Life History of Cones in the Rhodopsin-Mutant P23H-3 Rat: Evidence of Long-term Survival

Abstract

purpose. To follow the status of cones over the life of the P23H-3 transgenic rat, while the rod population is depleted.

methods. P23H-3 heterozygous and Sprague-Dawley (SD) control rats were raised in dim, cyclic light from postnatal day (P)10 to P540. Retinas were examined for cone density, cone outer segment (OS) length, cone axon and soma morphology, and the amplitude of rod and cone components of the electroretinogram (ERG) were determined.

results. In the P23H-3 retina, cone density followed a developmental pattern, increasing from P10 until P20, declining during early adult life (to P150), then steadying at levels found in the SD retina until P540. Cone OSs elongated to P30 and then slowly shortened during late adulthood; at P350 and P540, cone OSs were significantly shorter than in the background SD strain. Cone axons shortened slowly throughout adult life as the outer nuclear layer thinned. The rod a-wave declined steadily in the P23H-3 retina from P10, falling below amplitudes seen in the SD strain from early life. By contrast, the cone b-wave maintained amplitude at SD levels, until P380.

conclusions. Despite the ongoing loss of rod function and numbers, cone numbers in the P23H-3 retina were maintained at levels found in the SD rat to the oldest age examined, and cone function and OS morphology were maintained for approximately 1 year, indicating a long period of cone independence. The long period of cone survival creates an opportunity to induce self-repair, if the stress causing their dysfunction can be reduced.

The loss of cone vision is a progressive, debilitating, and currently untreatable feature of many retinal dystrophies. In many forms of human retinal dystrophy, cones degenerate even though the causal mutation is in a protein expressed specifically in rods. This dependence of cones on rod survival is devastating for the sufferer, and understanding and limiting this dependence has been identified as a priority of the development of treatment. 

The transgene in the P23H-3 transgenic rat was engineered to mimic a mutation of rhodopsin that causes autosomal dominant retinitis pigmentosa (RP) in humans and causes a similar, well-documented degeneration in the rat. In the P23H-3 rat, rods degenerate at a rate dependent on ambient illumination, and late loss of cone function has also been described. ^{1 2 3 4}  

Cones make up approximately 1% of the total photoreceptor population in the albino rat retina. ^{5 6} Electroretinography and behavioral and spectrophotometric studies have identified two classes of cones in rat retina, identified by their photopigment. In one class, the photopigment has a peak absorption at 509 nm (LM-sensitive), and a second has a peak absorption at 359 nm (UV sensitive). ^{7 8 9 10} Immunolabeling of the two rat cone classes has shown that approximately 90% of the cone population contain the LM-sensitive pigment and 10% express the UV-sensitive pigment. ^{6 11 12}  

To understand and test therapeutic interventions aimed at cones in the rat retina, a careful and detailed description of their life history is necessary. To date, the study of cones in the P23H-3 retina has been largely restricted to assessment by the photopic electroretinogram (ERG) and to the first 200 days of life. ^{3 13} Because the pathologic reduction in cone function in human disease may begin relatively late, it is important to understand the status of cones in the aging and aged retina. 

The present study assesses the life history of cones in the heterozygous P23H-3 rat retina, from before eye opening until late mature adulthood (18 months), in animals raised in scotopic ambient light. We paired electrophysiological measures of cone function with a morphologic analysis of cone density, distribution, and outer segment structure. We report that cone function is maintained, despite progressive loss of rod function and numbers, through the first year of life, and that the number of cones is maintained at normal levels into the oldest age examined. The results suggest that cones maintain their integrity for long periods, independent of rods, and that the late loss of function observed is due to outer segment shortening rather than a decrease in the number of cones. Possible mechanisms of this independence and eventual damage are discussed, and it is noted that the persistence in the number of cones allows the possibility of their cone, as reported previously. ¹  

Materials and Methods

Animals

All procedures were in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research, and with the requirements of The Australian National University Animal Experimentation Ethics Committee. Observations were made in P23H transgenic rats (Line 3; Beckman Laboratories, University of California, San Francisco, CA) and Sprague-Dawley (SD) control rats aged from postnatal day (P)10 to P540. The P23H animals used were heterozygotes, the offspring of mating P23H-3 homozygotes with SD albinos, and express a mouse P23H mutant opsin transgene in addition to the two endogenous opsin genes (Steinberg RH, et al. IOVS 1996;37:ARVO Abstract 3190). ^{14 15} Male and female animals of each strain were used equally throughout the study. All animals were born and raised in dim cyclic light (12 hours in 5 lux; 12 hours in dark), and housed in a room with an ambient temperature of 21°C where food and water were available ad libitum. 

