels are not immediately and/or fully destroyed by lated immune cells, supporting the interpretation that at 2 μM ouabain. These vessels could therefore possibly least a portion of them infiltrated the retina. Numbers of provide infiltrating immune cells access to damaged PCNA+ immune cells decrease at 72 hpi and instead retinal tissue. PCNA positive signal is present in non-immune cells (Fig. 4F, L asterisks, Fig. 5). Identity of S-phase cells at 72 h post-ouabain lesion Using the M-phase specific marker PH3, we did not Since Müller glia respond to retinal injury in zebrafish identify significant numbers of PH3+ cells at any of by re-entering the cell cycle [8, 11, 12, 42], and in order these time points (Additional file 1: Figure S4 and not to determine the identity of non-leukocyte PCNA+ cells shown). The absence of PH3+ cells could possibly be at 72 h post-ouabain injection, we co-stained immune due to rapid mitosis, making detection of mitotic cells in cells in retinal cryosections obtained at this time point fixed retinal tissue difficult. An alternative explanation, with Glutamine Synthetase (GS), which stains Müller that immune cells die after S phase and prior to M glia (Fig. 7). At 72 hpi ouabain injection, nearly all phase, is not supported by the observations of very few PCNA+ nuclei in central and central-peripheral retina L-plastin+, TUNEL+ cells (discussed above) and of the are associated with Glutamine Synthetase+ (GS, Fig. 7B’ Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 9 of 20 Fig. 4 (See legend on next page.) Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 10 of 20 (See figure on previous page.) Fig. 4 Distribution and proliferation markers in immune cells during the response to retinal damage. Images show staining of L-plastin (magenta), PCNA (green), and DAPI (blue) in cryosections of retinas injected with saline (A–C and G–I) or ouabain (E–F and J–L) at 24, 48, and 72 hpi. Images are from regions of peripheral retina (rows 1 and 2, A–F) or central retina (rows 3 and 4, G–L). PCNA+, L-plastin+ cells are indicated with white arrows, while PCNA+ non-immune cells are indicated with asterisks. In images of peripheral retina, regions corresponding to the ciliary marginal zone (CMZ) are within dotted ellipses. A–C and G–I Saline injection does not result in significant changes in L-plastin+ cell accumulation, location, or morphology and does not result in significant PCNA+ signal in L-plastin+ cells (A–C and G–I). D–F and J–L In ouabain-lesioned retinas, significant numbers of ameboid-shaped L-plastin+ cells are located in the region vitreal to the ganglion cell layer at 24 hpi and many are PCNA+ (D, J, arrows). At 48 hpi ouabain, L-plastin+ cells remain ameboid in shape, appear to accumulate at the region of the ouabain-induced lesion, and many are PCNA+ (E and K, arrows). By 72 hpi, PCNA+ signal is mainly localized to non-immune cells (asterisks*), although ameboid-shaped L-plastin+ cells remain in regions corresponding to the lesion (F and L). K, L Arrowheads indicate immune cells invading from regions posterior to the retina. G–L Regions denoted with dotted white lines indicate regions (corresponding to the lesion) used for quantification of L-plastin+ cells shown in Fig. 4.Scale bar in A (applies to all images) = 20 μm. ONL = outer nuclear layer, INL = inner nuclear layer, GCL = ganglion cell layer and C’, yellow arrows) or L-plastin+ cytoplasm (Fig. 7A’ reporter (Fig. 8D–F), supporting significant immune cell and C’, asterisk(*)), although we did occasionally detect infiltration to the damaged retina. Despite the absence PCNA+ nuclei in basal retina that were not associated of mpeg1 reporter expression, morphological features of with GS or L-plastin signal (pink arrow, Fig. 7C’). This the L-plastin+ immune cells in damaged retinal tissue indicates that by 72 hpi, Mülller glia have re-entered the are consistent with a macrophage identity, including ir- cell cycle in order to replace neurons lost to the ouabain regularly cell-shaped cytoplasm and the presence of nu- lesion. merous vacuoles (Fig. 8D–F, arrowheads). Using H&E staining, we confirmed that these immune cells are mac- Identities of immune cells in damaged retinas rophages (Fig. 8G–I and G’–I’). Histological features in- To determine the identities of the immune cells initially clude elongated shape, irregular borders, numerous responding to ouabain-induced neuronal cell death, we vacuoles (asterisks, Fig. 8G–I,G’–I’), and granular cyto- used cell lineage-specific markers of phagocytic immune plasmic inclusions with intensity and color similar to cells in cryosections. Mpeg1 expressing macrophages/ that of extracellular debris, evidence of phagocytosis, microglia are visible at all time points (Fig. 8D–F), but and lysosomal degradation (Fig. 8G–I and G’–I’,[43]). only a fraction of L-plastin+ immune cells present at any H&E staining also revealed numerous vacuoles within of these time points also express the mpeg1 transgenic macrophages that had phagocytosed pyknotic nuclei Fig. 5 Quantification of immune cells in cryosections of damaged retina. A Quantification of L-plastin+ immune cells per 1000 μm² of inner retina (5-μm-thick sections) from saline or ouabain injected eyes at 24, 48, and 72 hpi, representative images in Fig. 3G–L). For details, see the “Quantification of Immune Cells in Retinal Cryosections (24–72 h post-injection)” section. Regions depicted by dotted white lines in Fig. 3G–L are representative regions used for quantification. DAPI+ nuclei surrounded by L-plastin+ signal were counted within the region of interest and normalized per 1000 μm². p values comparing densities between 24 and 48 hpi and 48–72 hpi ouabain injection are shown on the graph; asterisks indicate the following − 12 − 12 p values comparing ouabain to saline injected at indicated time points: *p = 0.012, **p =7.84 × 10 , ***p =1.21×10 . B Since thickness of the INL changes due to ouabain-induced lesion, the number of L-plastin+ cells is