The Timecourses of Functional, Morphological, and Molecular Changes Triggered by Light Exposure in Sprague–Dawley Rat Retinas

Retinal neurodegeneration can impair visual perception at different levels, involving not only photoreceptors, which are the most metabolically active cells, but also the inner retina. Compensatory mechanisms may hide the first signs of these impairments and reduce the likelihood of receiving timely treatments. Therefore, it is essential to characterize the early critical steps in the neurodegenerative progression to design adequate therapies. This paper describes and correlates early morphological and biochemical changes in the degenerating retina with in vivo functional analysis of retinal activity and investigates the progression of neurodegenerative stages for up to 7 months. For these purposes, Sprague–Dawley rats were exposed to 1000 lux light either for different durations (12 h to 24 h) and examined seven days afterward (7d) or for a fixed duration (24 h) and monitored at various time points following the exposure (up to 210d). Flash electroretinogram (fERG) recordings were correlated with morphological and histological analyses to evaluate outer and inner retinal disruptions, gliosis, trophic factor release, and microglial activation. Twelve hours or fifteen hours of exposure to constant light led to a severe retinal dysfunction with only minor morphological changes. Therefore, early pathological signs might be hidden by compensatory mechanisms that silence retinal dysfunction, accounting for the discrepancy between photoreceptor loss and retinal functional output. The long-term analysis showed a transient functional recovery, maximum at 45 days, despite a progressive loss of photoreceptors and coincident increases in glial fibrillary acidic protein (GFAP) and basic fibroblast growth factor-2 (bFGF-2) expression. Interestingly, the progression of the disease presented different patterns in the dorsal and ventral retina. The information acquired gives us the potential to develop a specific diagnostic tool to monitor the disease’s progression and treatment efficacy.


Introduction
The quality of vision dramatically impacts the quality of human life [1]. Many ocular pathologies induced by several factors, such as gene mutations, environmental stresses, metabolic dysfunction, and aging, lead to reduced visual performance and eventually complete blindness caused by inflammation, mitochondrial dysfunction, synaptic remodeling, and neuronal death [2]. These pathologies include a heterogeneous group of photoreceptor degenerations whereby populations of rods and cones are distinctly affected, i.e., rods are primarily involved in retinitis pigmentosa (RP) [3], cones in age-related macular degenerations (AMD) [3], and both rods and cones simultaneously in Leber congenital amaurosis (LCA) [4] and Stargardt's disease (STGD1, autosomal recessive [5]). Although a variety 1.
To characterize and correlate early morphological and biochemical changes occurring in the degenerating retina with in vivo functional analysis of retinal activity.

2.
To investigate the progression of neurodegenerative stages.
In the current work, we used a consolidated rat model of light-induced retinal degeneration, previously shown to mimic some aspects of AMD (the presence of oxidative stress [34], inflammatory processes [35], and photoreceptor death [36]). Indeed, apart from the relevance of this model in AMD research, we exploited its well-known characteristics to follow a neurodegenerative process that starts in the dorsal retina and spreads over time. Specifically, by using two different protocols (<24 h and 24 h of light exposure) and a combination of histological and electrophysiological investigations, we first determined the minimal duration of bright light exposure (1000 lux, white light) necessary to induce detectable retinal functional alterations. Interestingly, our findings revealed that functional changes precede anatomical ones. In the second stage, we investigated the temporal course of the neurodegeneration following 24 h of light exposure while extending our analyses to the inner retina by relying on antibodies against cell-type-specific markers and trophic factors.
Altogether, the knowledges acquired will allow us to extrapolate information (i.e., emergent ERG signals) that might help to detect the pathology through functional tests before a large portion of photoreceptors is already injured, the retina severely damaged, and neuronal circuitry remodeled. Secondly, following alterations over time will provide valuable insights for dissecting general retinal degeneration pathways that might help develop treatments based on different dystrophy stages.

Electrophysiological Recordings and Data Analysis
We performed flash electroretinogram (fERG) recordings in controls and evaluated retinal function in each experimental group following the light exposure treatment. To minimize variability among groups, electrophysiological recordings were performed before exposing animals to LD (a pull of these recordings were used as HC, assuming no agerelated changes in our temporal window of interest [37,38]). Albino rats were previously dark-adapted overnight, and electroretinograms were recorded in a dark room [14]. Briefly, animals were anesthetized with an intraperitoneal injection of ketamine/xylazine (keta-vet 100 mg/mL, Intervet production Srl, and xylazine 1 g, Sigma Co., at 10 and 1.2 mg/100 g, respectively). Corneas were anesthetized with a drop of Novocaine, and pupils were dilated Cells 2021, 10, 1561 4 of 28 with 1.0% mydriacyl tropicamide. Animals were mounted on a stereotaxic apparatus and positioned inside the Ganzfeld dome's opening (Biomedica Mangoni, Pisa, Italy). Body temperature was maintained at 37.5 • C using a heating pad controlled by a rectal temperature probe. Recordings were carried out for both eyes simultaneously, with a platinum electrode loop being placed on each cornea, and individually considered. The reference electrodes were inserted subcutaneously in the proximity of the eyes, and the ground electrode was inserted in the anterior scalp, between the eyes. The standard ERG protocol advocated by the ISCEV (International Society for Clinical Electrophysiology of Vision) was used [39]. Responses were recorded at progressively brighter short flashes (0.001-100 cd*s/m 2 range, from scotopic to photopic) over 450 ms, plus 50 ms of pre-trial baseline. Responses were amplified differentially, bandpass filtered at 0.3-300 Hz, and sampled at 16.3 kHz. To reduce variability and background noise, responses were averaged (n = 3 per each luminance), with an inter-stimulus interval (ISI) ranging from 60 s for lower intensities to 5 min for the three brightest flashes. We evaluated the amplitudes (in µV) of the a-wave (baseline to the first negative deflection), b-wave (from a-wave peak to positive b-wave peak), and oscillatory potentials (the high-frequency wavelengths residing on the leading edge of the b-wave separated from the ERG using a high-pass filter. The amplitudes of wavelets OP1-4, measured from the trough to the peak of each response component, were summed) (see Figure 5a,f). For photopic recordings (light-adapted ERG, cone response), rats were light-adapted to background illumination (30 cd*s/m 2 , 10 min), and 20 replicate responses elicited by white light (100 cd*s/m 2 ) at a pulse frequency of 1 Hz were averaged. A 30 Hz photopic flicker response was also obtained using the same background illumination and stimulus intensity. One hundred replicate responses were averaged. The first and the second harmonic magnitudes were determined by performing FFT analysis with a compiled routine (see Figure 6a,c).
In some cases, measurements were obtained from the same animal at successive survival times following light exposure cessation (not more than three ERG sessions, including the initial one, were performed in total for a single animal). When the amplitude of either an a-wave or a b-wave was not higher than at least 1 SD (not measurable) across the entire recording session, the retina was discarded from the subsequent analyses (more common in the later stages of the degeneration, i.e., LD24h210d). Custom-written procedures in IGOR Pro 6.3 (Wavemetrics, Lake Oswego, OR, USA) software were used to analyze electrophysiological data. The distributions of ERG response across several experimental groups at different light intensities were described by means and standard errors. Two-way ANOVA followed by the Tukey post hoc test was used to compare the ERG parameters (a, b, and OP amplitudes) between groups of animals for each flash intensity used. One-way ANOVA followed by the Tukey post hoc test was used to analyze statistical significance in the photopic recording (photopic ERG and flicker ERG, a single flash intensity used).

