Glial Cell Responses and Gene Expression Dynamics in Retinas of Treated and Untreated RPE65 Mutant Dogs

Abstract

Purpose: The long-term evaluation of RPE65 gene augmentation initiated in middle-aged RPE65 mutant dogs previously uncovered notable inter-animal and intra-retinal variations in treatment efficacy. The study aims to gain deeper insights into the status of mutant retinas and assess the treatment impact.

Methods: Immunohistochemistry utilizing cell-specific markers and reverse transcription-quantitative PCR (RT-qPCR) analysis were conducted on archival retinal sections from normal and RPE65 mutant dogs.

Results: Untreated middle-aged mutant retinas exhibited marked downregulation in the majority of 20 examined genes associated with key retinal pathways. These changes were accompanied by a moderate increase in microglia numbers, altered expression patterns of glial-neuronal transmitter recycling proteins, and gliotic responses in Müller glia. Analysis of advanced-aged mutant dogs revealed mild outer nuclear layer loss in the treated eye compared to moderate loss in the corresponding retinal regions of the untreated control eye. However, persistent Müller glial stress response along with photoreceptor synapse loss were evident in both treated and untreated eyes. Photoreceptor synaptic remodeling, infrequent in treated regions, was observed in all untreated advanced-aged retinas, accompanied by a progressive increase in microglial cells indicative of ongoing inflammation. Interestingly, about half of the examined genes showed similar expression levels between treated and untreated advanced-aged mutant retinas, with some reaching normal levels.

Conclusions: Gene expression data suggest a shift from pro-degenerative mechanisms in middle-aged mutant retinas to more compensatory mechanisms in preserved retinal regions at advanced stages, despite ongoing degeneration. Such shift, potentially attributed to a number of surviving resilient cells, may influence disease patterns and treatment outcomes.

Leber congenital amaurosis (LCA) is a group of retinal diseases characterized by severe visual impairment.^1,2 One form is caused by mutations in the RPE65 gene, which encodes the retinal pigment epithelium 65 kDa protein (RPE65) participating in the retinoid cycle.³ Under physiological conditions, RPE65 converts all-trans-retinyl esters to 11-cis-retinol, the precursor of 11-cis-retinal, and the key chromophore of the opsin visual pigments. However, in the presence of non-functional or dysfunctional RPE65, the retinoid cycle is impaired, leading to compromised replenishment of 11-cis-retinal.^4,5 RPE65-LCA manifests as a complex disease, wherein vision loss stems from concurrent dysfunction and degeneration of photoreceptors. Rod function is primarily affected, followed by cone dysfunction.^2,6–8 RPE65 gene therapy yields rapid recovery of light sensitivity, a phenomenon observed in both humans and animal models, including mice and dogs.^2,9,10 The efficacy of photoreceptor protection from degeneration is notably heightened in retinal regions with relatively preserved photoreceptor populations at the time of intervention.^10,11 However, photoreceptor loss progresses in areas with reduced photoreceptor populations.¹⁰ 

Current understanding of the mechanisms governing death pathways in RPE65 disease remains incomplete. Prior mouse genetic studies imply that in the absence of 11-cis-retinal, it is the aberrant unliganded opsin activation of the phototransduction cascade that may contribute to photoreceptors’ demise, rather than the accumulation of all-trans-retinyl esters in the RPE.⁵ This suggestion is supported by double knockout experiments involving Rpe65^−/− and G protein subunit alpha transducin 1 (Gnat1)^−/−, demonstrating retinal rescue.⁵ A hallmark feature of visual signaling in mouse Rpe65^−/− photoreceptors is diminished intracellular calcium levels due to the closure of the cGMP-gated ion channels coupled with Ca²⁺ extrusion by the Na⁺/Ca²⁺ and K⁺ exchanger.⁵ Interestingly, insufficient Ca²⁺ levels have been previously proposed as a potential cause of photoreceptor death.¹² Data from a recent study using D- and L-cis enantiomers of the calcium channel blocker drug diltiazem, conducted in rd1 and rd10 mice, substantiate this premise. The application of L-cis diltiazem, which reduces the activity of cGMP-gated ion channels, resulted in a significant increase in photoreceptor cell death at high doses.¹³ Although the “low Ca²⁺ hypothesis” offers reasonable interpretation, there may be additional, yet unidentified mechanisms contributing to apoptotic triggers. 

