Genetic Ablation of Cone Photoreceptors Eliminates Retinal Folds in the Retinal Degeneration 7 (rd7) Mouse

Abstract

purpose. Folds or pseudorosettes in the outer retina are commonly observed in animals with genetic, viral, or chemically induced retinal degeneration. This study examined the effect of genetic ablation of cone photoreceptors on the production of retinal folds in two mouse retinopathy models. One is the rd7/rd7 retina, a model of enhanced S-cone syndrome, Goldman-Favre syndrome, and clumped pigmentary retinopathy that is also associated with an approximately twofold excess of S-cones. The other is postnatal day (P)0 N-methyl-N-nitrosourea (MNU) treatment.

methods. A transgene that directs cone-specific expression of the diphtheria toxin A chain was used to ablate cone photoreceptors. Retinal folds, numbers of photoreceptors, and numbers of cones were quantified in retina flatmounts or transverse sections, and photoreceptor apoptosis was quantified by immunostaining for activated caspase 3.

results. Cone ablation by the cone-DTA transgene eliminated folds in the rd7/rd7 retina, whereas chemical ablation of up to 30% of rods (by exposure to MNU at P13) had little or no effect on folds in the rd7/rd7 retina. Cone ablation by the cone-DTA transgene had no effect on retinal folds produced by P0 MNU treatment or on the progressive loss of rod photoreceptors in the rd7/rd7 retina.

conclusions. Despite their relatively low abundance, cones play a critical role in retinal folding in the rd7/rd7 retina. The relevant molecular and cellular mechanisms remain to be determined.

In diverse mammals, multiple microscopic retinal folds or pseudorosettes are associated with some inherited retinopathies, fetal or early postnatal exposure to cytotoxic chemicals or ionizing radiation, and a variety of fetal or early postnatal viral infections. These folds largely involve the photoreceptor layer and, to a lesser extent, the inner nuclear layer (INL). Inherited retinopathies with this phenotype are seen in several canine breeds, including Bedlington and Sealyham terriers, Australian shepherd dogs, and Labrador retrievers. ^{1 2 3} Retinal folds are also associated with numerous naturally occurring and genetically engineered retinopathies in the mouse, including loss-of-function mutations in the genes coding for Nr2e3, a rod photoreceptor-specific nuclear hormone receptor (the mouse retinal degeneration 7 (rd7) mutation) ^{4 5} ; Nrl, a basic motif-leucine zipper transcription factor ⁶ ; Crumbs-like-1 (Crb1), a large single-pass transmembrane protein (the mouse retinal degeneration 8 (rd8) mutation) ⁷ ; mammalian Hairy and enhancer of split homolog 1 (Hes1), a basic helix-loop-helix factor ⁸ ; and p56lck, a src-like nonreceptor protein tyrosine kinase. ⁹ Retinal folds are also associated with a transgenic model of Spinocerebellar ataxia type 7 (Sca7) caused by expansion of a polyglutamine tract in Ataxin 7, a cytosolic protein of unknown function, ^{10 11} with transgenic overexpression of E2F1, a transcription factor that controls cell proliferation and apoptosis, ¹² and with transgenic expression of Bcl-2 from the neuron-specific enolase promoter. ¹³ However, most inherited retinopathies in the mouse, whether spontaneous or engineered with transgenic or knockout technologies, are not associated with retinal folds, ¹⁴ indicating that folds are not an obligatory or a nonspecific consequence of progressive retinal degeneration. 

Chemical treatments that induce retinal folds include exposure to the antimitotic agents cytosine arabinoside (araC), ¹⁵ cisplatin, ¹⁶ and N-methyl-N-nitrosourea (MNU). ¹⁷ Retinal folds are observed only when exposure to these compounds occurs during the early postnatal period of rapid cell proliferation and not when exposure occurs later in life. A similar response is observed in monkeys and dogs after, respectively, intrauterine and neonatal exposure to ionizing radiation. ^{18 19} Interestingly, retinal folds are induced in the mature retina within several days of exposure to lovastatin, an inhibitor of cholesterol biosynthesis, or N-acetyl-S-trans,trans-farnesyl-l-cysteine (AFC), an inhibitor of protein prenylation, implicating prenylated proteins in the maintenance of outer retinal architecture. ²⁰  