Tissue Processing, Morphological Analyses, and Immunostaining Protocols
At the end of the last recording session, rats were sacrificed by CO 2 inhalation. Both eyes were enucleated and fixed in 4% paraformaldehyde for 6 h at 4 • C and washed in 0.1 M phosphate-buffered saline (PBS, pH 7.4). The eyes were further processed to obtain an "eyecup" consisting of the sclera, choroid, and retina. Eyecups were cryoprotected by immersion in 10%, 20%, and 30% sucrose solution overnight, embedded in OCT compound (Tissue Tek; Qiagen, Valencia, CA, USA), and frozen in liquid nitrogen. With eyecups marked to maintain proper orientation, cross-sections 20 µm thick were obtained for each retina (CM1850 Cryostat; Leica, Wetzlar, Germany), collected on gelatine poly-L-lysine coated slides, and stored at −20 • C until processed. The analyses were performed on the sections crossing the optic nerve in the dorsal-ventral direction (~8-10 central sections per retina were collected on different slides) to minimize variations in the retinal length and position.
The morphological analysis was carried out by labeling the nuclei with the DNAspecific dye bisbenzimide and measuring the thickness of the outer nuclear layer (ONL) across the entire retinal extension from dorsal to ventral crossing the papilla, following a previously described procedure [40]. Shortly, histological reconstructions were obtained by joining consecutive acquired micrographs (each long~380 µm, usually 10-12 per hemiretina). To facilitate comparison between retinas, each retina was divided into 10 dorsal and 10 ventral fields while taking the optic nerve as a reference. Analyses included: (1) the ratio of ONL to the total retinal thickness to measure ONL thickness (rather than the absolute thickness of the ONL (µm), to compensate for possible oblique sectioning); (2) the number of photoreceptor rows in the ONL; (3) damaged ONL length. ONL thickness and ONL rows were calculated as the averages of 4 equi-spaced measurements from each of the 20 fields (to account for slight differences within the same area). The averages of the ONL thickness and rows measured in the 10 dorsal or ventral fields were calculated for each section. We then evaluated the extent of damage in the photoreceptor layer by calculating the ratio between "damaged ONL length" (clearly detectable because of the presence of "rosette" structures [41]) and the total dorsal retinal length. Measurements were performed using ImageJ 2.0 software (Rasband, W.S., ImageJ, U. S. National Institutes of Health, Bethesda, MD, USA).
Together with nuclear labeling, the slides were immunolabeled with antibodies according to Table 2. Bovine serum albumin (BSA) (Sigma-Aldrich, St. Louis, MO, USA) or specific sera were used to block nonspecific binding sites, and Triton X−100 (Sigma-Aldrich) was used as a detergent, when appropriate. Primary antibodies were omitted to control for nonspecific binding of the secondary antibody. Only one of the two eyes was used for a single marker or histological analyses; therefore, the number of analyzed sections is equivalent to the number of animals (consequently, in figure legends, we reported N. as the number of analyzed retinal slices).

Microscopy and Image Analysis
Images of the retina were acquired with Leica TCS SP5 (Leica Microsystems, Wetzlar, Germany) or Nikon80i (Nikon, Tokyo), and acquisition parameters were kept constant throughout each imaging session. Measurements of fluorescence intensity and signal localization were made with ImageJ Fiji software (National Institutes of Health, Bethesda, MD, USA).
The L/M opsin signal was calculated following the method used by Rutar et al. [41]. Briefly, ten images (5 dorsal and 5 ventral) equally spaced were taken under 20× magnification from a total of 5 or 6 retinal sections from different rats for each experimental condition. We counted external outer segments (OSs) of cone photoreceptors. The number of positive cells is expressed as a mean of labeled cells per field or as their sum, respectively, for the dorsal and ventral retina.
Spatial distribution and the total number of Iba1 + microglial cells were analyzed in the ganglion cell layer and inner plexiform layer (GCL + IPL), in the inner nuclear layer, and outer plexiform layer (INL + OPL), and in the outer nuclear layer (ONL). Each retinal layer was determined according to the nuclear staining. The number of Iba1 + cells was manually counted throughout the entire section from dorsal to ventral (10 images per hemisphere acquired at 40× magnification, each micrograph covered~380 µm of retinal length). Results are given as the total number of microglial cells (Iba1 + ) per layer and per section.
Densitometric analysis of GFAP and bFGF-2 fluorescent signals (N = 3-6 and N = 5-7 retinal sections, quantified in Figures 3 and 7, respectively) was performed using customwritten algorithms in ImageJ Fiji software (NIH, Maryland, USA). The percentage of areas positive for bFGF-2 or GFAP over the total area of the ONL or the entire retinal area, respectively, was determined by applying a threshold and calculated as described in Antognazza et al. [44].
Cholinergic neurons in the retina, known as starburst amacrine cells (SBAC), were labeled with anti-ChAT antibody. The integrity of their processes, located in the inner plexiform layer (IPL), was quantified in two areas of the retina (dorsal and ventral, respectively) relatively to the total length (N = 5 retinal sections from different rats for each experimental condition; HC and 24 h light-exposed rats). Cell bodies density was also quantified (data not shown because of lack of statistical significance).
Synaptophysin, a protein expressed in the presynaptic vesicles, was qualitatively assessed (acquired under 60x magnification) in 5 retinal sections for each experimental condition, and representative immunofluorescence images are shown in the figure.

