Sectoral activation of glia in an inducible mouse model of autosomal dominant retinitis pigmentosa

Retinitis pigmentosa (RP) is a group of blinding disorders caused by diverse mutations, including in rhodopsin (RHO). Effective therapies have yet to be discovered. The I307N Rho mouse is a light-inducible model of autosomal dominant RP. Our purpose was to describe the glial response in this mouse model to educate future experimentation. I307N Rho mice were exposed to 20,000 lx of light for thirty minutes to induce retinal degeneration. Immunofluorescence staining of cross-sections and flat-mounts was performed to visualize the response of microglia and Müller glia. Histology was correlated with spectral-domain optical coherence tomography imaging (SD-OCT). Microglia dendrites extended between photoreceptors within two hours of induction, withdrew their dendrites between twelve hours and one day, appeared ameboid by three days, and assumed a ramified morphology by one month. Glial activation was more robust in the inferior retina and modulated across the boundary of light damage. SD-OCT hyper-reflectivity overlapped with activated microglia. Finally, microglia transiently adhered to the RPE before which RPE cells appeared dysmorphic. Our data demonstrate the spatial and temporal pattern of glial activation in the I307N Rho mouse, and correlate these patterns with SD-OCT images, assisting in interpretation of SD-OCT images in preclinical models and in human RP.

Retinitis pigmentosa (RP) is a group of monogenic blinding disorders characterized by a primary phase of rod photoreceptor apoptosis followed by cone photoreceptor death. The clinical effects of these mutations and the resulting pattern of photoreceptor loss are progressive, concentric loss of peripheral vision, and night vision, with complete blindness in advanced disease 1 . Mutations in rhodopsin (RHO) represent approximately twelve-percent of all RP cases and are usually inherited in an autosomal dominant pattern 2 . Over 200 mutations in RHO have been described to date 3 . RP remains uncurable and one barrier to the generation of therapeutic interventions is the paucity of animal models for preclinical efficacy studies since a variety of mechanisms lead to the diseased phenotype 4,5 .
The I307N Rho mouse model of light-induced adRP recapitulates some crucial aspects of patients with the B1 phenotype, including sectoral and variably penetrant retinal degeneration [6][7][8][9] . These characteristics lend to the use of the I307N Rho mouse in preclinical efficacy studies or mechanistic studies. Furthermore, the light-inducible and tunable nature of retinal degeneration in the I307N Rho mouse creates an opportunity for controlled experimentation when compared to other mouse models of RP that experience early retinal degeneration that may overlap with the final stages of retinal development in the postnatal period 7,10 .
The generation and characterization of the I307N Rho mouse model was conducted over a decade ago by Budzynski et al. and has gained popularity among research groups in recent years [6][7][8]10,11 . Gargini et al. were the first to report activation of microglia and Müller glia at an early time point following illumination of the I307N Rho mouse and suggested that morphologically distinct cells exist in the area of injury when compared to nearby, relatively unaffected retina. They proposed that the model could be an excellent resource for studying cone death, retinal inflammation, and inner retinal remodeling 8,10 . We recently characterized the time-course of light-induced retinal degeneration in this model with spectral-domain optical coherence tomography (SD-OCT). We demonstrated that a hyper-reflective signal and retinal swelling were evident in the acute period following

Results
Retinal glial cells respond to damage in a sectoral pattern within the first week after light damage. We previously showed that exposure of the awake I307N Rho mouse to overhead, bright whitelight results in inferonasal retinal degeneration 7 . Our rationale for employing overhead lighting rather than diffuse lighting via mirrored cages, which has been used elsewhere 6,8,10 , was to recapitulate the sectoral retinal injury seen in class B1 RP patients 9 . Gargini et al. have demonstrated that microglia progressed to an ameboid phenotype by day two after light exposure and that their activation occurred with regional dependence 8 . We thus expanded their analysis to determine whether microglia and Müller glia would follow a sectoral pattern of activation at other time points. To this end, we prepared frozen paraformaldehyde (PFA)-fixed cross-sections of the retina from I307N Rho mice before and one day through one month after light exposure. These sections were prepared along the superior-inferior axis to capture areas in the inferior retina that likely received the brunt of light exposure. Sections were subsequently stained with an antibody against CD45 (Fig. 1a), a receptor protein tyrosine phosphatase expressed in all nucleated hematopoietic cells 18 , or against GFAP, an intermediate filament associated protein that is expressed with Müller cell reactivity and in astrocytes 19 (Fig. 1b).