In dogs, recessively inherited retinopathy is caused by a frameshift mutation in RPE65 gene, resulting in a premature stop codon and a subsequent lack of functional protein.^7,14 The mutation was initially identified in Swedish Briard dogs^14,15 and later found in dogs of the same breed from the United States and other countries.⁷ The RPE65 deficiency in dogs is characterized by blindness in low-light conditions and varying degrees of visual impairment in bright light.^7,8,16–18 Residual electroretinogram (ERG) responses decreased with age,¹⁸ ultimately leading to a complete lack of recordable responses in untreated advanced-aged mutant retinas for 9.5 to 10.6 years.¹¹ 

RPE65 mutant dogs show an extended retina-wide dysfunction-only phase followed by degeneration onset between 5 and 8 years at the rate of −0.33 log10/year of outer nuclear layer (ONL) thinning.¹⁰ Early intervention in RPE65 mutant dogs during the dysfunction-only stage improved visual function and demonstrated notable preservation in photoreceptor layer measures and structure within the treatment area, persisting 4.8 to 10.9 years post-treatment.^8,16 However, treatment in mid-life during combined dysfunction and degeneration stages, improved visual function, yet photoreceptor loss persisted, similar as reported in human RPE65-LCA.¹⁰ Extended treatment follow-up (>4 years) revealed the presence of intraocular and inter-animal variations in the degree of retinal degeneration in RPE65 mutant dogs.¹¹ The factors accounting for the heterogeneity associated with RPE65 spatio-temporal disease patterns are not fully understood. One plausible explanation posits that the deficiency of 11-cis-retinal may trigger prolonged dysfunction and stress in photoreceptors. This could lead to the selective demise of some cells while inducing a dormancy-like state in others, as previously proposed.¹⁹ Dormancy, known for its heterogeneity and plasticity,^20,21 may influence disease progression and severity. Another potential factor influencing the retinal landscape in RPE65 disease is that initial dysfunction in photoreceptors may have triggered secondary consequences, such as localized neuroinflammation. This inflammation involves the directed and selective activation of glial cells, including microglia, astrocytes, and Müller cells, which have been previously demonstrated to impact the progression of retinal degenerative diseases.^22–26 

To gain deeper insights into the status of mutant retinas at the time of intervention and the treatment impact, we re-evaluated archival retinal tissues from our recent study.¹¹ We examined changes in the expression of key genes associated with phototransduction, neuroprotective, and immune pathways, along with retinal glial cell responses and synaptic remodeling in the outer retina in middle-aged untreated and advanced-aged RPE65 mutant dogs that received RPE65 gene therapy in one eye when they were middle-aged. 

Materials and Methods

Ethics Statement

The research was conducted in full compliance with the Association for Research in Vision and Ophthalmology (ARVO) Resolution on the Use of Animals in Ophthalmic and Vision Research. All the studies have been approved by the University of Pennsylvania Institutional Animal Care and Use Committee (IACUC). 

Study Samples

The study used archival canine retinal samples fixed in paraformaldehyde and embedded in optimal cutting temperature compound (OCT) under RNAse free conditions. OCT blocks were stored in a −80°C freezer. OCT-embedded retinal sections, obtained from RPE65 mutant dogs in a previously published study,¹¹ and normal retinal sections originating from past laboratory projects were used for immunohistochemistry and gene expression analyses. These sections had been stored at −20°C. Study samples are detailed in Supplementary Table S1. Please note that the selection of retinal sections for downstream gene and protein analysis was based on the assessment of retinal structure in tissue sections stained with hematoxylin and eosin (H&E), retained from the previous study.¹¹ In advanced-aged mutant dogs that received RPE65 gene therapy in one eye when they were middle-aged, gene expression analysis used only treated areas from the bleb region and corresponding retinal regions from untreated fellow eyes. 

RNA Extraction and First Round of cDNA Synthesis

After examining frozen retinal sections by light microscopy to assess tissue integrity, selected retinal regions were collected by trimming unwanted tissue with a razor blade, transferred into a 1.5 mL tube, and washed in sterile 1 × Dulbecco's Phosphate Buffered Saline (Thermo Fisher Scientific, Rockford, IL, USA) for 5 minutes to eliminate OCT. From each of the 10 dogs, retinal tissue from two 10 µm-thick sections was used for RNA isolation utilizing the RNeasy FFPE Kit (Qiagen, Germantown, MD, USA). TURBO DNase (Thermo Fisher Scientific) was used for removal of genomic DNA contamination. Reactions were purified with RNA Clean & Concentrator (Zymo Research, Irvine, CA, USA) according to the manufacturer's instructions and the RNA was eluted in 20 µL nuclease-free water. The final RNA yield ranged from 150 ng to 400 ng between samples. First strand cDNA was synthesized using the Maxima H Minus cDNA Synthesis Kit (Thermo Fisher Scientific) with minor modifications. Specifically, the first strand cDNA synthesis step utilizing T7-Oligo(dT) (refer to Supplementary Table S2) was complemented with the RNA poly(A) tailing process. Polyadenylation and reverse transcription (RT) were concurrently conducted in a single tube. In brief, 1 µL of 10 mM ATP and 3U of Poly(A) Polymerase (New England Biolabs, Ipswich, MA, USA) were added to the assembled RT reaction (excluding T7-Oligo(dT) and reverse transcriptase), followed by incubation at 37°C for 15 minutes. Subsequently, after incorporating the remaining components (reverse transcriptase and 2.5 pmol of the T7-Oligo(dT)), the RT reaction was carried out at 50°C for 1 hour in a total volume of 40 µL, following cDNA purification using DNA Clean & Concentrator (Zymo Research). 

