Age-Dependent Disease Expression Determines Remodeling of the Retinal Mosaic in Carriers of RPGR Exon ORF15 Mutations

Abstract

purpose. To characterize the retinal histopathology in carriers of X-linked progressive retinal atrophy (XLPRA1 and XLPRA2), two canine models of X-linked retinitis pigmentosa caused, respectively, by a stop and a frameshift mutation in RPGRORF15.

methods. Retinas of XLPRA2 and XLPRA1 carriers of different ages were processed for morphologic evaluation, TUNEL assay, and immunohistochemistry. Cell-specific markers were used to examine retinal remodeling events.

results. A mosaic pattern composed of patches of diseased and normal retina was first detected in XLPRA2 carriers at 4.9 weeks of age. A peak of photoreceptor cell death led to focal rod loss; however, in these patches an increased density of cones was found to persist over time. Patches of disease gradually disappeared so that by 39 weeks of age the overall retinal morphology, albeit thinner, had improved lamination. In older XLPRA2 carriers (≥8.8 years), extended regions of severe degeneration occurred in the peripheral/mid-peripheral retina. In XLPRA1 carriers, opsin mislocalization and rare events of rod death were detected by TUNEL assay at 20 weeks of age; however, only patchy degeneration was seen by 1.4 years and was still apparent at 7.8 years.

conclusions. The time of onset and the progression of the disease differed between the two models. In the early-onset form (XLPRA2) the morphologic appearance of the retinal mosaic changed as a function of age, suggesting that structural plasticity persists in the early postnatal canine retina as mutant photoreceptors die. In the late-onset form (XLPRA1), patches of disease persisted until later ages.

X-linked retinitis pigmentosa (XLRP) comprises some of the most severe forms of RP, and accounts for approximately 10% to 20% of familial cases. ^{1 2} Six loci (RP2, RP3, RP6, RP23, RP24, and RP34) have been associated by linkage analysis with XLRP; of these RP2 and RP3 are the two most common (www.retnet.org/ http://www.sph.uth.tmc.edu/RetNet; provided in the public domain by the University of Texas Houston Health Science Center, Houston, TX). Retinitis pigmentosa GTPase regulator (RPGR), the disease gene of the RP3 locus, accounts for >70% of the cases of X-linked retinitis pigmentosa (XLRP), ^{3 4 5 6} and exon ORF15, a mutational hot spot in RPGR, has been shown to be mutated in 22% to 60% of patients with XLRP. ^{3 7 8}  

Mutations in RPGR can cause a spectrum of retinal diseases that include cone–rod dystrophies, ⁹ macular degeneration, ¹⁰ and, more commonly, rod–cone degenerations. ¹¹ Males affected with RPGR-XLRP typically exhibit night blindness in the first decade of life, followed by reduction of the visual field and loss of central visual acuity. By the end of the fourth decade, most patients are legally blind. ¹² Female carriers, on the other hand, display a broad spectrum of disease severity. The age of onset and the extent of fundus pigmentary changes, visual field defects and ERG abnormalities vary considerably between and within families. ^{11 13 14 15} A characteristic of female XLRP carriers is the patchy nature of the disease expression when assessed by funduscopic examination, Goldmann perimetry, ^{11 16} multifocal ERG, ¹⁷ and histopathology. ^{16 18 19 20} This mosaic of normal and diseased retinal foci is best explained by Lyon’s hypothesis of random X-chromosome inactivation that occurs at an early stage of female embryogenesis. ²¹  

Histopathologic reports of the retinal lesions in XLRP carriers are scarce. Indeed, they are limited to that of four individuals, three of them carrying mutations in RPGR. ^{16 18 19 20} Since these studies were conducted on postmortem retinas of older (>75 years of age) patients, the findings represent histologic alterations observed at an advanced stage of disease. For obvious reasons, access to retinal tissues from younger patients is likely to be even more limited and currently hampers our understanding of the course of the disease in XLRP carriers. The use of animal models with naturally occurring forms of X-linked retinal degeneration represents a valuable tool for examining whether, in carriers, the mosaic pattern of degenerating and normal retinal foci remains unchanged or not throughout the course of disease. 

In the dog, two distinct forms of X-linked retinal degeneration, termed XLPRA1 and XLPRA2, have been characterized, and are caused, respectively, by a stop and a frameshift mutation in RPGR exon ORF15. ²² The frameshift mutation causes an early and severe loss of photoreceptors (XLPRA2) that begins a few weeks after birth, while the stop-mutation disease (XLPRA1) has a later onset (∼11 months of age) and more gradual progression. ^{22 23 24} In both diseases, multiple patches of degeneration have been reported in carriers. ^{22 24} In the present study, we have expanded the analysis of the retinal mosaic in XLPRA1 and XLPRA2 heterozygotes and find focal lesions of photoreceptor disease and degeneration similar to those observed throughout the retinas of homozygotes. In addition, we report that the retinal mosaic disease pattern changes as a function of age in XLPRA2 carriers, whereas no such retinal plasticity occurs in carriers of the later onset XLPRA1 disease. 

Materials and Methods

Animals

Twelve dogs were used in the study: 10 crossbred female XLPRA2 carriers (age range: 3.9 weeks to 10.3 years), and 2 crossbred female XLPRA1 carriers (age: 20 and 24 weeks). All dogs were bred at the Retinal Disease Studies Facility (RDSF, University of Pennsylvania, New Bolton Center, Kennett Square, PA), and their genotype was determined either from the known status of their progenitors or from genetic testing for the disease-causing mutation. ²² All animals were anesthetized by intravenous injection of pentobarbital sodium, the eyes were enucleated, and then the dogs were euthanatized. All procedures involving animals were in compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. In addition to the 12 dogs that were specifically bred for this study, archival retinal sections from 7 older carriers of XLPRA1 (age range: 1.4–7.8 years) were included (Table 1) . No experiments with tissues from wild-type (i.e., nonmutant) dogs were repeated in the present study as they had already been performed and reported in a previous publication. ²³  

