Changes in the Photoreceptor Mosaic of P23H-1 Rats During Retinal Degeneration: Implications for Rod-Cone Dependent Survival

Abstract

Purpose.: To investigate the spatiotemporal relationship between rod and cone degeneration in the P23H-1 rat.

Methods.: Control Sprague-Dawley (SD) and P23H-1 rats of ages ranging from P30 to P365 were used. Retinas were processed for whole mounts or cross sections and rods and cones were immunodetected. We used newly developed image analysis techniques to quantify the total population of L/M cones (the most abundant cones in the rat) and analyzed the rings of rod-cone degeneration.

Results.: In P23H-1 rats, rod degeneration occurs rapidly: first the rod outer segment shortens, at P30 there is extensive rod loss, and by P180 rod loss is almost complete except for the most peripheral retina. The numbers of L/M cones are, at all postnatal ages, lower in P23H-1 rats than in control SD rats, and decrease significantly with age (by P180). Rod and cone degeneration is spatiotemporally related and occurs in rings that appear already at P90 and spread throughout the entire retina. At P180, the rings of rod-cone degeneration are more abundant in the equatorial retina and are larger in the dorsal retina.

Conclusions.: This work describes for the first time that in the P23H-1 rat, rod and cone degeneration is spatiotemporally related and occurs in rings. Cone loss follows rod loss and starts very soon, even before P30, the first age analyzed here. The characteristics of the rings suggest that secondary cone degeneration is influenced by retinal position and/or other intrinsic or extrinsic factors.

Introduction

Retinitis pigmentosa (RP) comprises a heterogeneous group of inherited retinal degenerations that lead to irreversible blindness. The disease is characterized by a progressive death of photoreceptors that affects primarily rods, and secondarily cones, because the mutated genes are usually expressed in rods. ^1–4 However, although all rods share the genetic defect, rod degeneration starts in the equatorial retina and proceeds from there in a centripetal fashion. ⁵  

There are several animal models of RP. We used in this study the P23H-1 transgenic rat, an animal that bears an autosomal dominant mutation in the rhodopsin gene (proline-to-histidine substitution at codon 23 of the rhodopsin protein) that causes photoreceptor dystrophy and death. ^6–9 Rhodopsin mutations account for approximately 30% to 40% of the autosomal dominant RP cases and some infrequent autosomal recessive cases, ^10–12 and the P23H genetic defect is one of the most prevalent mutations associated with autosomal dominant RP in North America. ^10–12 How a rhodopsin mutation causes photoreceptor death is still a matter of investigation, although some mechanisms have been proposed. 

Both in RP patients and in animal models of RP, cones degenerate usually when there is a substantial loss of rods. ^13,14 Why cones degenerate secondarily to rods, when the genetic defects are specific of rods, ^3,12 is it still a matter of debate, but various mechanisms of cone-rod dependence for survival have been postulated: rods secrete an essential survival factor for cones ¹⁵ ; rod loss causes oxidative damage to cones ¹⁶ ; retinal and choroidal degenerative changes impede normal photoreceptor nutrition ^17,18 ; and degenerating rods generate a toxin that reaches cones by gap junctions, causing their death. ¹⁹  

Viable cones remain in the retina for long periods of time after rod death both in RP patients and in RP animal models. ^{13,14,20–25} Cone survival in RP is important, as it may preserve photopic vision. ^26–28 Recently, various authors have studied cone survival in animal models of RP and have documented different degrees of cone reorganization. ^{13,14,20,24,25,29–32} Some of these studies, carried out in the S334ter-3 rat, ^30–32 have showed rings of cone degeneration and have investigated their nature and possible relationship to rod death. 

In a previous study, we have documented that photoreceptor degeneration in transgenic P23H-1 rats is followed by degeneration of all the retinal layers and by retinal ganglion cell (RGC) death. ³³ We also established, using automated counting techniques in whole-mounted retinas, the total number and topographical distribution of S- and L/M-cones in the albino and pigmented rat retina. ³⁴ In this study, we use the above-mentioned techniques and newly developed software to document for the first time the changes occurring in the cone mosaic in the P23H-1 rat retina during retinal degeneration. We show that cone loss follows rod loss and happens in rings distributed distinctively in the retina (García-Ayuso D, et al. IOVS 2011;52:ARVO E-Abstract 4338). Furthermore, we find a spatiotemporal relationship between rod and cone degeneration in the P23H-1 retina. Our results also suggest that intrinsic and extrinsic factors may influence the genetically determined photoreceptor degeneration. 

Materials and Methods

Animal Handling

Postnatal day (P) 30 to 180 homozygous albino female P23H-1 and age-matched albino female Sprague-Dawley (SD) rats were used for this study. Transgenic homozygous P23H-1 animals were obtained from M. LaVail (University of California at San Francisco School of Medicine; http://www.ucsfeye.net/mlavailRDratmodels.shtml), and bred at the University of Murcia; SD animals were obtained from the breeding colony of the University of Murcia, Spain. Rats were housed in temperature- and light-controlled rooms with a 12-hour light/dark cycle (light from 8 AM–8 PM) and had food and water ad libitum. Light intensity within the cages ranged from 5 to 30 lux (scotopic to mesotopic conditions). Animal manipulations were carried out following the Spanish and European Union regulations for the use of animals in research (Council Directive 86/609/EEC) and the ARVO statement for the use of animals in ophthalmic and vision research. Adequate measures were taken to minimize pain or discomfort. 

