Quantification of Spatial Vision in the Royal College of Surgeons Rat

Abstract

purpose. To examine how spatial vision deteriorates in the RCS rat over time as a background to experimental studies aimed at limiting photoreceptor degeneration.

methods. The Visual Water Task was used to quantify the grating acuity of pigmented dystrophic RCS rats as they aged and to compare both grating acuity and contrast sensitivity in nondystrophic RCS rats with those parameters in normal pigmented laboratory rats (Long-Evans).

results. Nondystrophic rats had grating acuities and contrast sensitivity functions that were similar to those obtained from Long-Evans rats. The grating acuity of dystrophic rats deteriorated from 80% of normal at 1 month of age to blindness by 11 months. Acuity declined rapidly to 0.32 cyc/deg over the first 4 months, with a slower decline thereafter.

conclusions. Robust measures of vision can be achieved in RCS rats using the Visual Water Task, and with this test, no visual dysfunction can be detected in the background strain. The course of functional deterioration in dystrophic rats is highly predictable, allowing the approach to be used to explore the substrates of the deterioration in vision and to monitor the effects of therapeutic retinal interventions on spatial vision.

Retinitis pigmentosa (RP) and age-related macular degeneration (AMD) together are the leading cause of blindness in North America, with RP affecting 1 in every 3500 ^{1 2 3} and AMD affecting 30% of people over the age of 75. ^{4 5 6} Although RP is most commonly caused by defects intrinsic to photoreceptors, several cases have been identified with recessive mutations of the MERTK gene, ⁷ and these have an orthologue in the Royal College of Surgeons (RCS) rat. In this animal, which is the most studied of the available animal models of retinal dystrophy, the gene defect renders cells of the retinal pigment epithelium (RPE) unable to phagocytose shed photoreceptor outer segments at a normal rate. ⁸ This inability leads to photoreceptor cell death and visual dysfunction. Because AMD may also be attributable in many cases to RPE cell dysfunction, albeit of different etiology and with different associated disease, the RCS rat, with certain qualifications, serves as an indirect model when contemplating nongenetic interventions such as cell transplantation to alleviate the condition. 

Previous reports ^{8 9 10 11 12} of anatomic changes in the retina of pigmented RCS rats have shown that by 3 weeks of age, there is already significant disruption of the outer segments. By 1 month of age, this process is even more evident, and, in addition, the outer nuclear layer is beginning to thin. By 3 months of age, this layer has been reduced to a single layer of cell bodies with the debris zone occupying the former outer segment area, and by 6 months the inner nuclear layer is in direct contact with the RPE cell layer, with very few photoreceptors surviving. Secondary changes in the inner retina, including the loss of retinal ganglion cells from whole pie-slice segments of retina, as well as laminar disruption and changes in the inner plexiform layer, have been shown in animals more than 6 months old. ^{13 14 15}  

By contrast, a smaller number of studies have been undertaken to examine the effects of retinal degeneration on visual function, and most of these have focused on ERG and on automatic responses, such as pupillary light reflex, avoidance behavior, and physiological measures, such as adaptation and threshold responses recorded in the brain. Surprisingly little attention has been devoted to the effects of retinal degeneration on spatial vision. Two previous studies have shown that there is deterioration in spatial vision, but neither examined progress over time. ^{16 17} Furthermore, those researchers were using testing regimens in which the maximum performance was considerably below the visual thresholds of rodents. ¹⁸ In recent work ¹⁹ a new test was developed, the Visual Water Task, in which it is possible to achieve high-resolution performance in rodents with a limited training regimen, so that it is possible from an early age to study changes in spatial vision over time. In the present study, we applied this method to the RCS rat, examining first whether any visual defect can be detected in nondystrophic RCS rats and second how spatial vision deteriorates with time in dystrophic animals. 

The rationale of this work is that as potential treatments for retinal disease are developed, they are inevitably examined first in rodents such as the RCS rat, because larger animal models are few in number and are not always appropriate for the iterative trials that may be necessary in examining validity of a particular procedure. However, at present, the most important end-point assessment, the effect of the treatment on spatial vision, has been poorly studied in rodents or, for that matter, any other animal. As a background to such studies, it was important to establish the background visual capabilities of the RCS rat over time. 

