The Electroretinogram Components in Abyssinian Cats with Hereditary Retinal Degeneration

Abstract

purpose. To examine phototransduction using the a-wave and other aspects of retinal function with the intraretinal b- and c-waves at different stages of an inherited photoreceptor degeneration in Abyssinian cats.

methods. Vitreal and intraretinal ERGs were recorded from eight dark-adapted, anesthetized Abyssinian cats. Brief bright flashes were used to elicit vitreal a- and b-waves. Longer, weaker flashes were used to elicit intraretinal b- and c-waves. Stages 1 through 4 of the disease were characterized ophthalmoscopically. Parameters of the Lamb and Pugh a-wave model (a _max, A, and t _eff) for the Abyssinian cats were compared with those for normal cats. Light microscopy was used to count photoreceptor nuclei.

results. The maximum a-wave amplitude, a _max, was significantly smaller in stage 1, and continued to decrease (stage 1: 50% of normal, stage 2: 28%, stage 3: 27%; and stage 4: unrecordable). There was a small, but not significant, decrease in the amplification constant A from 0.24 ± 0.11 s⁻² in normal cats to 0.16 ± 0.08 s⁻² in Abyssinian cats. The intraretinal b- and c-wave amplitudes decreased most dramatically during the early stage of the disease. Affected animals had fewer photoreceptors than unaffected Abyssinians or control animals. The number of photoreceptors declined most rapidly in the inferior periphery.

conclusions. The amplitudes of all ERG components were already reduced significantly by stage 1 and progressively declined. The lack of major changes in a-wave model parameters indicates that the degeneration is probably not due to a mutation in transduction proteins. Losses of photoreceptor function were larger than losses of photoreceptor nuclei.

Astrain of Abyssinian cats has a slowly progressing retinal degeneration inherited by an autosomal recessive gene. ^{1 2} Histologically, both rods and cones degenerate; however, the rods degenerate before the cones do, and changes are observed first in the periphery. ³ Narfström et al. ⁴ demonstrated that affected cats show characteristics similar to those in a human form of retinitis pigmentosa (RP). The disease is divided into four stages of severity based on ophthalmoscopy and the electroretinogram (ERG). ^{2 5} Reductions in the scotopic a- and b-wave amplitudes were observed even in an early stage of the disease. A previous ERG study observed no significant changes in the c-wave amplitudes compared with those of normal animals in stages 1 and 2 of the disease. ⁵  

The genetic defect in these animals is not known. To further characterize the rod photoreceptor degeneration, we investigated ERG a-waves in response to bright flashes and used the Lamb and Pugh ⁶ model of phototransduction to examine the data. The fits of the Lamb and Pugh model allow an extraction of two a-wave model parameters: the overall amplification constant of transduction (A) and an effective delay before the response begins (t _eff). The parameters were obtained at various stages of the disease. We also examined vitreal and intraretinal recordings of the b- and c-waves in response to weaker flashes. The intraretinal responses were from central retina, and allowed a comparison with those generated by the entire retina. A preliminary report of these results has been presented in abstract form (Kang Derwent JJ, et al. IOVS 2000;41:ARVO Abstract 1268). 

Methods

Animals

All experimental procedures were in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Details of the animal preparation have been described elsewhere. ⁷ Briefly, adult cats were anesthetized initially with either pentothal (5%, 0.35 mL/kg intravenously) or ketamine-acepromazine (25 mg/kg and 0.12 mg/kg intramuscularly) if the cat was difficult to handle. Atropine (0.3 mg subcutaneously) was given after the induction. Urethane (200–400 mg/kg loading dose followed by 20–40 mg · kg⁻¹ · h⁻¹) was used throughout the experiment as the long-term anesthetic. The right eye was mounted on a stainless-steel eye ring by attaching the conjunctiva, and surgery was performed to allow the introduction into the vitreous of the electrodes and microelectrode. ⁸ Topical atropine (1%) and phenylephrine (1%) were administered to dilate the pupil. Topical flurbiprofen sodium ophthalmic solution (0.03%; a prostaglandin inhibitor) was given every 5 to 6 hours to prevent pupil constriction. The cat was paralyzed with pancuronium bromide (0.2 mg · kg⁻¹ · h⁻¹) and was artificially ventilated. The temperature was kept at 38 °C to 39°C by using a feedback-controlled heating pad. The electrocardiogram, arterial blood pressure, and blood gases were monitored throughout the experiment. 