Retinal Histology

The left eyes of eight of the XLPRA2 carriers were used for morphologic examination of disease expression after plastic or optimal cutting temperature (OCT) compound embedding (Table 1 , morphology). The tissues were processed for plastic embedding as previously reported, according to a similar protocol as was used for the archival tissues from seven XLPRA1 carriers. ²⁵ After enucleation, the globes, and then the posterior segments, were isolated and fixed after a sequential triple fixative protocol (3% glutaraldehyde-2% formaldehyde; 2% glutaraldehyde-1% osmium tetroxide; and 2% osmium tetroxide). The posterior segments were then trimmed into pieces that extended from the optic nerve head to the ora serrata along the superior and inferior meridians, dehydrated, and embedded in epoxy resin (PolyBed 812; Polysciences, Warrington, PA). The tissues were sectioned with glass knives at 1 μm with a microtome (2065 Reichert Jung Supercut; Leica, Deerfield, IL), stained with azure II-methylene blue with or without a paraphenylenediamine (PPDA) counterstain. 

Sections from both the superior and inferior meridians were examined with a 40× objective by light microscopy (Axioplan; Carl Zeiss Meditec, Oberkochen, Germany). The sections were examined in contiguous fields from the optic disc to the ora serrata, including evaluation of the retinal pigment epithelium (RPE), the rod and cone outer (OS) and inner (IS) segments, and the thickness and density of the outer nuclear and inner nuclear layers (ONL and INL). Quantitative measurements of the ONL were made in XLPRA2 carriers in the central (site S1: 2000 ± 500 μm from the optic nerve head), peripheral (site S3: 2000 ± 500 μm from the ora serrata), and mid-peripheral (site S2: midpoint ±500 μm between the optic nerve head and the ora serrata) regions. As most archival sections from the XLPRA1 carriers did not contain either the optic nerve head or the ora serrata, the measurements were made at the extremities of the section and referred to as central and peripheral retina. 

TUNEL Assay

In the eight XLPRA2 and two XLPRA1 carrier dogs, one eye was processed to perform TUNEL assays and/or immunohistochemistry (IHC: TUNEL, Table 1 ). After enucleation, a slit was made through the globe at the level of the ora serrata, and the entire globe was fixed in paraformaldehyde, cryoprotected, and embedded in OCT medium as previously reported. ²³  

Cryosections (7 μm thick) along the superior meridian were used for TUNEL assay according to the manufacturer’s protocol (In Situ Cell Death Detection kit; Roche, Indianapolis, IN), and stained with DAPI. Positive controls included sections pretreated with DNase I (3 U/mL in 50 mM Tris-HCl [pH 7.5] and 1 mg/mL BSA) for 10 minutes at room temperature. For negative controls the terminal transferase enzyme was omitted from the TUNEL reaction mixture. Sections were examined from the optic disc to the ora serrata by epifluorescence microscopy with the 40× objective. TUNEL-labeled cells in the ONL were counted throughout the entire length of the section (i.e., from disc to ora serrata). To determine the proportion of photoreceptor cells that undergo cell death as a function of time, we expressed our results as the number of TUNEL-labeled photoreceptor cells per 10⁶ μm² of ONL as previously reported. ²³ The area of the ONL of each section was obtained by measuring the entire length of the ONL from optic disc to ora serrata and multiplying it by the average thickness of the ONL throughout the section (mean value of the thickness measured in the three locations reported above). If focal thinning was present at the site, the closest adjacent site that was representative of the ONL thickness was measured. In areas of decreased photoreceptor density, this method may slightly underestimate the proportion of cells that are TUNEL positive, but was selected because individual cell counts could not be determined on 7-μm-thick DAPI-stained cryosections. For each dog, this assay was performed in triplicate with sequential sections from the superior meridian. The results were averaged and reported as the mean ± SD. 

Immunohistochemistry

Cryosections made along the superior retinal meridian of the eight XLPRA2 and two XLPRA1 carrier dogs were processed as described and were used for fluorescence IHC. We used a battery of cell-specific antibodies, most of which already have been shown to specifically label the canine retina (Table 2) . ²³ Since previous testing of the RPGR and RPGRIP antibodies conducted in our laboratory failed to show any cross-reactivity or specificity on the dog retina, we did not include them in the study. ²² Cryosections (7–10 μm thick) were incubated overnight with the primary antibodies after a blocking step with 10% normal serum from the appropriate species. The antigen–antibody complexes were visualized with fluorochrome-labeled secondary antibodies (Alexa Fluor, 1:200; Molecular Probes, Eugene, OR). DAPI stain was used to detect cell nuclei. Slides were mounted with a medium composed of polyvinyl alcohol and DABCO (Gelvatol; Sigma-Aldrich, St Louis, MO), and examined with an epifluorescence microscope (Axioplan; Carl Zeiss Meditec). The images were digitally captured (Spot 4.0 camera; Diagnostic Instruments, Sterling Heights, MI) and imported into a graphics program (Photoshop; Adobe, Mountain View, CA) for display. Confocal microscopy was also used on a subset of sections to image discrete structural changes occurring in the outer segments. The confocal images were acquired on a tunable confocal (TCS-SP5; Leica Microsystems, Wetzlar, Germany) and MP system with an upright microscope (DM 6000 CFS; Leica Microsystems) through a 40× oil lens (NA = 1.25). The specimens were excited at 488 nm supplied by an Argon laser. DAPI was excited with multiphoton at 720 nm produced by a Ti:sapphire pulse laser (Chameleon Ultra II; Coherent, Santa Clara, CA). 

Results

Multifocal Retinal Degeneration

Retinal Histopathology of XLPRA2 Carriers.