DNA Template Preparation, In Vitro Transcription and Second Round of cDNA Synthesis

Next, we streamlined the preparation of a double-stranded DNA template pool for in vitro transcription, considering that T7 RNA polymerase can generate RNA from synthetic short DNA templates containing a double-stranded promoter region and a single-stranded region downstream, as shown in previous studies.^27,28 DNA templates were prepared as follows: 150 ng of ssDNA from first round of cDNA synthesis and 2.5 pmol of the T7AS primer (see Supplementary Table S2) were added to GoTaq Long PCR Master Mix (Promega, Madison, WI, USA) followed by incubation at 54°C for 2.5 minutes. In vitro transcription reactions were performed using HiScribe T7 High Yield RNA Synthesis Kit (New England Biolabs) and DNA templates containing a double-stranded T7 RNA Polymerase promoter region and a single-stranded region downstream. The in vitro transcribed RNA was treated with TURBO DNase (Thermo Fisher Scientific), column-purified, and subsequently used for cDNA synthesis using the Maxima H Minus cDNA Synthesis Kit (Thermo Fisher Scientific). 

Relative Quantification (ddCt) Assay

Primer sequences for reverse transcription-quantitative PCR (RT-qPCR) experiments are listed in Supplementary Table S2. Each qRT-PCR was performed in a total volume of 25 µL containing GoTaq Long PCR Master Mix (Promega, Madison, WI, USA), cDNA generated from 75 ng DNAse-treated RNA, primer concentration of 0.4 µM (each), SYBR Green I (Biotium Inc., Fremont, CA, USA), and 1.125 µL ROX reference dye (Thermo Fisher Scientific). In addition, betaine solution (Sigma-Aldrich) was added to all reaction to final concentration 1.5 M. Reactions were run in triplicates on the Applied Biosystems 7500 Real-Time PCR System. The gene expression level of hypoxanthine phosphoribosyltransferase 1 (HPRT1) was utilized to normalize the cDNA templates and the ratio of mutant to normal expression levels was analyzed using the ddCt method.²⁹ Amplification data were analyzed with the 7500 Software version 2.0.1 (Applied Biosystems). Genes with P < 0.05 and fold changes (FC) > ± 2 were considered differentially expressed. Gene expression levels in untreated middle-aged RPE65 mutant dogs (5.2 years) were compared to those in the normal control set 1 (n = 3, aged 2, 4, and 5.6 years old). Gene expression levels in retinal sections from untreated and treated advanced-aged RPE65 mutant eyes (9.5–10.6 years) were compared to the normal control set 2 (n = 3, aged 4, 5.6, and 8.6 years old). Please note, to address the complexities introduced by intra-animal and intra-retinal variations in ONL thickness across different untreated and treated mutant retinal regions, we used RNA isolated from normal untreated retinal sections (tapetal region) to obtain reference values for comparative gene expression analyses in mutant retinas. This approach ensures consistency and minimizes ambiguity in data interpretation. Given the low variation observed in the expression levels of the examined genes among the control sets, the use of these normal untreated dogs as controls was deemed appropriate. 

Fluorescent Immunohistochemistry

The immunohistochemistry (IHC) staining of retinal cryosections was carried out as described in our previous publications.^30,31 Primary antibodies used in the study are listed in Supplementary Table S3. Appropriate fluorescent secondary antibodies (Alexa Fluor Dyes; Molecular Probes, Eugene, OR, USA) were used at 1:200 dilution. Staining was examined by epifluorescence microscopy with a Zeiss Axioplan microscope (Carl Zeiss Meditech, Oberkochen, Germany) and captured images were analyzed using Zeiss miscroscopy software Zen 3.1. Where applicable, confocal images were captured with a TCS-SP5 confocal microscope system (Leica Microsystems, Buffalo Grove, IL, USA) under identical conditions and imported into ImageJ software.³² Confocal images were presented as a z-stack of 12 z-steps, with a single z-step measuring 0.35 µm, unless otherwise specified. Maximum projection of all images was equally adjusted for contrast and brightness with ImageJ. As the injection blebs were in the tapetal regions, images taken in normal and untreated mutant retinas were captured in the same areas. 

Results

Our cohort of RPE65 mutant dogs from the previous work¹¹ included 4 advanced-aged RPE65 mutant dogs (BR363, BR275, BR361, and BR323, age = 9.5–10.6 years) treated in mid-life (5–6 years) and 2 untreated middle-aged RPE65 mutant dogs (BR304 and BR318, 5.2 years; see Fig. 1 and Supplementary Table S1 for the summary). 

Figure 1.