Three cats were in the earliest stages of the disease (stage 1), two cats were in the middle stage (stage 2), two cats were in the late stage (stage 3), and one cat was in the final stage (stage 4). The stage 3 animals were 3 and 5 years old, and the stage 4 animal was 6 years old. Briefly, stage 1 is characterized by a subtle tapetal fundus discoloration. In stage 2, the discoloration of the tapetum is more distinct and widespread and peripheral vessels on the tapetal retina appear slightly attenuated. In stages 3 and 4, vessel attenuation is prominent, the discoloration becomes darker and more mottled, and large areas of the tapetal fundus are hyperreflective. Electroretinograms from these animals were compared to those obtained previously from normal, non-Abyssinian cats. Separate normal data sets were used for the bright flash a-wave analyses ⁷ and for the intraretinal b- and c-wave analyses ⁹ because the two types of recording were not available from a single set of normal animals. 

Visual Stimulation

Animals were allowed to dark adapt for at least 1 hour before the ERGs were collected. Two types of visual stimulation were used to record all the signals of interest (Table 1) . First, bright flashes of diffuse white light, produced by a photoflash (photoflash model 283; Vivitar Corp., Santa Monica, CA) were used to elicit the a- and b-waves recorded with a vitreal electrode. Stimuli were reflected to the eye by a Ganzfeld hemisphere. Flash intensity was varied over 6 log units with neutral-density filters. At zero attenuation, the illumination at the cornea was 4.45 log scotopic cd-sec · m⁻² measured with a photometer (model 40X; UDT, Hawthorne, CA) in photopic units and converted to scotopic units by assuming the spectral distribution of a xenon-filled flashtube). ¹⁰ The conversion factor, K, from the corneal illumination to the number of photoisomerizations per rod per flash (Φ) followed Breton et al. (Ref. ¹¹ , equation 13) with a few corrections for the cat eye. ⁷ The value of K was 27.13 photoisomerizations · rod⁻¹ · scot td⁻¹ · s⁻¹. The unattenuated retinal illumination was 7.93 log photoisomerizations per rod per flash. 

A second type of stimulus was used to obtain intraretinal recordings of the b- and c-waves during penetrations with an oxygen microelectrode. ¹² The vitreal ERG was recorded simultaneously, so there were two sets of vitreal b-wave recordings, but no duplication of the recording of other signals. These recordings were performed after the bright-flash recordings and used diffuse white light from a halogen bulb, delivered to the eye via a fiber-optic light guide. Intraretinal recordings were made in response to flashes that were 4 seconds in duration at an illumination that was at or near rod saturation (∼8.6 log quanta, 555 nm · deg⁻² · s⁻¹), which generally produced maximum b- and c-waves (e.g., Ref. ¹⁴ ). This stimulator was also used at lower illuminations, to generate intensity–response functions for the vitreal ERG. 

Recording

The vitreal ERG was measured between a chlorided silver wire in the vitreous and a chlorided silver plate behind the eye. Vitreal recordings are functionally equivalent to those from the cornea, but offer more stability in long-term recordings and have slightly larger amplitude. Vitreal responses were amplified (DC to 3 kHz; model M4A; WPI, Sarasota, FL), digitized at 5 kHz, and stored in an IBM-compatible computer. The bright flashes were triggered by the computer, and ERGs were collected with EPIC-XL software (LKC, Gaithersburg, MD). After each flash, the eye was allowed to re–dark adapt for at least 3 minutes. Responses to individual flashes were analyzed. 

Intraretinal ERG recordings were made by the voltage barrel of a double-barreled oxygen microelectrode ⁸ referenced to the chlorided silver plate. The microelectrode was advanced through the vitreous in the dark until a transient voltage deflection signaled that the retina had been encountered. From this point on, the electrode was advanced in 3-μm steps, stopping every 30 μm to record the local ERG. The penetration ended when the microelectrode crossed the retinal pigment epithelium (RPE), as indicated by the transepithelial potential. After each flash, the eye was allowed to re–dark adapt for at least 1 minute. 