also reported per 100 μm of curvi-linear distance of retina using a line that follows parallel to the basal border of the ONL. p values comparing densities between 24 and 48 hpi and 48–72 hpi ouabain injection are shown on the graph; asterisks indicate the following p values comparing ouabain injected to saline injected at indicated time − 6 − 9 − 10 points: *p =3.84× 10 ,**p =8.48×10 ,***p =5.5 ×10 . C PCNA+, DAPI+ nuclei surrounded by L-plastin+ signal quantified in retinal cryosections from saline or ouabain injected eyes at 24, 48, and 72 h hpi. Graph shows the fraction of L-plastin+ cell bodies with nuclear co-label of DAPI and PCNA. Only images from central and central-peripheral retina were quantified, and only L-plastin+ cell bodies within − 12 retinal tissue boundaries were counted. Error bars indicate standard deviation. *p = 0.0005, **p =1.01×10 , comparing ouabain injected to saline injected at indicated time points Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 11 of 20 Fig. 6 Progression of retinal degeneration and immune cell response visualized by Hematoxylin & Eosin (H&E). H&E staining of cryosections to visualize progression of the ouabain-induced lesion and immune cell accumulation. A–C Saline-injected controls did not show significant changes in retinal structure; ramified microglia are not readily visualized due to their thin processes with little cytoplasm and integration into retinal layers. D–F By 24 h post-ouabain injection (24 hpi), the inner retina has begun to swell, pyknotic nuclei are present (asterisks), and ameboid immune cells can be detected within the GCL and INL (D, white arrows and G, black arrows). Immune cells are visible in the vitreal space at 24 hpi (black arrows, G), consistent with invasion of extra-retinal immune cells from the retinal vitreal face. Immune cells accumulate in degenerating retinal tissue, which contains many pyknotic nuclei (asterisks), through 48 and 72 hpi ouabain injection (E and F). However, immune cells are no longer observed in the vitreal space at 48 and 72 hpi ouabain (E, F, H, I). Immune cells present in the degenerating retinal tissue are highly phagocytic, demonstrated by cytoplasmic color, regions of space immediately surrounding them, and the progressive disappearance of extracellular matrix and debris over time (D, E, F). Immune cells appear to invade from structures apical to the retina at 48 and 72 hpi ouabain (white arrows, E and F). G–I Evidence of surviving blood vessels at all time points following ouabain injection: at 24 hpi, blood vessel structures at the anterior of the eye adjacent to the peripheral retina are present (dashed box, G). Sections of intact retinal blood vessels at the vitreal face of the retina are present at 48 and 72 hpi ouabain injection (dashed box and inset, H, and black arrows, I). Scale bar in A (applies to A–F)=20 μm. Scale bar in G (applies to G–I)=40 μm (arrow, Fig. 8H’). Very few mpx + neutrophils were de- revealed very few immune cells bearing features of neu- tected in cryosections at 24, 48, or 72 h post-ouabain in- trophils (banded or segmented nuclei and pink, granular jection (Additional file 1: Figure S5), and H&E staining cytoplasm [43]), indicating that the few neutrophils Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 12 of 20 Fig. 7 PCNA expression in immune cells and Müller glia at 72 h post-ouabain injection. Representative images of retinal cryosections at 72 h post-ouabain injection (72 hpi) stained for L-plastin to mark immune cells (green), glutamine synthetase to mark Müller glia (GS, red), PCNA (white), and DAPI (blue). A L-plastin and DAPI. B GS and DAPI. C L-plastin, GS, and DAPI. A’ L-plastin, PCNA, and DAPI. B’ GS, PCNA, and DAPI. C’ Four color merge of all stains. Nearly all PCNA+ nuclei can be attributed to immune cells or Müller glia at this time point. Asterisk (*) in A’ and C’ denotes L-plastin+, PCNA+ cell. Yellow arrows in B’ and C’ are provided to emphasize selected GS+ and PCNA+ cells. Occasionally, we observed PCNA+ nuclei at the vitreal surface at 3 dpi that could not be attributed to L-plastin+ or GS+ cells (C’, pink arrow). Scale bar in A (applies to all images) = 20 μm present in peripheral regions at 12 hpi did not signifi- Microglia morphologies and distributions in histologically cantly increase in number, and that neutrophils do not regenerated retinas comprise a significant portion of responding leuko- We next investigated microglia morphology and retinal cytes. However, it is possible that the few neutrophils distribution at 14–21 days following ouabain injection. present could represent some of the TUNEL+ immune Although there has yet to be a demonstration of func- cells at 24 or 48 hpi (Fig. 2, discussed above), as neu- tional restoration at these times, we refer to 14 and trophils undergo apoptosis shortly after completing 21 dpi retinal tissues as “histologically regenerated,” due their phagocytic functions [43, 44]. We conclude that to the presence of new neurons and the emergence of the immune cells infiltrating the retina in response to plexiform layers [9, 10]. These histologically regenerated ouabain-induced neuron death may be derived from retinas, however, do show lamination defects. Typical microglia and/or macrophages in nearby tissues and lamination defects are the presence of neuronal nuclei monocytes from circulation that rapidly differentiate within plexiform layers, particularly the inner plexiform to macrophages to perform crucial phagocytic func- layer; such defects have been referred to as “laminar fu- tions. The absence of detectable mpeg1:mCherry ex- sions” [10, 45, 46], some of which resolve over further pression could be explained by delays in expression of recovery time [10]. We reasoned that microglia are ex- the transgenic reporter, or alternatively, that mpeg1 cellent candidates for involvement in resolution of histo- expression is downregulated in this particular activa- logical errors and clearing non-functional or apoptotic tion state. neurons that may arise during retinal regeneration, and Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 13 of 20 Fig. 8 (See legend on next page.) Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 14 of 20 (See figure on previous page.) Fig. 8 Immune cells responding to ouabain-induced retinal degeneration identify as macrophages. A–F. Retinal cryosections from transgenic mpeg1:mCherry fish following intravitreal injection of saline (A–C) or ouabain (D–F) were immunolabeled with L-plastin (green) and counterstained with DAPI (blue). A–C. In saline-injected controls, all L-plastin+ immune cells (green) co-label with the macrophage marker mpeg1:mCherry (red); co-label is yellow (A–C), indicating ramified microglia. D–F. Following ouabain injection, ameboid, L-plastin+ immune cells accumulate (green, D–F), but only a subset of these also express the mpeg1:mCherry reporter (yellow, D–F). Arrowheads indicate vacuoles in selected cells. G–I High magnification images of retinal cryosections from ouabain damaged retinas stained with H&E reveal classical characteristics of phagocytic macrophages in accumulating immune cells (G–I and G’–I’), indicating that these accumulating immune cells are in fact macrophages. These features include irregular shapes of cell bodies and nuclei, cytoplasm similar in color to the environment, space immediately surrounding the cell borders, and the presence of vacuoles (asterisks). H’ Pyknotic nucleus within the cytoplasm of a macrophage (arrow), indicating phagocytosis of apoptotic cells. The red signal in the ONL in images A–F is due to autofluorescence from photoreceptors. Scale bar in A (applies to A–F)= 20 μm. Scale bar in G (applies to G–I and G’–I”)=10 μm so selected these time points for further analysis. By each parameter begin to shift towards those in control 14 dpi, all immune cells in the retina express the mpeg1: retinas (Fig. 10), indicating dynamic changes are occurring GFP reporter (Fig. 9B–B”). Densities of microglia in his- within the microglial population, possibly a trend towards tologically regenerated and saline-injected retinas were ramification. Collectively, the morphological features of calculated from whole, flat mounted retinas considering microglia in newly histologically regenerated retinas indi- all retinal layers (described in the “Methods” section). cate that rather then returning to a ramified state, micro- Densities of microglia in saline and histologically regen- glia remain dynamic as histological features of regenerated erated retinas were calculated from image stacks ob- retinas continue to improve. tained from whole, flattened retinas (representative images in Fig. 9A and B). Densities of microglia were Discussion slightly elevated in histologically regenerated retinas In this study, we examined the immune response to compared to saline controls, but the differences were ouabain-induced degeneration of inner retinal neurons not statistically significant (Fig. 9E). However, microglia in zebrafish, as well as morphology of microglia in histo- have not returned to their normal retina distribution logically regenerated retinas. Key findings include (1) patterns and instead remain localized to regions of histo- immune cells accumulate rapidly and in large numbers logically regenerated retina, particularly the ganglion cell following substantial retinal damage; (2) in addition to layer (Fig. 9D, F), suggesting that they are indeed per- local microglia, extra-retinally derived macrophages ap- forming functional activities in regions containing new pear to comprise a significant portion of responding im- neurons. mune cells; and (3) Microglia in histologically regenerated Further examination of microglia morphologies within retinas retain morphological features of activation. Col- the ganglion cell layer at 14 and 21 dpi reveal that lectively, these results indicate that microglia and the im- microglia retain morphological features of activation ra- mune system are dynamic during retinal degeneration and ther than ramification (Fig. 10), even though retinal layers regeneration and could play important roles in a success- have been histologically regenerated by 14 dpi [9, 10, 46]. ful regenerative response in the retina. In control retinas (saline injected), microglia display long, complex processes and appear ramified (Fig. 10A, B), as in- Response of immune cells to damaged retinal tissue dicated by long perimeters and increased Feret diameter We find that immune cells accumulate rapidly in re- (Fig. 10J and K). In histologically regenerated retinas, sponse to ouabain-induced retinal lesion. We conclude microglia remain ameboid in shape, especially at 14 dpi that these immune cells are composed of responding ouabain (Fig. 10C, D compared to Fig. 10A, B), as indicated resident microglia as well as extra-retinally derived mac- by more circular morphology with reduced perimeter and rophages, providing a clear identity to the abundant nuclei reduced Feret diameter (Fig. 10I, K). Some microglia are of unidentified cell type in the inner retina following oua- associated with multiple nuclei (Fig. 10C asterisk, and E–G), bain damage documented in previous studies [9, 10, 33]. possibly due to phagocytosis of improperly localized, non- Several publications to date indicate that the innate im- functional, or apoptotic neurons. In addition, chains of mune response, particularly macrophages, are crucial to re- rod-shaped microglia were observed in regenerated retinas generation of a variety of tissues in zebrafish [24–27, 48]. (Fig. 10D). Rod-shaped microglia have been observed fol- However, it remains unclear to what extent and in which lowing optic nerve transection, which may phagocytose context(s) recruited/peripheral immune cells are involved. retinal ganglion cell debris and appear to be highly prolifer- In contexts of peripheral nervous system degeneration and ative [47]. In histologically regenerated retinas, between 14 regeneration, two