 

Histopathology of selected retinal regions in middle-aged and advanced-aged RPE65 mutant retinas. Each image grouping displays H&E-stained sections (top) and immunocytochemistry sections labeled with anti-RPE65 antibodies (bottom). Untreated middle-aged mutant dogs show varied ONL preservation from nearly normal (A) to a more advanced disease stage with a thinner ONL (B). (C-F2) Images represent retinal regions from four advanced-aged mutant dogs that received AAV-RPE65 treatment in one eye at mid-life. For each dog, the left image shows the retinal area from the untreated fellow eye, corresponding to the selected retinal region in the treated eye (middle). The right image schematically indicates the treatment boundary (dotted lines) and the illustrated area (arrow) in the examined section. All selected retinal areas showed evidence of progressive degeneration in both treated and untreated retinas. Despite variability between individual retinas, the photoreceptor nuclear layer in each treated area was moderately thicker than in the corresponding untreated area. Although the level of RPE65 labeling varied among the examined treated retinal regions, it did not correlate with the treatment outcome, consistent with findings from our previous work.¹¹ Please note that all H&E-stained retinal sections are from our previous work,¹¹ where the tissue blocks were serial sectioned, and every tenth section was H&E stained. The C2, D2, E2, and F2 schematic illustrations showing the treatment areas/selection of area for assessment were modified from figure 4 in Gardiner et al.¹¹ The dashed line in F1 indicates that the retina and retinal pigment epithelium (RPE) layers were brought into close photographic proximity as the retina was artifactually separated during processing. Scale bar = 50 µm.

Briefly, BR304, one of 2 untreated middle-aged RPE65 mutant dogs, exhibited nearly normal ONL preservation (see Fig. 1A), but the second dog, BR318, also 5.2 years of age, had a more advanced disease stage with a thinner ONL (see Fig. 1B). These findings align with the onset of the photoreceptor degeneration in canine RPE65-disease occurring at approximately 5 years and highlight differences in disease progression rate in mid-life in the mutant dogs.^10,11 To enhance the precision of characterizing selected retinal regions, representative adjacent retinal sections from BR304 and BR318 were used to examine gene and protein expression. Furthermore, the results were compared to those obtained in the untreated eyes of advanced-aged RPE65 mutant dogs to get an insight into disease-associated pathways. 

The treated eyes in these 4 advanced-aged RPE65 mutant dogs exhibited varied responses to AAV-RPE65 therapy,¹¹ with the most robust positive response in BR363, the least response in BR275 and with moderate treatment responses in the remaining 2 dogs, BR323 and BR361. In each of the 4 treated mutant retinas, a subset of adjacent retinal sections from the bleb region, displaying the most preserved ONL and notable RPE65 expression (see Fig. 1), was selected for the downstream gene and protein expression analysis. By focusing on regions that responded most positively to the therapy we aimed to provide details of the treatment effects, potentially enhancing the sensitivity of the study to detect changes in gene expression and protein profiles associated with the therapeutic response. It also allowed for a more straightforward comparison of gene and protein expression patterns without the need to account for significant variations in the treatment response across different retinal regions. As seen in Figures 1C, 1D, 1E, and 1F, the ONL thickness is maximal in BR363, whereas BR275, BR323, and BR361 have a relatively thinner ONL. To detail the effects of the therapy in these regions, corresponding retinal areas from untreated-control fellow eyes (see Figs. 1C, 1D, 1E, and 1F) were used for comparative analysis. Each of these regions have moderately thinner ONL compared to the matched area in the treated eye. As reported in our previous study, at the time of histological examination, ERG responses were detected only in the treated retinas, but not in the contralateral eyes, of advanced-aged RPE65 mutant dogs.¹¹ 

Gene Expression Profile in Middle-Aged and Advanced-Aged RPE65 Mutant Retinas

We used an optimized protocol for gene expression analysis using RNA isolated from paraformaldehyde fixed, OCT-embedded retinal sections (see Methods for details). The expression level of HPRT1 was used as a reference for quantification of gene expression (Supplementary Fig. S1). To characterize activity of retinal processes in middle-aged untreated and advanced-aged RPE65 mutant dogs, both treated and untreated eyes, we examined expression levels of 20 genes associated with: photoreceptor maintenance and phototransduction pathways (CRX, RHO, OPN1MW, and RCVRN)³³; neuroprotection (MANF and FGF2)^34,35; immune response (TLR4, IL6, TNF, and IL1B)^22,25; neuron-microglia interaction (CX3CL1, CD200, and CD47)^36–38; complement regulation (CD59)³⁷; stress response pathway from the endoplasmic reticulum (DDIT3 and HSPA5)³⁹; glucose uptake (SLC2A1)⁴⁰ and apoptosis (BCL2, BAX, and BMF).^41,42 The results presented in the Table show gene expression profiles in each of the RPE65 mutant retinas compared to the corresponding wild type retinas. 

Table.

 

Non-Differentially and Differentially Expressed Genes in the Study Model

 

Briefly, in untreated middle-aged RPE65 mutant dogs, out of the 20 genes tested, 11 genes in BR304 and 15 genes in BR318 exhibited marked downregulation. Eight genes (CRX, RCVRN, CX3CL1, CD47, TLR4, IL1B, BCL2, and MANF) demonstrated the most pronounced decrease in expression levels in both BR304 and BR318 retinas compared to the normal control set (more than 6.6-fold reduction). DDIT3, FGF2, and TNF were undetectable. These data indicate abnormalities in the functioning of various retinal pathways potentially making middle-aged RPE65 mutant retinas susceptible to neuroinflammation and further damage. 