a-Wave Analysis

The Lamb and Pugh ⁶ a-wave model was fitted to the responses to all flash energies  

\[F(t)\ {=}\ \mathrm{exp}{[}{-}\ \frac{1}{2}{\Phi}A(t{-}t_{\mathrm{eff}})^{2}{]}\ \mathrm{for}\ t{>}t_{\mathrm{eff}},\]

where F(t) is the cGMP-activated current expressed as a fraction of its dark value, Φ is the number of photoisomerizations per rod, t _eff is a brief delay, and A is an amplification constant. When applied to the a-waves, F(t) is equivalent to the normalized a-wave from flash onset to the peak of the a-wave  

\[a(t)\ {=}\ {[}1\ {-}\ F(t){]}a_{\mathrm{max}},\]

where a(t) is the a-wave response and a _max is the maximum a-wave amplitude. Intraretinal recordings in cat showed that the portion of the response that was fitted is purely from photoreceptors and has no detectable contribution from the inner retina. ¹³ Fitting was performed on computer (Photran software; LKC). Equation 2was used to obtain both ensemble and individual fits to the data. Ensemble fitting was used to fit the model simultaneously to a family of responses to different intensities, to obtain values of A and t _eff that produced the best fit. The ERGs in response to the brightest flashes (5.93 log photoisomerizations per rod or higher) could not be included in the ensemble, because at high intensity the value of A declines substantially. ⁷ Fits of individual responses were used to obtain A as a function of Φ. 

Hill Equation Analysis

The a-, b-, and c-wave intensity series were fitted to a Hill equation  

\[R(I)\ {=}\ \frac{R_{\mathrm{max}}I^{n}}{I^{n}\ {+}\ {\varsigma}^{n}},\]

where R(I) is the a-, b-, or c-wave amplitude, I is intensity, R _max is the maximum amplitude obtained by the fit, ς is half saturation, and n is a parameter that adjusts the steepness of the function. The fitting was performed using SigmaPlot (ver. 9.0; SYSTAT Software, Inc., Point Richmond, CA). 

Histologic Preparation

After measurements, both eyes were enucleated, and retinas were fixed in 2% glutaraldehyde in 0.1 M cacodylate buffer and embedded in paraffin. They were then sectioned and stained with toluidine blue for light microscopy. Measurements of outer retinal thickness and photoreceptor nuclei per section were made in three sections per retinal location at 400× and 1000×, respectively. The counts reported were made at the higher power. Measurements were made in the area centralis and far peripheral retina, in both superior and inferior regions. The same techniques were used to evaluate one eye of each of 12 additional animals. Three were stage 1 Abyssinians not used in the electrophysiological studies, to control for possible changes in counts resulting from the long duration of anesthesia and swelling of the retina. Three were from heterozygous unaffected Abyssinian cats, and six were from adult domestic short-haired cats. These latter two groups were used to determine whether the stage 1 animals had lost photoreceptors relative to the heterozygous or normal cats. 

Results

Bright-Flash ERG Analysis

Figure 1shows examples of intensity series in the dark-adapted retina for individual normal, stage 1, and stage 3 cats in Figure 1A 1C and 1E , and the Lamb and Pugh ⁶ a-wave model fits for these animals in Figures 1B 1D and 1F . The a-wave amplitude was measured from the baseline to the a-wave trough, and the b-wave amplitude was measured from the a-wave trough to the b-wave peak. The a- and b-wave amplitudes of Abyssinian cats were reduced in comparison to those of normal cats, even in stage 1. ERGs were unrecordable from the stage 4 animal. The oscillatory potentials (OPs) were evident on the rising phase of the b-waves in the normal and stage 1 cats (Figs. 1A 1C) , but were reduced or absent in the stage 3 cat. 

Equation 2was applied to the data shown in Figures 1A 1C and 1Eto generate the fits shown in Figures 1B 1D and 1F . Before fitting, the family of a-wave responses was normalized to the maximum a-wave amplitude, a _max. The fitted a-wave model responses are shown as dashed lines. Parameter values are given in the figure legend. 

The parameter A, which was derived from fits similar to those in Figure 1and reflects the gain of phototransduction, is plotted as a function of the number of photoisomerizations, Φ_, in Figure 2for normal cats (Fig. 2A , data reproduced from Ref. ⁷ ), stage 1 cats (Fig. 2B) , stage 2 cats (Fig. 2C) , and one of the stage 3 cats (Fig. 2D) . Both A and t _eff were varied to obtain the best fits. In both normal and affected Abyssinian cats, the value of A was approximately independent of Φ, except at the higher intensities, when A declined. This effect has been demonstrated in humans, ^{6 11} rats, ¹⁵ and normal cats. ⁷  