Tissue Processing

Rats were given a lethal dose of sodium pentobarbital (Dolethal Vetoquinol, S.A., Lure, France) and perfused transcardially through the ascending aorta first with saline and then with 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). Some eyes were used for cross sections and others for retinal whole mounts, as follows. 

Cross Sections.

The eyes of 2 control SD rats (P60) and of 11 P23H rats (P30, n = 3; P90, n = 3; P180, n = 3; and P365, n = 2) were enucleated and their superior pole marked with china ink. The cornea and lens were removed and the resulting eyecups were postfixed for 1 hour, cryoprotected in 15% sucrose (Sigma, Alcobendas, Madrid, Spain) and embedded, with the superior pole in a known position, in optimal cutting temperature compound (Sakura Finetek, Torrance, CA) for cryostat sectioning (15-μm thick). 

Whole Mounts.

The eyes from P30 (n = 9), P90 (n = 9), or P180 (n = 12) P23H-1 rats and from P30 (n = 8) and P180 (n = 8) SD rats were enucleated and the retinas were dissected as whole-mounts by making four radial cuts in the superior, inferior, nasal, and temporal retinal quadrants. Retinal orientation was maintained by making the deepest radial cut in the superior retina. The retinas were postfixed in 4% paraformaldehyde for 1 hour, washed in PBS, and subjected to immunohistofluorescence. 

Immunohistofluorescence

Cross Sections.

Sections were washed three times with PBS, 10 minutes each to eliminate the embedding medium, incubated overnight at 4°C with a mixture of the primary antibodies diluted in blocking buffer (PBS containing 0.1% Triton X-100 and 5% normal donkey serum; Jackson ImmunoResearch, Inc., Cambridge, UK). 

Secondary detection was performed by incubating the sections for 1 hour at room temperature with a mixture of the secondary antibodies diluted in PBS containing 0.1% Triton X-100. Sections were then mounted with antifading medium (VectaShield Mounting Medium; Vector, Atom, Alicante, Spain) or with TO-PRO-3 iodide (1:1000 dilution; Molecular Probes, Invitrogen, Barcelona, Spain) to counterstain all retinal nuclei. 

Whole Mounts.

Retinas were washed and permeated as described. ^33–36 Primary and secondary detection was carried out as for cross sections with the exception that PBS containing 2% Triton X-100 was used. Finally, the retinas were washed in PBS and mounted on subbed slides vitreal side down, and covered with antifading mounting media (Vectashield Mounting Medium). 

Antibodies and Dilutions

Primary Antibodies.

S-opsin was detected with goat anti-OPN1SW (N-20; Santa Cruz Biotechnology, Heidelberg, Germany), which detects the UV-sensitive opsin (i.e., S-cones) diluted 1:1000; L/M-opsin was detected with rabbit antiopsin red/green (Chemicon-Millipore Iberica, Madrid, Spain), which detects the L/M-opsin in rodents (i.e., L/M-cones) diluted 1:1200; and rhodopsin was detected with mouse antirhodopsin (Sigma-Aldrich, Madrid, Spain), which detects rods, diluted 1:10,000. Recoverin was identified with a mouse antirecoverin antibody used at 1:2000 dilution (JF McGinnis, University of Oklahoma, Oklahoma City, OK) and transducin with rabbit antitransducin Gαc (Cytosignal, Irvine, CA) diluted 1:200. 

Secondary Antibodies.

All secondary antibodies were purchased from Molecular Probes (Invitrogen) and diluted at 1:500 to detect opsins and rhodopsin and at 1:100 to visualize recoverin and γ-transducin. To double detect S- and L/M-cones, Alexa Fluor 488 donkey antigoat IgG raised against Heavy and Light chains (H+L) and Alexa Fluor 568 donkey antirabbit IgG (H+L) antibodies were used. To analyze all cones and rods, S and L/M opsin were detected using two secondary antibodies conjugated with the same fluorophore (Alexa Fluor 488 donkey antirabbit IgG [H+L] and Alexa Fluor 488 donkey antigoat IgG [H+L]), and rhodopsin was identified with Alexa Fluor 568 donkey antimouse (H+L). To double detect recoverin and γ-transducin, Alexa Fluor 488 donkey antirabbit IgG (H+L) and Alexa Fluor 555 donkey antimouse IgG (H+L) were used. 

Retinal Image Analysis

Confocal Microscopy.

Retinal cross sections labeled with recoverin and γ-transducin were viewed using laser-scanning confocal microscopy on a Leica TCS SP2 system (Leica Microsystems, Wetzlar, Germany). ³⁷  

Optic Microscopy.