studies found a requirement for recruited and 21 dpi, distributions of microglia measurements for innate immune cells [24, 25]. In the zebrafish retina, Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 15 of 20 Fig. 9 Microglia distribute to regions containing histologically regenerated retinal neurons. A–B. Images (A and B) show z projections from saline injected (A–A”) or regenerated (B–B”) mpeg1:GFP (green) flat-mounted whole retinas labeled with L-plastin (magenta). By 14 dpi, all immune cells in the histologically regenerated retinas express the mpeg1:GFP reporter (B–B”), and the marker L-plastin can be used to label these mpeg1:GFP+ immune cells. C–D Resliced images from whole flat-mounted retinas labeled with L-plastin (magenta) are shown to demonstrate distribution of microglia in control (C) and histologically regenerated retinas at 14 dpi ouabain (D). E Densities of microglia in whole, flat mounted saline injected (n = 3) and histologically regenerated retinas (14 (n = 4) or 21 dpi (n = 3) ouabain). Densities were calculated by counting microglia in individual z stacks obtained from flat mounted retinas, which included retinal layers from the ganglion cell layer to the outer nuclear layer (representing ap- proximately 50–80 μm of depth; Fig. 7A, B in their entirety show representative regions that were quantified). Counts were normalized to 1000 μm² (area) rather than volume because regenerated retinas are thinner than controls. This area represents x and y directions parallel to the layers of the flattened retina. F Fraction of microglia in the ganglion cell layer of saline injected (n = 3) and histologically regenerated retinas (14 (n = 4) or 21 dpi (n = 3) ouabain). **p = 0.048 (two-tailed Student’s t test comparing regenerated to saline injected at indicated time point), effect size 1.56. Error bars indicate standard deviation. Scale bar in A” (applies to A–A” and B–B”)= 20 μm. Vertical scale bar in C and D =20 μm White et al. [26] found that resident microglia were mpeg1:GFP+ cells were sequenced. Further, the selective unique in their response to selective rod cell death; re- neuronal ablation used in the Oosterhof et al. [49]and cruited phagocytes did not enter the retina. However, White et al. [26] studies initiated neuronal cell death these studies were performed in a developmental context through cell-specific drug sensitivity, rather than a more (larvae), making it difficult to determine if findings may be wide-spread and tissue-disrupting chemical insult. In our tied to a developmental microenvironment or could be system, which results in neuron death throughout the en- shared with degeneration and regeneration in adult ani- tire inner retina of adult fish due to chemical insult, re- mals. Using adult zebrafish following a cell-selective abla- cruited/invading macrophages comprise a significant tion of brain neurons, Oosterhof et al. [49] concluded that portion of responding immune cells. Likewise, following peripheral macrophages did not infiltrate the injured cryolesion of the peripheral adult tench retina, infiltrating brain. This conclusion, however, was based on sequencing immune cells were documented [50]. Likely, the executed results which yielded mainly markers of microglia rather cell death pathway(s) and resulting molecules released than macrophages (based on mouse orthologues), but only into the tissue environment by dying neurons, as well as Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 16 of 20 Fig. 10 Morphological features of microglia in ganglion cell layer of histologically regenerated retinas. A–D Images show L-plastin+ microglia (magenta) in the ganglion cell layer (GCL) of saline injected (A and B) or histologically regenerated retinas at 14 or 21 dpi ouabain (C and D). Microglia in saline injected retinas appear ramified, displaying long complex processes with multiple tips (A and B). Microglia in regenerated retinas at 14 dpi appear rounded (ameboid) in shape and have few processes (C). In regenerated retinas at 21 dpi, microglia remain more ameboid than those seen in saline injected controls; however, some microglia begin to display more cellular processes (D). Microglia in regenerated retinas are occasionally associated with multiple nuclei (asterisk, C, resliced projections E–G). Chains of rod-shaped microglia are observed in regenerated retinas (D, arrowheads). E–G Individual channels in resliced projections of the microglia denoted by a asterisk in (C). Arrows indicate DAPI+ nuclei. H–K Violin plots show distributions of area (H), circularity (I), perimeter (J), and Feret diameter (K) of individually traced microglia located in the GCL at 14 and 21 dpi following injection of saline (red plots) or ouabain (regenerated, teal plots). Violin shapes show distribution of all measurements; circles within violin plots represent individual microglia. Area is reported in square microns; circularity as a value from 0 to 1.0 (1.0 indicating perfect circle); perimeter and Feret’s diameter are reported in microns. More information on morphological parameters/measurements is located in the “Methods” section. Indicated p values (two-tailed Student’s t test) in black compare measurements from regenerated to saline injected at the same time point; p values in purple (bottom of graphs) compare measurements from 21 to 14 dpi ouabain. Scale bar in A (applies to A–D)=20 μm; scale bar in E (applies to E–G)= 5 μm the extent of tissue degeneration, are important factors in PCNA+ immune cells detected at 24 h post-ouabain in- determining if peripheral immune cells are recruited to jection present at locations consistent with ramified ret- the retina. inal microglia (Fig. 4J), as well as those residing at the We also find that a significant portion of immune cells vitreal face of the ganglion cell layer (Fig. 4D, J), suggest enter the cell cycle upon ouabain-induced retinal injury, that both infiltrating immune cells and resident micro- similar to findings following acute injury in the