Downregulation of Cone-Specific Gene Expression and Degeneration of Cone Photoreceptors in the Rpe65 ^−/− Mouse at Early Ages

Abstract

purpose. RPE65 is essential for the generation of 11-cis retinal. Rod photoreceptors in the RPE65-knockout (Rpe65 ^−/−) mouse are known to degenerate slowly with age. This study was designed to examine cone photoreceptors and the expression of cone-specific genes in the Rpe65 ^−/− mouse.

methods. Gene expression changes were identified by microarray and confirmed by real-time RT-PCR. Cone photoreceptors were stained by peanut agglutinin (PNA) lectin in the flatmounted retina. The 9- or 11-cis retinal was supplied by intraperitoneal injections.

results. The short-wavelength (SWL) cone opsin mRNA was markedly decreased at 2 weeks of age, whereas the decrease in the middle-wavelength (MWL) cone opsin mRNA occurred relatively later in age. In contrast, the rhodopsin mRNA level did not show any significant change at all the ages analyzed. Consistent with the cone opsin changes, the cone transducin α-subunit mRNA decreased at both 4 and 8 weeks of age, whereas again the rod transducin α-subunit did not show any significant change. Rpe65 ^−/− mice showed significant cone loss in both the central and ventral retina between 2 and 3 weeks of age. Administration of 9- or 11-cis retinal to Rpe65 ^−/− mice 2 weeks of age increased cone density by twofold in these areas.

conclusions. In the Rpe65 ^−/− mouse, the expression of cone-specific genes is downregulated and is accompanied by cone degeneration at early ages. Early administration of 9- or 11-cis retinal can partially prevent cone loss, suggesting that the absence of 11-cis chromophore may be responsible for the early cone degeneration.

The RPE65 protein is an important component of the retinoid visual cycle and is expressed in the retinal pigment epithelium (RPE). ¹ In addition, we have shown that RPE65 is present in cone photoreceptors of several different species. ² It has been shown that mutations in the RPE65 protein are associated with several forms of inherited retinal dystrophies, such as retinitis pigmentosa, ³ Leber’s congenital amaurosis, ⁴ autosomal recessive childhood-onset severe retinal dystrophy, and early-onset severe rod–cone dystrophies, ^{5 6 7 8 9} all of which can cause severe vision loss. 

The homozygous RPE65-knockout (Rpe65 ^−/−) mouse lacks 11-cis retinal and accumulates excessive levels of all-trans retinyl esters in the RPE, ¹ suggesting that the RPE65 protein is essential for the isomerization of all-trans retinyl esters to generate 11-cis isomers. Electroretinogram (ERG) recordings in Rpe65 ^−/− mice revealed small rod but no cone responses in both young adult and aged animals. ^{10 11} The remaining rod response is supported by the generation of isorhodopsin. ^{10 11 12} Consistent with the functional analysis, the Rpe65 ^−/− retina is known to contain no detectable rhodopsin, whereas it has normal rod opsin structure and close to normal amounts of regenerable opsin, ¹⁰ suggesting that the absence of chromophore is responsible for the diminished ERG response in rods. Histologic results showed a gradual reduction of rod outer segment length and a decrease in number of photoreceptor nuclei in aged Rpe65 ^−/− mouse retinas. ^{1 13} Recent studies have shown that systemic administration of 9- or 11-cis retinal can partially restore rhodopsin regeneration and thereby improve rod responses in the Rpe65 ^−/− mouse. ^{10 11 12}  

In contrast to the knowledge about rods, little is known about cone structure and function in the Rpe65 ^−/− mouse. Although it has been suggested by an early study that cone function may be preserved in the 1- to 12-month-old Rpe65 ^−/− mouse, ¹⁴ experiments by Seeliger et al. ¹⁵ have demonstrated that the remaining photoresponse at 4 to 5 weeks is from rod photoreceptors. These researchers did not find evidence of cone function. 

The present study examined cones and cone opsin expression in the Rpe65 ^−/− mouse by combining immunohistochemical and molecular biological analyses to elucidate expression changes in cone-specific genes related to cone degeneration in this mouse model. 