Examination of gene expression changes in selected retinal regions from untreated-control eyes of advanced-aged RPE65 mutants, showed undetectable DDIT3 and FGF2 levels, whereas RHO and RCVN levels continued to be downregulated across all 4 retinas, consistent with their expression patterns in middle-aged mutant dogs. In contrast, expression levels of CRX, OPN1MW, TLR4, IL6, BCL2, and BAX were similar to those in the normal control group, which may indicate partial compensatory mechanisms or the stabilization of certain cellular pathways despite ongoing degeneration. In line with this suggestion, the increased expression of SLC2A1 observed in 2 untreated mutant retinas, while being similar to normal in the other 2, supports the presence of an adaptive response to metabolic stress. Furthermore, upregulation of CD200 and CD59 levels was detected in 3 out of 4 retinas, implying an attempt to mitigate immune activation and inflammation. However, some variability in the gene expression levels of immune regulators, including TNF and IL1B, among untreated control mutant retinas points to an inconsistent inflammatory response. In addition, the consistent lack of CX3CL1 and CD47 expression in all 4 untreated control advanced-aged mutant retinas could potentially lead to unchecked immune cell activity, promoting chronic inflammation and cell loss. 

In AAV-treated retinal regions from advanced-aged RPE65 mutant dogs, expression levels similar to those in matched retinal regions from untreated-control fellow eyes were observed for 5 genes (CRX, RHO, OPN1MW, IL6, and IL1B) across all 4 retinas and for 2 other genes (SLC2A1 and BAX) in 3 out of 4. Three genes (DDIT3, CX3CL1, and FGF2) remained undetectable in all treated retinas. The normalization of CD200 levels in all treated retinas as well as CD47 and CD59 in 3 out of 4 retinas may indicate a positive regulatory effect on immune responses, potentially preventing excessive inflammation. Notably, in addition to the upregulation of the anti-apoptotic gene BCL2 across all 4 AAV-treated retinas, increased expression of the protective genes HSPA5 and MANF was observed in 2 of the treated retinas. The highest expression levels of BCL2 and MANF were recorded in the BR363 retina, which showed the best therapeutic response,¹¹ with BCL2 reaching a 53.3-fold increase and MANF a 25.5-fold increase. 

Overall, the observed gene expression changes indicate a complex interplay among photoreceptor dysfunction, immune dysregulation, and apoptotic pathways contributing to RPE65-dependent retinal degeneration. 

Characterization of Glial Cell Responses in Untreated and Treated RPE65 Mutant Retinas

To examine whether a disturbance in retinal homeostasis induced by the RPE65 defect elicits reactive changes in glial cells, including microglia, Müller glia, and astrocytes, we analyzed the expression patterns of glial cell markers by IHC. 

First, we examined expression levels of the allograft inflammatory factor-1, also known as IBA1, a microglia/macrophage marker⁴³ and CD68 molecule, a marker of microglia/macrophage phagocytic activity.⁴⁴ Compared to the normal group, the number of IBA1⁺ microglia were moderately increased in untreated middle-aged mutant retinas and noticeably higher in untreated eyes of advanced-aged mutant retinas, indicating a heightened inflammatory response (Figs. 2A–E, 2G, 2I, 2K). In the treated retinal regions of advanced-aged mutant retinas, microglia numbers remained close to normal (Figs. 2F, 2H, 2J, 2L), representing a positive response to the treatment. As neurodegeneration is an ongoing process in RPE65 disease and thinning of the photoreceptor nuclear layer is readily identified in RPE65 mutant retinas, especially in advanced-aged retinas, both treated and untreated, we anticipated an increased number of phagocytic microglia in the retinal layers. As shown in Figure 2, the number and distribution of CD68⁺ phagocytic microglia approached near-normal values in untreated middle-aged, treated advanced-aged, and 3 out of 4 untreated advanced-aged mutant retinas. A mild-to-moderate increase in CD68-positive microglia was noted in one untreated advanced-aged retina (BR363; see Fig. 2E). 

Figure 2.

 

The phagocytic activity of microglia in untreated and treated RPE65 mutant retinas. Immunolabeling of microglia with anti-CD68 (green) and anti-IBA1 (red) antibodies in normal (A, B), middle-aged mutant retinas (C, D), and advanced-aged mutant retinas (E–L) indicates a progressive increase in the number of IBA1-positive mononuclear phagocytes in untreated mutant retinas with advancing disease stages. Despite that, phagocytic activity of microglia is within the normal range in untreated middle-aged, treated advanced-aged and in three out of four untreated advanced-aged mutant retinas. A moderate increase in CD68⁺ microglia was observed in one untreated advanced-aged retina (E). Please note that panels A to L display the merged image for two channels, with DAPI included on the left of each set, separated by the dashed line. Scale bar = 50 µm.

Examination of Müller glial cell responses revealed disease-associated changes in the expression pattern of glutamine synthetase (GS) and the type III intermediate filament protein vimentin (VIM) in untreated RPE65 mutant retinas (Fig. 3). In normal retinas, GS is expressed in Müller cell bodies and processes (see Figs. 3A, 3B). In both middle-aged RPE65-mutant BR304 and BR318 retinas, GS immunolabeling was noticeably reduced in Müller cell processes in the ONL that extended toward the outer limiting membrane (OLM; see Figs. 3C, 3D). Comparatively, GS protein levels varied in untreated advanced-aged mutant retinas, ranging from nearly undetectable in BR361 and BR323, low in BR275, and close to normal expression pattern in BR363 (see Figs. 3E, 3G, 3I, 3K). As seen in Figures 3F, 3H, 3J, and 3L, the GS immunoreactivity in Müller cell processes returned to normal levels after gene therapy across all 4 advanced-aged RPE65 mutant dogs. 