At 3.9 weeks of age, no morphologic abnormalities could be seen in the maturing photoreceptors (Fig. 1A) . The earliest alterations that could be detected by 4.9 weeks of age on 1-μm-thick plastic sections consisted of patchy areas of OS misalignment and disruption resulting in a weaker PPDA staining (Fig. 1B) ; occasional pyknotic nuclei also were seen. At 5.9 weeks, there was more prominent OS disorganization that resulted in a clearer demarcation between diseased and nondiseased foci (Fig. 1C) . In diseased areas, rod cell density was decreased, yet consistently there was an increase in ONL thickness due to an increase in internuclear distance. By 26.1 weeks of age (Fig. 1D1) , there was a generalized thinning of the ONL. Patches of disease surrounded by structurally normal retina persisted, but they appeared to be fewer and of smaller size (Figs. 1D2 , 1D3) . This was even more obvious by 39 weeks of age. In older dogs (8.8 and 10.3 years) multifocal patches were rarely found, but endstage peripheral retinal atrophy that extended toward the mid-peripheral region was observed (Table 3) . Differences were observed between the two aged dogs both in the location and the extent of the retinal damage. In the 8.8 year-old dog, severe peripheral retinal damage (total atrophy to one to two rows of nuclei in the ONL) was seen in the superior region extending up to 7000 μm from the ora serrata. The central retina had a thin ONL (four rows of nuclei), and the photoreceptors had enlarged and short IS, but maintained their OS. Inferiorly, severe retinal damage (0–2 rows of nuclei) was seen throughout the entire retinal section. No OS were visible inferiorly, and the RPE was either directly apposed to the external limiting membrane (ELM) or to shortened IS. In the 10.3-year-old dog the peripheral and mid-peripheral regions of the retina were also severely affected, yet photoreceptor loss was more prominent superiorly than inferiorly (Figs. 1E1 1E2 1E3) . 

Retinal Histopathology of XLPRA1 Carriers.

Examination of H&E-stained cryosections from a 20- and 24-week-old XLPRA1 carrier did not reveal any morphologic alterations compatible with an ongoing disease process (Figs. 2A1 2A2 2A3) . The earliest age at which foci of disease were found was at the next time point, 1.4 years. These were observed throughout the retinal length and characterized by a local loss of OS and a decrease in rod density and ONL thickness (Figs. 2B1 2B2 2B3) . Beginning at 4.9 years of age and increasing with time, foci of increased cone density and rod loss were seen throughout the length of the retina (Figs. 2C1 2C2 2C3) . Similar to that reported in affected animals, ²⁴ substantial differences in disease severity were observed between carrier dogs of similar ages (e.g., dogs H6 and H74; Table 4 ). 

Immunohistochemical Characterization of Retinal Disease in XLPRA Carriers

Multifocal Retinal Disease in XLPRA2 Carriers.

Cell-specific markers for rods and cones were tested to determine whether they were differentially expressed in normal and diseased areas of the retina (Fig. 3) . Immunolabeling with rod opsin revealed alterations as early as 3.9 weeks of age that were similar to that previously reported in affected animals. ²³ These consisted of rod opsin mislocalization to the IS and ONL, and rod neurite sprouting (Fig. 3A1) . However, in the case of the carrier dogs examined in this study, these abnormalities had a multifocal distribution that overlapped with the areas of ONL thickening. DAPI labeling confirmed that areas of increased ONL thickness corresponded to areas of diseased rods (Fig. 3A2) . At 39 weeks of age the number of patches of rod opsin mislocalization and ONL thickening were considerably reduced (Figs. 3B1 3B2) . By confocal microscopy, we also confirmed that in those regions in which opsin was mislocalized and the ONL thickened, rod OS were misaligned and disorganized (Figs. 3C1 3C2) . 

Since cone opsins have been shown to mislocalize during the early phase of the disease in affected XLPRA2 dogs, we investigated whether this also occurs in carriers (Fig. 4) . Double immunolabeling with rod opsin and M/L opsin antibodies showed mislocalization of M/L cone opsin in young animals (3.9–12 weeks of age) in the same retinal patches where rod opsin was mislocalized (Figs. 4A1 4A2) . A peculiar finding that was detected as early as 5.9 weeks of age was the increased density of M/L cones in the patches where rod opsin was mislocalized (Figs. 4C1 4C2 4C3)A similar observation was made in cones labeled with human cone arrestin antibody (data not shown). In older animals (39 weeks), patches of M/L opsin mislocalization persisted despite a decrease in rod opsin mislocalization. Thus, there were fewer patches of rod opsin mislocalization than patches of M/L cone opsin mislocalization (data not shown), suggesting a possible loss of mutant rods and the persistence of diseased cones. The temporal pattern of S opsin labeling was slightly different since mislocalization, which also occurred in the areas where rod opsin mislocalized (Figs. 4B1 4B2) , was rarely seen after 12 weeks of age. No clumping of S cones was seen as with M/L cones. This finding was not surprising, since there as fewer S cones than M/L cones in the canine retina. 

At 26.1 weeks of age, phagocytic cells were frequently seen in the subretinal space of the degenerating patches, but not in the surrounding intact areas. Double immunolabeling with RPE65 and CD18 antibodies demonstrated that these cells were macrophages rather that detached RPE cells migrating into the subretinal space (Figs. 5A 5B) . 

Because photoreceptor degeneration in affected XLPRA2 dogs is associated with inner retinal remodeling, we investigated whether focal retinal disease alters the underlying inner retina in XLPRA2 carriers. PKCα and Goα labeling showed a retraction of the dendrites of rod bipolar cells that synapsed with the diseased rods (Fig. 6A1 , asterisks). GABA (γ-amino butyric acid) IHC, which was used to label GABAergic amacrine cells as well as the inner plexiform layer (IPL), did not show any focal alterations in the areas of photoreceptor disease in either young or older animals (data not shown). Müller cell reactivity, which was evidenced by intense GFAP labeling, occurred in the patches of photoreceptor disease as early as 8 weeks of age and was still intense at 12 weeks (Figs. 6B1 , 6B2) . At 26.1 weeks of age, radial extensions of GFAP labeling were considerably reduced, and by 39 weeks labeling was limited to astrocytes, and Müller cell end feet in the nerve fiber layer (data not shown). 

Multifocal Retinal Disease in XLPRA1 Carriers.