Statistical Analysis
One or two-way ANOVA was used to evaluate the effects of different LD durations and recovery periods. When the significance level was 0.05 or less, post hoc pairwise comparisons were performed using Tukey's test. Data are reported as mean ± standard error of the mean (SEM). Values of p < 0.05 were considered to be statistically significant. All statistical analyses were performed using Prism7 software (Graph Pad, San Diego, CA, USA). A detailed description of each test can be found in figure legends.

Different Durations of Light Exposure: Discrepancies between Functional Output and the Extent of Structural Damage
This section examines the effects of different light exposure durations (i.e., 12 h, 15 h, 18 h, 24 h at 1000 lux) on retina functional, morphological, and molecular states, assessed 7 days after the treatment. For the sake of simplicity, we refer to 12 h as short-term exposure (compared to 24 h). . Surprisingly, our analyses revealed no major functional differences between light damaged experimental groups, with a few exceptions. Indeed, even the short-term exposure induced a significant reduction of a-wave and b-wave amplitudes compared to the healthy control (Figure 1a,c). Interestingly, only the a-wave amplitude, obtained in response to the brightest stimulus (100 cd*s/m 2 ), was better preserved in LD12h and LD15h compared to LD24h (Figure 1a,b for magnification; a-wave amplitude was~2.5 times higher in LD12h and LD15h vs. LD24h, p = 0.0011 and p = 0.0177 respectively). Instead, our evaluation did not show any significant difference between retinas exposed to light with regard to the b-wave amplitude; as expected, they were all different compared to the HC (Figure 1c,d; p < 0.0001 for all stimulus intensities). In particular, fERG b-wave amplitude decreased by 30% or 15% in rats exposed to 12 or 24 h, respectively, compared to HC (LD12h7d = 287 ± 26 µV and LD24h7d = 150 ± 13 µV vs. HC = 1061 ± 76 µV at 100 cd*s/m 2 ). Statistical significance is represented as follows: *, # , $ p ≤ 0.05, **, ## p ≤ 0.01, ***, ### p ≤ 0.001, **** p ≤ 0.0001; *, # , $ are in comparison to HC, LD12h7d, LD15h7d, and LD18h7d, respectively; two-way and one-way ANOVA with Tukey's post hoc test were used to assess differences in the panels (a-g), respectively. Values represent means ± standard errors of the means (SEM). Between 6 and 12 retinas were recorded per group.

Outer Nuclear Layer (ONL) Disruption Does Not Correlate with Detectable
Functional Alterations Figure 2a indicates that exposure to high intensity light for different durations of ti (from at least 12 h to 24 h) caused a gradual alteration of the photoreceptor layer integr in the dorsal retina, which has extensively been shown to be the more susceptible . A-wave amplitude in the HC was significantly higher than all groups exposed to light for all luminances brighter than 0.1 cd*s/m 2 . B-wave amplitude was significantly higher than all experimental groups at all luminances. (b,d) Panels b and d represent a-wave and b-wave amplitudes measured at 100 cd*s/m 2 . LD24h7d a-wave amplitude at 100 cd*s/m 2 was significantly reduced compared to LD12h7d. ## p = 0.0011, and LD15h7d, $ p = 0.0177; two-way ANOVA with Tukey's post hoc test. (e) Amplitudes of the OPs sum in HC and different experimental groups are plotted as functions of stimulus intensity (0.001-100 cd*s/m 2 ). On average, OPs sum amplitudes were significantly affected by all durations of light exposure. (f) Panel f represents a large magnification of panel (e). The effect of light exposure was milder after 12 h; indeed, OP amplitude was significantly preserved in LD12h7d vs. LD24h7d at intensities brighter than 0.1 cd*s/m 2 . (g) The values represent the mean response amplitudes of photopic b-waves. Statistical significance is represented as follows: *, # , $ p ≤ 0.05, **, ## , ++ p ≤ 0.01, ***, ### p ≤ 0.001, **** p ≤ 0.0001; *, # , $ , + are in comparison to HC, LD12h7d, LD15h7d, and LD18h7d, respectively; two-way and one-way ANOVA with Tukey's post hoc test were used to assess differences in the panels (a-g), respectively. Values represent means ± standard errors of the means (SEM). Between 6 and 12 retinas were recorded per group.
It is worth highlighting that by studying oscillatory potentials (OPs), an indicator of the inner retinal activity [45], some differences were unexpectedly found between the lightexposed experimental groups. As observed for the other two analyzed ERG parameters, the light damage paradigm caused a significant reduction of OP sum amplitude at all stimuli intensities (i.e., to~35%,~32%, and~44% respectively in LD15 h, 18 h, and 24 h groups compared to HC at 10 cd*s/m 2 ; LD15h7d = 250 ± 32 µV, LD18h7d = 276 ± 48 µV and LD24h7d = 203 ± 15 µV vs. HC = 896 ± 100 µV, in Figure 1e), but with a few exceptions. Indeed 12 h led to a significant reduction of the OP sum compared to HC only at stimuli intensities between 3 and 30 cd*s/m 2 (reduction by~22% at 10 cd*s/m 2 ; LD12h7d = 415 ± 45 µV and HC = 896 ± 100 µV). What was more interesting was the difference between the two groups LD12h7d and LD24h7d (Figure 1f for magnification; at light intensities higher than 0.1 cd*s/m 2 ), revealing evidence for a different time course and extent of light-induced lesions that interests the inner retina.
In contrast, no differences between groups (except for the reduction compared to HC) were observed when cone functionality was assessed by photopic-ERG. Indeed, 7 days after 12 h, 15 h, 18 h, and 24 h of light damage, the photopic b-wave amplitudes repre-sented~33%,~30%,~24%, and~26% of the HC one, respectively (LD12h7d = 42 ± 5 µV, LD15h7d = 38 ± 5 µV, LD18h7d = 31 ± 10 µV, LD24h7d = 33 ± 7 µV and HC = 128 ± 18 µV; all significant vs. HC, Figure 1g). Figure 2a indicates that exposure to high intensity light for different durations of time (from at least 12 h to 24 h) caused a gradual alteration of the photoreceptor layer integrity in the dorsal retina, which has extensively been shown to be the more susceptible to induced neuronal death [41,46]. To analyze tissue structure, the cross-sectioned retinas, from the dorsal to ventral quadrant, were stained with Hoechst and the ONL thickness relative to the entire retina, and the number of nuclei rows in the ONL were analyzed. In the representative retinal sections in Figure 2a (Figure 2d). It is plausible to think that the loss of a limited number of photoreceptors does not change the ratio of ONL/total retinal thickness; indeed, the photoreceptor nuclei rows count may be a more convenient measure of mild damage, though does not account for possible oblique sectioning. As expected, the increase in light exposure duration produced a more significant thinning of the outer retinal layers, specifically in the hot spot area in the dorsal hemiretina ( Figure 2c; LD24h7d significantly different compared to HC and LD12h7d, p = 0.004 and 0.0038, respectively).