Both retinal whole mounts and cryostat cross sections immunodetected for opsins and/or rhodopsin, and were examined and photographed under a fluorescence microscope (Axioscop 2 Plus; Zeiss Mikroskopie, Jena, Germany) following standard procedures in our laboratory. ^{33–36,38–42} The microscope is equipped with the appropriate filters, a digital high-resolution camera (ProgRes C10; Jenoptik, Jena, Germany) and a computer-driven motorized stage (Pro-Scan H128 Series; Prior Scientific Instruments, Cambridge, UK) controlled by IPP (IPP 5.1 for Windows; Media Cybernetics, Silver Spring, MD) with a microscope controller module (Scope-Pro 5.0 for Windows; Media Cybernetics). For cross-sections, individual images were taken. To make reconstructions of retinal whole mounts, retinal multiframe acquisitions were acquired in a raster scan pattern using a ×10 objective (Plan-Neofluar, 10×/0.30; Zeiss Mikroskopie). Single frames were focused manually prior to the capture of the digitized images. Because the frame size is 0.627mm²/image, we usually need to acquire 154 images to scan the entire retina. 

The images taken for each flat mount retina were saved as a set of 24-bit color image pictures and later, these images were combined into a single high-resolution composite image of the whole retina using IPP. Reconstructed images were further processed using Adobe Photoshop CS 6 (Adobe Systems, Inc., San Jose, CA) when needed. 

Automated Quantification of L/M Cones

Although in the whole mounts, both rods, L/M and S cones were labeled, we were able to quantify automatically only L/M cones due to the irregular expression pattern of S-opsin in the S-cone. 

To quantify the total numbers of L/M-cones, the individual images taken in each retina were processed by a specific cell-counting subroutine developed by our group. ³⁴ Briefly, we used the IPP macro language to apply a sequence of filters and transformations to each image to clarify cell limits and separate individual cells for automatic cell counting. Those procedures have been previously reported in detail. ³⁴  

Isodensity Maps

To demonstrate the spatial distribution of L/M-cones, we constructed an isodensity map for each retina, using the specific subroutine developed in our laboratory. ³⁴ Briefly, using IPP macro language, every frame was divided into 25 sampling areas, in which cone counts were obtained and cell densities calculated. These densities were represented as filled contour plots using graphing software (Sigmaplot; Systat Software, Inc., Hounslow, London, UK; for further details see Ortín-Martínez et al. ³⁴ ). 

Ring Analysis

Whole-mounted retinas showed rings devoid of S- and L/M-cones. The ring boundaries were manually outlined using Adobe Photoshop CS 6 (Adobe Systems, Inc.). A specific subroutine was developed to analyze the characteristics of the rings using the IPP macro language. This subroutine performed three tasks: quantification of the total number of rings in the retina; measurement of their area, perimeter, and diameter; and determination of the x-y coordinates of the epicenter of the rings (with respect to the optic disk). The data were automatically exported to an Excel document for storage and posterior analysis. 

To analyze the distribution of the rings, the retina was divided into dorsal and ventral nasal and temporal areas, and also in three areas depending on their distance to the optic disk: central (0–1750 μm), equatorial (2000–3750 μm), and peripheral retina (4000–5750 μm). 

Statistics

Statistical analysis was carried out using SigmaStat 3.1 for Windows (SigmaStat for Windows Version 3.11; Systat Software, Inc., Richmond, CA). The Kruskal-Wallis test was used to compare more than two groups and the Mann-Whitney test was used when comparing two groups only. Differences were considered significant when P was less than 0.05. 

Results

Cone and Rod Morphological Changes

In control SD retinas (P30 and P180), rhodopsin and S- and L/M-opsin are expressed in the outer segment (OS) of rods and S- and L/M-cones, respectively (Figs. 1A, 1D). In P23H1 rats at P30, the OS of rods and cones (Figs. 1B, 1E) was markedly reduced compared with SD. From P90 onward, there was an abnormal rhodopsin and opsin expression in the dystrophic retinas and a strong immunoreactivity was observed not only in the OS of rods and cones, but also in the cell bodies and inner segments or axons (Figs. 1F, 2). At P180, most surviving rods and cones had lost their OS and showed also changes in their longer axis orientation, which changed from a vertical to a horizontal position (Figs. 1C, 1F). Also, surviving rods and cones had sprouted new axonal branches that sometimes were seen to “connect” different rods (Figs. 2A–C) or cones (Figs. 2D–F). The aberrant opsin expression was more marked for the S- than for the L/M-opsin (Fig. 1F) and thus did not allow the automated quantification of S-cones (see below). 

Figure 1

 

Rod and cone degeneration in the P23H-1 rat: analysis in retinal cross-sections. Representative retinal cross-sections from a control SD rat (A, D) and two P23H-1 rats processed at P30 (B, E) or P180 (C, F). (A–C) Immunodetection of rods (rhodopsin, red) and all cones (S and L/M opsin detected with the same fluorophore, green). (D–F) Double immunodetection of L/M-cones (L/M opsin, red) and S-cones (S-opsin, green). As the retinal degeneration in the P23H-1 strain progresses, it is observed that the photoreceptors' outer segment shortens and that the orientation of both, cones and rods, changes from vertical to horizontal (compare [B, E] with [C, F], respectively). In (F) there are two cones showing opsin expression in their somas and inner segments (arrows). Scale bar: 100 μm.

Figure 2

 

Aberrant expression of rhodopsin and opsin and morphological changes in degenerating photoreceptors. High-power microphotographs of whole mounts from P23H-1 animals processed at P180 in which rods have been immunodetected with rhodopsin (red, [A–C]) and S- and L/M cones double immunodetected with S- (green, [D]) and L/M- opsin (red, [E, F]). Both rods and cones show aberrant expression of the three visual pigments that is no longer circumscribed to the outer segment but found in the soma, inner segment, and axons. In addition, the photoreceptors show axonal sprouting (arrows). Scale bar: 100 μm.