zebrafish glia enter the cell cycle in response to ouabain-induced brain [49] and selective cell death in the retina [26]. retinal lesion. As cell division in responding macrophages Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 17 of 20 weans by approximately 3 days post-injury, PCNA signal genes were among the top hits yielded by microarray is instead detected in non-immune cells, which we have analysis in a context of ongoing rod photoreceptor death identified to be almost entirely Müller glia (Fig. 7)that are and regeneration [60]. Further work is needed to identify likely entering the cell cycle to replace the damaged neu- key inflammatory signals that are detrimental to regener- rons. Collectively, this suggests that a relatively large and ation versus those that are supportive. Likely, spatio- robust response of phagocytic macrophages and microglia temporal control of both pro- and anti-inflammatory is required to clear dead cell bodies and debris before a re- mediators is important, and these signals could differ generative response begins. depending on the particular disease or damage system. Importantly, our results support that the immune envir- Infiltration of immune cells to degenerating retina onment in mammals could potentially be modulated to suggests systemic inflammatory signals promote retinal regeneration. The act of phagocytosis by macrophages does not neces- sarily initiate an inflammatory response [51]. However, because our data supports substantial levels of immune Morphologically, microglia in histologically regenerated cell infiltration into the damaged retina following a oua- retinas appear functionally active bain lesion, this indicates that at least some systemic in- To date, most studies of inflammation and/or microglia flammatory signals are received by extra-retinal immune in retinal regeneration have focused on effects on Müller cells, even before the peak of retinal cell death. Acute in- glia proliferation [22, 26]. Further, recent studies that de- flammation appears to be a crucial signal initiating CNS pleted microglia/macrophages prior to or during the ini- regeneration and specific inflammatory mediators alone tial immune response to neuronal degeneration altered may be able to trigger neuronal progenitor proliferation the dynamics and timing of regeneration [24, 26]. How- in zebrafish [27]. In our system of retinal damage in zeb- ever, regeneration of neurons, although delayed, was ul- rafish, as well as others [26, 52], responding microglia/ timately recovered. We therefore characterized microglia macrophages localize almost exclusively to regions of ret- in histologically regenerated retinas, after the peak of inal degeneration, suggesting that inflammatory signals Müller glia proliferation [9], at the time that plexiform are spatially controlled. In particular, the localization of layers become evident, but prior to documented restor- responding macrophages in our system moves in a wave ation of visual function [10]. We found that microglia from the central region originating from the optic nerve localize to regions of regenerated neurons, rather than head and the vitreal surface of the retina at the periphery, returning to their typical distributions. In addition, to the inner nuclear layer, which correlates temporally microglia retain morphology that indicates activation for with the progression of the ouabain-induced retinal lesion. several weeks following the initial lesion, suggesting that Further, the switch in proliferative activity from microglia are active and performing functional roles. It responding immune cells to Müller glia (discussed remains to be determined when, or if, microglia return above) suggests that inflammation is indeed acute and to their original distribution patterns and ramified state, temporally controlled. and whether this coincides with restoration of retinal Hallmarks of neurodegenerative disease include function. Although visually mediated behaviors and chronic inflammation and infiltration of immune cells; ERGs are restored by 60–80 dpi [10, 33], it is not known however, it is now recognized that these features may if ERGs are restored prior to this time point. It has been have both pathological and beneficial effects [53–56]. In documented that histological organization of the regen- the context of damaged zebrafish retina, however, in- erated retina improves over time [9, 10], and it is pos- flammation appears to be beneficial rather than detri- sible that this may correlate temporally with recovery of mental, which is in contrast to what has been described retinal function. Microglia/macrophages are excellent in mammals. It is possible that zebrafish have the ability candidates to facilitate improvement of histological to rapidly switch any harmful inflammatory signals to organization as they are mobile, highly responsive, and those that are supportive of regeneration. Alternatively, highly phagocytic making them readily able to locate the molecular aspects of inflammation in zebrafish may and eliminate apoptotic or non-functional neurons, ex- differ from that of mammals [57]. Interestingly, aspects cess neurons shown to be generated in zebrafish retinal of reactive gliosis still exist during retinal regenerative regeneration [10, 61], and/or prune regenerated synaptic responses in zebrafish [58], suggesting that aspects of processes that arise during retinal regeneration. There- harmful inflammation are not completely absent, but fore, functional roles of microglia/macrophages during supportive signals likely dominate. Consistent with this retinal regeneration in zebrafish should be explored. idea, both pro- and anti-inflammatory cytokines have These functions may include known roles that are essen- been found to be important to retinal regeneration in tial to central nervous system development and normal zebrafish following light damage [31, 59] and immune processes [62–69], yet it is also possible that