Outer Nuclear Layer (ONL) Disruption Does Not Correlate with Detectable Functional Alterations
In contrast, a significantly reduced number of nuclei rows compared to the HC was observed within the dorsal hemiretina also in the LD15h7d and LD18h7d groups (Figure 2e; LD15h7d, LD18h7d, and LD24h7d vs. HC, p = 0.0136, 0.0006 and <0.0001). This result is also significant when comparing LD24h7d to LD12h7d (respectively 4.6 vs. 8.4 rows, p = 0.0104). ONL disruption was then correlated with the a-wave amplitude (photoreceptors driven), and no correlation was found (Figure 2h, p = 0.20, and p = 0.6617 for the ONL thickness and N. rows vs. a-wave amplitude, respectively. Only 19 (LD12h7d = 5, LD15h7d = 4, LD18h7d = 4, and LD24h7d = 6) and 13 retinas (LD12h7d = 3, LD15h7d = 3, LD18h7d = 3, and LD24h7d = 4), for which both functional and morphological features were assessed, were included in this analysis. No HC was included to study the effect of light exposure). This result demonstrates that ERG prematurely diagnoses ONL matrix disorganization. Further analyses have been performed to measure the hot spot extension, evaluated as damaged ONL length/dorsal retinal length. Figure 2f presents a histological reconstruction of the dorsal hemiretina central section, with retinal damage extensions labeled with white arrows. We observed that the hot spot length increased with light exposure durations, covering up to 60% of the dorsal retina in the LD24h7d (Figure 2g). To evaluate the effect of different light exposure durations, corresponding retina areas (one in the dorsal and one in the ventral) were selected to determine the relative GFAPlabelled regions [47]. For this purpose, the images were processed with ImageJ software's threshold tool (see Materials and Methods). Areas of the retina marked with the threshold overlay (GFAP + astrocytes plus GFAP + end-foot of the MC; Figure 3a) were included in the measurement among study groups. Results obtained at 7d following 12 h, 15 h, 18 h, and 24 h light damage are shown quantitatively in Figure 3c. From the one-way ANOVA, performed separately for the dorsal and ventral retina, significant differences were found in the dorsal retina of all light-exposed groups compared to the HC (p = 0.0155, 0.0078, 0.0023, and 0.0018 for LD12,15,18, and 24 h vs. HC, respectively). Hence, even 12 h of light exposure was sufficient to produce retinal stress, in accordance with the ERG results. The upregulation of GFAP also extended in the ventral retina, which was significant only for the LD24h7d group compared to the HC (p = 0.0048). As expected, stressed retinas also increase the expression levels of trophic factors, such as bFGF-2. Notably, the release of bFGF-2 in the ONL negatively correlates with the b-wave amplitude [40]. In control retinas, bFGF-2 expression was confined to the Müller cell bodies in the INL (Figure 3b, first micrograph). Light damage induced its progressive release in the ONL (the protein was visible around photoreceptor cell bodies), and even though exclusively significant for the LD24h7d (p = 0.0288 vs. HC), it is noticeable that bFGF-2 levels increased in the dorsal retina after light exposure treatment. Indeed, a significant trend (p = 0.0168, one-way ANOVA with Trend) relates bFGF-2 release in the ONL to the duration of light exposure (no linear trend was revealed in the ventral side).