Figure 3shows two of the important parameters characterizing the a-wave: the average a _max, produced by a saturating flash of 6.93 log photoisomerizations per rod (Fig. 3A)and the amplification constant, A (Fig. 3B) . The value of a _max decreased significantly and progressively with disease stage, to 50% in stage 1 (P = 0.012 compared to normal by Wilcoxon test), 28% in stage 2 (P = 0.044 compared to stage 1 and to normal), 27% in stage 3, and 0% in stage 4 compared to the normal group. Figure 3Bshows the amplification constant A at 5.93 log photoisomerizations per rod per flash in individual cats, plotted as a function of the stage of the disease. Responses from one stage 3 animal (cat 302) were too noisy to be fitted, and so they are not represented in Figure 3B . For each stage, the average value of A is plotted as an open circle and these are connected by a dashed line. A was 0.24 ± 0.11 s⁻² in normal cats and 0.16 ± 0.08 s⁻² in six Abyssinian cats, but this was not a significant decrease (t-test, P = 0.09), and there was no consistent change with disease stage, in contrast to the decrease in a _max. The parameter t _eff, which reflects the effective delay of phototransduction, was compared at 5.93 log photoisomerizations per rod per flash. The average t _eff was 2.11 ± 0.21 ms (n = 8) in normal cats, 1.77 ± 0.22 ms (n = 3) in stage 1 cats, 1.50 ± 0.49 ms (n = 2) in stage 2 cats, and 0.93 ms (n = 1) in the stage 3 cat. This decrease in t _eff was not significant by the nonparametric Wilcoxon test. 

Figure 3Cshows the b-wave measured from bright-flash responses as a function of disease stage. As shown in Figure 1 , the bright flashes were generally at or above b-wave saturation, even when attenuated. The decrease in the vitreal b-wave was pronounced, but somewhat smaller than the decline in a _max. 

Vitreal Intensity–Response Functions

Figure 4shows the amplitudes of the a-wave (Fig. 4A) , b-wave (Fig. 4B) , and c-wave (Fig. 4C)amplitudes as a function of light intensity. The a-wave amplitudes were from data sets like those shown in Figure 1 , in response to brief bright flashes, and the b- and c-wave amplitudes were obtained using 4-second flashes of lower illumination. The 4-second flashes were used for the b-wave, because the entire stimulus–response relation was not obtained in response to the bright flashes. It can be appreciated from Figure 1that a flash producing a very small a-wave still produces a saturated b-wave. For the c-wave, its long integration time makes the longer stimulus more appropriate. 

The a-wave amplitudes were dramatically decreased at all flash intensities for the Abyssinian cats compared with the normal cats. The b-wave amplitudes were also affected across the range of flash intensities for all stages. The c-wave amplitude-response functions were also affected by the disease progression; however, variability in both normal cats and affected Abyssinian cats was large. The Hill equation (equation 3)was applied to the averaged intensity–response series for each stage, and the fitted parameters are listed in Table 2 . Only small changes were observed in the sensitivity parameter (ς) as a function of the disease stage, and these were not consistent across the different ERG components. For all ERG components, it was primarily the maximum amplitude parameter, R _max, that was affected by the disease. 

Intraretinal ERGs

Intraretinal ERGs have never been recorded in affected Abyssinian cats or any other animal model of RP. These provide useful information, because they are from the central retina, whereas the vitreal responses average activity across the retina in a way that is not precisely known. Because weaker flashes were used for these responses, the a-wave was not elicited, but the illumination was sufficient to saturate the b- and c-waves. Figure 5Ashows the intraretinal b-wave amplitude, measured just before the microelectrode encountered the RPE. This location was chosen for measuring the b-wave because the maximum b-wave amplitude, which occurs more proximally in the retina, was often masked by a large negative-going potential that is believed to be an intracellular recording from a damaged horizontal cell. Such contamination of the b-wave is not observed more deeply in the retina. The b-wave just outside the RPE is somewhat smaller than the b-wave at its maximum, ¹⁶ but recording at this location allowed consistency. The inset in Figure 5Ashows waveforms of intraretinal ERGs from stage 1 and -3 cats. For comparison, Figure 5Bshows the vitreal b-wave recorded simultaneously. 

The intraretinal b-wave declined with disease stage, following a similar function as the vitreal b-wave (Fig. 3C , bright-flash responses; Fig. 5B , weaker-flash responses). The function was not identical, however, because the intraretinal response appeared to be better preserved in stage 3 than the vitreal response was. This difference is probably because the funduscopically visible disease progresses from the periphery toward the central retina, but the change in the central responses in stage 1 indicate that functionally the central retina is affected early in the disease. 