Methods

Animals

Animals were kept in a 12-hour light–dark cycle with an ambient light intensity at the eye level of the mice of 85 ± 18 lux. Rpe65 ^−/− mice were genotyped as described previously. ¹ Wild-type (wt) C57BL/6 mice were purchased from Harlan (Indianapolis, IN). Care, use, and treatment of the animals were in strict agreement with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Microarray Procedures and Data Analysis

Age-matched wt and Rpe65 ^−/− animals (postnatal day [P]60) were killed by decapitation and the retinas isolated. All mice used for the gene expression analysis were euthanatized at 11 AM to eliminate any potential circadian rhythm effects. Retinas from four animals per genotype were pooled, and data points were examined in duplicate. Total RNA was isolated (TRIzol; Invitrogen-Life Sciences, Gaithersburg, MD), followed by a clean-up with minicolumns (RNAeasy; Qiagen, Valencia, CA). The quality of the RNA was examined by gel electrophoresis (quantity and integrity of the 18s and 28s ribosomal RNA bands) and spectrophotometry (260/280 nm ratio). 

Five micrograms of total RNA was used to generate double-stranded cDNA (Invitrogen, Carlsbad, CA), which then served as a template for the generation of biotinylated cRNA (BioArray HighYield RNA transcript labeling kit; Enzo Diagnostics; Farmingdale, NY). The labeled, purified probes were fragmented in 8 M Na⁺-citrate buffer and used for hybridization on U74A oligonucleotide arrays (Affymetrix, Santa Clara, CA). Hybridization and readout were performed by the DNA Microarray Core Facility at the Medical University of South Carolina (Affymetrix Fluidics Station, used according to the instructions indicated in the Affymetrix Expression Analysis Technical Manual). 

Microarrays were scanned (Affymetrix scanner and Microarray Suite 5.0 software; Affymetrix). The expression data were normalized using Dchip, a model based program that allows for the comparison of multiple arrays. ¹⁶ Data from duplicates were averaged, expressed as mean ± upper and lower boundaries and filtered with respect to the multiples of change (x-fold). 

Quantitative Real-Time Reverse Transcription–PCR

To verify the data obtained from microarrays, mRNA levels of the short- (SWL) and middle-wavelength (MWL) cone opsin, rhodopsin, and cone and rod transducin α-subunits were analyzed by quantitative real-time RT-PCR using specific primer pairs (Table 1) . By using the Primer3 software (http://www-genome.wi.mit.edu/cgi-bin/primer/primer3.cgi/ provided in the public domain by the Whitehead Institute, Massachusetts Institute of Technology, Cambridge, MA), all primer sets were designed from mRNA sequences spanning big introns to avoid amplification from possible genomic DNA contamination. The primer sequences were checked by a BLAST search to assure sequence specificity. 

Four mice of each age group in each genotype were used for real-time RT-PCR. RNA was isolated from the two retinas of each mouse individually. RT reaction was performed as described previously. ¹⁷ The RT products were diluted to 1:10, and 2 μL of each of the diluted RT products, and 3 picomoles of primers was used for the two-step PCR in a final volume of 25 μL. The PCR included a denaturation and hot start at 95°C for 10 minutes, followed by 45 cycles with melting at 95°C for 15 seconds and elongation at 60°C for 60 seconds. Fluorescence changes were monitored after each cycle (SYBR Green; Applied Biosystems, Inc., Foster City, CA). Melting curve analysis was performed (0.5°C/s increase from 55°C to 95°C with continuous fluorescence readings) at the end of 45 cycles to ensure that specific PCR products were obtained. Amplicon size and reaction specificity were confirmed by electrophoresis on a 2.5% agarose gel. All reactions were performed in triplicate. Results were evaluated by computer (SmartCycler II software; Cepheid, Sunnyvale, CA). The average C _T (threshold cycle) of fluorescence units was used for analysis. Each mRNA level was normalized by the 18s rRNA levels. Quantification was calculated using the C _T of the target signal relative to the 18s rRNA signal in the same RNA sample. Effects were quantified and expressed by the x-fold change method calculated as  

\[x-\mathrm{Fold\ Change}{=}2^{{-}{\Delta}({\Delta}C_{\mathrm{T}})}\ \mathrm{with}\ {\Delta}C_{\mathrm{T}}\ {=}\ C_{\mathrm{T,target}}\ {-}\ C_{\mathrm{T,18s\ rRNA}}\ \mathrm{and}\ {\Delta}({\Delta}C_{\mathrm{T}})\ {=}\ {\Delta}C_{\mathrm{T,wt}}\ {-}\ {\Delta}C_{\mathrm{T,RPE65}^{{-}/{-}}}.\]

The mRNA levels were averaged in the four Rpe65 ^−/− mice of each age group and compared to those of the age-matched wt control. 