Tissue Collection

Animals were euthanatized with an overdose of pentobarbital sodium (>60 mg/kg, intraperitoneal). The eyes were marked at the superior aspect of the limbus for orientation, enucleated, and immersion-fixed in 4% paraformaldehyde in 0.1 M phosphate-buffered saline (PBS) at pH 7.4 for 3 hours. One eye from each animal was processed for cryosectioning, while the fellow eye was processed for wholemounting. For cryosectioning, eyes were rinsed twice in 0.1 M PBS and left in a 15% sucrose solution overnight to provide cryoprotection. The eyes were embedded (Tissue-Tek OCT Compound; Sakura Finetek, Tokyo, Japan) and snap frozen in liquid nitrogen before being cryosectioned at 12 μm (CM1850 Cryostat; Leica, Tokyo, Japan). Sections were mounted on gelatin and poly-l-lysine-coated slides and dried overnight at 50°C before being stored at −20°C. For retinal wholemounts, the retina was dissected from the eye cup, flattened by making radial incisions, gently sandwiched between two glass slides, and immersed in 4% paraformaldehyde at 4°C for up to 2 weeks before immunolabeling. 

Immunohistochemistry of Retinal Wholemounts

To assess the status of cones during the life of the P23H-3 rat, we immunolabeled retinal wholemounts for the long-medium wavelength–sensitive (LM) cone outer segment (OS) opsin, the dominant cone opsin in the rat retina. Flattened retinas were dehydrated and rehydrated in ethanol before being blocked with normal goat serum and incubated overnight with rabbit polyclonal antibodies to LM opsin (1:1000; Chemicon, Temecula, CA). After they were washed with 0.01% Triton-PBS, the retinas were incubated for 24 hours with an antibody to rabbit IgG conjugated to Alexa Fluor 488 (1:1000; Invitrogen-Molecular Probes, Eugene, OR) and subsequently mounted and coverslipped on glass slides. To control for nonspecific labeling, some retinal wholemounts were prepared as negative control specimens (i.e., they were taken through all the steps just described, except that the primary antibody was omitted). 

LM opsin-labeled cone OSs were assessed for both density and length. The samples were visualized by fluorescence microscopy (10× objective, Axioplan 2; Carl Zeiss Meditec, Jena, Germany), and images of the OS layer were captured with a digital camera (AxioCam MRc; Carl Zeiss Meditec). We systematically reconstructed the whole retinal surface by splicing 100 to 200 separate digital images into a montage (with the Panorama feature of Axiovision suite software; Carl Zeiss Meditec). Using the digital montage, LM opsin-labeled OSs were counted over areas of 0.01 mm² at 0.5-mm intervals across the entire retinal surface (∼100–250 fields per retina). At each counting field, the average length of cone OSs was also recorded. Counts of OSs per 0.01 mm² were averaged across the entire retina to give an overall cone density for each sample. To assess regional variations in OS density and length the retinal surface area was divided into superior and inferior regions, into nasal and temporal regions, and into central and peripheral regions (all relative to the optic nerve head). 

Immunohistochemistry of Retinal Sections

Cryosections were labeled with rabbit polyclonal antibodies to LM opsin (1:1000; Chemicon). Sections were washed for 15 minutes in 75% ethanol, followed by a 5-minute wash in distilled water and two 5-minute washes in 0.1 M PBS. The sections were blocked with 10% normal goat serum for 1 hour before being incubated for 24 hours at 4°C with primary antibody. After they were washed in 0.1 M PBS, the sections were treated with an antibody to rabbit IgG conjugated with Alexa Fluor 488 (1:1000; Invitrogen-Molecular Probes) for 24 hours at 4°C before incubation for 2 minutes with the DNA-specific dye bisbenzamide (1:10 000). To control for nonspecific labeling, some sections were prepared as negative control specimens (i.e., they were taken through all the above steps, except that the primary antibody was omitted). 