Figure 5Cshows the intraretinal c-wave amplitude, recorded in the subretinal space near the RPE, and Figure 5Dshows the corresponding vitreal c-wave amplitudes. Narfström et al. ⁵ reported that the vitreal c-wave was better preserved than the a- or b-wave during the early stages of the disease. This is consistent with the present work; the difference between the normal and stage 1 animals was not significant (P = 0.47) for the vitreal responses. However, the difference in the intraretinal c-wave between normal and stage 1 animals was significant, as would be expected from the dependence of the c-wave on photoreceptor responses. The reason that the amplitude of the vitreal c-wave can give a misleading impression about the underlying intraretinal events is that the vitreal c-wave is the difference between an RPE c-wave component and a Müller cell component, slow PIII. ^{17 18} If these decline together, as would be expected from the dependence of both events on photoreceptors, ^{19 20} the vitreal responses can be relatively unaffected. ¹⁸  

Photoreceptor Counts and Histology

Figure 6shows examples of the histology from Abyssinian cats, on which measurements of photoreceptors and retinal thickness were based. Precise correlation of the central histologic sections and the intraretinal recording location was not possible, but both were from the area centralis. Counts from the far peripheral retina were obtained because the vitreal ERG is summed across the whole retina, making counts from both central and peripheral regions relevant. Figure 7shows photoreceptor counts as a function of the disease stage for the three investigated areas. This figure also shows data for three additional groups of animals that were not used for electrophysiological measurements: six normal domestic short-haired cats, three heterozygous Abyssinians in which the fundus appears normal, and three additional early stage 1 animals, to control for the possibility that the histology from stage 1 animals used for electrophysiology was distorted by the extended period of anesthesia or poor fixation of the tissue. 

Photoreceptor numbers were lower in stage 1 Abyssinians than in heterozygous Abyssinians or normal cats, but then did not decline further until stage 3. In stage 3, the central and superior retina still had 40% of the normal number of photoreceptors, but the inferior periphery had <10% of the normal number. The difference between stages 3 and 4 for central and superior retina was dramatic, with complete photoreceptor loss by stage 4. Outer retinal thickness is shown in the right column of Figure 7 . Percentage changes in thickness were smaller than changes in cell number, which may occur because there is disorganization and loss of tight packing accompanying degeneration. 

Discussion

In the eight Abyssinian cats studied herein, ERG changes were similar to the results previously reported. ^{4 5 21} The a- and b-wave amplitudes in Abyssinian cats with photoreceptor degeneration were reduced, with prominent changes beginning in stage 1. In earlier studies, the amplitude of the corneal b-wave in stage 1 was 60% of the amplitude in unaffected Abyssinian cats, ⁵ and 57% of the amplitude in normal (non-Abyssinian) cats. ²¹ In the present study, the vitreal b-wave in stage 1 was 62% of that in normal (non-Abyssinian) animals when bright flashes were used, and 55% of normal when the weaker stimulus was used. The losses continued as the disease progressed, and the ERG was flat in stage 4 animals, again consistent with previous work. ⁵  

Even in stage 1, the Abyssinian cats have fewer photoreceptors than unaffected Abyssinians or normal domestic short-haired cats, which must explain some of the ERG changes. However, the ERG changes were greater than the loss of photoreceptors. In stage 1, photoreceptor numbers were reduced to approximately 70% of those in the normal retina, whereas the a-wave maximum amplitude was reduced to 50% of normal. This suggests some dysfunction in the remaining rods. In stages 2 and 3 the a-wave maximum amplitude continued to decline, but the photoreceptor count did not decrease further until stage 3, and so again the ERG changes preceded the loss of rods. Changes in the ERG measures were similar to the changes in another functional measure, photoreceptor oxygen utilization, as demonstrated in the companion work. ¹² Another finding of the present work was that the a-wave was reduced more than the b-wave at the earliest stages of disease (stages 1 and 2). This has also been shown recently by using component factor analysis ^{22 23} and by graphic representation of a-and b-wave wave amplitudes in homozygous affected cats compared with heterozygous, ophthalmoscopically normal cats. ²⁴ The same trends have been observed in P23H transgenic rats. ²⁵  