Cone Density Analysis

Wholemounted Retina Preparation.

Retinas of the experimental animals were prepared as described previously. ² Briefly, the retina–lens complex was fixed in 4% formaldehyde solution in phosphate-buffered saline (PBS, pH 7.4). After several washes in PBS, FITC-conjugated peanut agglutinin (PNA; Sigma-Aldrich, St. Louis, MO) was added for overnight incubation. After several washes in PBS, the lens was detached from the retina, which was flatmounted and covered by a coverslip after the application of several drops of anti-fade solution (Prolong; Molecular Probes, Eugene, OR). 

Microscopic Analysis.

The samples were analyzed with a fluorescence microscope (Axioplan II; Carl Zeiss Meditec, Inc., Jena, Germany) equipped with a digital camera. Images were captured (Spot-RT Camera, with Spot software, ver. 3.0; Diagnostic Instruments, Sterling Heights, MI) and processed (Photoshop; Adobe Systems, Mountain View, CA). 

Cell Count.

Cones were counted in the following retinal areas: central (within 500 μm of the optic nerve) and peripheral (between the central area and retinal edge). Peripheral, ventral, and dorsal areas were counted separately. Four micrographs each, in the central and the peripheral (two dorsal + two ventral) areas, were taken in each animal, at 1000× magnification. The number of cells was averaged and the data analyzed by Student’s t-test. 

Preparation and Injection of Retinals

The two retinal isomers, 9-cis retinal (Sigma-Aldrich) and 11-cis retinal ¹⁰ were prepared for injection under dim red light. The retinal was dissolved in absolute ethanol (10% final concentration) to which vehicle solution (10% BSA in 0.9% NaCl) was added (see Ref. ¹³ for more details). Each litter of Rpe65 ^−/− mice was randomly divided into noninjected control animals and those that received either 9- or 11-cis retinal injections. Animals were injected intraperitoneally three times every third day, starting at P14 at a dose of 0.25 μg/g body weight. Fourteen days after the final injection, improvement of retinal function was confirmed with electroretinogram recordings (data not shown), and eyes were collected for histology. Mice that received retinal injections together with their respective control littermates were moved into the dark after the first injection, until the end of the experiment. Experiments were performed using at least three animals for each group. To exclude possible influences from the solvent of the retinals, a separate control experiment was conducted that included a noninjected control group, a vehicle injection control group, and the 11-cis retinal injection group, using the same injection route and schedule. Cone opsin mRNA levels were evaluated by real-time PCR. 

Results

Downregulation of Cone-Specific Gene Expression in the Rpe65 ^−/− Mouse at Early Ages

Microarray experiments were performed at 2 months of age, a time when the retina is mature, but significant rod photoreceptor degeneration has not yet occurred in the Rpe65 ^−/− mouse. ¹³ The resultant expression levels of the normalized but unfiltered data (wt C57BL/6 against Rpe65 ^−/− mouse retina) are shown in Figure 1A , demonstrating that the two data sets were closely correlated. Rather than finding changes in rod-specific genes, this screen identified that three cone photoreceptor-specific genes were significantly downregulated: SWL (UV-sensitive) cone opsin, MWL (green-sensitive) cone opsin, and GNAT-2 (cone transducin α-subunit; Fig. 1B ). No further cone-specific genes were identified on the array, and for the purpose of this project, no further analysis of the microarray data was performed. 

Real-time RT-PCR analysis was used to confirm the expression levels of the same subset of photoreceptor-specific genes, SWL and MWL cone opsin, rhodopsin, and rod and cone transducin α-subunits (GNAT-1 and GNAT-2, respectively). The mRNA levels of these genes were normalized against 18s rRNA levels. As shown in Figures 2A 2B 2C (representative RT-PCR amplicons) and 2D 2E 2F(average mRNA levels as a percentage of that in wt), the expression levels of the SWL cone opsin (Figs. 2A 2D)in Rpe65 ^−/− retina were significantly decreased at the age of 2 weeks, compared with that in the wt of the same age. By the age of 4 and 8 weeks, SWL cone opsin expression was almost completely eliminated. In the same Rpe65 ^−/− retinas, the reduction of MWL cone opsin (Figs. 2B 2E)progressed more slowly than that of the SWL cone opsin. Rhodopsin mRNA levels (Figs. 2C 2F)remained unchanged in the same Rpe65 ^−/− retinas at all the ages analyzed, consistent with rod degeneration only at late ages in Rpe65 ^−/− mice. ^{1 13}  

Similarly, the cone transducin α-subunit mRNA showed a significant decrease by 4 weeks of age in the Rpe65 ^−/− mice, whereas the rod transducin α-subunit mRNA did not show any significant change in Rpe65 ^−/− mice at all ages analyzed (2–8 weeks of age; Fig. 3 ). 