Microglia Increased in the Dorsal Retina after Exposure to Constant Light
The immune response is triggered along with the degeneration of the outer retina and macrogliosis (assessed by anti-GFAP staining) [48]. Thus, we characterized the effect of light exposure on microglia distribution in the retina using Iba1 as a marker. Representative images of the dorsal retina from HC and the four experimental groups (different light exposure durations) are shown in Figure 4a (Iba1 + cells in green). As expected, [49], under physiological conditions, microglia that showed typical resting characteristics were located in the inner retina in HC (Figure 4b,d, GCL + IPL, and INL + OPL), and no cells were found in the ONL in the dorsal and ventral retina (GCL + IPL > ONL + OPL > ONL; p < 0.01). Exposure to light promoted their morphological change (from ramified to ameboid, depicted in Figure 4a with the orange and white arrows, respectively) and migration to the ONL (green dashed arrow in Figure 4a, LD12h7d) [50]. Detailed analyses of Iba1 + cells in the dorsal and ventral retina and across retinal layers are reported in Figure 4b-e. In the retinas of LD groups, we found that Iba1 + cells accumulate significantly in the ONL in the dorsal retina after 7 days from 18 h or 24 h of light exposure (LD18h7d vs. HC, p = 0.0013; LD24h7d vs. HC, p < 0.0001; one-way ANOVA). Overall, when the total number of microglia across all layers was assessed, a significant increase in Iba1 + cells was found only in the dorsal retina in LD24h7d compared to HC, LD12h7d, and LD15h7d (p < 0.0001, p = 0.0052, p = 0.0287, respectively). Only 24 h of light exposure induced a significant difference between the dorsal and ventral retina (t-test, LD24h7d dorsal vs. ventral, p = 0.04). The immune response is triggered along with the degeneration of the outer retina and macrogliosis (assessed by anti-GFAP staining) [48]. Thus, we characterized the effect Statistical analyses were performed with the analysis of variance (one-way ANOVA) followed by a Tukey's post hoc test; one-way ANOVA with Trend was also used to analyze data in panel d. Statistically significant differences are represented as follows: * p ≤ 0.05, ** p ≤ 0.01, respectively, compared to HC. Data are shown as the means ± SEM of at least N = 3 retinas from different rats for each experimental condition. ANOVA). Overall, when the total number of microglia across all layers was assessed, a significant increase in Iba1 + cells was found only in the dorsal retina in LD24h7d compared to HC, LD12h7d, and LD15h7d (p < 0.0001, p = 0.0052, p = 0.0287, respectively). Only 24 h of light exposure induced a significant difference between the dorsal and ventral retina (ttest, LD24h7d dorsal vs. ventral, p = 0.04).  , d data). Subretinal Iba1 + cells were not included in the count. Statistical analyses were performed with the analysis of variance (one-way ANOVA) followed by a Tukey's post hoc test. Statistically significant differences are represented as follows: *, # , $ p ≤ 0.05, **, ## , $$ , ++ p ≤ 0.01, ***, +++ p ≤ 0.001, #### p ≤ 0.0001; *, # , $ vs. HC, LD12h7d, LD15h7d, and LD18h7d, respectively; % was used to depict differences within layers. Data are shown as means ± SEM of at least N = 3 retinas from different rats for each experimental condition.

Long-Term Morpho-Functional Changes in the Neurodegenerative Progression
Given that 24 h of white light exposure resulted in well-established retinal structural and functional impairment after 7 days, we decided to investigate how the neurodegenerative process evolves using this experimental protocol. After inducing damage (24 h, 1000 lux), critical stages were determined within different retinal regions  , d data). Subretinal Iba1 + cells were not included in the count. Statistical analyses were performed with the analysis of variance (one-way ANOVA) followed by a Tukey's post hoc test. Statistically significant differences are represented as follows: *, # , $ p ≤ 0.05, **, ## , $$ , ++ p ≤ 0.01, ***, +++ p ≤ 0.001, #### p ≤ 0.0001; *, # , $ , + vs. HC, LD12h7d, LD15h7d, and LD18h7d, respectively; % was used to depict differences within layers. Data are shown as means ± SEM of at least N = 3 retinas from different rats for each experimental condition.

Long-Term Morpho-Functional Changes in the Neurodegenerative Progression
Given that 24 h of white light exposure resulted in well-established retinal structural and functional impairment after 7 days, we decided to investigate how the neurodegenerative process evolves using this experimental protocol. After inducing damage (24 h, 1000 lux), critical stages were determined within different retinal regions for up to 7 months.

The Ventral Retinal Structural Layer Organization Was Better Preserved Compared to the Dorsal One
The loss of photoreceptors induced by 24 h light exposure accompanied a progressive decrease in the ONL thickness over time. Figure 7 displays the histological appearance of the dorsal (a) and ventral (b) retina along the vertical meridian from control rats and those exposed to 24 h of high-intensity light at different time-points following the light exposure. The dorsal retina was more severely affected than the ventral one, whose retinal layered structure appeared better preserved (no rosette structures or matrix disorganization). Photoreceptor nuclei were few and dispersed in the dorsal side. The INL directly neighboring the RPE, and several holes were present in the retina 30 days after light exposure (see also Figure 9a). The average ratios and the mean number of photoreceptor rows for all twenty quadrants (Figure 7d,f) also decreased in the ventral side in LD animals.
Furthermore, despite photoreceptor rows' progressive reduction (Figure 7f; all LD groups are significantly different vs. HC), relative ratios in the ventral retina were conserved, underling a homogeneous thinning among retinal layers up to 60 days of recovery (mean ONL/total retinal thickness overall ns vs. HC till later stages of neurodegeneration). A further analysis was done to measure the extent of damage, evaluated as damaged ONL length/dorsal retinal length. In the central section, crossing the optic nerve, the amount of dorsal ONL damage increased over time, reaching 80% of the dorsal retina after 60 days (Figure 7g).