Retinal folds are also seen after viral infection during retinal development in a variety of species. Examples include infection of neonatal rats with lymphocytic choriomeningitis virus or simian virus 40, ^{21 22} intrauterine infection of lambs with bluetongue virus, ²³ and neonatal infection of kittens with feline panleukopenia virus. ²⁴  

At present the molecular and cellular basis of retinal folding remains poorly understood. However, this phenomenon warrants further study because it may be relevant to the distinctive retinal thickening and dysplasia present in humans with mutations in CRB1 ²⁵ or NR2E3 ²⁶ and the visual dysfunction in humans with SCA7. ²⁷ It could also be relevant to the folds and rosettes seen in human retinal dysplasias associated with a variety of chromosomal anomalies or environmental insults ^{28 29 30} and with the rosettes that are typically seen in retinoblastoma. ³¹ As retinal folding necessarily involves a local detachment, the underlying mechanism(s) may also shed light on the biology of adhesion between the RPE and the retina. 

Materials and Methods

Mice

rd7/rd7 Mice (Nr2e3 ^rd7/Nr2e3 ^rd7) were obtained from the Jackson Laboratory (Bar Harbor, ME) and were mated to cone-DTA mice. ³² The resultant rd7/+;cone-DTA/+ mice were mated to rd7/rd7 mice to obtain the four experimental genotypes as littermates. Cone-DTA mice were mated to C57BL6 to obtain transgenic and nontransgenic littermates for MNU injection. Tail DNA was prepared and genotyped by PCR, as described. ^{32 33} All mouse procedures were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and the guidelines of the Animal Care and Use Committee of the Johns Hopkins Medical School. 

Immunostaining of Retinal Sections

Enucleated mouse eyes, with the dorsal-ventral position marked, were punctured at the cornea and fixed with 4% paraformaldehyde/PBS at room temperature for 1 hour. Fixed eyes were cryoprotected in PBS with 20% sucrose at 4°C overnight and then embedded in optimal cutting temperature compound (OCT; Tissue-Tek, Tokyo, Japan). Immunostaining of frozen sections was carried out essentially as described ³⁴ with the following antibodies: rabbit anti–short wavelength opsin (S-opsin, JH455) and anti–middle wavelength opsin (M-opsin, JH492) ³⁵ ; rabbit anti–cone transducin alpha (I-20, Santa Cruz Biotechnology, Santa Cruz, CA); rabbit anti–cleaved caspase-3 (#9661; Cell Signaling Technology, Danvers, MA); and rabbit anti–GFAP (Z0334; Dako, Carpinteria, CA). The number of cleaved caspase-3–positive cells was counted for at least seven adjacent 10-μm sections through or immediately adjacent to the optic disc. Images were captured on a confocal microscope (LSM510; Zeiss, Oberkochen, Germany). 

Immunostaining of Wholemount Retinas

After brief fixation of enucleated mouse eyes (with the dorsal position marked) in 4% paraformaldehyde/PBS at room temperature, the retinas were dissected and fixed with 4% paraformaldehyde/PBS for an additional hour at room temperature. After three washes in PBS, the retinas were incubated with a blocking buffer (PBS with 5% normal goat serum and 0.3% Triton X-100) and then incubated with primary antibodies in blocking buffer at 4°C overnight. The following day, the retinas were washed in wash buffer (PBS with 1% Triton X-100 and 1% Tween-20) for 4 hours at room temperature and then were incubated with secondary antibodies diluted in blocking buffer at room temperature for 2 hours. The retinas were then washed as described, and images were captured with the use of a microscope (Axiophot; Zeiss, Oberkochen, Germany) and modular imaging software (Openlab 3.1.2; Improvision, Lexington, MA). 