Distribution of Cones in Rpe65 ^−/− Mouse Retina

To determine whether the reduction in cone-specific genes is caused by cone photoreceptor cell loss, we examined cone density and distribution patterns in Rpe65 ^−/− mice. The Rpe65 ^−/− and control mouse retinas were stained by fluorescent PNA, which has been shown to label both MWL and SWL cones. ^{18 19} At 2 weeks of age, there were no detectable differences in cone density and distribution pattern between the Rpe65 ^−/− and wt mice (Figs. 4A 5) . However, by 3 weeks of age, fewer PNA-positive cones were observed in the central and ventral areas of the Rpe65 ^−/− mice, suggesting that cone inner and outer segment degeneration or cone cell death had occurred in these areas. In the central retina of the Rpe65 ^−/− mice at 3 weeks of age, the density of PNA-positive cones decreased to ∼32% of that in the same area of the Rpe65 ^−/− mouse at 2 weeks of age, and to ∼29% of that in the wt C57BL/6 mouse at 3 weeks of age (Fig. 5) . By 4 weeks of age, large areas in the central and ventral retina became almost “cone free” (Fig. 4B) . The remaining cones were found in the dorsal retina, 571 ± 46 μm from the optic nerve and around the edges of the retina (Fig. 4D , above the dotted line). Cone distribution and density in older mice (9 months of age) remained similar to that at 4 weeks of age (Fig. 4C) , suggesting that cone degeneration in the peripheral retina was significantly slower than in the central area. As shown in Figure 5 , the major cone loss occurred in the central area between 2 and 3 weeks of age. 

The Effect of 9- and 11-cis Retinals on the Cone Degeneration in the Rpe65^−/− Retina

To determine whether the early cone degeneration may be due to the absence of 11-cis retinal and impaired generation of cone pigments, Rpe65 ^−/− mice were treated with 9- or 11-cis retinal, starting at 2 weeks of age with a total of three injections. This regimen has been shown to lead to the regeneration of rhodopsin to ∼30% of that in retina by 5 weeks-of-age (data not shown). Likewise, these three consecutive 11-cis retinal injections also significantly increased cone opsin mRNA when compared with the vehicle control (P < 0.01). The retinas from the treated and control animals were stained with FITC-PNA. As shown in Figure 6 , the PNA-positive cone density was increased by ∼2-fold over the age-matched, noninjected Rpe65 ^−/− control in the central and ventral areas after the 11- and 9-cis retinal treatments (P < 0.05). Both isomers were equally effective in this respect. 

Discussion

The Rpe65 ^−/− mouse is considered a good model for certain forms of retinal dystrophy. Previously, it has been shown that rods degenerate slowly with age in the Rpe65 ^−/− mouse ^{1 13} and that rod function can be partially restored by the administration of 9- or 11-cis retinal. ^{10 11} Cones have not been examined histologically in prior studies, and the expression of cone-specific genes has not been analyzed in this mouse model. The effect of the administration of 11-cis retinal on cone function has likewise not been reported. In this study, we report that in the Rpe65 ^−/− retina, certain cone-specific genes are downregulated, and cone degeneration starts at an early age. We have also shown that cone degeneration can be slowed down by early administration of 9- or 11-cis retinal, suggesting that the absence of chromophore may be responsible for cone degeneration. To our knowledge, the Rpe65 ^−/− mouse retina appears to be a unique model that shows early cone loss, whereas rods remain intact at an early age. Other factors that affect both rods and cones (e.g., light damage, age) appear to impair rod survival before survival in cones. ^{20 21 22} These observations may greatly influence our understanding of photoreceptor degeneration in human photoreceptor dystrophies associated with a loss of RPE65 and thus should be addressed in these patients. 