Retinal Thickness Measurements

To assess retinal thickness, we labeled cryosections with the DNA-specific dye bisbenzamide. The sections were thawed at room temperature before they were washed in 70% ethanol for 15 minutes, followed by a 5-minute wash in distilled H₂O and two 5-minute washes in 0.1 M PBS. The sections were then incubated for 2 minutes with bisbenzamide (1:10,000), washed in 0.1 M PBS, and coverslipped with a glycerol/gelatin medium. Retinal thickness measurements were made on digital images of stained cryosections. At each measurement location, the thickness of the outer nuclear layer (ONL) as well as the thickness of the retina, from inner to outer limiting membrane (ILM–OLM), was recorded. The ratio of ONL to ILM-OLM was used for analysis to account for obliquely cut sections. In at least two sections per animal, we took four measurements, approximately 100 μm apart, from both the superior and inferior mid-peripheral areas of the retina (a total of at least 16 measurements per animal). Results from five animals at each time point were averaged and analyzed by the statistical method described later. 

Electroretinography

The function of photoreceptors was assessed by the flash-evoked ERG. The animals were dark-adapted overnight and prepared for recording in dim red illumination as described previously. ⁴ As in previous reports, ¹⁶ responses to a standard test flash (44.5 cd · s · m⁻²) were considered to be mixed with contributions from rods and cones. Responses to the test flash preceded, by 395 ms, by a conditioning flash (12 cd · s · m⁻²) were considered those of cones. By subtracting the cone response from the mixed response, we isolated the rod response. Using this method, three measurements of amplitude were used for analysis: rod a-wave, rod b-wave, and cone b-wave. The standard flash stimulus was of sufficient intensity to elicit saturated a- and b-wave responses. 

Statistical Analyses

Data were analyzed by using a two-tailed Student’s t-test with P < 0.05 considered to represent a statistically significant difference. All data are presented as the mean ±1 SEM. 

Results

Quantitative Analysis of Cones

The relatively low number of cones in the rat retina meant that individual OSs could be clearly distinguished in LM opsin-labeled wholemounts (Fig. 1) . This form of preparation enabled us to assess the number, distribution, and OS length of LM cones across the entire retinal surface at 11 ages during the life of the P23H-3 rat (Figs. 1A 1B 1C 1D 1E 1F 1G 1H 1I 1J 1K) , from before eye opening (P10) until the late stages of the rodent’s lifespan (P500). 

Cone Density.

Cone density in the P23H-3 retina was calculated by gecounting the number of LM-labeled OSs across the surface of the flatmounted retinas shown in Figure 1 . The density of LM cones increased by 26% between P10 and P16, reaching a peak of 4481/mm² (Fig. 2A , Table 1 ), a time during which the eyes of the rat open and terminal photoreceptor differentiation occurs. Over the next 10 days of retinal development (P20–P30) when the photoreceptor population in the normal retina undergoes a period of developmental culling, the number of P23H-3 cones dropped by 20%. The P23H-3 dystrophy begins during this period when the onset of cell death in the ONL matches that of the nondegenerative retina but is several orders of magnitude higher. ¹³ Between P30 and P115, as the rat reached adulthood, the population of LM cones declined steadily, decreased by 40% before appearing to stabilize. At all ages studied over the next 13 months, from P115 to P500, cone density in the P23H-3 retina remained steady at approximately 2000/mm². 

Cone Outer Segment Length.

The oblique orientation of immunolabeled OSs on flatmounted retinas enabled us to assess their length. To ensure there was no mechanical damage or distortion to OSs during the preparation of wholemounts, we also immunolabeled retinal cryosections for three selected ages and compared the histology. The length of immunolabeled cone OSs on flatmounted retinas and sections were in good agreement (compare Figs. 1O 1P 1Qwith Figs. 1B 1G 1K ), confirming that wholemounts allowed the reliable quantification of OS length. 