Light-Induced Cone-Photoreceptor Damage
Following light exposure, the first sign of damage is observed at the level of photoreceptor OSs. Retinal sections were stained with an antibody for L/M opsin, allowing differentiation of individual cones for measuring cone density. L/M cones (in general also referred to as L-cones, because of the opsin they express) are mainly localized in the medial and central retina, where light exposure effects are more pronounced, and they are more abundant than S cones (densest in the retinal rims and periphery [51]). Figure 8a shows a composite of photographs taken from the control with uniform cone photoreceptor staining and light-exposed retinas characterized by a diminished density of cones and almost an absence of their corresponding external segments (cone OSs) on the dorsal side. As previously shown [41], the L/M opsin immunostaining revealed a sparse and disorganized population of surviving cones in the hot spot. After 7days, the dorsal retina's remaining cones lost their normal elongated morphology and were immunoreactive to L/M opsin throughout the soma and axon terminals [52].
The  they are more abundant than S cones (densest in the retinal rims and periphery [51]). Figure 8a shows a composite of photographs taken from the control with uniform cone photoreceptor staining and light-exposed retinas characterized by a diminished density of cones and almost an absence of their corresponding external segments (cone OSs) on the dorsal side. As previously shown [41], the L/M opsin immunostaining revealed a sparse and disorganized population of surviving cones in the hot spot. After 7days, the dorsal retina's remaining cones lost their normal elongated morphology and were immunoreactive to L/M opsin throughout the soma and axon terminals [52]. The remaining L and M-cones in retinal sections are plotted in Figure 8b,c. Twentyfour hours of light damage resulted in the loss of approximately 75% of the L/M-cone population by 45 days: detected cones were 134.3 ± 41.81 (N = 3) at 45 days compared to 535. 40 ± 62.55 (N = 5) in the control (p = 0.0014) (Figure 8c). Sometimes, even though it was difficult to identify associated Hoechst-stained cell nuclei within the ONL, some residual L/M opsin immunoreactivity was present. Overall, in the ventral retina, no significant differences were found between HC and experimental groups up to the time point evaluated; a decrease was detectable only in the first two acquired fields (1-and 2-V) of ventral hemiretina at LD24h45d, in proximity to the damaged area (Figure 8b,c). Animals sacrificed after 45d were not included in cone density analysis. The number of lost cones progressed further from 7 to 45 days after light damage, without any significant rescue. Outer segments (OSs) appear normally distributed throughout the retina in the control condition. Light-exposed retinas immediately showed a partial reduction of cone density (0 d) depicted by the length of the orange bar and redistribution of the signal into the soma (*, orange) in the dorsal retina, shortening the remaining OSs in the ventral at 7d. Abbreviations: ONL: outer nuclear layer; OSs: outer segments. Scale bar = 50 µm. (b,c) Numbers of cones distributed along the retinal sections passing through the optic nerve (ON), the totals on the dorsal and ventral sides, and their sum for HC, LD24h0d, LD24h7d, LD24h15d, LD24h30d, and LD24h45d. Statistical significance is represented as follows: *,ˆp ≤ 0.05, **,ˆˆp ≤ 0.01, *** p ≤ 0.001, ****,ˆˆˆˆp ≤ 0.0001; * andˆrefer to HC and LD24h0d, respectively; two-way ANOVA with Tukey's post hoc test. Data represent the means ± SEM of at least N = 5 retinas from different rats, for each experimental condition.

Light Exposure Induced Modifications in the Inner Nuclear Layer (INL), Detectable at Least after 15 Days Following Light Exposure
Morphological analyses mostly underlined a longitudinal extension of the hot spot area. However, as previously described from the ratio analysis (ONL thickness/total retinal thickness) and in other studies [28,29,34,53], the inner retina is also severely implicated. Figure 9a illustrates representative nuclei-stained retinal cross-sections after 7, 30, and 60 days post light exposure treatment. Gradually, with increasing time (after 7 days), the INL structure appeared highly disorganized, considerably thinned, and characterized by the presence of hole-like areas (Figure 9a; see also Figure 8a,b for a comparison between dorsal and ventral). Furthermore, in light of our functional data indicating a partial recovery of the a-wave and the b-wave, peaking at 45 days, we speculated that there could be compen-satory retinal changes to account for the functional recovery. To assess this possibility, we studied the expression pattern of synaptophysin (SYN), an integral membrane protein expressed in presynaptic vesicles. In control retinas, SYN was expressed in the IPL and OPL, whereas after light exposure, the OPL signal was fainter compared to control (Figure 9b). Accordingly, we looked for other biochemical and immunohistochemical changes in the inner retina. Therefore, we aimed to establish whether neuronal modifications in the inner retina occurred in the light damaged rat model using an amacrine-specific marker. We stained retinal sections with an antibody against choline-acetyltransferase (ChAT), revealing the so-called ON and OFF starburst amacrine cells. These are critical elements in the encoding of motion, heavily contributing to direction-selective ganglion cells properties. We observed that although there was no reduction in the number of marked cell bodies either in the INL or GCL (data not shown), their dendritic arborization was disrupted 30 days after light damage (Figure 9c,d; significant at 210d vs. all experimental groups assessed at an earlier time point). As with other markers, ChAT expression in the ventral retina was mainly unaffected by neurodegenerative progression (data not shown).