currently Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 18 of 20 unknown or novel functional roles may be involved and of Health under Grant #P20GM103408 (DMM), start-up funds from the University of Idaho (DMM), and NIH R21EY026814 (DLS). NIH Grant S10 are yet to be discovered. OD0108044 (DLS) funded purchase of the Nikon Andor spinning disk confocal microscope and camera. We are grateful to Ann Norton (Director, University of Idaho IBEST Optical Imaging Core), Dr. Michael Redd (University Conclusions of Utah) for the L-plastin antibody, the Zebrafish International Resource Center Interactions of the immune and central nervous system (ZIRC) for transgenic zebrafish lines and antibodies, Dr. Onesmo Balemba have seen a surge of interest, supporting the idea of im- (University of Idaho) for the helpful discussion of H&E, and members of the Stenkamp laboratory for the review of figures. mune system influence on regenerative potential [70]. With goals to regenerate damaged retinal tissue in Funding humans, further work in investigating neuro-immune in- Faculty Seed Grant from the University of Idaho (DMM). Start-up Funds, University of Idaho (DMM). teractions during successful retinal regeneration, which Institutional Development Award (IDeA) from the National Institute of occurs in a variety of vertebrate organisms including General Medical Sciences of the National Institutes of Health under Grant teleost fish, is crucial. The present work provides a #P20GM103408 (DMM). NIH R21EY026814 (DLS). framework for future studies and confirms that the zeb- NIH S10 OD0108044 (DLS). rafish retina is an excellent model to study immune- The funders had no role in the study design, data collection/analysis, or neuron interactions during neuronal regeneration. This writing of the manuscript. will allow us to uncover key inflammatory signals and Availability of data and materials the function of the innate immune system in the initi- The datasets used and/or analyzed during the current study are available ation and successful execution of retinal regeneration. from the corresponding author upon reasonable request. Further, this will allow us to probe the functional roles Authors’ contributions of microglia in the retina, both in normal and regenera- DMM designed and performed experiments, collected/analyzed the data, tive contexts. and wrote the manuscript. AGL performed the experiments and collected/ analyzed the data. DLS designed the experiments, analyzed the data, and wrote the manuscript. All authors read and approved the final manuscript. Additional file Ethics approval Additional file 1: Figure S1.A–A”. To assess co-label of mpeg1:GFP and All procedures using zebrafish were performed in compliance with protocols mpeg1:mCherry transgenes in retina, double transgenic embryos from a approved by the University of Idaho Animal Care and Use Committee cross of gl22 Tg x gl23 Tg fish at 3 dpf were fixed, washed, and eyes re- (IACUC). No human subjects were used in this work. moved then mounted for imaging. A z series (5 μm step size) was ob- tained. Images show selected z-projections from whole eyes; white lines Competing interests indicate the eye boundary. A. mpeg1:GFP signal. A’ mpeg1:mCherry sig- The authors declare that they have no competing interests. nal. A” Merge to show colabel. Scale bar in A” =20 μm. Whole retinas from adult mpeg1:GFP and mpeg1:mCherry fish were stained for L- Publisher’sNote plastin. Images show expression of individual transgenes with L-plastin Springer Nature remains neutral with regard to jurisdictional claims in (magenta, B’, and C’). Essentially all transgene signal coincides with L- published maps and institutional affiliations. plastin (B” and C”). Scale bars in B” and C= 100 μm. Figure S2. Retinal cryosections corresponding to peripheral (A) or central regions (B), adja- Received: 7 November 2017 Accepted: 30 April 2018 cent to the optic nerve head (onh, denoted by **) at 12 h post-injection (12 hpi) saline. Cryosections were stained for PCNA (green), L-plastin (ma- genta), and DAPI (blue). Scale bar in B = 20 μm, applies to both images. References Figure S3. Retinal cryosections corresponding to peripheral (A) or central 1. Bringmann A, Iandiev I, Pannicke T, Wurm A, Hollborn M, Wiedemann P, regions (B, adjacent to onh, denoted by **) at 12 hpi ouabain from Osborne NN, Reichenbach A. Cellular signaling and factors involved in mpeg1:mCherry transgenic fish stained for mpx (green) and DAPI (blue). Müller cell gliosis: neuroprotective and detrimental effects. Prog Retin Eye Scale bar in B = 20 μm, applies to both images. Figure S4. Images show Res. 2009;28:423–51. retinal cryosections from mpeg1:mCherry fish following intravitreal saline 2. Karlstetter M, Scholz R, Rutar M, Wong WT, Provis JM, Langmann T. Retinal (A) or ouabain (B) injection at 72 hpi. Cryosections were labeled with microglia: just bystander or target for therapy? Prog Retin Eye Res. 2015;45: anti-phosphorylatedhistone 3 (PH3, green) and DAPI (blue). Arrows indi- 30–57. cate PH3+ nuclei. Red signal in the outer retina is autofluorescence from 3. Wan J, Goldman D. Retina regeneration in zebrafish. Curr Opin Genet Dev. photoreceptors. Scale bar in A = 20 μm, applies to A and B. Figure S5. 2016;40:41–7. Images show retinal cryosections following intravitreal injection of oua- 4. Hamon A, Roger JE, Yang X-J, Perron M. Müller glial cell-dependent bain at 24, 48, and 72 hpi. Cryosections were stained for mpx (red) and regeneration of the neural retina: an overview across vertebrate model DAPI (blue). Scale bar = 20 μm, applies to all images. (PDF 1222 kb) systems. Dev Dyn. 2016;245:727–38. 5. Stenkamp DL. Neurogenesis in the fish retina. Int Rev Cytol. 2007;259:173–224. Abbreviations 6. Lenkowski JR, Raymond PA. Müller glia: stem cells for generation and CMZ: Ciliary marginal zone; DAPI: 4′,6-diamidino-2-phenylindole; dpi: Days regeneration of retinal neurons in teleost fish. Prog Retin Eye Res. 2014;40: post-injection; GCL: Ganglion cell layer; H&E: Hematoxylin and Eosin; 94–123. hpi: Hours post-injection; INL: Inner nuclear layer; MG: Müller glia; ONL: Outer 7. Goldman D. Müller glial cell reprogramming and retina regeneration. Nat nuclear layer; PCNA: Proliferating cell nuclear antigen; PKC: Protein kinase c; Rev Neurosci. 