Figure 3.

 

Müller cell and astrocyte responses in untreated and treated RPE65 mutant retinas. Representative epifluorescence microscope images showing normal (A, B), middle-aged mutant retinas (C, D), and advanced-aged mutant retinas (E–L) immunostained with anti-glutamine synthetase (GS; green), anti-vimentin (VIM; red), and with anti-glial fibrillary acidic protein (GFAP; silver) antibodies. Immunolabeling reveals a moderately elevated number of astrocytes in both untreated middle-aged mutant retinas (C, D) and in two out of four untreated advanced-aged mutant retinas (E, K), with their normal localization observed in the IPL and, to a lesser extent, in the OPL. There is markedly reduced density of GS protein in the apical region of Müller cells in middle-aged mutant retinas (C, D, indicated by blue arrows) versus normal retinas (A, B). In untreated advanced-aged mutant retinas GS labeling nearly undetectable in two retinas (I, K), low in BR275 (G), and similar to normal in BR363 (E). Treated mutant retinas show no abnormalities in GS patterns (F, H, J, L). In comparison to normal, VIM is notably upregulated in the inner half of the Müller cells and their endfeet across all RPE65 mutant retinas. Please note that panels A to L show the merged image for all channels, with DAPI included on the left of each set, separated by the dashed line. Part of the image (separated by the dashed line) is presented by omitting a color layer to better illustrate a particular feature. Scale bar = 20 µm.

Furthermore, we observed upregulation of VIM in the inner half of the Müller cells and their endfeet across all RPE65 mutant retinas, both untreated and treated, indicating activation of Müller glia-associated gliotic responses⁴⁵ (see Fig. 3). Immunolabeling of astrocytes with ant-glial fibrillary acidic protein (GFAP)⁴⁶ showed a moderate increase of GFAP⁺ astrocyte numbers in both middle-aged retinas (see Figs. 3C, 3D) as well as in two out of four untreated advanced-aged mutant retinas (see Figs. 3E, 3K). Despite that, there was a normal distribution of these cells, predominantly present in the ganglion cell layer (GCL), to a lesser extent in the inner plexiform layer (IPL), and occasionally in the outer plexiform layers (OPL; see Fig. 3). GFAP immunolabeling in treated advanced-aged mutant retinas was nearly normal (see Figs. 3F, 3H, 3J, 3L). 

Considering the complex GS protein expression pattern in the examined retinal layers of middle-aged mutant dogs, we further detailed the status of these retinas using additional cellular markers. Confocal imaging of labeled BR304 and BR318 retinal sections revealed that the expression pattern of excitatory amino acid transporter 1 (EAAT1), a glial-neuronal transmitter,⁴⁷ exhibited disease-associated changes similar to those of the GS protein (Supplementary Figs. S2A–S2C). Additionally, we confirmed the upregulation of VIM protein in GS⁺ Müller glia, as opposed to GFAP (Supplementary Figs. S2D–S2F). We also observed a prominent upregulation of the roundabout guidance receptor 1 (ROBO1) protein in the inner half of the VIM⁺ Müller cells and their endfeet (Supplementary Figs. S3A–S3C). Notably, such changes in ROBO1 expression were concomitant with the upregulation of the slit guidance ligand 2 (SLIT2) in paired box 6 (PAX6)-positive amacrine cells (Supplementary Figs. S3D–S3F). In our previous work, we highlighted the correlation between dysregulation of the axon guidance ROBO/SLIT pathway in X-linked retinal degeneration and retinal cell remodeling.⁴⁸ The potential relevance of the ROBO/SLIT pathway to RPE65 disease pathogenesis requires further investigation. 

In summary, untreated RPE65 mutant retinas show activation of Müller glial gliotic responses and potential disturbances in the glutamate recycling pathway. Additionally, the progressive increase in the number of IBA1-positive cells as the disease progresses indicates persistent retinal inflammation. The data also suggest that GFAP-positive astrocytes respond to retinal degeneration with moderate gliosis, indicated by increased astrocyte numbers in a subset of examined loci but normal distribution. Treated advanced-aged mutant retinas demonstrated normal microglial and astrocytic responses, as well as normal GS patterns in Müller cells. Despite ongoing retinal degeneration, microglial cell phagocytic activity remained close to the normal range in both untreated and treated mutant retinas. 