MNU Injection

MNU was injected into the dorsal fat pad at postnatal day (P)0 (the day of birth) at 60 mg/kg body weight or at P13 at doses of 10, 20, 30, 40, or 50 mg/kg body weight, as described. ¹⁷ For each genotype and MNU dose, eyes from at least four MNU-treated mice were examined at one month of age for P0 injections, and eyes from at least three MNU-treated mice were examined at 3 weeks of age for P13 injections. 

Quantification

To quantify the ablation of cones by the cone-DTA transgene, 325-μm × 257.5-μm images of flatmounted retinas stained with Gnat2 were captured roughly halfway between the optic disc and either the dorsal or ventral edges of the retina (Fig. 1 ; Table 1 ). Images were divided into four equal sectors, and the number of cones was counted and averaged from a 100-μm × 100-μm area of each sector. Given that Gnat2 is expressed in rods at a low level relative to cones in rd7/rd7 and rd7/rd7;cone-DTA retinas, only strongly stained (i.e., cone) outer segments were counted. For quantifying the effects of the cone-DTA transgene and P13 MNU exposure on the total number of photoreceptors (see Fig. 4G ), the numbers of nuclei in the INL and outer nuclear layer (ONL) were counted in transverse sections and averaged from four confocal images of 100-μm width derived from zones that were approximately halfway between the optic disc and the edge of the retina. Calculating the ratio of ONL to INL nuclei controlled for any deviation of the sectioning angle with respect to the plane of the retina. 

Results

Cone-DTA Transgene Eliminates Retinal Folds in rd7/rd7 Mice

The rd7/rd7 retina has approximately twofold more S-opsin–positive photoreceptors (S-cones) than the wild-type (WT) mouse retina. ^{5 34} Like WT S-cones, the rd7/rd7 S-cones are enriched in the ventral retina. However, the cell bodies of the extra rd7/rd7 S-cones are distributed throughout the inner two thirds of the ONL, in contrast to the normal location of WT cone cell bodies in the outer one third of the ONL. M-cones in the adult rd7/rd7 retina resemble WT M-cones in number and dorsal enrichment and in the location of their cell bodies within the outer third of the ONL. However, at P14, when WT M-cones are found at their adult locations, rd7/rd7 M-cone cell bodies are distributed throughout the ONL. ³⁴ The adult rd7/rd7 retina is also characterized by derepression of multiple cone-specific genes in rod photoreceptors. ^{34 36}  

In the mouse, cones constitute approximately 3% of photoreceptors, with rods constituting the remaining 97% of photoreceptors. ³⁷ Macroscopic defects that characterize the rd7/rd7 retina—retinal folding and progressive photoreceptor degeneration—affect cones and rods and could arise from defects in either rods or cones or both. To separate the contributions of cone and rod photoreceptors to retinal folding and progressive photoreceptor degeneration, we genetically ablated cone photoreceptors with a cone-specific diphtheria toxin A transgene (referred to as cone-DTA). ³² This transgene is controlled by the human L-opsin promotor, which, in the mouse, is expressed in M- and S-cones beginning at approximately P5. ³⁵  

Consistent with the findings of Soucy et al., ³² the cone-DTA transgene eliminates nearly all M-cones and most S-cones in the phenotypically normal rd7/+ retina, with a few S-cones surviving in the ventral retina (Figs. 1A 1B 1C 1D 1E 1F 1G 1H and 2A 2B 2C 2D 2E 2F ; Table 1 ). Similarly, in the rd7/rd7 retina, the cone-DTA transgene eliminates nearly all M-cones (Figs. 1I 1J 1K 1L 1M 1N 1O 1P and 2G 2H 2I 2J 2K 2L ; Table 1 ). Interestingly, a larger proportion of S-cones survive in the rd7/rd7;cone-DTA retina than in the rd7/+;cone-DTA retina (compare Figs. 1F and 1Hversus 1N and 1P and Figs. 2D 2E 2Fversus 2J–L). Cell bodies of these surviving S-cones are located in the outer one third (normal) and the inner two thirds (ectopic) of the ONL, as judged by in situ hybridization with an S-opsin probe (data not shown). Independent of the presence of the cone-DTA transgene, the genes coding for cone transducin alpha (Gnat2) and multiple other cone-specific proteins are derepressed in rods in the rd7/rd7 retina (Figs. 2G 2J , and data not shown). (We note that Gnat2 is expressed at a lower level in rods than in cones, and in retina flatmounts Gnat2 immunostaining in rods appears as a uniform low-level signal [Fig. 1and Chen et al.³⁴]. The Gnat2 immunostaining in Figure 1shows the strongly labeled M- and S-cone outer segments above the rod signal.) Consistent with the cone specificity of the cone-DTA transgene and the earlier observation that cone visual pigments are not among the genes derepressed in rd7/rd7 rods, ^{34 36} we observe no disorganization or loss of rod photoreceptors in rd7/rd7;cone-DTA retinas beyond the slow rod loss characteristic of the rd7/rd7 retina. 