Spatial Differences in Glial Cells' Reaction, bFGF-2, and Microglia Modulation: Non-Neuronal Cells' Involvement in Retinal Degeneration
Light-induced gliosis was assessed by studying the evolution of astrocyte alterations during retinal degeneration through immunocytochemical localization of GFAP. Figure 10a illustrates the representative distribution of the GFAP signal in the marginal hot spot area (penumbra [41]). Due to the highly disorganized structure typical of the chronic phase, GFAP expression measurements were performed in the penumbra area. Quantification of the signal across all retinal layers (Figure 10c) discloses a significant increase in GFAP immunoreactivity after 7 days both on the dorsal and ventral sides compared to the control. Anyway, in the ventral retina, whose structure was better preserved, the GFAP + area peaked at 30 days. It slightly decreased by 45 days, with significant differences between LD24h7d, 15d, and 30d vs. LD24h90d (one-way ANOVA, Tukey's test, p = 0.0067, 0.0192 0.0265, respectively). Notice that at 90d GFAP + signal in the dorsal retina was significantly higher compared to the ventral retina (p = 0.0032).
On the contrary, in the penumbra (dorsal retina), glial cell reaction was less exacerbated at 15 d, 30 d, and 45 d than at 60 d and 90 d. Müller cells' support to photoreceptors was monitored by analyzing the distribution and the protein level of bFGF-2 in the ONL using immunohistochemistry, at successive survival times after 24 h LD (Figure 10b). No significant differences among groups were detected, although bFGF-2 expression was linearly augmenting in the dorsal retina up to 60d (Figure 10d; one-way ANOVA with Trend, p = 0.0099. Non-significant trend in the ventral side was found, p = 0.2592). On the ventral side, bFGF-2 expression increased immediately after LD, became prominent at about 15 d, and then decreased to control levels by 60d (indeed LD24h60d dorsal is significantly different compared to LD2460d ventral). The immune response also accompanies neurodegeneration for a considerable time. The onset of photoreceptor degeneration activated the recruitment of retinal microglia from the plexiform layer to the injured site (see Figure 3). The analysis of Iba1 + cells distribution among layers described a consistent migration of resident microglia from the inner retina (GCL-IPL and INL) to the ONL, which required more than 24 h to occur (indeed LD24h0d not significant vs. HC). Iba1 + cells (both microglia and macrophages) decreased in the inner retina during more extended recovery periods from the onset of neurodegeneration and increased in the ONL. Following the initial typical acute neuroinflammatory response, from 30 to 60 days, we noticed a significant decrease of microglial cells in the dorsal retina (Figure 11d). No differences were detected in the ventral retina when microglia cells were summed across layers (Figure 11f).

Discussion and Conclusions
Most neurodegenerative retinal diseases are due to progressive degeneration of rods or cones. Current treatment strategies and paradigms manage to slow down photoreceptor loss or replace them using prosthetic devices or stem cells [6][7][8][9]11,12]. The therapeutic efficacy of these approaches depends on the diagnosis of the disease, the timing of which is usually related to the symptom's occurrence. At later neurodegenerative stages, however, therapeutical approaches are less effective. Moreover, early dysfunctions of retinal diseases might be overcome by compensatory mechanisms of the retina, which could silence pathological signs [54,55]. Therefore, it is paramount to achieve rapid diagnosis and effective, targeted treatment at the earliest possible stage.
Aiming to understand better the critical steps involved, we exposed adult albino rats to bright light to induce retinal neurodegeneration and followed the disease's morphological and functional progression. It was previously shown that this model mimics some aspects of human retinal degenerative disorders [56]. Our experiments first goal was to determine the minimal duration of bright light exposure necessary to induce retinal damage. Likewise, the second goal was to study the pathophysiological changes of the diseased retina over time, extending the focus to the inner retina.

Different Durations of Bright Light Exposure Provided Insights into Early Events in the Pathology of Light-Induced Retinal Damage
The first aim of our experiments was to characterize retinal pathophysiological processes following different durations of light exposure. We revealed that 12 h of homogeneous light exposure is sufficient to severely reduce the retinal response to light (assessed by in vivo analyses). However, this was not associated with critical morphological changes. Indeed, functional changes precede anatomical ones.
Specifically, the dark-adapted response is more severely impaired than the lightadapted response, in agreement with the higher survival rate of cones compared to rods, confirming previous studies about the etiology (rhodopsin-mediated [57,58]) of this lightinduced retinopathy [59]. Interestingly, OP sum amplitude analysis revealed that the deterioration of OPs correlated better with the duration of light exposure than did the amplitudes of the a-wave and b-wave. The OPs are described as low-voltage, high-frequency components of the ERG [45] and are believed to reflect the synaptic activity of inhibitory feedback processes generated mainly by the amacrine cells. However, contributions of other inner retinal cells cannot be excluded [45]. Our findings indicated that 12 h of bright light exposure did not strongly affect the inner retinal circuitry, as suggested by the more preserved OP amplitude at LD12h7d (at least at specific stimuli intensities) despite continuing a-wave and b-wave amplitudes reduction.
Remarkably, by analyzing morphological features, we demonstrated that a minimum of 15 h is required to trigger the development of a hot spot (matrix disorganization and rosette structures formation) in the dorsal retina at 7 days after damage [41], which characterizes the model of light damage. Conversely, retinal morphology seemed better preserved after 12 h of light exposure, given the more subtle changes in photoreceptor rows and microglia numbers. Aside from the absence of a prominent hot spot in the LD12h7d group, the functional response (a-wave and b-wave amplitude) was similar to those of all other groups. Therefore, histological results at the ONL level did not correlate with those achieved at the functional level. Given this, we examined how different mechanisms might occur even before ONL disorganization and could influence the functional in vivo test. One such factor, the cytoskeletal type III intermediate filament protein GFAP of astrocytes, is considered an early marker for retinal injury and is commonly used as an index of gliosis-hypertrophy [60,61]. Twelve hours of exposure to light was enough to activate Müller cells, which are well-known to control several retinal trophic factors [62]. Simultaneously, activation of self-protective mechanisms might impinge on the retinal response to light, as supported by studies showing that an increase in bFGF-2 causes a reduction of the b-wave [40,63]. Moreover, it was interesting to notice that while the architecture of the ventral retina is affected more moderately than the dorsal retina (the layered structure of the retina was maintained), the inflammatory markers are upregulated compared to their baseline values. The upregulation of a range of protective factors might contribute to photoreceptor stability and protection; however, this does not extend to retinal performance, as ERG reduction shows.
This study disclosed that 12 h of light damage was sufficient to elicit clear signs of retinal dysfunction with minor morphological signs of damage in the dorsal hot spot area. For some parameters, such as photopic ERG, the functional impairment was irrespective of the exposure duration. Overall, this indicates that functional alterations precede morphological ones. These findings could be used to develop a diagnostic test that highlights an inflammatory condition in the retina that impinges on retinal functional output, as demonstrated. Indeed, ERG responses, recorded with an active extracellular electrode, according to Ohm's law, might be affected by a change in the resistance. We are currently investigating whether neuroinflammation can bias the ERG output by changing the resistance of the eye. Consequently, the substantial initial reduction in retinal function is likely due to both electro-mechanical problems and the neuroinflammatory state that causes a change in tissue resistances.