Retinal cross sections from aging P23H1 rats were also doubly immunolabeled for transducin and recoverin, and counterstained with TOPRO3 to identify all retinal nuclei (Fig. 3). Recoverin is expressed by rods, cones, and two bipolar cell subtypes ⁴³ and γ-transducin by cones and at least two subtypes of cone bipolar cells. ⁴⁴ These proteins are expressed in the whole cell (i.e., soma and axons) and in photoreceptors, also in the OS. Thus, this analysis permitted assessment of changes in the photoreceptor morphology that might be missed when using OS-specific markers, such as rod and cone opsins. 

Figure 3

 

Temporal changes of recoverin and γ-transducin expression in the P23H-1 rat retina. Double immunodetection of recoverin (red) and γ-transducin (green) in retinal cross-sections from P23H-1 rats analyzed at P30 (A, E), P90 (B, F), P180 (C, G), and P365 (D, H). Recoverin is expressed in both rods and cones and two bipolar cell subtypes and γ-transducin labels cones and some subtypes of bipolar cells. Note the progressive cell loss and morphological changes with the degeneration. Arrowheads in (E) and (F) point to disorganized, tortuous, short and swollen cone outer segments. Arrows in (E) point to abnormal cone pedicles that have lost their typical triangular morphology and that spread into the outer plexiform layer. (G) and (H) show that the remaining photoreceptor cells at old ages are cones and in (G) and (H), arrows point to the few remaining synapses between bipolar cells and cones (yellow signal). All retinal nuclei have been counterstained with TO-PRO. Scale bar: 20 μm.

The first observation was that the thickness of the outer nuclear layer (ONL), where the photoreceptor nuclei lay, decreased with age (Fig. 3). Although the normal thickness of the ONL in SD rats is 10 to 14 photoreceptor nuclei thick, it was only 4 to 5 nuclei thick in P30 P23H-1 rats (Figs. 3A, 3E) and has almost disappeared by P365 (Figs. 3D, 3H), in agreement with previous reports. ^33,35 Therefore, at P30 there is already extensive rod loss, and cones begin to degenerate later (see below). 

At P30, in cross sections there was no indication of cone cell death, but the cone OS were disorganized, thicker, tortuous, and shorter (Figs. 3A, 3E, arrowheads). Cone pedicles had lost their typical triangular morphology and were spreading into the OPL (Fig. 3E, arrows). Three-month-old P23H-1 rats showed further changes in cone morphology: cone cell size was reduced, most cell bodies were aligned close to the outer plexiform layer, many had lost their axons, and their outer segments appeared to be swollen (Figs. 3B, 3F, arrows). At P365, the few remaining recoverin-positive photoreceptors were also transducin positive, suggesting that the only photoreceptors that had survived were cones. All cones had lost from P180 their typical morphology and alignment and most of their processes were horizontally oriented in the sections (Figs. 1C, 1D; 3D, 3H) as shown before. In addition, there were areas of contact between cones and cone bipolar cells (Figs. 3G, 3H, arrows) that could correspond to remaining synaptic contacts. 

Cone and Rod Spatial Reorganization: Rings of Rod-Cone Degeneration

In control SD rats, rhodopsin signal in flat-mounts had a grasslike appearance throughout the whole retina, both at P30 and at P180 (not shown). Although in P23H-1 retinal sections, rod loss was observed at P30, in whole-mounted retinas rod loss was evident only at P90. We know that at this later time the great majority of the population of rods has degenerated (Figs. 1, 3 ³³); however, in whole mounts, we could still distinguish many preserved rods that were denser in the ventral than in the dorsal retina (Figs. 4B–F). By P180, rods were abundant only in the most peripheral retina (Figs. 4A, 4G–J). At this time, rod loss was almost complete across the retina and, in the ventral retina, the few remaining rods were rearranged in ringlike-shaped areas (see below, Figs. 4G–J). 

Figure 4

 

Rod and cone degeneration in the P23H-1 rat: analysis in flat-mounted retinas. (A) Photomontage of a whole-mounted retina from a P23H-1 animal processed at P180 showing rods (rhodopsin signal in red) and all cones (S- and L/M- opsins detected with the same fluorophore, green). At this postnatal time, most rods have disappeared from the retina, except at the extreme retinal periphery. Throughout the retina, degeneration rings delineated by rods and cones are observed. (B–J) Magnifications from whole-mounted P23H-1 retinas immunodetected as retina in (A) and analyzed at P90 (B–F) or 180 (G–J). (B, C, G, H) Images were acquired from the ventral retina and (E, F, I, J) from the dorsal retina. At both postnatal times, rhodopsin was more abundant in the ventral (B, C, G, H) than in the dorsal (E, F, I, J) retina. Rings delineated by rods and cones are already observed at P90, but at P180 in the ventral retina they have rhodopsin in their border (G, H) while in the dorsal retina (I, J) rhodopsin expression is lost. Scale bars: 1 mm in (A) and 100 μm in (B–J).