2014;15:431–42. TUNEL: Terminal deoxynucleotidyl transferase dUTP nick end labeling 8. Gorsuch RA, Hyde DR. Regulation of Muller glial dependent neuronal regeneration in the damaged adult zebrafish retina. Exp Eye Res. 2014; Acknowledgements 123(c):131–40. Funding for this work was provided by a Faculty Seed Grant from the 9. Fimbel SM, Montgomery JE, Burket CT, Hyde DR. Regeneration of inner University of Idaho (DMM), an Institutional Development Award (IDeA) from retinal neurons after intravitreal injection of ouabain in zebrafish. J Neurosci. the National Institute of General Medical Sciences of the National Institutes 2007;27:1712–24. Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 19 of 20 10. Sherpa T, Lankford T, McGinn TE, Hunter SS, Frey RA, Sun C, Ryan M, 35. Ellett F, Pase L, Hayman JW, Andrianopoulos A, Lieschke GJ. mpeg1 Robison BD, Stenkamp DL. Retinal regeneration is facilitated by the promoter transgenes direct macrophage-lineage expression in zebrafish. presence of surviving neurons. Dev Neurobiol. 2014;74:851–76. Blood. 2011;117:e49–56. 11. Bernardos RL, Barthel LK, Meyers JR, Raymond PA. Late-stage neuronal 36. Raymond PA, Reifler MJ, Rivlin PK. Regeneration of goldfish retina: rod progenitors in the retina are radial Müller glia that function as retinal stem precursors are a likely source of regenerated cells. J Neurobiol. 1988;19:431–63. cells. J Neurosci. 2007;27:7028–40. 37. Herbomel P, Thisse B, Thisse C. Ontogeny and behaviour of early 12. Fausett BV, Goldman D. A role for alpha1 tubulin-expressing Müller glia in macrophages in the zebrafish embryo. Development. 1999;126:3735–45. regeneration of the injured zebrafish retina. J Neurosci. 2006;26:6303–13. 38. Le Guyader D, Redd MJ, Colucci-Guyon E, Murayama E, Kissa K, Briolat V, 13. Thummel R, Kassen SC, Montgomery JE, Enright JM, Hyde DR. Inhibition of Mordelet E, Zapata A, Shinomiya H, Herbomel P. Origins and unconventional Müller glial cell division blocks regeneration of the light-damaged zebrafish behavior of neutrophils in developing zebrafish. Blood. 2008;111:132–41. retina. Dev Neurobiol. 2008;68:392–408. 39. Renninger SL, Gesemann M, Neuhauss SCF. Cone arrestin confers cone vision 14. Nagashima M, Barthel LK, Raymond PA. A self-renewing division of zebrafish of high temporal resolution in zebrafish larvae. Eur J Neurosci. 2011;33:658–67. Müller glial cells generates neuronal progenitors that require N-cadherin to 40. Herbomel P, Thisse B, Thisse C. Zebrafish early macrophages colonize regenerate retinal neurons. Development. 2013;140:4510–21. cephalic mesenchyme and developing brain, retina, and epidermis through 15. Raymond PA, Barthel LK, Bernardos RL, Perkowski JJ. Molecular a M-CSF receptor-dependent invasive process. Dev Biol. 2001;238:274–88. characterization of retinal stem cells and their niches in adult zebrafish. BMC 41. Mathias JR, Dodd ME, Walters KB, Yoo SK, Ranheim EA, Huttenlocher A. Dev Biol. 2006;6:36. Characterization of zebrafish larval inflammatory macrophages. Dev Comp 16. Thummel R, Kassen SC, Enright JM, Nelson CM, Montgomery JE, Hyde DR. Immunol. 2009;33:1212–7. Characterization of Müller glia and neuronal progenitors during adult 42. Lenkowski JR, Raymond PA. Muller glia: stem cells for generation and zebrafish retinal regeneration. Exp Eye Res. 2008;87:433–44. regeneration of retinal neurons in teleost fish. Prog Retin Eye Res. 2014; 17. Bringmann A, Pannicke T, Grosche J, Francke M, Wiedemann P, Skatchkov 40(C):94–123. SN, Osborne NN, Reichenbach A. Müller cells in the healthy and diseased 43. Kumar V, Abbas AK, Aster JC. Robbins Basic Pathology. Ninth edition. retina. Prog Retin Eye Res. 2006;25:397–424. Philadelphia, PA: Elsevier/Saunders; 2012. 18. Jadhav AP, Roesch K, Cepko CL. Development and neurogenic potential of 44. Kolaczkowska E, Kubes P. Neutrophil recruitment and function in health and Müller glial cells in the vertebrate retina. Prog Retin Eye Res. 2009;28:249–62. inflammation. Nature Publishing Group. Natur Rev Immunol. 2013;13:159– 19. Fischer AJ, Bongini R. Turning Müller glia into neural progenitors in the 75. retina. Mol Neurobiol. 2010;42:199–209. 45. Hitchcock PF, Lindsey Myhr KJ, Easter SS, Mangione-Smith R, Jones DD. Local 20. Wang M, Ma W, Zhao L, Fariss RN, Wong WT. Adaptive Müller cell responses regeneration in the retina of the goldfish. J Neurobiol. 1992;23:187–203. to microglial activation mediate neuroprotection and coordinate 46. Sun C, Galicia C, Stenkamp DL. Transcripts within rod photoreceptors of the inflammation in the retina. J Neuroinflammation. 2011;8:173. zebrafish retina. BMC Genomics. 2018;19:127. 21. Wang M, Wang X, Zhao L, Ma W, Rodriguez IR, Fariss RN, Wong WT. 47. Yuan T-F, Liang Y-X, Peng B, Lin B, So K-F. Local proliferation is the main Macroglia-microglia interactions via TSPO signaling regulates microglial source of rod microglia after optic nerve transection. Sci Rep. 2015;5:10788. activation in the mouse retina. J Neurosci. 2014;34:3793–806. 48. Petrie TA, Strand NS, Tsung-Yang C, Rabinowitz JS, Moon RT. Macrophages 22. Fischer AJ, Zelinka C, Gallina D, Scott MA, Todd L. Reactive microglia and modulate adult zebrafish tail fin regeneration. Development. 2014;141:2581–91. macrophage facilitate the formation of Müller glia-derived retinal 49. Oosterhof N, Holtman IR, Kuil LE, van der Linde HC, Boddeke EWGM, Eggen progenitors. Glia. 2014;62:1608–28. BJL, van Ham TJ. Identification of a conserved and acute 23. Wynn TA, Vannella KM. Macrophages in tissue repair, regeneration, and neurodegeneration-specific microglial transcriptome in the zebrafish. Glia. fibrosis. Immunity. 2016;44:450–62. 2016;65:138–49. 