The total number of CD45-positive cells were counted in the major retinal lamina of either the inferior retina or the superior retina (Fig. 1c). Importantly, CD45 staining co-localized with the ionized calcium-binding adapter molecule 1, an accepted marker of microglia 20 (see Supplementary Fig. 1a). We chose to utilize the CD45 marker since it yielded more consistent results. Microglia showed a sectoral pattern of activation, as two-way repeatmeasures ANOVA demonstrated that retinal sector is a significant independent factor and the total number of microglia was significantly higher only in the inferior retina, but not superior retina, compared to baseline during post-hoc testing. Furthermore, there were dramatic shifts in microglia localization within the retina. With the majority of ONL already degenerated by day eight, the microglial response appeared to resolve towards the baseline distribution. Microglia did not migrate to the ONL or subretinal space (SRS) in wild-type littermates exposed to the same light damage protocol one day prior (see Supplementary Fig. 1b).
To quantify the differences in GFAP-positivity in the inferior versus superior retina, we measured the mean fluorescence intensity (MFI) of each image relative to baseline (Fig. 1d). The statistical peak of GFAP accumulation occurred in the inferior retina on days eight and fifteen after illumination. Furthermore, the independent variable of retinal sector showed significance in two-way repeat-measures ANOVA, again suggesting regional differences in glial activation. Wild-type animals exposed to the light damage protocol did not develop GFAPpositive Müller cell processes (see Supplementary Fig. 1b). These results indicate that retinal degeneration in the I307N Rho mice induced both sectoral and temporal activation of microglia and Müller glia.
Microglia undergo rapid and dramatic morphological changes and expand in number after monocyte infiltration. Next, we sought to improve our understanding of microglia in the acute phase after light challenge. I307N Rho mice were exposed to light and PFA-fixed retinal flat-mounts were prepared from early time points (e.g., within hours of the cessation of illumination) and up to one month thereafter. Flatmounts were then stained with anti-CD45, anti-GFAP, anti-Iba1 and/or isolectin B4 (IB4) and CD45-positive cell counts were performed.
Microglia migrated from the inner retina (IPL and GCL), beginning as early as two hours after light exposure, followed by repopulation of the inner retina on day three (Fig. 2a,b). When microglia at the interface of the OPL and ONL were manually isolated for 3D-reconstructions, dendrites could be seen infiltrating the ONL as early Scientific Reports | (2020) 10:16967 | https://doi.org/10.1038/s41598-020-73749-y www.nature.com/scientificreports/ as two hours and four hours (Fig. 2c); individual phagocytic events were also apparent at high magnification at these time points (see Supplementary Fig. 2a). Continued phagocytosis from twelve hours through one day (Fig. 2a, white arrows) coincided with retraction of microglia dendrites until a fulminant ameboid morphology was attained by day three, likely after profound phagocytosis (Fig. 2c). Indeed, at the day three time point, a large fraction of the nuclei appeared pyknotic ( Supplementary Fig. 2a). As time extended from day eight through one month after light exposure, microglia returned to a ramified morphology (see Supplementary Fig. 2b). Again, retinal microglia in wild-type littermates that were exposed to the same light damage protocol did not show evidence of morphological change (see Supplementary Fig. 2c). Furthermore, prominent co-localization of Iba1 and CD45 signal was observed in samples belonging to I307N Rho mice, offering further evidence that the CD45 signal was derived from microglia or macrophage (see Supplementary Fig. 2d). CD45-positive monocytic cells were visualized within the inner retina, particularly on day three (Fig. 2a, red arrows). Since monocyte-derived macrophage infiltrate the retina of other mouse models of retinal degeneration 21-24 , we counted CD45-positive monocytic cells (Fig. 3a). The highest mean infiltration of CD45positive monocytic cells occurred on days one and three post-light exposure when compared to baseline, though Figure 1. Activation of glial cells occurs in a sectoral pattern in the induced I307N Rho mouse with a more robust response in the inferior retina than superior retina. I307N Rho mice were exposed to thirty minutes of 20,000 lx of light and their eyes were extracted one day, three days, eight days, fifteen days or one month thereafter for IF-staining of PFA-fixed frozen sections. Sections were stained with (a) CD45 (green) and DAPI (blue) or (b) GFAP (green) and DAPI (blue). Images with an original magnification of 40 × were captured in the superior and inferior retina. (c) The number of microglia that appeared in the subretinal space (SRS), ONL, OPL, and inner retina (INL, IPL and GCL) were counted in two 40 × images (725 μm of length × 16 μm of thickness) in the inferior or superior retina per mouse at each time point and summed. Counts were performed by two masked observers then averaged (R 2 = 0.8286). Bar graphs represent the mean count ± s.e.m. (n = 3 per time