Cone distribution in SD rats has been recently described. ³⁴ S-cones are more abundant in the retinal rim, show medium densities in the ventral retina and very low densities in the dorsal retina. L/M-cone distribution is complementary to that of the S-cones: they are scarce in the retinal rim, show higher densities in the dorsal retina, and medium densities in the ventral retina. When both cones are identified, their distribution is more or less even in the retina (Figs. 5A–C). In the P23H-1 rat, the normal photoreceptor mosaic is disrupted at P90 by the appearance of rings delimited by both types of cones (Figs. 4B–F, 5D–F) and rods (Figs. 4G, 4H). At P180, these rings could be seen all throughout the retina (Fig. 4A) and had a more or less circular shape that was delineated by the strong fluorescence of the thickened OS of cones and rods (when still present, i.e., at P180 only in the ventral retina) (Figs. 4A, 4G–J). Inside the rings, only a few photoreceptors remained, they had lost their OS but could be detected because of the abnormal expression of rhodopsin/opsin in their somas and axons. It was in the center of the rings where the aberrant axonal sprouting was observed both in rods and cones (see above; Figs. 1–3). 

Figure 5

 

S- and L/M-cones outline the degeneration rings in theP23H-1 retinas. Double immunodetection of S- (green) and L/M- (red) opsins in P180 SD (A–C) and P23H-1 (D–F) rats. (C) and (F) correspond to the merged images. In SD retinas, cones are homogeneously distributed, whereas in the P23H-1 retinas, both (S- and L/M-) delineate the degeneration rings (D–F). In (E) and (F) are observed some L/M cones inside the rings. Scale bar: 100 μm.

Qualitative, Quantitative, and Topographical Analyses of the Degeneration Rings

The degeneration rings were first observed at P90, but it is possible that they appear earlier, between P30 and P90. At P90, the rings were small and sometimes not clearly delimited (Figs. 4B–F), but at P180, when rhodopsin immunoreactivity was almost absent (Figs. 4A, 4G–J, 5E, 5F, 6A), they could be seen more clearly all across the retina. Thus, the number, size, and distribution of the rings as well as their distance to the optic disk were analyzed at P180 (Fig. 6; Table 1). The total number of rings observed and analyzed was 754 (n = 10). The mean number of rings per retina (±SD), their diameter and area were 75 ± 46, 210 ± 85 μm, and 40,012 ± 33 μm², respectively (Table 1). 

Figure 6

 

Topography of degeneration rings in P180 P23H-1 rats. (A) Photomontage of a whole-mounted retina from a P23H-1 animal processed at P180 where S- and L/M- cones have been double immunodetected. In photomontages like these, the degeneration rings were manually outlined and colored in blue ([D, E]: note that [D] corresponds to photomontage in [A]). The characteristics of the rings were analyzed (Table 1). Graph in (B) shows the number of rings at a given distance from the optic disk, whereas graph in (C) shows the number of rings in each retinal region. Scale bar: 100 μm.

Table 1. 

 

Number, Area, and Diameter of Degeneration Rings in P180 P23H-1 Rats

	Total	Retinal Region
		Ventral	Dorsal	Nasal	Temporal
Number	75 ± 46	19 ± 10	18 ± 15	18 ± 10	19 ± 16
Area, μm2	40,012 ± 33,183	35,395 ± 27,674	44,298 ± 37,144	40,230 ± 33,534	39,801 ± 32,838
Diameter, μm	210 ± 85	198 ± 82	222 ± 87	212 ± 81	209 ± 89

 First column shows the mean ± SD number of rings, their area and diameter per retina. Second to fifth columns show the same data in each retinal region: ventral, dorsal, nasal, or temporal, respectively.

Most of the rings were located at the equatorial retina (P < 0.05, Mann-Whitney test) and the mean distance from the epicenter of the rings to the optic disk was 2516 ± 789 μm (Figs. 6B, 6D, 6E). Ring distribution was similar in the dorsal and ventral retina (Fig. 6C) and in the four retinal quadrants (P = 0.782 and P = 0.787, Mann-Whitney test; Table 1). There was, however, a difference in the size of the rings in the different retinal regions. Mean area and diameter of rings were significantly larger in the dorsal than in the ventral retina (P < 0.001, Mann-Whitney test; Table 1). However, no difference was found between the number of rings in the temporal and nasal regions (P = 0.477 and P = 0.481, Mann-Whitney test; Table 1). 

In conclusion, in the P23H-1 retina, there are rings of rod-cone degeneration that appear before P90 and are more distinct at P180. The rings are more abundant in the equatorial retina and larger in the dorsal than in the ventral retina. The size of the rings, thus, depends on retinal location, and possibly on other factors (see Discussion). 

Total Number of L/M-Cones

It has been reported that in P23H rats, although cones are reorganized in rings, they do not die. ¹³ To investigate this fact in depth, we automatically quantified the whole population of L/M-cones in P30 and P180 SD control rats, and in P30, P90, and P180 P23H1 rats (Table 2). 