A new result of the present study is the small, nonsignificant changes in A and t _eff, which characterize the activation steps in phototransduction according to the Lamb and Pugh ⁶ model. Another new result is the lack of substantial change in the half-saturating intensity, ς, of the a-wave. As has been previously found in affected Abyssinan cats up to 2 years old, ⁴ ς for the b-wave was also not markedly changed. Together, these findings suggest that the decrease in amplitude of the ERG is not caused by a defect in the biochemistry of the transduction process. Caution is required, however, because cases are known where there is a mutation in rhodopsin, and the activation stage of transduction is still normal. ²⁶  

The results presented herein reinforce the similarity between the slow degeneration in Abyssinian cats and that in some types of human RP. Breton et al. ¹¹ reported that a _max decreased in a patient with RP_, but the amplification constant A and delay t _eff were indistinguishable from their mean normal values. This led them to suggest that the rods in human RP were either shortened or partly absent, yet had normal phototransduction. Rods in Abyssinian cats may eventually be shortened in length or absent, as shown by earlier ultrastructural studies. ³ Using an equivalent model of phototransduction, Hood and Birch ²⁷ reported that some forms of RP can alter the activation stage of transduction, but found other patients with small a-waves and normal transduction. They and Shady et al. ²⁸ suggested that patients with RP often have normal amplification early in life, but the degeneration of rods (i.e., further damage to the outer segments) may lead to reduced A. There is no evidence in cats that reduced photoreceptor numbers lead to a change in A, but based on the suggestions from human studies, A would be expected to change in cats late in stage 3; we studied too few animals at that stage to be certain on this point. Consistent with the Hill analysis performed herein, the rod a- and b-waves in patients with cone–rod degenerations were reduced in amplitude, but generally did not show reduced sensitivity. ²⁹  

The present results also provide the only intraretinal recordings of the ERG in any animal model of RP and show clearly that the area centralis is affected early in the disease, even though the central fundus looks relatively normal compared with the periphery. The comparison of the intraretinal b-wave with the vitreal b-wave indicates that later in the disease the central retina is protected somewhat relative to the peripheral retina. These results provide independent confirmation of those of Seeliger and Narfström, ³⁰ who used a different technique, the multifocal ERG, to analyze the spatial distribution of disease in the Abyssinian cats. They also found a generalized loss of function in the early stage and a relative preservation of central retina late in the disease. 

The intraretinal results also provide new insight on the c-wave, and show that it declines with a time course similar to that of the b-wave. This had not been apparent from previous corneal recordings ⁴ or our vitreal recordings, which did not show a dramatic change in the c-wave (to 81% of control in the previous work and 70% of control in the current study during stage 1). We can now assert that the lack of change in the vitreal c-wave occurs because both the RPE and Müller cell components decrease, leaving the vitreal or corneal response, which is the difference between these components, relatively unchanged. 

A substantial amount of the change in the ERG results from loss of photoreceptors rather than simply dysfunction. This is also true in rodent models of degeneration in which relative cell loss and ERG loss have been investigated. ^{25 31} However, it appears that the percentage changes in the photoreceptor counts and in the a-wave amplitude can follow somewhat different time courses. In the present study, comparison of Figure 3A and 7Dindicates that the a-wave was reduced to half of normal at a time when at least 70% of the photoreceptors remained in most of the retina (stage 1), and to approximately 30% of normal when there were still 40% of the photoreceptors present (stage 3). The larger decrement in the a-wave is consistent with subtle changes that are visible at the electron microscopic level. Electron microscopic changes include disorganization and vacuolization of the outer segments as well as complete degeneration of some entire photoreceptors. ³ These are more prominent in the periphery, but are present at all locations. They suggest a defect in the gene for a structural protein, but this is not certain. Conceivably, this would affect the rods in several ways, at least one of which must be linked to membrane voltage responses. For instance, the a-wave would be smaller if the gene defect made the rods generally more hyperpolarized in the dark, so that when the light-dependent cation channels closed during illumination, the rod membrane potential would have less range before it reached full hyperpolarization. 