Degeneration of Cones Versus Rods in the Rpe65 ^−/− Mouse

The Rpe65 ^−/− retina has diminished rod photoresponses and appears to lack cone responses in the ERG. ^{14 15} Histologically, rod degeneration occurs at late ages and progresses slowly. By 7 weeks of age, the Rpe65 ^−/− mouse has the same number of rows of photoreceptor nuclei as the wt mouse, ¹ whereas by 12 months of age, ∼35%, and, by 17 months of age, ∼50% of the nuclei are lost. ¹³ Administration of 11- or 9-cis retinal can partially restore rod function in Rpe65 ^−/− mice over a broad age range (1–18 months-of-age have been tested thus far). ^{10 11} The present study showed that massive cone degeneration occurs between 2 to 3 weeks of age, and that most of the cones have degenerated by 4 weeks of age in the central area of the retina of this mouse. 

Consistent with the different degeneration rates between rods and cones observed by histology, cone opsin and cone transducin expression is reduced at early ages, whereas rhodopsin and rod transducin do not show any significant changes at early ages. The decreased mRNA levels of the cone-specific genes may suggest that cone degeneration is not limited to the outer segments, as the transcription of several cone-specific genes were markedly decreased in cone cells of the knockout mouse. Yet, if 11- or 9-cis retinal was administered before the onset of cone degeneration, a certain percentage of cones could be prevented from degenerating. This result seems to suggest that the absence of 11-cis retinal is responsible, directly or indirectly, for cone degeneration. The fact that the cone loss precedes the rod loss suggests that cones are more susceptible to damage from chromophore deprivation than are rods. This could be explained by in vitro studies showing that cone opsin is less stable than rod opsin in the absence of 11-cis retinal. ²³ However, the chicken cone opsins generating longer-wavelength pigments are in turn less stable than the shorter-wavelength pigments, which is opposite to the results we obtained on the two mouse cone opsins. ²⁴ Alternatively, our group has shown that RPE65 is expressed in cones but not in rods. ² Thus, the absence of this protein in cones may be another contributing factor to early cone degeneration; however, the function of RPE65 in cones is still not understood. 

Cone Degeneration in Rpe65 ^−/− Mouse: SWL Versus MWL Pigment

The mouse retina contain two types of cone pigments, MWL and SWL, which have characteristic regional distributional patterns. ^{25 26} First, unlike other species, the majority of cones in the mouse retina express both cone pigments; however, the ratio of SWL to MWL cone opsins varies in a dorsoventral manner. ²⁶ MWL cone opsin is expressed in every cone of the retina, establishing a dorsoventral gradient, whereas SWL cone opsin is distributed in an opposite gradient, sparing the most dorsal rim of the retina. Second, the mouse retina contains approximately three times more SWL than MWL cone mRNA transcripts. ²⁶ Although it has not yet been shown at the protein level that there is more SWL cone opsin in the mouse retina, electrophysiological experiments using electroretinograms have shown that the absolute sensitivity of the of the UV peak is approximately four times higher than the MWL cone peak. ²⁷  

Our molecular results demonstrated that mRNA transcripts for SWL cone opsin decreased earlier than those for MWL cone opsin in the Rpe65 ^−/− retina. By 2 weeks of age, SWL cone opsin transcript levels dropped to ∼50% of wt levels, whereas those for MWL cone opsin decreased relatively later. In addition, the loss of outer segments appeared to start from the central and ventral retina (which has a higher SWL-to-MWL cone opsin ratio) and progress from there to the dorsal region (which has a lower SWL-to-MWL cone opsin ratio). The differential susceptibility of cone photoreceptors across the retina to outer segment loss suggests that different molecular or biochemical events control cone opsin expression and outer segment degeneration in SWL and MWL cones. ²⁸  

In conclusion, our experiments suggest that if retinas are exposed to exogenous 9- or 11-cis retinal before histologic evidence of cone degeneration, some fraction of cones can be protected from degeneration. These results indicate that the absence of 11-cis chromophore may have a role in the early onset of cone degeneration. 

 Supported in part by National Eye Institute Grants EY12231 and EY015650 (J-xM), EY04939 (RKC), and EY13520 (BR) and a grant from the Foundation Fighting Blindness, the American Diabetes Association, Juvenile Diabetes Research Fund, and the Kirchgessner Foundation. RKC is a Research to Prevent Blindness Senior Scientific Investigator.

 Submitted for publication June 4, 2004; revised December 1, 2004; accepted December 8, 2004.

 Disclosure: S.L. Znoiko, None; B. Rohrer, None; K. Lu, None; H.R. Lohr, None; R.K. Crouch, None; J.-x. Ma, 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.