Analyses Performed after Different Recovery Periods Following 24 h of Bright Light Exposure Showed Discrepancies between the Morphological and Functional Timecourses of the Disease
To achieve our second goal, retinas exposed for 24 h to light at 1000 lux were analyzed at different time points after the cessation of light exposure (acute phase)-up to 7 months (chronic phase). This study allowed us to categorize the pathophysiological process at different stages. The immediate consequence of light exposure was characterized by a progressive and irreversible loss of photoreceptors and a significant attenuation of retinal function in response to light. The chronic phase was characterized by a significant partial recovery, maximal at 45 days, despite the hot spot region gradually enlarging. Our findings suggest that during this stage, the retina attempted to recover from the light insult. Functional rescue following a phototoxic insult is a well-known phenomenon in light-damaged rats [36,64]. It was shown by Rutar et al. [41], as well as others, that once the animal is returned to standard lighting conditions, regrowth of the outer segment occurs, although we found a reduction in the total number of cones expressing L/M-opsin up to 45 days after light exposure. Schremser and Williams [65] showed that growing outer segments to their maximal length could take 20 days, but our study observed the maximal ERG amplitude value at 45 days. While the a-wave's (linked to photoreceptor function) maximum amplitude recovered to~34% (LD24h45d a-wave = 159 ± 21 µV and HC = 466 ± 33 µV, at 10 cd*s/m 2 ) [66], the b-wave's (second-order neurons and Müller cells) amplitude recovered to~45% of its pre-LD value (LD24h45d b-wave = 503 ± 49 µV and HC = 1098 ± 74 µV, at 10 cd*s/m 2 ). Conversely, OP amplitudes recovered significantly (~50% at 10 cd*s/m 2 ), indicating a more pronounced rescue of the inner retina at 45 days.
We recorded ERG responses for up to 7 months and observed that retinal function deteriorates again after 45 days. As already described, astrocyte retinal response is triggered after bright light exposure and is followed by an upregulation of neurotrophic factors [67]. We demonstrated that after 24 h of light damage, the retinas of adult rats enhanced the expression of GFAP. This increase was maintained for up to 7 days in the dorsal side and up to 30 days in the ventral, and there was a subsequent ventral reduction at 45 days (ONL structure was more preserved) that became significant at 90 days, suggesting that GFAP may contribute, along with bFGF-2, to improving retinal function. bFGF-2 expression increased over time (significant trend) in the dorsal ONL for up to 60 days, but in the ventral side, its increase started immediately after LD and became prominent at approximately 15 days. Subsequently, it declined to control levels by 60 days (indeed LD24h60d significantly different only compared to LD24h15d). This slower downregulation of bFGF-2 in the ventral retina correlated with the increased b-wave amplitude at 45 days, as reported in Valter et al. [40]. More studies will be necessary to unveil the mechanisms by which worsening of the retinal function occurred after this time point.
Thus, the initial decrease in retinal function might have been due to a combination of OS shortening, the inhibitory effect of bFGF-2, and the change in retinal resistance due to gliosis. In other words, as the decreases in the quantified GFAP, bFGF-2, and Iba1 + cells accompanied an increase in OS length, a gradual gain in function occurred. Afterward, in the last stage (at 210 days), severe photoreceptor degeneration seemed to overcome retinal functional restoration. Interestingly, the inner retina appeared disarranged at later stages; synaptophysin expression was reduced, and dendritic arborization of ChAT + cells became disorganized. These results suggest that the damage initially triggered by the bright light exposure was too severe such that the intrinsic protective mechanisms of the retina were insufficient to stop its deterioration. Interestingly, the evaluation of synaptophysin expression was discordant with the results obtained in a mouse model by Luis Montalbán-Soler et al. [64]. This difference could be explained by species-dependent differences or by different mechanisms activated through various light damaging protocols; it must be noted that mice (differences exist among strains) do not present a hot spot.
Nonetheless, it is worth noting that most of the retinal disorganization seen in our model was observed in the dorsal retina for up to 7 months, with a rapid loss of photoreceptors followed by substantial disorganization of the inner retina. Instead, the ventral side showed parallel reductions in the thickness of both the ONL and the inner retina. In this chronic phase (i.e., months after the cessation of light exposure), the degree of retinal damage was proportional to the length of the recovery period. The surviving photoreceptors were responsible for maintaining the a-wave. Future studies should look more profoundly at the synaptic and cellular reorganization in the animals inner retina to better understand the retinal network changes. If the outer retina is affected, the signals sent to the inner retina would also be influenced, and consequently, the response of secondorder neurons would be impaired. These modifications must lead to substantial changes in the receptive field properties of retinal ganglion cells that are ultimately reflected in altered visual performance [26,68]. Are these changes occurring only at a later stage of the pathology? Are these permanent changes, or can they be restored after the substitution of defective photoreceptors?
It is clear that the ability to detect these pathologies at an early stage is essential for any attempt at restoring retinal function, reducing the impairment of the inner retina, or increasing the chance of successful response to future therapies. Additionally, our data may provide valuable information for developing retinal prosthetics more efficiently and improving their integration with the patient's retina. Furthermore, it will be fundamental to adopt a more realistic approach that considers the degeneration in the outer retina and the rewiring of the remaining circuit.

Data Availability Statement:
The data that support the findings of this study are available on request from the corresponding author, S.D.M.

Conflicts of Interest:
The authors declare no conflict of interest.