Methods

Animals

Animals were cared for in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Four Long-Evans hooded rats, six nondystrophic RCS rats (rdy ⁻, p ⁺), and six dystrophic RCS rats (rdy ⁺, p ⁺) were used in this study. Long-Evans rats were bred from stock originally obtained from Charles River (St. Constant, Quebec, Canada) and raised in the Canadian Centre for Behavioral Neuroscience (CCBN) vivarium. RCS breeding pairs were obtained from the Lund laboratory at the University of Utah, and their offspring were born and raised at the CCBN. Long-Evans rats and nondystrophic RCS rats of both sexes were tested as adults, 6 to 12 months of age. Dystrophic RCS rats of both sexes were tested monthly beginning at 1 month of age. For the duration of the experiment, all animals were housed in the same room with an ambient temperature of 21°C, 35% relative humidity, and 12-hour light–dark cycle, where food and water were available ad libitum. The housing consisted of transparent Plexiglas cages (35 cm long × 20 cm wide × 13 cm high) hanging on a rack with other cages. 

Visual Water Task

The Visual Water Task has been described in detail previously. ^{19 20 21 22} Apparatus consisted of a trapezoid-shaped tank containing water, with two computer monitors facing through a clear glass wall into the wide end of the pool. Visual stimuli (sine-wave gratings and gray) were generated and projected on the screens using a computer program (Vista; CerebralMechanics; available commercially at http://www.cerebralmechanics.com). On the screens, the black level was 0.05 cd/m² and the white level was 72.8 cd/m², when measured with a light meter (model LS-110; Minolta, Osaka, Japan) positioned at the point where the animal makes a choice. This point is defined by a 46-cm midline divider that extends into the pool from between the monitors, creating a Y-maze with a stem and two arms. A moveable, transparent Plexiglas escape platform (37 cm long × 13 cm wide × 14 cm high) was always submerged directly below whichever monitor displayed the grating. An LRLLRLRR sequence, a pattern the animals could not memorize, was used for the location of the gratings. 

Rats are instinctive swimmers, and the Visual Water Task capitalizes on their natural inclination to escape from water to a solid substrate, the location of which is directly paired with a specific visual stimulus—in this case, a sine wave grating. The other arm of the maze has no platform, and the monitor displays a uniform gray of the same mean luminance. Animals are released into the pool from the wall opposite the monitors, and the end of the divider within the pool sets a choice point for the rats that is as close as they can get to the visual stimuli without entering one of the two arms. The length of the divider, therefore, sets for the animals the effective spatial frequency of the visual stimuli. The animals usually stop at the end of the barrier and inspect both screens before choosing a side. If the animals swim to the platform below the grating without entering the arm with the monitor showing a uniform gray, the trial is considered correct. If they swim into the arm of the maze that contained the gray stimulus, the trial is recorded as an error. 

Animals were first trained to discriminate between a low spatial frequency (∼0.1 cyc/deg), vertical sine wave grating (+ stimulus; 100% contrast), and uniform gray of the same mean luminance (36.2 cd/m² at the choice point). The animals were tested in groups of five or six, with 15 to 20 interleaved trials each, with each session lasting 45 to 60 minutes. No more than two sessions, separated by at least 1 hour, were performed in a single day. All trials were run with the room lights off. Once animals achieved near-perfect performance (90% or better over at least 40 trials), the acuity was then measured. A flexible method-of-limits procedure was used in which incremental changes in the spatial frequency of the sine wave grating were made until choice accuracy fell below 70%. Movies 1, 2, and 3 display an animal performing the task at approximately 0.25, 0.5, and 0.8 cyc/deg, respectively. Accuracy for a given frequency was measured in blocks of 10 trials when near threshold and shorter blocks at the low spatial frequencies, thereby minimizing the number of trials far away from threshold. A preliminary grating threshold was established when animals failed to achieve 70% accuracy at a spatial frequency. To assess the validity of this estimate, the spatial frequency of the grating was reduced by 3 to 4 cycles, and the experimental procedures were repeated until a stable pattern of performance was established. The performance at each spatial frequency was averaged for each animal and a frequency-of-seeing curve was constructed. The point at which the curve intersected 70% accuracy was recorded as the grating acuity. Dystrophic RCS rats were tested to the age at which they could not discriminate a black screen from a white screen (0.015 cyc/deg). 