At every retinal location where cone density was quantified (100–250 fields per retina), the OSs were also measured. At the earliest age studied (P10) and before eye opening, the entire LM cone population had OSs <8 μm in length (Figs. 2B 3) . At P16, 2 days after eye opening and after final photoreceptor differentiation, 10% of LM cone OSs had elongated to 8 to 15 μm. By P30, 90% of all LM cones had OSs greater than 15 μm in length. From P30 until adulthood (P150), the OSs of the cone population remained stable at an average length of 16 μm. Beyond 150 days of age, cone OSs steadily shortened until, by P500, 40% of all LM cone OSs were <8 μm in length and only 5% remained longer than 15 μm. 

Topography of Cone Density and Outer Segment Length.

Cone density and OS length were not uniform across the P23H-3 retina. There was no significant difference in the density or OS length of LM cones between the superior and inferior, or temporal and nasal, regions of retina (data not shown). However, cone density was significantly (P < 0.05), and on average 20%, higher in the central retina at all ages from P16 to P500 compared with the peripheral retina (Fig. 2A) . As well as being at a higher density, cones in the central retina had significantly (P < 0.05) longer OSs than those in the retinal periphery at every age above P20 (Fig. 2B) . 

Soma and Axon Morphology of Cones

We used the tendency of LM opsin to accumulate in the membrane of stressed cones ^{17 18 19} to demonstrate the morphology of the somas and axons of cones in the P23H-3 retina (Fig. 4 , left), and in the aged SD retina (Fig. 4 , right). Because the somas of LM cones are located at the OLM, the process between the soma and the OS corresponds to the inner segment, while the process between the soma and the OPL is the cell’s axon (diagram in Fig. 4 ). Although not all cones in these retinas were labeled in this full-length way, those that did label may be representative. Cone soma morphology in the P23H-3 retina appeared constant during adulthood, but the length of the axon was steadily reduced after P30. This shortened the cone cell in a striking way, but may be an adaptive change (to the thinning of the ONL), rather than a degenerative change. In the SD retina by contrast, this form of labeling was rare before P500 (Fig. 4right), and the length of the cone cell was maintained. 

Cone and Rod Function

The dark-adapted ERG was recorded in P23H-3 animals across ages 16 to 500 days and compared to those in age-matched SD control rats (Figs. 5 6) . Descriptions are available ^{3 20 21} of ERG responses in the developing and adult P23H rat retina, but a systematic investigation extending to the later stages of the animals’ life spans has not been reported. 

Rod-Mediated Responses.

In the nondegenerative SD retina, rod ERG responses were robust and well developed at P16, only days after eye opening. The amplitudes of the SD rod a-wave and rod b-wave remained virtually constant until 230 days of age, after which both amplitudes began to deteriorate at a similar rate, dropping by 20% to 30% between P230 and P540 (Figs. 6A 6B , Table 2 ). 

In the degenerative P23H-3 retina, like the SD, rod ERG responses were apparent and well-developed at P16. After this point, the amplitude of the rod a- and b-wave steadily deteriorated with age (Figs. 6A 6B , Table 2 ). The a-wave amplitude underwent a more severe decline than the b-wave, decreasing by 45% between P16 and P120, and a further 60% between P180 and P540. This compared to drops of 20% and 50% in the amplitude of the rod b-wave over comparable periods. The overall drop in amplitude over 540 days of life in the P23H-3 rat was 80% for the rod a-wave and 55% for the post-receptoral b-wave. Preservation of the rod b-wave after deterioration of the a-wave has been described previously in the P23H-3 retina, ³ and is thought to be due to remodelling in rod bipolar cell pathways. ²²  

After P16, early in retinal development, differences in rod-mediated ERG responses between the rhodopsin-mutant P23H-3 retina and the nondegenerative SD retina were apparent (Figs. 5 6A 6B) . At 20 and 24 days of age the average amplitudes of the rod a- and rod b-waves were smaller in the P23H-3 than the SD rats and this separation became significant (P < 0.05) at P30, as reported previously. ³ The amplitude of rod responses remained significantly (P < 0.01) lower in the P23H than in the SD at all ages studied. The difference in amplitude between the two strains of rat widened with age, and the effect was more pronounced in the a-wave than for the b-wave. The a-wave amplitude in the P23H-3 retina was 72% of the SD value at P30, 56% at P180, and 28% at P540. Comparable values for the rod b-wave amplitude were 74%, 62%, and 51%. 