Synaptic Remodeling and Compensatory Postsynaptic Sprouting in the Outer Retinas of Advanced-Aged RPE65 Mutant Dogs

Finally, we examined potential synaptic remodeling in photoreceptors, as functional consequences of RPE65-dependent retinal degeneration. For this purpose, photoreceptor terminals were labeled with an antibody directed against postsynaptic density protein 95 (PSD95), a marker for both cone pedicles and rod spherules,⁴⁹ and C-terminal-binding protein 2 (CtBP2/RIBEYE), a component of photoreceptor ribbons.⁵⁰ Postsynaptic sites were labeled with antibody against Parvalbumin (PVALB), a marker of horizontal cells.⁴⁸ 

Representative images of retinal sections from untreated middle-aged retinas (BR304 and BR318) showed that labeled photoreceptor presynaptic sites and horizontal cell processes were confined to the OPL similar to normal control (Figs. 4A–4D). Despite differences in the density of synaptic sites between BR304 and BR318 retinas, attributed to BR304 having a thicker and presumably more disorganized OPL than BR318, the density remains within the normal range. 

Figure 4.

 

Assessment of photoreceptor synaptic remodeling in untreated and treated RPE65 mutant retinas. Representative confocal imaging of normal canine retinas (A, B), untreated middle-aged mutant retinas (C, D), and both treated and untreated retinas of advanced-aged mutant dogs (E–L), immunostained with the retinal ribbon marker C-terminal-binding protein 2 (CtBP2; red), photoreceptor terminal marker postsynaptic density protein 95 (PSD95; silver), and the horizontal cell marker Parvalbumin (PVALB; green). In normal retina (A, B) and untreated middle-aged mutant retinas (C, D), ribbons labeled by CtBP2 staining indicates normal stratification of synaptic contacts in the OPL. A number of retracted ribbon-containing synaptic terminals (yellow arrow) coupled with PSD95 labeling are seen in all untreated advanced-stage disease retinas (E, G, I, K) and one treated retina (J). Ribbon-containing photoreceptor synaptic terminals, including retracted ones, are closely juxtaposed with sprouted processes of horizontal cells (selectively exemplified in inserts with 2 times magnification). Please note that the images are presented by omitting a color layer (separated by a dashed line) to better illustrate a particular feature. Confocal images are presented as a z-stack of 12 z-steps, with a single z-step measuring 0.2 µm. Scale bar = 20 µm.

In untreated advanced-aged mutant retinas, a number of presynaptic sites retracted into the ONL (Figs. 4E, 4G, 4I, and 4K, yellow arrowheads). This was accompanied by the sprouting of horizontal processes into the ONL and their invagination into the retracted presynaptic sites. In treated advanced-aged mutant retinas, only one retina (BR361) out of the four showed retraction of photoreceptor synapses into the ONL, accompanied by aberrant outgrowth of horizontal cell processes into the ONL (Figs. 4F, 4H, 4J, and 4L). Compared to middle-aged mutant retinas, the density of CtBP2-positive presynaptic sites was noticeably reduced in the examined retinal regions of all four untreated advanced-aged mutant eyes and in three out of four AAV-treated eyes. 

In summary, all untreated advanced-aged mutant retinas demonstrated the presence of remodeling in the outer retina, including reduced density of synaptic terminals and abnormal stratification of both presynaptic and postsynaptic sites. These defects were not noted in the retinal regions of untreated middle-aged mutant dogs. In treated advanced-aged mutant retinas, retraction of photoreceptor axon terminals was infrequent, suggesting stabilization of the remodeling processes. Regardless of mid-life treatment, photoreceptor synaptic site loss was evident in the majority of examined treated regions of advanced-aged mutant retinas. Despite evident photoreceptor loss in advanced-aged mutant retinas, the degree of synaptic site loss does not solely correlate with ONL thickness; it may also be influenced by the severity of retinal cell dysfunction. 

Discussion

The long-term outcomes of RPE65 gene augmentation therapy targeting RPE65-LCA remain controversial. Therapeutic effects exhibit patient variability, with short-term improvements in visual function but an observed inability to decelerate the rate of retinal degeneration.^2,9,10,51 To enhance treatment efficacy, it is crucial to gain further insights into retinal dysfunction in RPE65 disease. Human patients experience a more rapid disease progression when compared to mouse and dog animal disease models at the time of treatment.^2,10,52 Consequently, the optimal modeling of human RPE65-LCA involves older animals that manifest both dysfunction and degeneration phases of the disease, and this is best seen in the canine model. 

In this study, we used archival retinal tissues from our prior research, investigating long-term (>4 years) effects of RPE65-AAV therapy in RPE65 mutant dogs treated at mid-life.¹¹ At the time of treatment, this period coincides with profound visual dysfunction and varying degrees of photoreceptor degeneration, mirroring patient-relevant disease stages. Data from our previous work indicate that different disease stages may exist intra-retinally in neighboring regions in RPE65 mutants.¹¹ Hence, utilizing adjacent tissue sections for both gene and protein expression analysis, as applied in this study, enhances the precision of characterizing distinctive features in selected retinal regions. We deliberately refrained from using total RNA isolated from the entire retina to examine gene expression dynamics, as opposed to the previous study in the RPE65 knock out (KO) model.⁴² 

In retinal sections from 2 untreated middle-aged RPE65 mutant dogs, a notable decrease in the expression levels of genes linked to essential retinal pathways was observed, potentially leading to a compromised cellular response to stress and damage. These alterations in gene expression coincide with reduced density of glial-neuronal glutamate transmitter recycling proteins in the apical region of Müller cells. Based on continuous opsin activation and the “low Ca²⁺ hypothesis,”^5,12 this observation may be partially attributed to a sustained reduction in glutamate release at the photoreceptor synapses. Additionally, prolonged dysfunction and stress could compromise photoreceptor integrity, resulting in damage to cellular structures and membranes. This damage may consequently impact the molecular interplay between photoreceptors and Müller cells. 