 

Primers for Real-Time PCR

Gene	Forward Primer (5′–3′)	Reverse Primer (5′–3′)	Product Size (bp)	Location	GenBank No.
SWL opsin	tgtacatggtcaacaatcgga	acaccatctccagaatgcaag	153	Exons 3–4	AF190670
MWL opsin	ctctgctacctccaagtgtgg	aagtatagggtccccagcaga	154	Exons 4–5	AF190672
Rhodopsin	caagaatccactgggagatga	gtgtgtggggacaggagact	136	Exons 4–5	NM_145383
GNAT1	gaggatgctgagaaggatgc	tgaatgttgagcgtggtcat	209	Exons 2–4	NM_008140
GNAT2	gcatcagtgctgaggacaaa	ctaggcactcttcgggtgag	192	Exons 2–4	NM_008141
18s rRNA	tttgttggttttcggaactga	cgtttatggtcggaactacga	199	890–1088	MMRNA18

Figure 1.

 

Gene expression analysis comparing age-matched wt (C57BL/6) and Rpe65 ^−/− mice. (A) Normalized expression levels of the averaged C57BL/6 and Rpe65 ^−/− data plotted against each other revealed a close correlation between the two data sets. (B) Expression data for rod- and cone-specific pigments and transducins are expressed as x-fold-change ± upper and lower boundaries of change (vertical lines). Rho, rhodopsin.

Figure 2.

 

Comparison of mRNA levels of cone and rod opsins between wt and Rpe65 ^−/− mice. Equal amounts of mRNA from the Rpe65 ^−/− and C57BL/6 mice at the ages indicated were used for real-time RT-PCR and normalized by 18s rRNA levels. Representative real-time RT-PCR amplicons for the (A) SWL cone opsin, (B) MWL cone opsin, and (C) rhodopsin, respectively, from wt and Rpe65 ^−/− at 4 weeks of age. Average mRNA levels of (D) SWL cone opsin, (E) MWL cone opsin, and (F) rhodopsin. Each mRNA level in Rpe65 ^−/− mice was expressed as a percentage of the respective value of the age-matched wt mice (mean ± SD, n = 4). *P < 0.05, **P < 0.01.

Figure 3.

 

Comparison of mRNA levels of rod and cone transducin α-subunits in wt and Rpe65 ^−/− mice. Equal amounts of mRNA from the Rpe65 ^−/− and C57BL/6 mice at the ages indicated were used for real-time RT-PCR and normalized to 18s rRNA levels. Average mRNA levels of (A) cone transducin α-subunit (GNAT2) and (B) rod transducin α-subunit (GNAT1). Transducin levels in Rpe65 ^−/− mice are expressed as percentages of respective levels in age-matched wt mice (mean ± SD, n = 4). *P < 0.05, **P < 0.01.

Figure 4.

 

Progression of cone degeneration in the Rpe65 ^−/− retina. Cones were labeled with FITC-PNA in flatmounted retinas. Representative images from the central areas in the retinas of Rpe65 ^−/− mice at the ages of (A, a) 2 and (B, b) 4 weeks, respectively. (C, c) Central retina from Rpe65 ^−/− mice at 9 months of age. Note that blood vessels become visible through the thinning retina in the old Rpe65 ^−/− mice. (D) Cone degeneration pattern in the entire retina: montage of four micrographs from the same retina as in (B). Dorsal area is at the top. (D, dotted line) Demarcation line of cone degeneration. ON, optic nerve. Scale bars, 100 μm.

Figure 5.

 

Cone density in the aging Rpe65 ^−/− retina. Retinas from Rpe65 ^−/− and C57BL/6 control mice of different ages were stained with FITC-PNA, and cone photoreceptors were counted in central (A) and peripheral (B) areas of the retina. Data are expressed as the average number of cones per image field (mean ± SD, n = 3). *P < 0.05, **P < 0.01.

Figure 6.

 

Injection of 9- and 11-cis retinal partially prevented cone loss in Rpe65 ^−/− retina. Rpe65 ^−/− mice received injections of 9- or 11-cis retinal. PNA -positive cones were counted in each region of the retina, averaged in each group, and compared to the respective control. Data are expressed as x-fold increases of cone density over respective noninjected Rpe65 ^−/− control retinas (mean ± SD, n = 3). All the cone numbers presented are significantly higher than the respective noninjected control (P < 0.05).

The authors thank Michael Redmond (National Eye Institute) for kindly providing Rpe65 ^−/− mice and for a critical review of the manuscript.