Cone-Mediated Responses.

It has been shown that the photopic b-wave in P23H-3 animals is normal until 200 days of age. ³ Present data confirm this finding and extend it, showing that cone-driven ERG responses in P23H-3 animals closely matched those of SD control animals up to 380 days of age (Figs. 5A 5B 5C 6C , Table 2 ). Only at the oldest age studied (P540) could we see a significant (P < 0.01) separation in postreceptoral cone responses between the nondegenerative and degenerative retinas (Figs. 5D 6C) . Between P380 and P540, the b-wave amplitude remained unchanged in the SD retina but dropped by 45% in the degenerative P23H-3 retina. 

Retinal Thickness

To track total (rod and cone) photoreceptor loss over the time course of this study, measures of ONL thickness were made in the P23H-3 retina and compared to age-matched SD controls (Fig. 7) . In both the degenerative and nondegenerative retina there was a 20% thinning of the ONL as the retina matured between P16 and P90, presumably reflecting developmental culling of the photoreceptor population. In the SD retina, no further ONL thinning was detected up to 540 days of age, while in the P23H-3 retina, cell loss in the ONL continued steadily and by 230 days of age the layer had thinned to 68% of the SD value. These findings are consistent with previous studies ^{3 23} and the thinning of the P23H-3 retina can be explained by excess photoreceptor loss during development and a continuous low level of cell death during adulthood. ¹³ We extended the analysis of the ONL to the aged P23H-3 retina; the ONL remained approximately half the thickness of control retinas, but did not reduce measurably in thickness after P230. 

Cones in the Mature Adult Retina

The status of cones in the aged P23H-3 retina, at a time when both rod and cone ERG responses are severely depressed, was compared to that in aged SD retina. There was no significant difference in the average density of LM cones between the P23H-3 retina and the SD retina at P350 or P500 (Fig. 8A , Table 1 ). However, while their numbers were comparable, cones in the aged P23H-3 retina had substantially shorter OSs than those in the SD aged retina (Figs. 1M 1N , Fig. 8B ). 

Discussion

Maintenance of Cone Numbers and Late Failure of Cone Function

From late retinal development to early adulthood (P16–P120), the excess rod loss that characterizes the P23H-3 rat retina begins, peaks, and declines to low, but still abnormal, adult levels. ¹³ Clinically, this period corresponds to the initial development of symptoms of impaired rod function, such as night blindness. ²⁴ The current results show a 23% thinning of the ONL, and 45% and 20% reductions in the amplitude of the rod a- and b-wave during this period, supporting the histologic and functional findings of rod loss previously reported. ³  

The novel data presented herein show a 50% drop in cone density in the P23H-3 retina between 16 and 120 days of age. This reduction in density may not result, however, from the death of cones. The growth in the area of retina continued to as late as P70, ²⁵ while cone generation ended before birth ²⁶ and, further, the amplitude of the cone b-wave increased during this period. This increase may be partly or wholly due to the substantial lengthening (by 80%) of cone OSs throughout early adulthood. The amplitude of the cone b-wave in the P23H-3 retina was maintained at levels found in the SD rat from P16 to P120, again suggesting that cone function is not affected by the decrease in rods and signaling during this period. There is thus no evidence in these data that cone loss in this period was greater in the P23H-3 strain than in the nondegenerative SD strain. 