Remarkably, rd7/rd7;cone-DTA retinas completely lack folds (Figs. 1 2) . In multiple crosses between rd7/rd7; cone-DTA and rd7/rd7 parents, retinal folds were found in all rd7/rd7 progeny, as expected (n > 10) but were never observed in rd7/rd7;cone-DTA littermates (n >10). This comparison among littermates rules out the possibility that the absence of folds arises from differences in genetic background (except in the unlikely event that the relevant background difference happens to be tightly linked to the cone-DTA transgene). Because retinal folds in rd7/rd7 mice gradually disappear between 5 and 16 months of age, ⁴ the analysis was performed on retinas from mice at 1 month of age, a stage when retinal folds are seen in all rd7/rd7 mice. These data imply that in the rd7/rd7 retina, the presence of normal or ectopic cones is required for retinal folding. 

Cone-DTA Transgene Does Not Block the Retinal Degeneration in rd7/rd7 Mice

Next, we investigated the effect of eliminating retinal folds on degeneration in the rd7/rd7 retina. Although the rd7/+ and rd7/+;cone-DTA retinas show little GFAP activation in Müller glia (Figs. 3A 3D) , a surrogate marker of neuronal disease and injury, ³⁸ in the rd7/rd7 retina there is prominent GFAP activation, especially around retinal folds (Fig. 3B) , in agreement with a previous report of localized Endothelin-2 transcript accumulation at these sites. ³⁹ Concomitant with the elimination of retinal folds in the rd7/rd7;cone-DTA mice, GFAP activation is much reduced and the remaining GFAP activation suggests a uniform and low-level stress response (Fig. 3E) , possibly related to rod photoreceptor abnormalities. 

In mature rd7/+ and rd7/+;cone-DTA retinas, as in the mature WT mouse retina, virtually no apoptosis occurs (Fig. 3Gand data not shown). However, significant numbers of apoptotic photoreceptors are seen in the rd7/rd7 retina, as judged by anti–cleaved caspase-3 immunostaining (Fig. 3C) ; somewhat surprisingly, these apoptotic photoreceptors are not enriched in the folded (and detached) regions. Similar numbers of apoptotic photoreceptors are observed in the rd7/rd7;cone-DTA retina (Figs. 3F 3G) . Thus, the progressive loss of rd7/rd7 rods is unaffected by the elimination of most of the cone photoreceptors and by the elimination of all retinal folds. Most likely, the progressive loss of rd7/rd7 rods is caused by intrinsic transcriptional abnormalities resulting from the absence of Nr2e3. ^{34 36}  

Unlike Cone-DTA–Dependent Ablation of Cones, MNU-Induced Elimination of Photoreceptors Only Weakly Reduces the Number of Retinal Folds

It is possible that the elimination of folds in the rd7/rd7;cone-DTA retina is caused not by ablation of cones per se but by the modest reduction in total photoreceptor cell number irrespective of whether the photoreceptors are rods or cones. In particular, if retinal folding represents a response to overcrowding of photoreceptors, any treatment that decreases the number of photoreceptors should also decrease the number of folds. Consistent with this idea is the progressive disappearance of almost all folds in rd7/rd7 retinas between ages 3 weeks and 6 months, concomitant with the slow loss of approximately 20% rods (Fig. 4G) . This overcrowding model predicts that eliminating at least 20% of photoreceptors should also eliminate most of the retinal folds from the rd7/rd7 retina. 