IHC experiments were limited to the OCT-embedded retinas of two dogs (20 and 24 weeks old). A patchy pattern of rod opsin and M/L cone opsin mislocalization was seen at both ages (Fig. 7) . Although rod neurite sprouting was clearly visible, we did not observe a thickening of the ONL at the sites of disease as in the XLPRA2 carriers. It is important to note that mislocalization of rod and M/L cone opsins occurred at an age when there are no structural abnormalities visible in the retina by light and electron microscopy in hemizygous males. ²⁴ S opsin mislocalization was seen in both animals but was limited to the peripheral retina (data not shown). No abnormal pattern in PKCα or Goα labeling was observed either, and GFAP labeling was limited to astrocytes and Müller cell end feet (data not shown). 

Kinetics of Photoreceptor Cell Death

A TUNEL assay was performed on XLPRA2 carrier retinas to determine the course of photoreceptor cell death. Quantification of TUNEL-labeled photoreceptors in XLPRA2 carriers showed a similar burst of cell death at 5.9 weeks of age (241–302 TUNEL-labeled cells per 10⁶ μm² of ONL) as previously reported in affected XLPRA2 dogs (Fig. 8Aand Ref. ²³ ). Between 12 and 26.1 weeks of age, there was a significantly reduced and more constant rate of cell death. By 39 weeks of age only two to eight TUNEL-labeled cells per 10⁶ μm² of ONL were counted. To confirm that cell death was occurring within the foci of retinal disease, we examined sections that were TUNEL-labeled and processed for rod opsin IHC. TUNEL-positive photoreceptors were only found in areas where rod opsin was mislocalized to the ONL (Fig. 8B) . In the 20- and 24-week-old XLPRA1 carriers, a limited number of TUNEL-labeled rods were observed and were restricted to the foci of rod opsin mislocalization (Fig. 8C) . 

Remodeling of the Retinal Mosaic in XLPRA2 Carriers

Evident thinning of the ONL was seen in older carriers of XLPRA2, and, in addition, there also seemed to be fewer disease patches (Fig. 9) . To further substantiate the observation that the retinal mosaic appeared to remodel in the young adult dog, we determined the proportion of retinal length that was occupied by foci of retinal degeneration as a function of age. This appraisal was performed, when possible, on sections from both the superior and inferior meridians. Our findings (Figs. 9A1 , 9A2 9B)show that at 5.9 weeks of age, large and numerous foci of disease occupied >40% of the retinal sections, whereas their number and size decrease markedly with age. By 39 weeks of age, less than 10% of the retina appeared to contain patches of disease. With time, the irregular waviness of the vitreous aspect of the ONL disappeared (compare Fig. 9A1with 9A2 ), but a substantial thinning was noted. At 39 weeks of age, the ONL thickness (superior: 36 ± 1 μm; inferior: 38 ± 1 μm) was reduced to approximately 44% to 57% of that measured at 3.9 weeks of age (superior: 82 ± 4 μm; inferior: 67 ± 2 μm; Fig. 9C ). Further reduction in ONL thickness in a 10.3 year-old carrier suggests that rare events of photoreceptor cell loss persist beyond the 39-week time point (Fig. 9C) . 

Immunohistochemistry confirmed these results by showing that in animals older than 26 weeks of age there was a significant decrease in the number and size of the patches of rod opsin mislocalization (compare Fig. 3B1with 3A1 ). It also confirmed that the ONL was thinner but appeared more uniform in older animals (compare Fig. 3B2with Fig. 3A2 ). In older animals, the cone mosaic remained irregular, and focal areas of increased cone density were frequently found (Fig. 4D1) . As early as 12 weeks of age, some of these patches of high cone density were characterized by the clumping of the cell somas, which often resulted in multiple layers of cones in the ONL (Figs. 4D2 4D3) . 

In the inner retina, the number of patches of dendritic retraction followed a similar time course as that of the patches of rod opsin mislocalization. By 39 weeks of age, these had become very rare, and most of the bipolar cells had a normal appearance (compare Fig. 6A2with 6A1 ). 

No such remodeling of the retinal mosaic was seen in young XLPRA1 carriers, which express a later onset and more progressive form of the disease. 

Discussion

A single report on the use of multifocal electroretinography in carriers of XLRP has shown that these patients experience a mosaic pattern of retinal dysfunction. ¹⁷ A similar multifocal pattern of disease has been confirmed by several histologic observations of retinas from older heterozygotes. ^{16 18 19 20} The present study examined the retinas in carriers of two canine models of RPGR-XLRP, and showed significant changes (summarized in Table 5 ) in the appearance of the retinal mosaic as a function of age, particularly in the early onset model. These remodeling events suggest that the young adult canine retina retains some degree of morphologic plasticity. 

In carriers of XLPRA2, the time of disease onset was similar to that previously reported in affected animals (∼4–5 weeks of age) and was also followed by a burst of rod cell death at 6 to 7 weeks of age. This further confirms our conclusion ²³ that a model of constant or exponentially decreasing risk of cell death ²⁶ does not apply to the kinetics of photoreceptor degeneration in XLPRA2. 

Surprisingly, the peak count of TUNEL-positive cells in heterozygotes was very close to that in affected animals (female homozygotes and male hemizygotes), both in timing and amplitude. In the context of random X-chromosome inactivation, we expected that this value would be approximately 50% of that in affected dogs. This finding may, therefore, reflect the occurrence of primary skewed X-inactivation, a process that has been implicated in female carriers with X-linked mental retardation disorders and hemophilia A (for review, see Ref. ²⁷ ). A noteworthy observation is that counts of TUNEL-positive cells in affected and carrier animals were similar but only at the time of occurrence of the peak of cell death. Indeed, at later time points, values in carriers were ∼50% of that reported in affected dogs. Thus, an alternative explanation could be that “mutant” photoreceptors undergoing degeneration may cause, through a bystander effect, the death of neighboring “normal” photoreceptors that are expressing the wild-type allele. Such a non–cell-autonomous mechanism of degeneration has been suggested to occur in the chimeric retina of a transgenic mouse line expressing both wild type and mutant (P347S) rhodopsin. ²⁸ This degeneration may be caused by the release of a diffusible apoptosis-promoting factor. ²⁹ If so, during the burst of photoreceptor death, there may be involvement of both intrinsic (mutation dependent) and extrinsic (toxic/apoptotic inducing factors) mechanisms of cell death. 