As the laboratory rat rarely lives beyond 2 years, the period from 120 to 540 days (18 months) of age covers mature and late adulthood. Data are available on photoreceptor status in the P23H-3 strain up to 200 days of age. ^{3 13} Present data show that the loss of rod ERG responses in the P23H-3 retina, both the a-wave and the postsynaptic b-wave, continued to P540. Cone function, by contrast, was maintained at levels seen in the SD until P365, falling below SD levels only at the oldest age examined (P540). This late period of cone failure in the P23H-3 rat retina may correspond to mature adulthood in humans with the same rhodopsin mutation, when the cone visual field constricts and can fail completely. ²⁴  

Previous work has shown that visual performance in rat models of retinal dystrophy cannot always be predicted from ERG amplitudes or ONL quantification (McGill TJ, et al. IOVS 2007;48:ARVO E-Abstract 3448). The present evidence that cones maintain their signaling level and OSs longer than rods and then survive as cells even when their signaling falls and OSs shorten, provides a basis for the persistence of cone-based behavior. We are currently testing behavioral vision as a function of age in the P23H-3 rat using the optomotor head-turning response. 

The steady decline of rod function and the late decline of cone function in the P23H-3 retina cannot be accounted for in terms of cell death. The thickness of the ONL does not decrease measurably after P200 and the density of cones also does not fall measurably during this period. For cones, the loss of b-wave amplitude can be explained, at least in part, by a substantial (20%) shortening of cone OSs. Thus, it appears that both rods and cones survive to these late stages, but are damaged morphologically and functionally. In patients with rhodopsin-mutant RP, cone function remains normal until >75% of rod OSs are lost, and cone dysfunction declines three to four times more slowly than rods. ²⁷ The present finding that cone function in the P23H-3 rat persists undiminished for a year after the loss of rod function can be detected, indicates that this strain is a good model of cone degeneration in comparable human retinal dystrophies. 

The accumulation of LM opsin in the inner segment, soma, processes, and synaptic terminals, seen in a minority of cone cells in the P23H-3 retina at all ages from P16 to P500, and in aged SD retinas, allowed the overall morphology of cones to be followed. This redistribution of opsins has been described in rods in the rd mouse, ²⁸ rds mouse, ^{29 30} P23H rat, ³¹ RCS rat, ³² an experimental model of retinal detachment, ³³ a light-induced damage model, ³⁴ and the developing rat retina, ³⁵ and in cones in human retinas with enhanced S-cone syndrome, ¹⁸ X-linked cone degeneration, ¹⁹ and an autosomal dominant cone dystrophy. ¹⁷  

It has been postulated that redistribution of opsins results from either misplacement of newly synthesized opsin molecules or back diffusion of old opsin molecules from damaged OSs. ^{28 32} In primates, cone opsins are present throughout the entire cell body when they are first expressed in the fetal period and become restricted to the OS later in development. ³⁶ In the present context, these observations show that cones adapt to the thinning of the ONL (as rods die) by reducing the length of their processes. The shortening of OSs can be considered a degenerative change, because it is associated with a reduction in ERG amplitude, but the overall shortening of cone cell length may be adaptive. 

The Limits to Cone Survival

The survival of cones in retinas degenerating because of rod-specific mutations may depend on the rate and extent of rod dysfunction. The P23H-3 heterozygous rat used in the present study displays a relatively slow rate of rod degeneration, and cone function is not affected until P540. In the P23H-1 and homozygous P23H-3 strains, rod degeneration is more rapid, and cone dysfunction in these animals is evident much earlier, at P56 ³ and P21, ^{20 21} respectively. In the rd/rd mouse, which carries a mutation in the gene for cGMP phosphodiesterase β-subunit, rod degeneration is particularly rapid. Cone degeneration is marked (25%) by 7 weeks of age, when 99.7% of rods have degenerated. ³⁷ We are currently testing whether cone dysfunction can be induced by increasing rod degeneration rates in the P23H-3 heterozygote. The present results suggest that the cone ERG can fall 50% in amplitude, without a measurable loss of cone cells. If this is true in humans with comparable dystrophies, the repair of cones, induced by reducing the damage due to stress, may restore cone function. 