Table 2

 

Total Number of L/M-Cones in SD and P23H-1 Rats at Different Postnatal Days

Rat	SD				P23H-1
	P30		P180		P30		P90		P180
	Left	Right	Left	Right	Left	Right	Left	Right	Left	Right
1	225,175	229,219	210,803	198,995	186,117	179,691	179,749	181,072	138,670	146,577
2	221,574	206,704	193,277	244,041	178,981	166,952	178,078	165,049	90,096	93,244
3	208,208	216,367	210,033	268,211	191,944	160,459	177,346	172,827	169,055	163,905
4	229,284	236,622	223,940	255,318	150,142	172,552	170,801	179,473	127,372	158,434
5	243,251	227,895	271,672	192,141	155,789	178,030	175,542	164,409	155,465	152,843
6			235,871	210,630	172,253	168,072	161,759	165,511	144,381	142,756
7									163,700	157,925
8									141,205	127,323
Mean ± SD	225,498 ± 11,348	223,361 ± 10,555	224,266 ± 24,922	228,22 ± 29,008	172,537 ± 15,191	170,959 ± 6626	173,879 ± 6681	171,390 ± 7545	141,243 ± 23,197	142,875 ± 21,590
Mean ± SD	224,430 ± 11,011		226,244 ± 27,114		171,748 ± 11,745*		172,634 ± 6918*		142,059 ± 22,423*†

 Number of L/M-cones quantified in each of the analyzed retinas. Penultimate row shows the mean ± SD of the left or right retinas. Because there were not differences among them, the last row shows the mean ± SD of both retinas for each group.

*  Statistically significant compared with control (Mann-Whitney test).

†  Statistically significant compared with previous postnatal time (Kruskal-Wallis test).

In control SD rats, the mean (±SD) number of L/M-cones at P30 and P180 was 224,430 ± 11,011 and 226,244 ± 27,114, respectively. These numbers were not significantly different from those found in control SD rats in previous studies of this laboratory (231,736 ± 14,517; n = 14³⁴). 

In P23H-1 rats, the number of L/M-cones was 171,748 ± 11,745, 172,634 ± 6918, and 142,059 ± 22,423 at P30, P90, and P180, respectively (Table 2, Fig. 7A). Comparison of the mean numbers of L/M-cones among both rat strains (SD and P23H-1) revealed that the P23H-1 strain had a significantly lower number of L/M-cones than the SD strain at all the postnatal times analyzed (Mann-Whitney test, P < 0.001; Table 2, Fig. 7A). 

Figure 7

 

Percentage and topographical loss of L/M cones in the P23H-1 rat. (A) Percentage of L/M cone loss in the P23H-1 rat. At P30, the P23H-1 rat has 24% fewer L/M-cones than SD rats (the average number of L/M cones in P30 and P180 SD rats has been considered 100%; data from Table 1). From P30 to P90, no cone loss is observed, but from P90 to P180, there is 17% further loss. *Statistically significant (Mann-Whitney test, P < 0.001). This is graphically observed in the L/M cone isodensity maps shown in [B–D]. The number of L/M-cones counted in each map is shown at the left bottom of each retina. Compared with SD (B), L/M cone density in the P23H-1 strain is lower at P30 (C) and P180 (D). Furthermore, at P180 the degeneration rings are observed ([D], arrows). Color scale ranges from 0 (purple) to 6500 or more L/M-cones/mm² (red). D, dorsal; T, temporal; V, ventral; N, nasal. Scale bar: 1 mm.

There were no significant differences (Mann-Whitney test) between the number of L/M-cones counted in the left and right retinas processed at any given postnatal age. However, comparison of the number of L/M-cones quantified at different survival times showed that there were significantly fewer L/M-cones in the retinas processed at P180 than at earlier postnatal times (Kruskal-Wallis, P < 0.001; Table 2, Fig. 7A). 

These quantitative data are graphically observed in the isodensity maps showing the distribution of L/M cones (Fig. 7). In control SD retinas, there was an area of highest L/M-cone density just above the optic disk along the nasotemporal axis (Fig. 7B). This distribution was maintained up to P180, which was the oldest age analyzed. In P23H-1 animals, at P30, the youngest rats analyzed, the area of highest cone density was not observed (Fig. 7C). The isodensity maps of older P23H-1 rats also showed the age-dependent loss of L/M-cones, which is evidenced by the loss of the higher L/M-cone densities (warm colors) across the retina (Fig. 7D). 

In conclusion, P23H-1 rats have lower numbers of cones than SD rats at P30, do not show an area of highest density above the optic disc, and lose cones between P30 and P180. 

Discussion

We have used photoreceptor identification and newly developed image analysis techniques developed in our laboratory to study the effect of rod degeneration on the cone population in the P23H line 1 homozygous albino rat. These rats suffer a rhodopsin mutation that it is one of the most prevalent mutations associated with autosomal dominant RP in America. ¹⁰ The mutant P23H rhodopsin has been documented to be misfolded and retained in the endoplasmic reticulum, causing endoplasmic reticulum stress and rod death, ⁴⁵ although other mechanisms, such as membrane desestabilization ⁴⁶ and toxicity, have been proposed (reviewed in Ref. 47). 