Contrast sensitivity was assessed using similar procedures, except the minimal contrast required to differentiate between the screens at different spatial frequencies was measured. Contrast thresholds were measured after grating acuity was assessed, so minimal retraining was required. Seven spatial frequencies were tested; 0.059, 0.119, 0.208, 0.297, 0.505, 0.712, and 0.890 cyc/deg. At each spatial frequency, trials were initiated at 100% contrast and the contrast was decreased systematically until performance fell below 70% accuracy. The contrast threshold was measured independently at least three times, after which final values were computed from frequency-of-seeing curves of the combined data. 

Results

Training

All rats readily learned to associate swimming to the platform with escape from the water. Approximately 100 trials were required for animals to reach 90% accuracy over 40 trials, and there were no obvious group differences in this ability. All animals learned serendipitously to grasp the end of the divider and inspect both screens before making their choices. 

Spatial Vision of Congenic RCS and Long-Evans Rats

During the course of testing, both nondystrophic RCS and Long-Evans rat strains performed the Visual Water Task with near-perfect accuracy until the spatial frequency was increased to approximately 0.80 cyc/deg. At around this point, both strains began to make errors in discriminating the stimuli, but still performed above 70% accuracy until approximately 1.0 cyc/deg. The average grating acuity of the strains was slightly higher than 1.0 cyc/deg, (Long-Evans, 1.03 ± 0.004 cyc/deg [SE]; c-RCS, 1.01 ± 0.0009 cyc/deg [SE]) and did not differ statistically (F = 0.286, P = 0.607). The acuities of the Long-Evans animals were comparable to what we and others have reported previously. ^{19 21 23}  

Contrast sensitivity measures of Long-Evans and nondystrophic RCS rats revealed typical inverted U-shaped functions ²⁴ that did not differ significantly, one from the other (F = 0.000, P = 0.986). Both strains had peak sensitivities at 0.208 cyc/deg, and hypothetical intersections of the curves with spatial frequency at approximately 1.0 cyc/deg. Figure 1 illustrates the contrast sensitivity curves for both Long-Evans Rats and nondystrophic RCS rats. 

There was no evidence of strain-related differences in performance in the tests that might reflect different cognitive capabilities. 

Acuity of Dystrophic RCS Rats with Age

The grating acuity of dystrophic RCS rats was tested each month over the course of 11 months. Acuity decreased from 0.82 cyc/deg at 1 month of age to 0.32 cyc/deg by 4 months of age. The acuity deteriorated thereafter at a slower rate until the 10th month of age when the best animals could only distinguish a black screen from a white screen (one cycle displayed across the two screens, 1 cycle/60° = 0.015 cyc/deg). By the 11th month, none of the animals were able to distinguish between a black and a white screen and all were thus considered blind. Figure 2 illustrates the deterioration of visual acuity of dystrophic RCS rats over the course of 11 months. The competence of animals in performing the task did not appear to change as their vision deteriorated. Each animal, regardless of its acuity, swam to the barrier (choice point) before making a choice. In addition, the behavior of dystrophic animals in the task was indistinguishable from that of nondystrophic animals, except that their grating threshold declined progressively with age (Fig. 3) . The stage at which the dystrophic animals were no longer able to discriminate between black and white screens was evident, however, because they no longer swam to the barrier to make their decisions. Instead, each animal swam randomly to one side of the pool and followed the perimeter of the pool wall with their vibrissa until reaching the platform. 

Discussion

The results show that congenic nondystrophic RCS rats have acuity and contrast sensitivity measures that are indistinguishable from Long-Evans strain rats; a strain known to have normal spatial vision. ^{19 21} An acuity of approximately 1.0 cyc/deg has been achieved in normal pigmented rodents in several behavioral studies, using quite different but often more time-consuming testing regimens, ¹⁸ not suitable for the current studies. In addition, physiological recording from cells in the primary visual cortex of normal pigmented rats have given spatial-resolution thresholds of approximately 1.2 cyc/deg, suggesting that this is the true optimal performance that can be expected in rodents. 

The contrast sensitivity curve of the RCS rats was essentially identical with that of Long-Evans. A similar U-shaped function has also been characterized, ²⁴ but the low-frequency fall-off in both strains at 0.059 cyc/deg may have been exaggerated by the small number (2) of cycles on the screen. 