The Mechanism of Late Cone Damage

The question arises as to what causes cone dysfunction in the P23H-3 retina, given that cone numbers remain constant. As just noted, the shortening of cone OSs is an obvious cause of the functional decline, but the trigger for this shortening remains unclear. It has been suggested that dying rods release substances that are damaging to cones, ³⁸ that alternatively cones are dependent on a rod-derived trophic factor ^{39 40 41} and that cones are injured by oxidative damage due to decreased oxygen consumption by rods. ^{42 43} In aged human RP retinas with various rhodopsin mutations, including the P23H, shortening of cone OSs is accompanied by downregulated expression of specialized cytoplasmic proteins such as calbindin, cone arrestin, and cytochrome oxidase. ^{44 45} Such protein expression changes may contribute to cone dysfunction in the late stages of RP. Further, it has been suggested that a general decrease in renewal of structural proteins may directly result in cone OS shortening. ⁴⁵ Upregulation of stress-induced factors known to suppress the ERG, such as FGF-2 ^{46 47} and CNTF, ^{48 49} are also likely to contribute to reductions in rod and cone responses in the aged P23H-3 retina. 

Clinical Implications

The present results provide a baseline for assessing the impact of rod loss on cone integrity. Separately, we report on the long-term impact of different levels of ambient illumination on the stability and performance of the adult P23H-3 retina ⁵⁰ and are currently testing the damaging effects of rapid rod loss on cones in the P23H-3 retina. In the current study we noted that, when ambient conditions were held steady and low, cones were stable in function for a long period, during which rods failed. Two avenues for optimizing the status of the adult retina can be identified. First, adult status is improved (more photoreceptors surviving and functional) if ambient conditions are optimized throughout life. Second, because many cones survive but suffer OS damage and shortening, the opportunity to create conditions for their self-repair ¹ remains. 

 Supported by Retina Australia, Grant 268060 from the National Health and Medical Research Council of Australia, and Grant CE561903 from the Australian Research Council.

 Submitted for publication October 14, 2008; revised November 23 and 26, 2008; accepted March 12, 2009.

 Disclosure: V. Chrysostomou, None; J. Stone, None; K. Valter, 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.

 

Immunohistochemical labeling of LM cone photoreceptors in P23H-3 and SD retinas. In wholemount preparations (A–N) individual cone OSs (green) could clearly be distinguished, allowing us to quantify both cone density and OS length. (A–K) Representative images from the superior central P23H-3 retina at successive ages spanning the lifetime of the rodent; from before eye opening at P10 to late adulthood at P500. (L) Anterior edge of the young adult P23H-3 retina. (M, N) Age-matched SD controls for the two oldest ages studied (P350 and P500). (O–Q) LM opsin-labeled retinal sections, used to confirm the OS length seen in wholemount preparations. Scale bar, 20 μm.

Figure 2.

 

Cone density and OS length in the P23H-3 retina across ages 10 to 500 days. The average density of LM cones (A) and their OS length (B) was determined by assessing immunolabeled OSs across the surface of retinal wholemounts. The retinal surface was divided equally in two, into central and peripheral areas, to assess regional variations. For each age studied, retinas from three different animals were assessed. Data are presented as the mean ± SEM. *Data points that are significantly different (P < 0.05) by a Student’s t-test.

Table 1.

 

Cone Density across Ages in the P23H-3 Rat

Age (d)	LM Cones/mm2
10	3322 ± 3.50
16	4481 ± 103
20	4409 ± 156
25	4003 ± 68.5
30	3571 ± 185
70	3117 ± 15.1
115	2196 ± 69.3
150	2273 ± 31.7
260	1989 ± 16.7
350	2111 ± 127
500	1992 ± 75.5
SD P350	2178 ± 81.8
SD P500	2163 ± 102

  n = 3 at each age.

 Data are expressed as the mean ± SEM.

Figure 3.

 

Variations in the length of LM cone OSs in the P23H-3 retina across age. The histograms show the proportion of the cone population at each age with OSs that are <8 μm, 8–15 μm, and >15 μm in length. Data are presented as the mean ± SEM of results in three animals.

Figure 4.