When comparing the histopathologic patterns between the XLPRA1 and XLPRA2 carriers, we noted that the mosaic gradually disappeared in XLPRA2 between 4.9 and 39 weeks of age, whereas it persisted, until at least 7.8 years of age, in XLPRA1. This time course may correlate with the time of onset of degeneration in these two diseases (XLPRA2: early onset; XLPRA1 late onset), and the ability of the retina to undergo a “corrective” remodeling event. We hypothesize that this remodeling may occur in the XLPRA2 carrier retina through tangential dispersion of photoreceptors expressing the wild-type RPGR allele into the degenerating patches. Such an event may only occur during the short time window that follows photoreceptor maturation in the dog and may no longer be present in the nonplastic adult retina such as that of a 1.4-year-old XLPRA1. It is to be determined whether postdevelopmental tangential dispersion of photoreceptors also occurs in human carriers of XLRP, as this may have practical implications for corrective gene therapy. Indeed, if rods and cones have the capacity to move to neighboring regions of diseased retina, then there is a risk that a treated region may end up being depleted of its rescued photoreceptors to the benefit of surrounding nontreated areas. This could have applications for the treatment of carriers of XLRP by opting for panretinal rather than localized delivery of the gene therapy construct. 

The early-onset and rapidly progressing disease that occurs in carriers and affected XLPRA2 dogs (homozygous females and hemizygous males) is thought to result from a toxic gain of function. The two nucleotide deletion (del1084-1085) in RPGRORF15 causes a frameshift that modifies significantly the amino acid sequence by adding 34 basic residues, increases the isoelectric point, and prematurely terminates translation 71 amino acids downstream. ²² The five nucleotide deletion (del 1028-1032) in XLPRA1 introduces a premature stop codon that truncates the RPGROR15 isoform from its 230 C-terminal amino-acids and slightly decreases the isoelectric point. ²² The current hypothesis is that this leads to a loss of function of RPGRORF15, and that this isoform may not be critical for retinal development and maturation. A similar correlation between disease phenotype and individual mutations in RPGR has been reported in two mouse models of RPGR-XLRP. ^{30 31}  

Besides the difference in the severity and time course of the patchy degeneration, XLPRA1 carriers differed from XLPRA2 carriers by exhibiting variation in disease severity at similar ages. This has also been found in XLPRA1-affected hemizygous male dogs ²⁴ and is most likely caused by the effect of modifier genes. No such modifiers have yet been identified; however, the candidate genes RPGRIP1, RANBP2, NPM1, PDE6D, NPHP5, and ABCA4 have been excluded. ³²  

A question that was not addressed in the present study is whether the remodeling of the cytoarchitecture of the carrier retina is paralleled by some functional plasticity. By 39 weeks of age, despite a loss of ∼50% of rods, the lamination of the ONL and the morphology of rod bipolar cells appear to have recovered from the earlier phase of multifocal degeneration. This potentially suggests that rewiring of second-order neurons to surviving (nonmutant) photoreceptors can occur. Assessing whether structural recovery leads to a period of improved retinal function is a worthwhile subject for future investigation. 

Several immunohistochemical and morphologic alterations observed in both the XLPRA1 and XLPRA2 carrier dogs have also been reported in human patients. ^{33 34} This includes the mislocalization of rod and cone opsins ^{16 19 20} which may underline the putative role of the RPGRORF15 isoform in trafficking proteins from the IS to the OS. Double fluorescent IHC showed that mislocalization of rod and cone opsins occurred within the same patch of disease. This finding was unexpected since, during retinal development, rods and cones that are clonally derived from the same photoreceptor progenitor cell have a different pattern of dispersion. Rods get distributed along a radial column whereas cones are tangentially dispersed. ^{35 36} The mislocalization of cone and rod opsins within the same patch could be explained by a defective developmental tangential dispersion of cones expressing the mutant RPGR allele. Alternatively, cone opsin mislocalization may be the consequence of a toxic bystander effect of the surrounding “mutant” rods on cones that express the wild-type RPGR allele. The lack of an antibody raised against the mutant canine RPGR isoforms currently precludes us from further investigating this question. 

Another common finding that was observed in RPGRORF15 carrier dogs and in a human patient is the focal aggregation of cones. ¹⁹ This was easily recognized in older animals and was often the only remaining histologic criteria that could be used to define a patch of disease. Careful examination of the cone distribution showed an increased density of cones in the patches of disease as early as 5.9 weeks of age. 

In XLPRA2, a “corrective” remodeling phase follows the initial burst of cell death, and leads to preserved retinal lamination despite a loss of approximately 50% of the rods by 39 weeks of age. However, this event is only transient and topographically restricted as large regions of degeneration/atrophy are found in the peripheral and mid-peripheral retina of older animals. A similar topographical distribution has been described in a human carrier. ¹⁹ Several mechanisms that are not mutually exclusive could explain this persistent loss of photoreceptors, and its regional predilection. It has recently been reported that in human fibroblasts at least 15% of all X-linked genes can escape random X chromosome inactivation. ³⁷ This secondary skewing of X inactivation has been shown to be associated with aging, although it is likely to be species-, tissue-, and locus-specific. ^{38 39} Thus, it is possible that in older carriers of XLPRA2, reactivation of the silenced mutant RPGR allele leads to a second wave of photoreceptor death. Another potential explanation is that there is a minimal number of rows of photoreceptors necessary to sustain viability of these cells. If such is the case, then gradual thinning of the ONL throughout the retinal surface is bound to have a more severe impact on peripheral regions where the ONL is several rows of ONL thinner than centrally. Finally, this late phase of peripheral degeneration could also be explained by the generation of new photoreceptors from progenitor cells located at the retinal margin. Several reports have suggested that this mechanism of photoreceptor renewal may not be limited to fish, amphibians, and birds, but also may be found in mammals. ⁴⁰ If this is the case in the canine retina, then newly formed photoreceptors expressing the mutant RPGR allele would not only undergo cell death but could also cause a detrimental bystander effect on nonmutant photoreceptors located in the peripheral retina. 