The retinal dystrophy in the RCS rat was first recognized on a pink-eyed tan hooded rat, ²⁵ crossbred to produce a strain with pigmented eyes. ²⁶ This strain was subsequently outbred ²⁶ and sent to the Institute of Ophthalmology (London, UK) where it was maintained as an inbred colony and provided the source of the present animals. Previous studies ^{21 23} have shown that genetic mutations, such as those of albinism, can produce large changes in visual acuity. The normality of the nondystrophic RCS rat demonstrated in this study is important, because it establishes that it can serve as a suitable control in many experiments and that comparative data from Long-Evans rats is relevant. With respect to the present studies, abnormalities in the visual function of dystrophic RCS rats (rdy ⁺, p ⁺) are exclusively the consequence of the RPE abnormality and not of the irrelevant background effects resulting from inbreeding. This study also shows that spatial vision deteriorates in dystrophic RCS rats in two phases: (1) a phase of rapid deterioration already underway by 1 month, when the acuity is 0.82 cyc/deg, and continuing until 4 months of age, by which age the acuity has reached 0.32 cyc/deg; and (2) a phase of slow deterioration of responsiveness that begins at 4 months and gradually increases in slope, leading to eventual total loss of spatial vision by 11 months. This sequence and indeed the levels of performance possible at even 6 months of age might not be predicted from previous anatomic and functional studies. 

Anatomic investigation has shown that although a near complete photoreceptor complement is present at 1 month of age, there is nevertheless considerable disruption of outer segments. Over the next months, there is a decline in photoreceptor density from 300 photoreceptor cell nuclei per midsagittal section at 2 months of age (approximately two nuclei thick) to 100 at 3 months (a single cell layer). ^{8 12} After that, only a discontinuous layer of single cells is seen, although a few photoreceptors, which appear to be cones, are still present at 1 year of age. Because rods account for more than 95% of photoreceptors in the rat, any raw count of photoreceptors without distinction of type is likely to bias toward rods. There has at this point been no systematic study of changes in cone numbers with age. With time, there are changes in the inner retina. ^{13 14 15} Some of the changes result from abnormal vascular formations and from 6 months onward, there is progressive loss of retinal ganglion cells. 

ERG studies ^{27 28} indicate that the a-wave, an indicator of rod function, is lost by day 55, and the remaining b-wave which is largely cone dependent disappears by 80 to 100 days. Adaptation studies (Girman S, et al. IOVS 2003;44:ARVO E-Abstract 482) indicate that rod function is severely compromised as early as 3 weeks of age, although there is indication that a slowly adapting response is still present up to 3 months, but this would be unlikely to play any role in discriminations under the testing conditions used in this study. Physiological studies (Girman S, et al. IOVS 2003;44;ARVO E-Abstract 482) of single-unit responses in the cerebral cortex of RCS rats show that with time, units become less well tuned to specific stimuli and by 7 months, units can no longer be isolated that respond to visual stimulation. Multiunit recording studies from the superior colliculus under mesopic conditions (background luminance of 0.02 cd/m²) show that thresholds increase with age and that by 6 months it is hard to get responses to focal stimulation although responses can still be elicited to full-field stimulation. Certain reflex responses such as the pupillary light reflex can be elicited albeit with higher threshold levels up to at least 12 months of age. ^{29 30} However, it has been shown that such responses may be driven by melanopsin-containing ganglion cells and may not need photoreceptors. ^{31 32}  

Head-tracking to moving stripes is lost in untreated dystrophic rats by 8 weeks, ¹⁷ whereas normal rats can track at better than 0.5 cyc/deg. Previous work ¹⁷ testing acuity in dystrophic RCS rats also showed deterioration in performance with time, but under the testing conditions used, best performance in nondystrophic animals was approximately 0.38 cyc/deg and by 6 months. Dystrophic rats from the same strain as that used in the current study were unable to discriminate stripes of 0.1 cyc/deg. The present test gives optimal performance in nondystrophic rats at levels similar to those found in other experiments, ^{18 33} and, at 6 months, dystrophic rats can still perform at 0.28 cyc/deg. The increased sensitivity of the present method allows better analysis of change with time and titration of the dynamics of such change. 