To test this prediction, we took advantage of the remarkable photoreceptor-specific cytotoxicity of MNU when delivered systemically to mice between P11 and P20. ¹⁷ One week after injecting WT or rd7/rd7 mice with a single MNU dose of 10, 20, 30, 40, or 50 mg/kg at P13, we observed a dose-dependent elimination of photoreceptors with little or no change in the number of cells in the INL in WT or rd7/rd7 retinas (Fig. 4) . Despite the elimination of approximately 30% of photoreceptors after a 30-mg/kg MNU injection, the number of folds in the rd7/rd7 retina was unchanged (Figs. 4E 4G) . However, the elimination of approximately 75% of photoreceptors, after a 50-mg/kg MNU injection, reduced the number of folds in the rd7/rd7 retina by approximately 70% (Figs. 4F 4G) . By contrast, the number of photoreceptors ablated by the cone-DTA transgene, in either WT or rd7/rd7 retinas, is so small as to be below the limit of reliable quantification when counting ONL/INL cell ratios (Figs. 2 4G) . Thus, though massive photoreceptor loss after MNU exposure at P13 can lead to a decrease in the number of retinal folds, the mechanisms associated with this process cannot account for the loss of retinal folds that arises from the ablation of a far smaller number of cone photoreceptors by the cone-DTA transgene. 

Cone-DTA Transgene Does Not Eliminate Retinal Folds Induced by MNU Exposure at P0

Finally, to examine the generality of cone photoreceptor ablation in eliminating retinal folds, we tested the ability of the cone-DTA transgene to eliminate retinal folds that had been induced chemically. In contrast to the uncomplicated photoreceptor loss that follows MNU treatment of WT mice at P13 (Figs. 4A 4B 4C) , MNU treatment at P0 produces retinal dysplasia with multiple folds. ¹⁷ As shown in Figure 5 , after a single 60-mg/kg dose of MNU at P0, ¹⁷ retinas from WT and cone-DTA mice at 1 month of age contained equal numbers of folds (n = 4 mice of each type), despite an efficiency of cone ablation by cone-DTA that was comparable to that seen in retinas not treated with MNU (compare Fig. 5Bwith Figs. 2E 2F ). The failure of cone-DTA–induced cone ablation to ameliorate P0 MNU-induced retinal folds suggests that retinal folds induced by MNU and by loss of Nr2e3 involve mechanistically distinct processes. Consistent with this view, we note that in the rd7/rd7 retina, the density of folds is highest near the center of the retina, whereas after MNU treatment at P0, the density of folds is highest in the retinal periphery. 

Discussion

The principal results of this study are that the folds present in the rd7/rd7 retina could be completely eliminated by early ablation of most cone photoreceptors, that the elimination of as many as 30% of the rods by MNU exposure at P13 had little or no effect on the number of folds in the rd7/rd7 retina, and that retinal folds associated with P0 MNU exposure were unaffected by cone ablation. We also observed that the slow and progressive loss of rods in the rd7/rd7 retina was unaffected by cone ablation and that Müller cell activation in the rd7/rd7 retina was largely correlated with the presence or absence of retinal folds and not with the slow photoreceptor degeneration. Although it was not the focus of this study, we note that the disappearance of folds in the rd7/rd7 and P0 MNU retinas over the first several months of postnatal life implies a dynamic restructuring of the retina-RPE interface to accommodate new zones of retinal attachment. 

In the rd7/rd7 retina, it is interesting that folds are present at equal density in the dorsal and the ventral parts of the retina, even though excess S-cones are largely confined to the ventral retina. It is also interesting that the folds are completely eliminated in the ventral retina despite only a twofold reduction in S-cone number by the cone-DTA transgene. The difference between M-cone and S-cone ablation efficiencies in the retinas of WT and rd7/rd7 mice most likely reflects differences in the strength of transgene expression in the two cell types. In the cone-DTA transgene used here, the DTA open reading frame is under the control of the human L-opsin promoter, which has been shown to direct expression to both M- and S-cones in the mouse but with a higher expression level in M-cones. ³⁵ It is also possible that the S-cones in the rd7/rd7 retina are less sensitive to ablation by the cone-DTA transgene than are normal S-cones. 