A significant finding of this study is that the morphologic appearance of the retinal mosaic changes with time and, as a consequence, so do the histologic criteria used to define patches of degeneration. This effect was clearly observed in XLPRA2, and may also occur in human carriers of early-onset forms of XLRP. The clinical implication of the remodeling events that occur in the carrier retina is that, depending on the stage of the disease, it may be challenging to diagnose, both in animal models and human patients, the mosaic pattern via in vivo imaging techniques such as optimal coherence tomography. In the case of XLPRA2, OCT imaging in a 5.9-week-old carrier would identify foci of increased and decreased ONL thickness corresponding respectively to patches of diseased and normal retina. On the other hand, examination at later stages would not show any mosaicism and would probably reveal either normal retinal lamination with generalized and uniform thinning of the ONL (e.g., 39 weeks of age), or prominent and extended retinal thinning with loss of lamination in the peripheral retina of much older animals (e.g., 10.3 years of age). 

In summary, characterization of the morphologic changes that occur in the carrier retina of two canine models of XLRP established that prominent remodeling events occur throughout the course of the disease and demonstrated that the retinal mosaic shows a dynamic and evolving response pattern. It also supports the evidence that heterozygotes for RPGR-XLRP mutations may experience severe visual impairment. ^{19 41} These results provide new insight to gene therapy approaches that can be used in both affected and carrier retinas. 

 Supported by National Institutes of Health Grants EY-06855, EY-13132, and EY-17549; the Foundation Fighting Blindness Individual Investigator Award and Center grants; a Fight for Sight Nowak family grant; Vision Research Center Grant P30 EY-001583; Hope for Vision Foundation; the ONCE International Prize for Biomedicine and R&D for New Technologies for the Blind; the Van Sloun Fund for Canine Genetic Research; and the University of Pennsylvania Research Foundation. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Eye Institute or the National Institutes of Health.

 Submitted for publication December 29, 2008; revised February 10, 2009; accepted May 21, 2009.

 Disclosure: W.A. Beltran, None; G.M. Acland, None; G.D. Aguirre, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

List of Dogs Used for the Morphologic, TUNEL, and Immunohistochemistry Studies

ID	Age	Morphology	TUNEL	IHC
XLPRA2 carrier
Z277	3.9 weeks	EPON	+	+
Z278	4.9 weeks	EPON	+	+
Z279	5.9 weeks	EPON	+	+
Z274	7.1 weeks	OCT	+	+
Z280	7.9 weeks	OCT	+	+
Z225	12 weeks	EPON	+	+
Z184	26.1 weeks	EPON	+	+
Z199	39 weeks	EPON	+	+
Z87	8.8 Years	OCT
Z24	10.3 Years	EPON
XLPRA1 carrier
H408	20 weeks	OCT	+	+
H414	24 weeks	OCT	+	+
H237 (40)*	1.4 Year	EPON
H6 (3)*	3.5 Years	EPON
H74 (18)*	3.8 Years	EPON
H43 (8)*	4.9 Years	EPON
H44 (7)*	5.7 Years	EPON
H25 (4)*	6 Years	EPON
H4 (2)*	7.8 Years	EPON

 Morphologic assessment was performed on either epoxy resin (EPON)-embedded sections or OCT-embedded cryosections. +, assay performed.

*  Number in parentheses corresponds to identification used in Zeiss et al. ²⁴

Table 2.

 

List of Primary Antibodies Used in This Study

Antigen	Host	Source, Catalog No. or Name	Working Concentration	Normal Retinal Localization
RPE65	Rabbit	T. Michael Redmond, NEI, Bethesda, MD	1:10,000	Retinal pigment epithelium
Human cone arrestin	Rabbit	Cheryl Craft, USC, Los Angeles, CA; LUMIF	1:10,000	Cone photoreceptors
M/L cone opsin	Rabbit polyclonal	Chemicon, Temecula, CA; AB5405	1:500	OS of M/L cones
S cone opsin	Rabbit polyclonal	Chemicon; AB5407	1:500	OS of S cones
Rod opsin	Mouse monoclonal	Chemicon; MAB5316	1:1,000	OS of rods
Synaptophysin	Rabbit polyclonal	DakoCytomation; Carpinteria, CA A0010	1:100	Neuron synapses, OPL, IPL
Protein kinase C (PKCα)	Mouse monoclonal IgG2b	BD Biosciences, Franklin Lakes, NJ; 610107	1:100	Rod bipolar cells
Goα	Mouse monoclonal IgG1	Chemicon; MAB3073	1:5,000	ON (rod and cone) bipolar cells
γ-amino butyric acid (GABA)	Rabbit polyclonal	Chemicon; AB5016	1:50	GABAergic amacrine cells
Glial fibrillary acidic protein (GFAP)	Rabbit polyclonal	DakoCytomation; Z0334	1:1,000	Astrocytes, Müller cells (reactive)
Glutamine synthetase	Mouse monoclonal	Chemicon; MAB302	1:20,000	Müller cells
CD-18	Mouse monoclonal IgG1	Leukocyte Antigen Laboratory, UC Davis, Sacramento, CA	1:10	None unless disease

Figure 1.

 

Retinal morphology in carriers of XLPRA2. (A) Normal maturation of the retina at 3.9 weeks. (B) Early patch of OS misalignment (below white bar) and ONL thickening in a 4.9-week-old carrier. (C) Prominent foci of OS disruption in a 5.9-week-old carrier. Note the increased thickness of the ONL, decreased cell density, and occasional pyknotic figures (arrows) in these patches. (D ₁ ) Multifocal retinal degeneration in a 26.1-week-old carrier of XLPRA2. Foci of retinal degeneration (arrows) are surrounded by areas of normal retinal morphology. (D ₂ ) High magnification of a patch of retinal degeneration. Note the misalignment and loss of photoreceptor OS, the shortening and broadening of the IS, and the decrease in cell density in the ONL. (D ₃ ) High-magnification at the transition zone of a patch of retinal disease. Note the progressive disorganization of the IS and OS and the decrease in chromatin compaction of the rods as one moves from a normal to a diseased area. (E ₁ –E ₃ ) Retinal disease in a 10.3-year-old carrier of XLPRA2. (E ₁ ) Despite a reduction in ONL thickness (∼5–6 nuclei thick) in the central region, the retinal morphology appears well preserved in comparison to the severe retinal degeneration and complete retinal atrophy observed, respectively, in the mid-peripheral (E ₂ ) and peripheral (E ₃ ) regions (epoxy resin, 1-μm sections). Scale bars: (A–C, E ₁ –E ₃ ) 20 μm; (D ₁ ) 80 μm. GCL, ganglion cell layer.