24. Carrillo SA, Anguita-Salinas C, Peña OA, Morales RA, Muñoz-Sánchez S, 50. Lillo C, Velasco A, Jimeno D, Cid E, Aijón J, Lara JM. Non-neuronal cells Muñoz-Montecinos C, Paredes-Zúñiga S, Tapia K, Allende ML. Macrophage involved in the degeneration and regeneration of the fish retina. J recruitment contributes to regeneration of mechanosensory hair cells in the Neurocytol. 2001;30:475–91. zebrafish lateral line. J Cell Biochem. 2016;117:1880–9. 51. Hochreiter-Hufford A, Ravichandran KS. Clearing the dead: apoptotic cell 25. Villegas R, Martin SM, Donnell KCO, Carrillo SA, Sagasti A, Allende ML. sensing, recognition, engulfment, and digestion. Cold Spring Harb Perspect Dynamics of degeneration and regeneration in developing zebrafish Biol. 2013;5:a008748. peripheral axons reveals a requirement for extrinsic cell types. Neural Dev. 52. Craig SEL, Calinescu A-A, Hitchcock PF. Identification of the molecular 2012;7:1–1. signatures integral to regenerating photoreceptors in the retina of the 26. White DT, Sengupta S, Saxena MT, Xu Q, Hanes J, Ding D, Ji H, Mumm JS. zebra fish. J Ocul Biol Dis Inform. 2008;1:73–84. Immunomodulation-accelerated neuronal regeneration following selective 53. Glass CK, Saijo K, Winner B, Marchetto MC, Gage FH. Mechanisms underlying rod photoreceptor cell ablation in the zebrafish retina. Proc Natl Acad Sci U inflammation in neurodegeneration. Cell. 2010;140:918–34. S A. 2017;114:E3719–28. 54. Amor S, Puentes F, Baker D, van der Valk P. Inflammation in 27. Kyritsis N, Kizil C, Zocher S, Kroehne V, Kaslin J, Freudenreich D, Iltzsche A, neurodegenerative diseases. Immunology. 2010;129:154–69. Brand M. Acute inflammation initiates the regenerative response in the 55. Amor S, Woodroofe MN. Innate and adaptive immune responses in adult zebrafish brain. Science. 2012;338:1353–6. neurodegeneration and repair. Immunology. 2014;141:287–91. 28. Sieger D, Peri F. Animal models for studying microglia: the first, the popular, 56. Rezai-Zadeh K, Gate D, Town T. CNS infiltration of peripheral immune cells: and the new. Glia. 2012;61:3–9. D-day for neurodegenerative disease? J NeuroImmune Pharmacol. 29. Langmann T. Microglia activation in retinal degeneration. J Leukoc Biol. 2009;4:462–75. 2007;81:1345–51. 57. Kyritsis N, Kizil C, Brand M. Neuroinflammation and central nervous system 30. Kettenmann H, Hanisch U-K, Noda M, Verkhratsky A. Physiology of microglia. regeneration in vertebrates. Trends Cell Biol. 2014;24:128–35. Physiol Rev. 2011;91:461–553. 58. Thomas JL, Ranski AH, Morgan GW, Thummel R. Reactive gliosis in the adult 31. Nelson CM, Ackerman KM, O'Hayer P, Bailey TJ, Gorsuch RA, Hyde DR. zebrafish retina. Exp Eye Res. 2016;143(C):98–109. Tumor necrosis factor-alpha is produced by dying retinal neurons and is 59. Lenkowski JR, Qin Z, Sifuentes CJ, Thummel R, Soto CM, Moens CB, required for Muller glia proliferation during zebrafish retinal regeneration. Raymond PA. Retinal regeneration in adult zebrafish requires regulation of J Neurosci. 2013;33:6524–39. TGFβ signaling. Glia. 2013;61:1687–97. 32. Salter MW, Stevens B. Microglia emerge as central players in brain disease. 60. Morris AC, Forbes-Osborne MA, Pillai LS, Fadool JM. Microarray analysis of Nat Med. 2017;23:1018–27. XOPS-mCFP zebrafish retina identifies genes associated with rod 33. McGinn TE, Mitchell DM, Meighan PC, Partington N, Leoni DC, Jenkins CE, photoreceptor degeneration and regeneration. Invest Ophthalmol Vis Sci. Varnum MD, Stenkamp DL. Restoration of dendritic complexity, functional 2011;52:2255–66. connectivity, and diversity of regenerated retinal bipolar neurons in adult 61. Powell C, Cornblath E, Elsaeidi F, Wan J, Goldman D. Zebrafish Müller glia- zebrafish. J Neurosci. 2018;38:120–36. derived progenitors are multipotent, exhibit proliferative biases and 34. Westerfield M. The zebrafish book. A guide for the laboratory use of regenerate excess neurons: Nature Publishing Group. Scientific Reports. zebrafish (Danio rerio). 5th ed. Eugene: University of Oregon Press; 2007. 2016;6:24851. Mitchell et al. Journal of Neuroinflammation (2018) 15:163 Page 20 of 20 62. Paolicelli RC, Bolasco G, Pagani F, Maggi L, Scianni M, Panzanelli P, Giustetto M, Ferreira TA, Guiducci E, Dumas L, Ragozzino D, Gross CT. Synaptic pruning by microglia is necessary for normal brain development. Science. 2011;333:1456–8. 63. Stevens B, Allen NJ, Vazquez LE, Howell GR, Christopherson KS, Nouri N, Micheva KD, Mehalow AK, Huberman AD, Stafford B, Sher A, Litke AM, Lambris JD, Smith SJ, John SWM, Barres BA. The classical complement cascade mediates CNS synapse elimination. Cell. 2007;131:1164–78. 64. Schafer DP, Lehrman EK, Kautzman AG, Koyama R, Mardinly AR, Yamasaki R, Ransohoff RM, Greenberg ME, Barres BA, Stevens B. Microglia sculpt postnatal neural circuits in an activity and complement-dependent manner. Neuron. 2012;74:691–705. 65. Sipe GO, Lowery RL, Tremblay M-È, Kelly EA, Lamantia CE, Majewska AK. Microglial P2Y12 is necessary for synaptic plasticity in mouse visual cortex. Nat Commun. 2016;7:10905. 66. Sierra A, Encinas JM, Deudero JJP, Chancey JH, Enikolopov G, Overstreet- Wadiche LS, Tsirka SE, Maletic-Savatic M. Microglia shape adult hippocampal neurogenesis through apoptosis-coupled phagocytosis. Cell Stem Cell. 2010; 7:483–95. 67. Huang T, Cui J, Li L, Hitchcock PF, Li Y. The role of microglia in the neurogenesis of zebrafish retina. Biochem Biophys Res Commun. 2012;421: 214–20. 68. Mazaheri F, Breus O, Durdu S, Haas P, Wittbrodt J, Gilmour D, Peri F. Distinct roles for BAI1 and TIM-4 in the engulfment of dying neurons by microglia. Nat Commun. 2014;5:1–11. 69. Peri F, Nüsslein-Volhard C. Live imaging of neuronal degradation by microglia reveals a role for v0-ATPase a1 in phagosomal fusion in vivo. Cell. 2008;133:916–27. 70. Kokaia Z, Martino G, Schwartz M, Lindvall O. Cross-talk between neural stem cells and immune cells: the key to better brain repair? Nat Neurosci. 2012; 15:1078–87.

Journal

Journal of NeuroinflammationSpringer Journals

Published: May 28, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off