point). (d) GFAP expression was measured as the mean fluorescence intensity (MFI) within a polygon encompassing the area between the NFL and photoreceptor inner segments using ImageJ. The relative MFI was calculated as the MFI for a given image of the retina divided by the average MFI associated with the images of the inferior retina at baseline. Bar graphs represent the mean relative MFI ± s.e.m. (n = 3 per time point). Two-way repeated-measures ANOVA with matching across the retinal sector (mixed-model ANOVA) followed by Sidak's multiple comparisons tests was performed to compare the total number of microglia or GFAP MFI at each time point to its corresponding within sector baseline. Normality was tested with the Shapiro-Wilk test of residuals. Of note, statistics were performed on raw MFI data. ns. = not significant; *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. Statistical data can be viewed in Supplementary Table S1 and S2.
Scientific Reports | (2020) 10:16967 | https://doi.org/10.1038/s41598-020-73749-y www.nature.com/scientificreports/ the magnitude of infiltration was highly variable (Fig. 3b). Indeed, the extent of CD45-positive monocytic cell infiltration likely reflects the severity of retinal degeneration, which can range from mild to severe even with the same dose of light 7 . We also observed that the CD45-positive monocytic cells appeared to infiltrate via the optic nerve head (ONH) and major retinal veins and noted that those in the day three samples appeared to be more dispersed when compared to day one (Fig. 3a).

Microglia and Müller glia organize across the front of ONL hyper-reflectivity on SD-OCT.
We previously reported that a hyper-reflective signal develops in the ONL of the I307N Rho mouse during retinal degeneration on day one 7 . Therefore, we sought to determine if the area within which the microglia manifest an activated morphology coincides with the area of hyper-reflectivity on SD-OCT at the four hour time point when we both expected a mature hyper-reflective signal and distinct microglia morphology. To this end, we utilized the retinal vasculature as a reference to produce an overlay of a CD45 and Isolectin-B4 (IB4) co-stained retinal flat-mount and its corresponding SD-OCT hyper-reflective signal (Fig. 4a). Strikingly, morphologically distinct populations of microglia were observed across the hyper-reflectivity front in the XY-dimension (Fig. 4b), which was also seen in the Z-dimension as the microglia within the focus of light injury projected into the photoreceptor layer, beyond the interface of the OPL and ONL (Fig. 4c). The co-existence of morphologically distinct populations of microglia across the boundary of degeneration was evident at every time point up to one month (see Supplementary Fig. 3). Furthermore, CD45 and GFAP signal co-modulated across the front of injury on day eight when we expected intense microglia and Müller activation, suggestive of a coordinated response (Fig. 4d).
Development of hyper-reflective SD-OCT signal, and transient retinal thinning, precedes the phase of profound microglial migration. The development of a hyper-reflective signal in SD-OCT scans is a common feature of models of light-induced retinal degeneration in the acute phase of injury 26 . Interestingly, we observed that the hyper-reflective signal was well-developed by two hours after illumination when only microglial dendrites, and not cell bodies, had permeated the ONL (Fig. 2c), suggesting that the hyperreflectivity was not derived from microglia in the ONL by themselves. To better address the contribution of microglia, if any, to the hyper-reflective signal, we performed serial SD-OCT scans of I307N Rho mice early after light exposure (Fig. 5a). Unexpectedly, we observed that the retinal thickness decreased in the inferonasal retina soon after the cessation of light exposure. We thus measured the total retinal thickness (TRT) in the superotemporal and inferonasal retina where mild to no degeneration and maximal degeneration were expected, respectively. A significant reduction in TRT of the inferonasal retina was detected from thirty minutes to two hours after illumination. In support of our previous work 7 , TRT then increased from four hours to one day after light challenge before decreasing up to day eight (Fig. 5b). We measured ONL hyper-reflectivity in parallel with ImageJ (https ://image j.nih.gov/ ij/) for each SD-OCT scan. We found that reflectivity appears as early as fifteen minutes after light exposure but achieves a statistically significant value only at forty-five minutes post illumination before continuing to increase up to day one (Fig. 5c). Of note, we previously demonstrated that the day one time point achieved the maximal hyper-reflective signal during a three day time course 7 . The magnitude of changes to TRT (both shrinking and swelling) was qualitatively proportional to the development of hyper-reflective signal (see Supplementary  Fig. 4a). Importantly, wild-type littermates subjected to the same protocol did not exhibit significant changes in TRT or hyper-reflectivity (see Supplementary Fig. 4b). Taken together, the radical shifts in TRT may be the main contributor to producing the hyper-reflective signal rather than the infiltration of microglia of the ONL.