It appears then that there are several variables that affect measures of performance. First is the question of whether rods, cones, or some other cell type such as the melanopsin-containing ganglion cells are responsible for behavior. The present study was conducted under photopic conditions in which rods are likely to be saturated. Furthermore, adaptation studies (Girman S, et al. IOVS 2003;44ARVO E-Abstract 482) and the age at which the ERG a-wave is lost ^{27 28} suggest that much of the testing in dystrophics is done under conditions in which the rods are likely to be severely dysfunctional or nonfunctional. The luminance levels that drive melanopsin-mediated responses ³¹ are much higher than those used in our study. For these reasons, it is most likely that the test reflects the abilities of cones. The second variable is the sensitivity of the particular test. It is known for example that in patients with retinal disease, the ERG response can be flat, but the patients may still have a considerable degree of vision. ³⁴ In dystrophic RCS rats, the cone-based ERG fails at approximately 14 weeks, ^{27 28} which is close to the end of the first phase of deterioration in acuity. Comparisons between our results and those attained previously ¹⁷ show that even minor changes in methodology and testing conditions can influence performance measures considerably. A third potential variable is the condition of the inner retina, whether changes in cell patterns and synaptic details and, after 6 months of age the loss of retinal ganglion cells, might all contribute to the decline over time. Finally, there is the issue of where in the central nervous system (CNS) the input signals are being processed to effect the behavior. Based on previous observations, ³⁵ it is likely that high-resolution responses are cortically mediated. However, it cannot be excluded that at lower spatial frequencies, noncortical mechanisms may also play a role. 

In summary, these results show that the Visual Water Task provides a highly discriminative method for assessing visual performance in normal and dystrophic rats. It shows the progressive deterioration in vision in dystrophic RCS rats, continuing over a time frame that is more extended than would be expected from previous studies, and it shows two distinct phases in the deterioration curve. This prompts further investigations of the substrates of the changes, focusing particularly on cone-mediated channels, and the limiting factors in performance, including the potential for gain change with time. Most important, this work provides the substrate for further studies ³⁶ (Prusky GT, et al. IOVS 2003;44:ARVO E-Abstract 512) investigating the effects of transplantation and other procedures designed to slow the progress of retinal degeneration on visual performance. The fact that spatial vision as assessed in this study does not correlate easily with indices, such as counts of cells in the outer nuclear layer and full-field ERG recordings, emphasizes the importance of studying it as an endpoint in potential treatments for retinal disease, since the goal of such studies is clearly to improve vision above all. 

 Supported by a National Sciences and Engineering Research Council (NSERC) research grant (GTP) and by funding from Foundation Fighting Blindness (RDL), Johnson and Johnson Focused Giving Program (RDL), and National Eye Institute Grant EY14038 (RDL). RDL is the Calvin and JeNeal Professor of Ophthalmology.

 Submitted for publication September 3, 2003; revised October 17, 2003; accepted December 4, 2003.

 Disclosure: T.J. McGill, None; R.M. Douglas, (P); R.D. Lund, None; G.T. Prusky, (P)

 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.

 

Contrast sensitivity of Long-Evans and nondystrophic control (c) RCS rats measured in the Visual Water Task. Each strain displayed a characteristic inverted U-shaped function that peaked near 0.2 cyc/deg. There were no significant strain differences in the contrast sensitivity functions.

Figure 2.

 

RCS rat grating acuity as a function of age. There was a rapid decrease in the grating acuity until 4 months of age. A slower decline in acuity then occurred until blindness at 11 months. *Point at which the animal could not determine a black computer screen from a white one. A: age at which RCS rats lose the retinal a-wave (Sauve Y, et al. IOVS 2003;44:ARVO E-Abstract 485); B: age when the b-wave is lost (Sauve Y, et al. IOVS 2003;44:ARVO E-Abstract 485); C: age when there is a loss of responses in the superior colliculus to small spots (Girman S, et al. IOVS 2003;44:ARVO E-Abstract 482).

Figure 3.

 

Frequency-of-seeing curves used to calculate the acuity of an individual dystrophic RCS rat at 1, 2, 4, and 8 months of age (arrows: grating threshold). A high level of performance was maintained at spatial frequencies below threshold as the animal aged and its acuity decreased.

Supplementary Materials

Movie 1 - (20.1 MB) Animal performing task at approximately 0.25 cyc/deg 

Movie 2 - (18.8 MB) Animal performing task at approximately 0.5 cyc/deg 

Movie 3 - (16.0 MB) Animal performing task at approximately 0.8 cyc/deg 

The authors thank Tatianna Arjannikova, Yelena Arjannikova, and Shaomei Wang for valuable technical assistance.