Table 3.

 

Thickness of the ONL in the central (S1), Mid-peripheral (S2), and Peripheral (S3) Retinas of XLPRA2 Carriers

ID	Age	S1*	S2*	S3*	Retinal Length (μm)
Z277	3.9 Weeks
Superior		ND	ND	ND	ND
Inferior		11–12	11	10	7,500
Z278	4.9 Weeks
Superior		11–12	10–11	9–10	12,600
Inferior		12	10	9	8,750
Z279	5.9 Weeks
Superior		10	8	7	11,700
Inferior		8	8	7	9,200
Z225	12 Weeks
Superior		11	10	ND	14,000
Inferior		9	8	7	10,700
Z184	26.1 Weeks
Superior		6	6	5	16,400
Inferior		6–7	6–7	5–6	11,000
Z199	39 Weeks
Superior		7	7	7	15,000
Inferior		8	7	6	12,000
Z87	8.8 Years
Superior		5–6	4–5	0	15,600
Inferior		3	3	0	11,400
Z24	10.3 Years
Superior		6	0	0	13,700
Inferior		7	5	5	12,000

*  Data are the mean number of rows of ONL nuclei. ND, not determined.

Figure 2.

 

Retinal morphology in carriers of XLPRA1. (A₁–A₃) H&E-stained retinal cryosections along the superior meridian of a 24-week-old carrier of XLPRA1. No morphologic abnormalities were detected at this age in the central (A₁), mid-peripheral (A₂), and peripheral (A₃) regions. (B₁–B₃) Epoxy resin retinal section of a 1.4-year-old carrier of XLPRA1. (B₁) Focal patches (arrows) of retinal degeneration were found throughout the retina. In these areas, there was a loss of the OS, a shortening and broadening of the IS, and a decrease in ONL cell density (B₃) and contrast with surrounding areas (B₂) that have a better preserved morphology. (C₁–C₃) Epoxy resin retinal section of a 7.9-year-old carrier of XLPRA1. (C₁) A generalized thinning of the ONL was seen throughout the retina, yet remaining photoreceptors maintained their normal IS and OS ( ). (C₂) Focal area of severe ONL thinning. (C₃) Higher magnification of (C₂) inset. In this foci, note the increased number of cone IS and nuclei and the reduced density of rods. Scale bars: (B₂, B₃) 10 μm; (A₁–A₃, C₁–C₂) 20 μm; (B₁, C₃) 40 μm. GCL, ganglion cell layer.

Table 4.

 

Central and Peripheral ONL Thickness along the Superior Meridian of XLPRA1 Carrier Retinas

ID	Age	Central	Peripheral
H408	20 Weeks	9	7
H414	24 Weeks	9	6–7
H237	1.4 Years	9	3–4
H6	3.5 Years	9	3
H74	3.8 Years	5	0
H43	4.9 Years	6–7	3–4
H44	5.7 Years	5	2–3
H25	6 Years	5	1–2
H4	7.8 Years	5	4–5

 Data are the mean number of rows of ONL nuclei per site.

Figure 3.

 

Disease and remodeling of the rod photoreceptor mosaic in carriers of XLPRA2. (A ₁ , A ₂ ) Double-fluorescence IHC in a 4.9-week-old XLPRA2 carrier showed numerous patches of rod opsin (green) mislocalization to the inner segments and ONL, but no abnormalities in cone arrestin labeling (red; A ₁ ). Colabeling with DAPI nuclear stain (blue) shows that patches of rod opsin mislocalization and rod neurite sprouting (inset) occur in areas of ONL thickening (A ₂ ). Double-fluorescence IHC in a 39-week-old XLPRA2 carrier showed a single patch of rod opsin (green) mislocalization to the inner segments and persistent normal expression of cone arrestin (B ₁ ). DAPI nuclear stain showed that the overall ONL thickness was reduced (B ₂ ). The inner margin of the ONL was less wavy than at earlier ages. However, persistent focal regions of ONL thickening were seen at the sites of rod opsin mislocalization (B ₂ ). Rod opsin and DAPI labeling in a 5.9-week-old XLPRA2 carrier imaged by confocal microscopy (maximal projection reconstituted image) (C ₁ ). Note the disorganized OS (arrowheads) in the patches of opsin mislocalization and ONL thickening, which contrast with the aligned rod OS (arrows) in adjacent areas of nondiseased retina. (C ₂ ) A single-plane image at higher magnification of the inset in (C ₁ ). Scale bars: (A ₁ , A ₂ , B ₁ , B ₂ ) 40 μm; (C ₁ ) 20 μm.

Figure 4.

 

Disease and remodeling of the cone photoreceptor mosaic in carriers of XLPRA2. Double-fluorescence IHC in a 4.9-week-old XLPRA2 carrier shows M/L opsin mislocalization to the inner segments, ONL, and cone pedicles (A₁) in the same patches where rod opsin was mislocalized ( ; A₂). (The labeling of the inner retina with the M/L opsin antibody was not specific, and the yellow in the cone OS was the result of a postacquisition increase in the green signal of the overlay image, to improve visualization of the rod opsin mislocalization; it does not represent colocalization). In a 5.9-week-old XLPRA2 carrier, faint S opsin mislocalization to the inner segments and ONL was seen in some cones (arrows; B₁). These cones are located within patches of rod opsin mislocalization (B₂). Immunolabeling in a 5.9-week-old XLPRA2 carrier showed an increased density in M/L cones in the patches of rod and M/L opsin mislocalization (C₁–C₃). (D₁) Plastic-embedded retinal section from a 26.1-week-old XLPRA2 carrier showed focal accumulation of large euchromatic nuclei resembling that of cones in an area of rod loss. (D₂, D₃) Double-fluorescence IHC in a 39-week-old XLPRA2 carrier showed increased density of cones that was either distributed in a monolayer (D₂) or clumped together (D₃). There was no rod opsin mislocalization in the patches of increased cone density. Scale bars: (A₁, A₂, B₁, B₂, C₂, C₃, D₂, D₃) 20 μm; (C₁) 40 μm; (D₁) 10 μm.