Microglia migrate and transiently adhere to the RPE after light exposure. We previously showed that the RPE in the I307N Rho mouse remains morphologically and functionally intact fifteen days after light Figure 2. Microglia in the I307N Rho mouse migrate to the outer retina, change to an ameboid morphology, and expand in number after induction. I307N Rho mice were exposed to 20,000 lx of light for thirty minutes and eyes were enucleated before and between two hours and three days thereafter for preparation of PFA-fixed retinal flat-mounts and subsequent IF-staining for CD45 and DAPI. (a) Full-thickness images were captured in the area of light damage and Z-projections were created with ImageJ to depict CD45-positive cells (green) in the outer (OPL and ONL) or inner retina (IPL and GCL). Microglia responded by progressively changing from a ramified to an ameboid configuration. Isolated phagocytic events were observed at twelve hours and one day (white arrows). CD45-positive monocytic cells were seen in the inner retina on day three (red arrow). Original magnification = 10x. (b) Microglia were counted in the inner and outer retina in two separate images for a given animal and summed. The bar graph represents the mean count ± s.e.m. (n = 3 per time point). Microglia distribution changed as early as two hours after light damage. One-way ANOVA followed by Dunnett's multiple comparisons tests was performed to compare the inner retina, outer retina, and total counts to baseline. Normality was tested with the Shapiro-Wilk test of residuals. Statistical symbols: black circle = inner retina; white circle = outer retina; * = total count. One-symbol = p < 0.05; Two-symbols = p < 0.01; Threesymbols = p < 0.001; Four-symbols = p < 0.0001. (c) 3D reconstructions (CD45 = green; ONL = blue; INL = purple) of isolated microglia (original magnification = 100x) at the interface of the OPL and ONL were created with ImageJ. Microglia dendrites rapidly infiltrated the ONL and began to retract as early as two hours and twelve hours, respectively. A frank ameboid morphology is achieved by day three. A 125 µm segment of the SD-OCT B-scan is provided to show the development of hyper-reflectivity alongside the microglia morphological changes (ONL: blue bar; INL: purple bar). The choroid and nerve-fiber layer (NFL) are positioned superiorly and inferiorly, respectively. Statistical data can be viewed in Supplementary Table S3.   11 . In addition, subretinal microglia may protect the RPE during light-induced retinal degeneration 27 . Given these observations, we sought to determine if microglial migration to the RPE is an early event and if microglia establish permanent residence at the RPE after light exposure. We thus exposed I307N Rho mice to light and extracted their eyes on day three and up to one month after that for preparation of IF-staining of RPE flat-mounts with anti-CD45, anti-Iba1, or anti-zonula occludins-1 (ZO1), a junctional protein expressed at the periphery of RPE cells 28 (Fig. 6). The day one time point was excluded since significant amounts of degenerated retina co-segregated with RPE during dissection. We observed that microglia established transient and late residence on the RPE, particularly on day eight (Fig. 6a-e). Microglial adherence to RPE also occurred in a sectoral pattern with an inferonasal bias (Fig. 6f). The co-localization of CD45 and Iba1 signal on RPE flat-mounts again demonstrated that these cells were likely microglia or macrophage (see Supplementary Fig. 5a). The RPE appeared grossly dysmorphic on day three after light exposure with both enlarged, irregular cells and small cells. Interestingly, dysmorphia existed predominantly along the edge of light damage (see Supplementary Fig. 5b-d). By day eight and afterward, the RPE resumed a more normal cobblestone appearance with isolated areas of irregular cells. These observations suggested that RPE partially recover histologically after an initial insult.