It is reasonable to speculate that folding in the rd7/rd7 retina arises from a greater surface area of photoreceptors and Müller cells at the external limiting membrane compared with the surface area of the RPE. Under this model, the retina folds to fit its larger surface area onto a limited RPE surface. This simple model nicely accounts for the progressive loss of retinal folds as the rd7/rd7 rods slowly die, a process that would be predicted to relieve overcrowding in the photoreceptor layer. By the same logic, this model could also account for the elimination of retinal folds after ablation of most of the cones. 

However, quantitative analysis of our experimental data provides three arguments against the simplest version of a “photoreceptor crowding” model in which the only relevant variable is the total number of photoreceptors. First, at 3 weeks of age, when the rd7/rd7 retina displays numerous folds, WT and rd7/rd7 retinas have nearly identical numbers of photoreceptors (Figs. 4A 4D 4G) . Second, in the rd7/rd7 retina, the total number of cells ablated by the cone-DTA transgene represents only approximately 3% to approximately 4% of photoreceptors. (This estimate comes from the following calculation: in the WT mouse, cones account for approximately 3% of photoreceptors. ³⁷ Table 1shows that in the rd7/rd7 retina, the number of cones has increased by approximately 160%, equivalent to approximately 5% of photoreceptors. Of these, two thirds—i.e., approximately 3% to approximately 4% of all photoreceptors—are ablated by the cone-DTA transgene, including nearly all cones in the dorsal retina and approximately 50% of cones in the ventral retina.) Third, eliminating up to 30% of rods by P13 MNU exposure had essentially no effect on retinal folds in the rd7/rd7 retina. Taken together, these observations suggest that cones confer a propensity for retinal folding that is greater than their small numbers might suggest. Consistent with this idea, the all S-cone retina of the Nrl KO mouse also exhibits folds. ⁶  

What might be the mechanism(s) by which excess cones induce retinal folding? One possibility is that cones form points of contact with extracellular matrix, neighboring rods, or Müller glia to control expansion or contraction of the photoreceptor surface of the retina. A second, and not mutually exclusive, possibility is that excess cones disrupt the columnar organization of rod cell bodies in a manner that leads to expansion of the photoreceptor surface. 

The work reported here represents an initial step in exploring the mechanistic basis of retinal folding. Although the relevant molecular and cellular mechanisms remain to be determined, the data indicate that in some contexts cones play essential roles in retinal folding. It will be interesting to determine whether alterations in cone numbers or cone properties play a role in other retinal disorders that include folds. 

 Supported by the National Eye Institute and the Howard Hughes Medical Institute.

 Submitted for publication August 4, 2006; revised January 23, 2007; accepted April 2, 2007.

 Disclosure: J. Chen, None; J. Nathans, 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.

 

Figure 1.

 

Flatmounted retinas from rd7/+, rd7/+;cone-DTA, rd7/rd7, and rd7/rd7;cone-DTA mice. Flatmounted retinas from 1-month-old littermates of the indicated genotypes—rd7/+ (A–D), rd7/+;cone-DTA (E–H), rd7/rd7 (I–L), and rd7/rd7;cone-DTA (M–P)—were immunostained for cone transducin alpha (Gnat2) and were viewed under differential interference contrast (DIC) mode focusing on the ONL (A, C, E, G, I, K, M, O) or under fluorescence mode focusing on photoreceptor outer segments (B, D, F, H, J, L, N, P). Given the reciprocal gradient of the two types of cones across the mouse retina with the highest density of M- and S-cones in the dorsal and ventral parts of the retina, respectively, both dorsal (A, B, E, F, I, J, M, N) and ventral (C, D, G, H, K, L, O, P) regions are shown. Retinal folds are only observed in retinas from rd7/rd7 mice and are seen as round objects of approximately 50-μm diameter in DIC mode and as clusters of Gnat2-stained outer segments in fluorescence mode. Almost all of the intense Gnat2 staining in retinas with the cone-DTA transgene comes from surviving S-cones. In rd7/rd7 retinas, Gnat2 is also present at a lower level in rod outer segments. The high intensity of Gnat2 immunostaining in the folds, and especially at the annulus of each fold, reflects the vertical superposition of large numbers of immunoreactive cone outer segments (and, at lower intensity, rod outer segments) as seen in Figures 2G and 2H . Scale bar: 50 μm.