Another problem that does not have a clear resolution is the relation between responses of individual rods and the size of the a-wave, which is the summation of the individual responses. One hypothesis would be that, at a particular stage of the disease, some rods produce little or no response and healthier ones produce full responses, yielding a smaller a-wave, and the other hypothesis is that the a-wave is reduced because the response of each rod is reduced to some extent. One argument weakly favors the latter interpretation, but another argument tends to favor the former hypothesis. The point in favor of the latter hypothesis is that it is now clear that there are functional defects across the entire retina, and so it cannot be that the ERG decreases because the peripheral rods are affected but the central rods are normal. This reasoning is weak because it could also be that selected cells in all locations are affected, whereas other cells in the same regions are normal. One of the arguments in favor of a patchy depression of rod responses is that the electron microscopy showed some defective cells and some apparently normal cells in all locations. There is no evidence of a general shortening of outer segments sufficient to reduce their light capturing volume uniformly, commensurate with the large change that occurs by stage 1. The disease progression would be related to an increase over time in the number of histologically abnormal cells, and presumably therefore, electrically abnormal cells. 

One way to resolve some of these issues would be to obtain recordings from individual isolated photoreceptors from Abyssinian cats and reconstruct an a-wave from the population of those individual responses. Such experiments are feasible, but difficult. The other approach, which has been unsuccessful so far, would be to find the gene defect, and reason forward from that. 

In summary, the ERG a-wave is attenuated, even in the earliest stage of the retinal degeneration in Abyssinian cats. The amplitude change is not caused by a defect in the transduction cascade, but rather in the number of active photoreceptors and/or the ability of each to capture light. The changes in photoreceptor signaling propagate to the b- and c-waves, although the changes in these second-order signals are somewhat smaller than those in photoreceptor responses. Although the funduscopic changes and photoreceptor counts indicate that the peripheral retina is affected earlier than the central retina, intraretinal recordings and photoreceptor counts show that major functional changes also occur in the central retina in stage 1. As far as can be determined, the results in the cat are consistent with major findings in human RP. This suggests that other attributes of the condition in the cat, which can be investigated with more invasive studies, can also provide information about the human disease. 

 Supported by National Eye Institute Grants R01EY05034 (RAL) and T32 EY07128 (LP-S), the Foundation Fighting Blindness (KN), Swedish Medical Research Council Grant 14X-12209 (KN), and the National Cancer Institute (24XS028) (KN).

 Submitted for publication September 28, 2005; revised February 27, 2006; accepted June 6, 2006.

 Disclosure: J.J. Kang Derwent, None; L. Padnick-Silver, None; M. McRipley, None; E. Giuliano, None; R.A. Linsenmeier, None; K. Narfström, 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.

 

Signals Recorded with the Two Types of Stimuli Described in the Methods

Recording	Stimulus
		Bright, brief	Dim, 4 second
Vitreal a-wave	✓
Vitreal b-wave	✓	✓
Vitreal c-wave		✓
Intraretinal b-wave		✓
Intraretinal c-wave		✓
Control data	Reference 7	Reference 9

Figure 1.

 

The ERG of a normal, dark-adapted cat retina (A), and stage 1 (C), and stage 3 (E) Abyssinian cats in response to bright, brief stimuli. (A) The flash intensity varied from Φ = 2.93 to 7.93. The intensity Φ is given in log photoisomerizations per rod per flash. The a-wave amplitude saturated at −699 μV (cat 236). (B) The a-wave responses of Figure (A) normalized to the saturated a-wave amplitude. The Lamb and Pugh ⁶ model, applied to the data, is shown as dashed lines. Responses to Φ = 2.93 to 4.93 were fitted simultaneously. A and t _eff were 0.83 s⁻² and 4.13 ms, respectively. Responses to the three brightest flashes (Φ = 5.93–7.93) were fitted individually. A and t _eff were, respectively, 0.17 s⁻² and 2.23 ms (Φ = 5.93); 0.04 s⁻², 1.75 ms (Φ = 6.93); and 0.02 s⁻², 1.73 ms (Φ = 7.93). (C) The flash intensity varied from Φ = 2.93 to 7.93 in a stage 1 cat (cat 249). (D) The a-wave responses normalized to the saturated a-wave amplitude, −293 μV. A and t _eff were, respectively, 0.62 s⁻² and 2.93 ms for Φ = 2.93 to 4.93; 0.16 s⁻², 1.94 ms (Φ = 5.93); 0.05 s⁻², 1.34 ms (Φ = 6.93); 0.02 s⁻², 1.36 ms (Φ = 7.93). (E) The flash intensity varied from Φ = 3.93 to 6.93 for a stage 3 cat (cat 268). (F) The a-wave responses normalized to the saturated a-wave amplitude, −111 μV. A and t _eff were, respectively, 0.38 s⁻² and 0.82 ms for Φ = 3.93 to 4.93; 0.11 s⁻², 0.93 ms (Φ = 5.93); 0.05 s⁻², 1.00 ms (Φ = 6.93).