Discussion
Our goal in utilizing the I307N Rho mouse has been to recapitulate the B1 phenotype of human adRP, which is characterized by the coexistence of areas of intense retinal degeneration with areas of healthy retina. Since many humans with the B1 phenotype experience more severe disease in the inferior and nasal retina, it has been proposed that asymmetric illumination from above (e.g. sunlight, room lighting) yields this sectoral pattern 9,17 . We previously demonstrated that exposure of the I307N Rho mouse to a bright overhead light induces retinal swelling, outer retina hyper-reflectivity, and eventual retinal thinning in SD-OCT scans with an inferonasal bias 7 . Our analysis here suggests that the inferonasal patterning is preserved at the histological level as a more profound increase in microglia and Müller glia reactivity occurred in the inferior retina. Moreover, the inferonasal hyper-reflectivity coincided with the region of the retina that contained morphologically active microglia. Our time course analysis further demonstrated the major events and timing involved in the response of microglia after light treatment, which can be correlated with the phases of retinal shrinkage, retinal swelling, and ONL loss on SD-OCT (Fig. 7).
Other research groups have induced the I307N Rho mouse with short durations of relatively low intensity, but diffuse, lighting in mirrored cages 6,8,10 . In our previous study, we found that longer durations and higher intensities were required to produce retinal degeneration when using overhead light 7 . The light damage protocol employed in the present study (30' of 20,000 lx of light) was selected because it generated severe ONL loss,  with reference to the IB4-positive retinal vasculature (red). The image was captured such that the vasculature in the fundus image and retinal flat-mount overlapped. Of note, the PFA-induced contraction and coverslipping of the flat-mount prevented optimal alignment; therefore, the scale was carefully adjusted for improved alignment. Bottom left: the background SD-OCT and hyper-reflectivity signals from the top left panel were converted to blue and purple, respectively, using ImageJ. A Z-projection was then generated to depict the area of hyper-reflectivity in a fundus view. Bottom right: the front of hyper-reflectivity was isolated and merged with the CD45-signal (green) of the flat-mount using ImageJ, which demonstrated that morphologically distinct microglial populations were separated by the hyper-reflectivity front.  26 , exhibit activated microglia with a qualitatively large size and a frank ameboid morphology, apparent on day three in our study. The rd1 and rd10 mouse models exhibit slower, but still rapid retinal degeneration that occurs in the weeks after eye opening and Total retinal thickness (TRT) was measured as the distance from the superficial aspect of the NFL to the deep aspect of the RPE. The TRT was recorded at six predetermined positions relative to the ONH, three in the superotemporal and inferonasal retina each, which were then averaged within and between eyes. (c) Longitudinal reflectivity profiles were obtained with ImageJ to measure reflectivity, which was recorded at the same three positions in the superotemporal and inferotemporal retinas as for the TRT and with the same averaging scheme. Measurements were not performed beyond the day one time point since we previously determined this to be the maximal response 7 . The bar graphs represent either the mean TRT or mean reflectivity ± s.e.m. (n = 4). For both TRT and reflectivity, two-way repeated-measures ANOVA for matching data (for both time and retinal sector variables) followed by Tukey's multiple comparisons tests were performed to determine statistical significance. The multiple comparisons tests compared each retinal segment at a given time point to its corresponding baseline (reported as asterisks above each bar graph) and to the opposite side of the retina at the same time point (reported as asterisks above each horizontal line segment). The factors of time and retinal sector, as well as their interaction, were statistically significant. Normality of residuals was assessed with Shapiro-Wilk test. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. Statistical data can be viewed in Supplementary Table S4 30,31 . In the more slowly progressive P23H RHO rat and P23H Rho mouse, microglia retract their dendrites yet maintain an intermediate ramified phenotype 27,32,33 . Thus morphological activation of microglia towards a large, ameboid phenotype likely occurs in proportion to the size of the phagocytic bolus introduced by retinal degeneration, and may have important implications for cell signaling, such as the production of IL-1β 34 . The boundary between damaged and preserved retina is a unique feature in the induced I307N Rho mouse and showed a steep gradient in microglia morphology, GFAP-expression by Müller glia, and partially reversible RPE injury. This organization points to potential, unique interactions among these cell types in close space 35 that may not occur in models of global retinal degeneration that exhibit a more uniform distribution of microglia, like the P23H RHO rat 36 . Though we established correlation between Müller glia and microglia, we did not investigate the mechanisms of interaction between these cell types. However, crosstalk between Müller glia and microglia has been reported to help determine photoreceptor survival in other models 16,[37][38][39] . Müller glia are capable of providing crucial neuroprotective support to photoreceptors, for example, by secreting LIF 40 , but can also contribute to the inflammatory response via secretion of cytokines 41,42 . Microglia similarly have shown evidence for neuroprotective and neurotoxic effects on photoreceptor survival 13,43 . Even more, subretinal microglia may protect the RPE during rapid light-induced retinal degeneration 27 . The seemingly opposing capabilities of microglia and Müller glia may appear incongruous during pan-retinal degeneration, but may be reconciled in the context of the lesion boundary in the I307N Rho mouse where pro-survival pathways and clearance of unsalvageable retina must co-exist in close but distinct spaces. Interestingly, Stefanov et al. noted that horizontal cell remodeling is most severe at the boundary of retinal degeneration in the I307N Rho mouse, which is suggestive of unique signaling events 10 .
Our data indicate that the number of microglia and/or macrophage increases after the infiltration of CD45positive monocytic cells via the inner retinal veins and ONH. Our analysis, however, did not differentiate between endogenous microglia, infiltrating macrophage, and monocyte-derived macrophage [22][23][24][44][45][46] . Genetic techniques to differentiate microglia or monocyte-derived macrophage based on endogenous fluorescence have been employed by others 21,22 . These methodologies alongside single-cell RNA-seq have demonstrated that microglia and monocyte-derived macrophage may behave differently during retinal degeneration, both in terms of tropism www.nature.com/scientificreports/ within the retina, as endogenous microglia have displayed a bias towards populating the SRS in the P23H Rho knock-in mouse, and gene expression profiles 22,23,27 . Future studies with the I307N Rho mouse ought to examine the relative abundance of endogenous microglia versus monocyte-derived macrophage across the boundary of light-damage and assay for differences in phenotype 31 . We found that microglia in the I307N Rho mouse reduce in number, move towards homeostatic positioning among the retinal layers, and exhibit a more ramified phenotype in the weeks following light exposure.  Fig. 4 (red and blue). The dashed curves are based on previous observations 7 while the dotted curves represent predictions. The data used for the red and blue curves were transformed to fit within the range of the green curve. SD-OCT pathology can be organized into phases of retinal shrinkage, retinal swelling, and ONL loss, wherein all phases exhibit hyper-reflectivity. Micrographs of microglia in the outer retina demonstrate the typical morphology at each of the time points. (b) The major events of the microglia response are listed in the table provided and the shaded blocks (green) represent the presence of an event at one or multiple time points. When possible, a gradient was applied to the shaded area to indicate an increasing (white to green) or decreasing (green to white) effect. After light treatment, microglia (1) orient their dendrites towards photoreceptors, penetrate the ONL, and initiate phagocytosis as early as two hours and through four hours, (2) progressively migrate to the outer retina within two hours and through one day, (3) retract their dendrites and continue phagocytosis by twelve hours and through one day, (4) attain an ameboid morphology on day three and continue to phagocytose material until the ONL is cleared, (5) increase in number on days three through eight, (6) transiently adhere to the RPE on days eight through fifteen, and (7) return to a ramified phenotype beginning on day eight. Other important events include infiltration of the retina by CD45-positive monocytic cells on days one and three, and reconstitution of the inner retina population of microglia/macrophages on day three. These events predominantly overlap with retinal swelling and ONL loss phases on SD-OCT. The figure was created with Adobe Illustrator CC 2019 (www.adobe .com).