Because the rat has a low density of cones that can therefore be quantified, ³⁴ we thought that the P23H rat could be a good model to study the kinetics of rod-induced cone degeneration. There are three lines of P23H rats that have a different expressions of the transgene and, thus, different rates of degeneration. ^8,48 We have used the subline 1, which shows the fastest degeneration rate. In this study, we show for the first time in P23H-1 rats that cone loss occurs in rings scattered throughout the retina that can be seen when most rods have degenerated and that are more abundant in the equatorial retina and larger in the dorsal retina. Therefore, it appears that cone degeneration is secondary to rod degeneration and might be also influenced by other intrinsic or extrinsic factors (see below). 

Rod Degeneration in P23H-1 Rats

The rat retina is a rod-dominated retina, because cones represent only 1% of the photoreceptor population. ^49,50 We have found that in the P23H-1 rats, there is a quantitative and qualitative diminution of rhodopsin expression, even at P30, the earlier age studied (Fig. 1). In whole-mounted retinas, the normal evenly distributed mosaic formed by the rod outer segments was disrupted because the numbers of rods were diminished, and in whole mounts and cross sections, there was also a shortening of the rod outer segments and rod loss with age (Figs. 1–4). 

Rod loss in P23H-1 rats is a consequence of the dystrophy and has been documented to commence as soon as P16 ^8,27,43,51 and to be accelerated by light exposure. ¹³ Because the P23H mutation affects the rhodopsin gene, it is thus expected that this retinal dystrophy causes first rod loss. In a previous study, we showed that at P30, less than half of the population of photoreceptors remained in the retina. ³³ In this study, we show that rod degeneration progresses rapidly with age, it is more severe in the dorsal retina, and by P180 (Fig. 4) normal rod densities are found only in the extreme retinal periphery. This area, called retinal rim or ciliary margin zone (CMZ) has been described in humans ^52,53 and in rats ⁵⁴ and it has been shown to serve as a neurogenesis zone. ^55–58  

Total Number of L/M-Cones in the P23H-1 Rat

In retinal degenerations due to mutations of rod proteins, it has been shown that cones survive for long periods of time after rod degeneration. ^{22–24,28,50,59,60} In the P23H rats and other rats with rhodopsin mutations, the kinetics of cone degeneration have been investigated, ^13,20,29,32 but to our knowledge, the topographical distribution of cone loss in the P23H rat had not been described in detail. 

Due to technical limitations, we could only count automatically the numbers of L/M-cones, but not of S-cones or rods. The total numbers of L/M-cones in control SD rats at P30 and P180 were similar to those found in previous studies of our laboratory. ³⁴ However, at all postnatal times analyzed, the P23H-1 strain had a significantly lower number of L/M-cones than SD rats, suggesting that in the P23H-1 rat cone cell death may start immediately after the initiation of rod death, probably before P30. Also, in this strain there is an age-related cone loss, since at P180 the number of L/M cones was significantly lower than at earlier ages. 

Apart from our previous study in SD rats, ³⁴ there are no other studies analyzing the total density of cones in the rat retina. However, there are some studies that have quantified the density of L/M- or S-cones per millimeter squared in normal SD rats ¹⁴ or in animal models of RP, such as the S334ter rats ^29,31 or the P23H-3 rat. ^13,20,61 In these studies, the authors sampled the retina and averaged the data to obtain mean densities, whereas we analyzed the whole retina, and therefore we cannot compare our numbers with the number obtained in these studies. Chrysostomou et al. ¹³ showed that in P23H-3 rats the densities of L/M-cones declined significantly between P16 and P120 but not thereafter and concluded that this loss was developmental and could not be due to cone death because they did not find a significant difference between aged P23H-3 and SD retinas. Based on data presented here, we think that the loss of cones in the P23H-1 rat is not due to an aging process but rather to the P23H-1 degeneration, because we do not find a decrease with age in the numbers of cones in control SD rats. 

Cone Topography in the P23H-1 Rat Retina: Degeneration Rings

Normal SD rat retinas showed the highest densities of L/M-cones in a horizontal area just above the optic disk, in agreement with our previous observations. ³⁴ However, this pattern was not conserved in the P23H-1 rat retinas. At P30, the first age studied, the densest area was not present and the L/M-cone isodensity maps showed fewer warm colors than those of SD rats. At P90, we observed ring-shaped areas lacking both rods and cones all throughout the retina. Similar rings of rod/cone degeneration have been recently reported in animal models of RP the 334ter-line-3 retina. ^30,31,32,62 This rat suffers a rhodopsin mutation that impedes the trafficking of rhodopsin to the outer segment and causes early photoreceptor degeneration. ^44,63  

We have first observed rings of rod-cone degeneration at P90, whereas in the S334-ter-line-3 rat, these rings have been reported to appear at earlier ages, approximately P10 to 20. ^30,32,62 However, it is possible that in P23H-1 rats the rings appear earlier, between P30 and P90, because we did not investigate retinas between these ages. Also, Ji et al. ³² reported that in the S334ter-line-3 rat the rings became widespread at P90 (when we first observed them in the P23H-1 strain) and begin to disappear at P180 (when they become widespread in the P23H-1 rat). In our model, the analysis of the rings revealed that they were more abundant in the equatorial retina and larger in the dorsal than in the ventral retina. However, in S334ter-line-3 rats, the rings appeared in the periphery of the retina and were much more abundant in the dorsal than in the ventral retina (138 over 54³²). We wonder whether these spatiotemporal differences between these two animal models are due to differences in the nature of the genetic defect or secondary to other causes, such as different rates of rod degeneration. Indeed, the P23H-1 rats suffer a slower degeneration than the S334ter-line-3 rats. ^8,64  