In the examined retinal regions of the untreated fellow eyes from advanced-aged mutant dogs, it was surprising to observe that the expression levels of a subset of genes, which had been downregulated in middle-aged mutant retinas, returned to levels comparable to those in normal controls. Additionally, we detected upregulation of several genes, including CD200 and CD59, in 3 of the 4 untreated advanced-aged mutant retinas, and SLC2A1 in 2 of these retinas. This observation could be attributed to cellular selection at earlier disease stages. In middle-aged RPE65 mutant retinas with both dysfunction and degeneration, photoreceptor cells may undergo varying degrees of damage, with some experiencing irreversible damage. The varying capacity of photoreceptor cells to engage stress response mechanisms may contribute to their heterogeneous behavior, leading to continued cell death in some cells while others survive until this advanced disease stage. Potentially, the restructured outer retinal regions composed of surviving resilient cells may reflect more “normal” expression levels of certain genes. It can be argued that similar processes may occur in AAV-treated retinal regions, rather than being solely attributed to the treatment's impact. This could explain the similarity in the expression levels of a subset of genes between AAV-treated retinal regions and those in untreated fellow control eyes. 

Notably, retinal regions in both untreated and treated eyes of advanced-aged dogs displayed some changes that were initially observed in middle-aged mutant retinas. Specifically, Müller glia stress response continued, and there was no improvement in expression levels for a subset of genes associated with phototransduction pathway (e.g. RHO), ER stress response (DDIT3), neuron-microglia interaction (CX3CL1), and neuroprotection (FGF2). Comparatively, remodeling defects in the outer retina, including abnormalities in synaptic density and disrupted stratification of presynaptic and postsynaptic sites, were evident in all advanced-aged untreated mutant retinas but not in untreated middle-aged mutant retinas. Such events are relatively common during the progression of primary photoreceptor degenerative diseases.^48,53,54 However, the pattern of synaptic remodeling and the stage at which it becomes noticeable are influenced by the underlying disease mechanism. 

Interestingly, despite all treated advanced-aged eyes displaying distinct ERG responses under dark- and light-adapted conditions,¹¹ noticeable loss of photoreceptor synaptic terminals was observed in three out of four examined AAV-treated retinas, similar to the levels in the untreated fellow eye. Synaptic sites were relatively well preserved only in the examined treated retinal regions of BR363, aligning with our previous findings, which showed that BR363 exhibited the most robust ERG responses.¹¹ These data imply that the therapy may not prevent synaptic defects in those photoreceptor cells that are severely compromised at the time of treatment. 

It remains unclear which molecular signals exacerbate photoreceptor synapse loss in RPE65 disease. Previous studies highlight the role of transmembrane receptor CD47 (“don't eat me” signal) in protecting synapses from abnormal microglial pruning.^55,56 CD47 expression decreased in middle-aged RPE65 mutant retinas, which were still morphologically similar to normal, and reached undetectable levels in all four untreated advanced-aged mutant retinas. Although increased microglial phagocytosis was not evident in the retinal regions of middle-aged and advanced-aged mutant dogs, decreased levels of CD47 cannot be excluded as a potential contributing factor to photoreceptor synapse loss that may occur between these two time points. Further research is required to address this point. 

Finally, this study demonstrated positive treatment effects, including normalized GS protein levels in Müller glia across all four treated retinas, as well as astrocyte and microglia cell numbers and distribution similar to normal. In contrast, GS protein levels in Müller cells were notably lower in three mutant retinas of the untreated fellow eyes, indicating potential issues with glutamate metabolism. As well, the increased numbers of IBA1-positive cells in these retinas suggests ongoing inflammation. 

Overall, the study describes a previously unknown layer of complexity in RPE65 disease mechanisms. Our data suggest significant alterations in retinal function and homeostasis in middle-aged RPE65 mutant retinas. The complexity of these alterations at this disease stage, including potential heterogeneity in photoreceptor cell damage, likely contributes to the variability in disease manifestation observed in advanced-aged mutant retinas and treatment outcomes. 

Acknowledgments

The authors express their gratitude to Raghavi Sudharsan and Jacob Appelbaum for their helpful discussions and comments, and to Emily Katz (summer student) for her assistance in IHC experiments. The treated and untreated sections from RPE65 mutant dogs were part of a slide collection set produced by Kristin L. Gardiner.¹¹ 

Supported by United States National Eye Institute/National Institutes of Health (grants R01-EY006855, R01-EY017549, and P30-EY001583) and the Foundation Fighting Blindness (FFB). 

Disclosure: T. Appelbaum, None; E. Santana, None; D.A. Smith, None; W.A. Beltran, None; G.D. Aguirre, None