 

Immunofluorescence labeling of LM opsin in the P23H-3 retina (left) from the juvenile (P16) to late adulthood (P500), and in the SD retina (right) during adulthood (P150–P500). The images have been arranged so that the OPL is aligned between images; the length of the image then approximates the overall length of the cone cells. LM opsin is strongly expressed in OSs in both strains and at all ages. In some P23H-3 cones at all ages and some SD cones at P500, LM opsin was detected throughout the cell membrane, allowing visualization of the process between the soma and the OS (the inner segment, IS) and of the process between the soma and the OPL (the axon). Left: our interpretation of the labeling observed, for one of the cells labeled in the P23H-3 P16 section and for one of the cells labeled in the P23H-3 P500 section. In the P23H-3 retina, cones shortened strikingly with age, adapting to the thinning of the ONL as rods died. No comparable thinning of the ONL was apparent in age-matched SD retinas. Scale bar, 10 μm.

Figure 5.

 

Sample waveforms of rod and cone ERG responses recorded from P23H-3 transgenic and age-matched SD control rats. The unconditioned responses (left) were recorded to a single flash stimulus of 44.5 cd · s · m⁻² and include contributions from both rods and cones. The conditioned responses (right) were recorded to a flash (44.5 cd · s · m⁻²) given 395 ms after a rod-saturating conditioning flash, so that the responses represent cone activity.

Figure 6.

 

Comparison of rod and cone ERG responses between P23H-3 and SD retinas across ages 16 to 540 days. In response to a stimulus flash of 44.5 cd · s · m⁻², three measures of amplitude were taken: (A) rod a-wave, (B) rod b-wave, and (C) cone b-wave. Rod-mediated responses were significantly lower in the P23H-3 retina than the SD retina at all ages above P30, while there was no separation of cone function between the two strains up to P380. Data are presented as the mean ± SEM of 5 to 12 animals (see Table 2 ). *Data points that are significantly different from age-matched SD control subjects (P < 0.05) by a Student’s t-test.

Table 2.

 

Comparison of ERG Responses across Age between the SD and P23H-3 Rat

Age (d)	n			Rod a-Wave (μV)*			Rod b-Wave (μV)*			Cone b-Wave (μV)*
		SD	P23H-3	SD	P23H-3	SD	P23H-3	SD	P23H-3
16	4	7	492 ± 82	565 ± 61	611 ± 73	669 ± 67	228 ± 40	234 ± 45
20	5	6	593 ± 65	507 ± 51	798 ± 113	659 ± 91	243 ± 22	265 ± 28
25	5	5	569 ± 91	488 ± 92	921 ± 174	706 ± 142	287 ± 56	333 ± 77
30	5	6	630 ± 94	452 ± 56	1010 ± 126	747 ± 120	323 ± 53	277 ± 47
60	9	6	537 ± 48	375 ± 48	775 ± 135	510 ± 58	391 ± 71	305 ± 56
90	5	6	614 ± 53	450 ± 42	869 ± 87	620 ± 61	421 ± 21	404 ± 28
120	8	10	498 ± 46	310 ± 17	721 ± 75	578 ± 53	325 ± 39	366 ± 23
155	11	5	490 ± 28	315 ± 15	733 ± 50	542 ± 42	309 ± 24	346 ± 21
180	5	6	518 ± 12	292 ± 13	845 ± 50	588 ± 28	303 ± 18	279 ± 12
230	5	5	540 ± 30	225 ± 20	737 ± 36	460 ± 40	311 ± 14	282 ± 14
380	5	12	342 ± 45	166 ± 10	461 ± 60	305 ± 29	210 ± 42	232 ± 16
540	5	6	392 ± 33	111 ± 13	599 ± 65	303 ± 22	254 ± 28	132 ± 15

 Tabulated values give mean ± SEM.

*  Amplitude of response to a flash stimulus of 44.5 cds/m².

Figure 7.

 

Measurements of ONL thickness in the P23H-3 and SD rat retina across ages 16 to 540 days. Data are presented as the mean ± SEM of results in five animals. *Data points that are significantly different from age-matched SD controls (P < 0.05) by a Student’s t-test.

Figure 8.

 

Comparison of cone density and OS length between the aged P23H-3 and aged SD retina. The mean density of LM cones (A) and the length of their OSs (B) in P23H-3 and SD retinas at 350 and 500 days of age was quantified using immunolabeled wholemounts. Data are presented as the mean ± SEM of results in three animals.