Table 1.

 

Efficiency of Cone Ablation by the Cone-DTA Transgene

Genotype	Cone Outer Segments per 10,000 μm2
		Dorsal Retina	Ventral Retina
rd7/+	179 ± 11	156 ± 3
rd7/+; cone-DTA	1 ± 1	14 ± 1
rd7/rd7	240 ± 14	309 ± 16
rd7/rd7; cone-DTA	3 ± 1	149 ± 19

 All values are mean ± SD (n = 4). Cone outer segments were visualized in flatmounted retinas using Gnat2 immunostaining (see Fig. 1 ).

Figure 2.

 

Retinal sections from rd7/+, rd7/+;cone-DTA, rd7/rd7, and rd7/rd7;cone-DTA mice. Retinal sections from 1-month-old littermates of the indicated genotypes—rd7/+ (A–C), rd7/+;cone-DTA (D–F), rd7/rd7 (G–I), and rd7/rd7;cone-DTA (J–L)—were immunostained for Gnat2, M-opsin, or S-opsin, and nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI). Dorsal (A, B, D, E, G, H, J, K) and ventral (C, F, I, L) regions are shown. Vertical brackets (right) show the extent of the ONL. Scale bar: 20 μm.

Figure 3.

 

GFAP activation and photoreceptor apoptosis. Retinal sections from 1-month-old littermates of the indicated genotypes—rd7/+ (A), rd7/rd7 (B, C), rd7/+;cone-DTA (D), and rd7/rd7;cone-DTA (E, F)—were immunostained for GFAP or cleaved Caspase-3 (Casp3; arrowheads), and nuclei were counterstained with DAPI. Vertical brackets, right: extent of the ONL. (G) Quantification of cells in the ONL labeled with anti–cleaved Caspase-3 per 10-μm retinal section. Circles: numbers of anti–cleaved Caspase-3–positive cells per section immediately adjacent to the optic disc. Scale bar: 20 μm.

Figure 4.

 

Ablation of photoreceptors by MNU exposure at P13 does not eliminate retinal folds in rd7/rd7 mice. DAPI-stained retinal sections from P20 WT (A–C) or rd7/rd7 (D–F) mice injected with MNU at the indicated dose at P13. Vertical brackets, right: extent of the ONL. (G) Quantification of the number of retinal folds per 10-μm section (cut through or immediately adjacent to the optic nerve; upper panel) and the ratio of the nuclei in the ONL and the INL (lower panel) of mice of the indicated age, treatment, and genotype. Shaded bars: indication of severity of retinal folding (lower panel). The ratio of ONL to INL nuclei in the uninjected +/+ (i.e., WT) retina (7th column) is slightly lower than that in the uninjected rd7/+ retina (1st column), possibly because the uninjected +/+ mice were raised together with MNU-injected littermates and therefore might have been exposed to a low level of MNU. NA, not applicable. Scale bar: 40 μm.

Figure 5.

 

Genetic ablation of cone photoreceptors does not eliminate retinal folds induced by MNU exposure at P0. Retinal sections from 1-month-old +/+ (i.e., WT) (A) or cone-DTA (B) littermates that were injected with MNU at P0 were immunostained for S-opsin and nuclei were counterstained with DAPI. Vertical brackets: extent of the ONL. Scale bar: 20 μm.

The authors thank Samuel Jacobson, Amir Rattner, and Yanshu Wang for advice, and Hugh Cahill and an anonymous referee for helpful comments on the manuscript.