Figure 5.

 

Focal invasion of the subretinal layer by phagocytic cells in carriers of XLPRA2. (A) Plastic-embedded retinal section of a 26.1-week-old XLPRA2 dog showing a phagocytic cell (arrow) in the subretinal space overlying a patch of degeneration. The multiple cytoplasmic inclusions were suggestive of phagocytosed rod disc material. (B) Double-fluorescence IHC in this same animal labels showed that these cells (arrows) were labeled with CD18 but not with RPE65 antibodies, confirming their macrophage origin. Scale bar: 10 μm.

Figure 6.

 

Disease and remodeling of the inner retinal mosaic in carriers of XLPRA2 at 7.9 and 39 weeks of age. Double-fluorescence IHC in a 7.9-week-old XLPRA2 carrier showed focal dendritic retraction of rod bipolar cells (yellow) beneath the patches of rod opsin mislocalization ( ; A₁). Absence of rod bipolar dendritic retraction was noted in the 39-week-old XLPRA2 carrier (A₂). (B₁, B₂) GFAP IHC in a 7.9-week-old XLPRA2 carrier showed patchy labeling of the radial extensions of reactive Müller cells (B₁). Double IHC confirmed that these GFAP labeled radial extensions occurred in Müller cells located beneath the larger patches of rod opsin mislocalization (B₂). Scale bars: (A₁, A₂) 20 μm; (B₁, B₂) 40 μm.

Figure 7.

 

Photoreceptor mosaic in the XLPRA1 carrier. (A₁, A₂) Double fluorescence IHC in a 24-week-old XLPRA1 carrier showed numerous patches of rod opsin (green) mislocalization to the inner segments and ONL ( ), but no abnormalities in cone arrestin labeling (red; A₁). Colabeling with DAPI nuclear stain (blue) showed patches of rod opsin mislocalization and rod neurite sprouting (arrows; A₂). (B₁, B₂) Double-fluorescence IHC in a 24-week-old XLPRA1 carrier showed M/L opsin mislocalization to IS, ONL, and cone pedicle (B₁) in the same patches where rod opsin is mislocalized (B₂). M/L opsin labeling of the inner nuclear layer and ganglion cell layer was nonspecific. Scale bars: (A₁, A₂) 20 μm; (B₁, B₂) 40 μm.

Figure 8.

 

Photoreceptor cell death in carriers of XLPRA. (A) Kinetics of photoreceptor cell death in the superior meridian of XLPRA2 carriers. Inset: previously published ²³ kinetics of cell death in affected XLPRA2 dogs. (B) TUNEL labeling of rod nuclei (arrows) in a 4.9-week-old XLPRA2 carrier was seen exclusively in the patches of rod opsin mislocalization. (C) Single TUNEL-labeled rod (arrow) in a 24-week-old XLPRA1 carrier was located in a patch of rod opsin mislocalization.

Figure 9.

 

Remodeling of the retinal mosaic in carriers of XLPRA2. (A ₁ ) Numerous foci of retinal disease (horizontal black bars) occupy a significant proportion of the retinal length as seen in the mid-peripheral region of the superior retina of a 5.9-week-old XLPRA2 carrier. (A ₂ ) Infrequent and small foci of retinal disease were found at later ages, as illustrated in the mid-peripheral region of the superior retina of a 39-week-old XLPRA2 carrier. (B) Percentage (mean ± SD) of the total retinal length along the superior and inferior meridians that was occupied by foci of retinal disease in carriers of XLPRA2 of different ages. (C) Mean (±SD) ONL thickness in the mid-peripheral region of the superior and inferior meridians in XLPRA2 carriers of different ages. ONL thickness was measured in areas between patches of disease. Scale bars, 10 μm.

Table 5.

 

Summary of Morphologic Changes in Carriers and Affected Dogs of Two Models of XLRP

	XLPRA1	XLPRA2
Mutation in RPGRORF15 and putative effect	Del (1028–1032) causes a premature stop; loss of function	Del (1084–1085) causes a frameshift; toxic gain of function
Carriers (current study)
Time course of patchy disease/degeneration	Observed between 20 weeks and 7.8 years	3.9–39 weeks
Peak of cell death	ND	5.9 weeks
ONL thickness in patches of disease
Early	Unchanged	Increased
Late	Decreased	Uniform decrease (no patchy disease)
Opsin (rod, M/L, S) mislocalization	Yes	Yes
Clumping of cones in patches of disease	Yes	Yes
Inner retina remodeling below patches of disease	ND	Yes (observed at 7.9 weeks)
Tangential displacement of nonmutant rods into patches of disease.	Limited	Yes
Phagocytic cells in subretinal space	ND	Yes
Mid-peripheral retinal atrophy at later stages	Yes	Yes
Affected 22 23 24 42
Onset of disease	11 months	3.9 weeks
Peak of cell death	ND	6–7 weeks
Progression of disease	Slow	Rapid
Opsin (rod, M/L, S) mislocalization	ND	Yes
Inner retina remodeling	ND	Yes (observed at 11.9 weeks)
Phagocytic cells in subretinal space	Yes	Yes

 ND, not determined.

The authors thank Cheryl Craft (University of Southern California, Los Angeles) for providing the human cone arrestin antibody, Lingli Zhang (School of Veterinary Medicine, University of Pennsylvania, Veterinary imaging core facility) for help with confocal microscopy, Julie Jordan for technical assistance, and Mary Leonard (University of Pennsylvania Biomedical Art and Design facility) for the illustrations.