Scientific Reports
| (2020) 10:16967 | https://doi.org/10.1038/s41598-020-73749-y www.nature.com/scientificreports/ Interestingly, microglia are capable of defaulting to the baseline, homeostatic transcriptome weeks after acute inflammation in endotoxin uveitis 47 . On the other hand, microglia in slowly progressive forms of RP may remain pro-inflammatory even after significant photoreceptor loss, which has been demonstrated in the P23H RHO rat 48 . Data from our present and past work with the I307N Rho model demonstrate that the ONL within the focus of light-induced retina can stably persist when mild retinal degeneration occurs 7 . Given this finding, determining whether microglia in the I307N Rho mouse can re-establish a homeostatic transcriptome after light exposure and what signals are required for such a transition in phenotype may be of value in illuminating novel therapeutic targets that could be exploited for the treatment of slowly progressive forms of RP. As we suggested in our previous work, the origin of the outer retina hyper-reflectivity may illuminate potential therapeutic targets 7 . Though it is well-documented that retinal swelling occurs during light-induced retinal degeneration 26,49,50 , we found that TRT decreased when hyper-reflectivity initiated in the first hour after cessation of light exposure. Light adaptation in mice induces changes in retinal structures that are apparent via SD-OCT, including expansion of the SRS and photoreceptor outer segments, in part due to changes in osmotic forces and water movement during phototransduction [51][52][53][54] . Since overactivation of phototransduction may underlie retinal degeneration in the I307N Rho mouse 6 , the massive reduction in retinal thickness and hyper-reflectivity that occurs very early after light exposure could involve dysregulated water movement after a period of increased respiratory demand 54 or G-protein-induced swelling of outer segments 53 . Fluid balance and retinal reflectivity may also be altered by Müller glia activity via modulation in the expression of aquaporin-4, the Kir4.1 potassium channel, or transient receptor potential isoform 4 (TRPV4) as well as changes to the integrity of the blood-retinalbarrier changes 49,55,56 . To this end, the I307N Rho mouse may help provide a link between novel pathological findings on SD-OCT, including the origin of retinal hyper-reflectivity, shrinking, and swelling, and associated mechanisms during retinal degeneration.
Our study presents an in-depth time course of the glial response during light-induced degeneration in the I307N Rho mouse and correlates these findings with SD-OCT imaging. Continued characterization of the I307N Rho mouse may uncover new insights regarding the B1 phenotype of adRP associated with mutations in RHO in human. Furthermore, comparison of the glial response after light-induction in the I307N Rho mouse and during retinal degeneration in animal models of slowly progressive and/or global RP could provide new perspectives for evaluating glial biology. Light damage protocol. A detailed protocol for light damage can be found in Massengill et al. 7 . Male and female mice were used for experimentation. Briefly, eight-to twelve-week old mice were dark adapted overnight in preparation for light damage. The following day, phenylephrine (Paragon BioTeck Inc., Portland, OR, USA) and atropine (Akorn, Lake Forest, IL, USA) were applied to the eye followed by a fifteen minute incubation, phenylephrine was then applied a second time followed by a second fifteen minute incubation, and phenylephrine was applied a third time immediately prior to challenging the mice with light. Ambulatory mice with maximally dilated eyes were placed into a cage with LED-lights fastened to its roof and the light was modulated to an intensity of 20,000 lx. Mice were exposed to the light for a thirty minute period and were utilized immediately for an experiment or returned to the 12:12-h dim-red light: dark lighting cycle. Light damage occurred between 4:00 PM and 12:00 AM.   Statistical analysis. Graphs were generated and statistical analyses were performed using GraphPad Prism Software, Version 8 (GraphPad Software Inc., San Diego, CA, USA). Normality of raw populations and residuals was assessed with the Shapiro-Wilk test. Two-way repeated measures ANOVA with matching for one factor (mixed-model ANOVA) followed by Sidak's multiple comparisons test was employed to compare differences in retinal sector within the same animals, but differences across time between separate animals (Fig. 1c,d). Twoway repeated measures ANOVA with matching for both factors (retinal sector & time) followed by Tukey's multiple comparisons test was used to compare differences in retinal sector across time within the same animal (Fig. 5b,c). Parametric one-way ANOVA with Dunnett's multiple comparisons test (Fig. 2b) and non-parametric one-way ANOVA (Kruskal-Wallis test) with Dunn's multiple comparisons test (Fig. 3b) was employed to compare differences in independent samples across time when normality was or was not met, respectively. An alpha level of 0.05 was employed.

Spectral-domain optical coherence tomography (SD-OCT). Eyes
Scientific Reports | (2020) 10:16967 | https://doi.org/10.1038/s41598-020-73749-y www.nature.com/scientificreports/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.