The rings of photoreceptor degeneration are mainly localized in the equatorial retina but are larger in the dorsal than in the ventral retina. We postulate that the rings represent retinal areas with more advanced degeneration and that the dorsal retina may degenerate earlier and/or faster than the ventral retina. Why should cone death in P23H-1 rats be more advanced in the equatorial dorsal retina? We do not know, but is a pattern that this rat strain shares with human RP patients, in which for unknown reasons rod loss also starts in the equatorial retina. ⁵ In rats, it has been documented that the superior retina is more sensitive to light than other retinal regions, ^35,65–67 and it has been proposed that is due to increased light exposure or other intrinsic retinal factors. Whether rod and cone degeneration in the P23H-1 rat are more accelerated in the dorsal retina due to intrinsic or extrinsic retinal factors, is a question that remains for future studies. 

Effect of Rod Degeneration on Cones

In inherited retinal degenerations, it is the loss of cones that is the cause of irreversible blindness. Because many times the genetic defect affects rods, it is possible that cones degenerate secondarily to rods and thus it may be possible to act at any point of the disease to prevent cone and vision loss. ^2,68 The center of the rings were mostly devoid of S- and L/M-cones and also devoid of rods, both in this study and in previous studies in the S334ter-line-3 rats ^30,32,62 and, therefore, it is possible that the rings of rod-cone degeneration are secondary to rod degeneration, as it has been proposed in the S334ter rat. ⁶²  

Because the centers of the rings were devoid of immunoreactive cones, the rings appeared when rods were almost absent, and the rings were delineated by both rods and cones, it is possible that the rings are the result of cone death secondary to rod death, as it has been postulated by other authors. ^15–19 In accordance with this, the numbers of L/M-cones decreased with age in P23H-1 rats and this loss was statistically significant at P180, when the rings became widespread throughout the retina. Therefore, we conclude that ring formation it is due to cone death secondary to rod loss. Because the rings seem to increase in size with time (Fig. 4), we propose that photoreceptor cell death proceeds centrifugally from the center of the rings and that cones remain viable on the periphery of the rings because they obtain some trophic support from a peripheral source, ¹⁵ are physically displaced (i.e., by Müller cells ^32,62 ), or are protected from a toxic effect present in the center of the rings. ^16–19 Other authors, however, have postulated that the rings found in S334ter-line-3 rats are due to cone migration ³² or to reversible loss of the cone outer segment, because it can be prevented with ciliary neurotrophic factor treatment ³⁰ or activation of the STAT3 pathway in Müller cells. ^31,62  

The rings were delimited by the degenerating rod and cone outer segments. The ring interior contained only a few somas and axons of rods and cones that delineated the rings and that were horizontally oriented toward the center of the rings. Sometimes, the cones had lost their outer segments. The somas and inner segments of rods and cones could be immunodetected from P90 onward due to the abnormal expression of opsin and rhodopsin in the inner segment, soma, processes, and synaptic terminals of degenerating rods and cones. This abnormal pattern has been previously seen in animal models of retinal degeneration; both induced and inherited. ^{14,29,32,69–71} As degeneration goes by, outgrowth of new processes from both rods and cones situated in the center of the rings can be seen, and this phenomenon has been previously observed in cones from some models of retinal degeneration such as the rd1 mouse, ^24,59 the S334ter-3 rat retina, ²⁹ and in human RP. ⁷²  

In conclusion, our study documents for the first time the total number and distribution of L/M-cones in the P23H-1 rat and shows that cone cell loss may start before P30 and continues through the life of the animal. We also document that rod degeneration progresses rapidly and is almost complete at P180, except for the retinal rim. We show that rod-cone degeneration occurs in ring-shaped areas that increase in number and size between P90 and P180, and at this latter time are more abundant in the equatorial retina and larger in the dorsal retina. Because the rings appear when most rods have degenerated and because rod loss is more rapid and severe in the dorsal than in the ventral retina, cone and rod loss are spatiotemporally related. Furthermore, because the rings are larger in the dorsal retina, we postulate that photoreceptor loss may be influenced by other retinal intrinsic or extrinsic factors and deserves further investigation. This work would allow us in the future to examine the causes of cell death and also to investigate neuroprotective strategies in this model. 

Acknowledgments

Supported by the Spanish Ministry of Economy and Competitiveness Instituto de Salud Carlos III-Fondo Europeo de Desarrollo Regional PI10/01496, PI10/00187, FIS PS0901854, SAF-2012-38328, BFU2012-36845, RD12/0034/0010, and RD12/0034/0014; Fundación Séneca de la Región de Murcia 04446/GERM/07; and Organización Nacional de Ciegos de España. 

Disclosure: D. García-Ayuso, None; A. Ortín-Martínez, None; M. Jiménez-López, None; C. Galindo-Romero, None; N. Cuenca, None; I. Pinilla, None; M. Vidal-Sanz, None; M. Agudo-Barriuso, None; M.P. Villegas-Pérez, None 

Footnotes

 DG-A and AO-M contributed equally to the work presented here and should therefore be regarded as equivalent authors.