Figure 2.

 

The parameter A obtained from fits of the type shown in Figure 1plotted as a function of log intensity (photoisomerizations). (A) Normal cats; (B) stage 1 cats; (C) stage 2 cats; (D) stage 3 cat. The data in (A) were reproduced from Reference ⁷ .

Figure 3.

 

(A) The average a _max in response to the bright, brief stimuli as a function of the stage of the disease. (B) The amplification constant A from individual cats (filled symbols) as a function of the stage of the disease. The average value of A is shown as an open circle (except for the stage 3 cat). (C) The average b-wave amplitude as a function of the disease stage. The b-wave amplitudes were obtained in response to bright flashes. The amplitude was measured from the a-wave trough to the peak of the b-wave.

Figure 4.

 

Intensity–response functions. (A) The a-wave responses to bright flashes were obtained. The normal response function was obtained from eight cats. (B) The b-wave responses as a function of intensity. The normal response function was obtained from seven cats. (C) The c-wave response as a function of intensity. The normal response function was obtained from seven cats. The b- and c-waves were obtained in response to 4-second flashes. Solid lines: the fits of the Hill equation. The b- and c-waves in response to the brightest flashes (either 0.5 or 0 log unit attenuation) decreased and were omitted from the fitting (except for the b-wave response function from normal cats).

Table 2.

 

Parameters of the Hill Equation Fitted to the a-, b-, and c-Waves of Normal and Abyssinian Cats

a-Wave	R max (μV)	n	ς(log Φ)
Normal	585	0.57	4.25
Stage 1	265	0.61	4.65
Stage 2	162	0.99	4.70
Stage 3	160	0.59	4.12

b-Wave	R max (mV)	n	ς (log unit)
Normal	1.28	0.48	2.76
Stage 1	0.71	0.60	2.82
Stage 2	0.79	0.59	2.50
Stage 3	0.39	0.61	3.24

c-Wave	R max (mV)	n	ς (log unit)
Normal	1.44	0.60	2.81
Stage 1	0.95	0.64	2.47
Stage 2	0.87	0.59	2.24
Stage 3	0.45	0.68	3.09

 Units of ς are log photoisomerizations per rod per flash for the a-wave, and relative log attenuation for the b- and c-waves, the same units as are used in Figure 4 .

Figure 5.

 

The intraretinal and vitreal b- and c-wave amplitudes in response to 4-second flashes as a function of disease stage. (A) The intraretinal b-wave amplitude as a function of disease stage. The intraretinal b-wave amplitude was measured just before the electrode encountered the RPE to minimize horizontal cell contributions to the recordings. Inset: sample intraretinal ERGs from stage 1 and 3 cats. (B) Vitreal b-wave amplitudes as a function of disease stage. The b-wave amplitudes were obtained at a saturating light intensity. Inset: sample of vitreal ERGs from stage 1 and 3 cats. (C) The intraretinal c-wave amplitude as a function of disease stage. The amplitude was obtained in the subretinal space near the RPE. (D) The vitreal c-wave amplitude as a function of disease stage.

Figure 6.

 

Light micrographs of the retina in stage 1 and 3 cats. (A) Central retina from a stage 1 cat. (B) Superior peripheral retina from a stage 1 cat. (C) Central retina from a stage 3 cat. (D) Inferior peripheral retina from a stage 3 cat. Scale bar, 20 μm.

Figure 7.

 

The number of photoreceptor nuclei and the outer retinal thickness as a function of disease stage. Shown are photoreceptor nuclei per high-power microscope field in central (A), superior (B), and inferior (C) areas. Each point is the average of counts from three sections in one eye. Animals used for electrophysiological measurements (•); Another group of stage 1 Abyssinians and a group of heterozygous Abyssinians ( ); Normal domestic short-haired cats (○). (D) Averages of data in (A), (B), and (C) normalized to the number of nuclei in the normal cats, where normal includes both domestic short haired and heterozygous Abyssinians. Groups are indicated in the key. (E–H) Measures of outer retinal thickness (OPL to RPE) from the same sections used to make the cell counts in (A–D).

The authors thank Michael Breton for the use of the Photran program and A/D board; Norbert Wangsa-Wirawan, Gulnur Birol, Ewa Budzynski, Christina Enroth-Cugell, John J. K. Derwent, and Yun Kim for their assistance; and Karen Holopigian for comments on a draft of the manuscript.