Photoreceptor Degeneration in the RCS Rat Attenuates Dendritic Transport and Axonal Regeneration of Ganglion Cells

Abstract

purpose. Photoreceptor loss in the Royal College of Surgeons (RCS) rat deprives the retinal ganglion cells (RGCs) of sensory input, which could interfere with RGC physiology. Whether axonal and dendritic transport is altered, and whether RGCs retain their capacity to regenerate their axons, both in vivo and in culture, was ascertained.

methods. The study was conducted at postnatal days (P) 30 (while most photoreceptors are still intact), P90 (photoreceptors being almost completely absent), and P180 (approximately 3 months after photoreceptor disappearance). RGCs were studied with retrograde transport of the fluorescent dye 4Di-10ASP. Dendritic transport was also studied with 4Di-10ASP that is transported from the cell bodies into the RGC dendrites. Regeneration of RGC axons in vivo was monitored in the grafting paradigm of replacing the cut optic nerve (ON) with a sciatic nerve (SN) piece. Cell counts were performed in retinal wholemounts. Axonal regrowth in vitro was assessed in organotypic cultures of retinal stripes.

results. Photoreceptor dystrophy did not adversely affect retrograde axonal transport but attenuated dendritic transport compared with the wild-type control rats. Axons of RGCs were able to regenerate if provided with a SN graft, and regeneration was observed to be similar between RCS and wild-type rats at P30 but differed significantly at P90 and P180. In addition to an age-dependent decline in the regenerative ability, seen also in control animals, the number of RCS RGCs able to regenerate declined drastically beginning at 3 months. It is plausible that the intraretinal reorganization, as a consequence of photoreceptor disappearance, interferes with the regenerative ability of the RGCs.

conclusions. The findings suggest for the first time that diminution of photoreceptor sensory input does not induce detectable death of RGCs until P180, but that it attenuates certain ganglion cell functions like intraretinal dendritic transport and propensity for axonal regeneration.

Our understanding of the mechanisms involved in processes of hereditary degeneration relies substantially on studies using animal models like rd mice ¹ and the Royal College of Surgeons (RCS) rat, ² which is characterized by an inability of retinal pigment epithelial cells to phagocytose photoreceptor outer segments. ^{3 4 5} Although not completely homologous to the human disease retinitis pigmentosa (RP), the deficient phagocytosis leads to insufficient clearance of photoreceptor outer segment debris produced during outer segment shedding. ^{3 6} Thus, photoreceptors degenerate during the first 3 months of life and are cleared away by intraretinal resident microglial cells and immigrating macrophages. ⁷ Initial signs of degeneration first become evident at the end of the second week of life (period of eye opening). Histologic sections show that dystrophy reduces the 10 to 12 rows of photoreceptor cell nuclei seen at the second week of life, to about a single row at the end of the second month; virtually no photoreceptor cells are detectable at around 90 days of life. ⁸ Because of photoreceptor loss, no photoreceptor response can be recorded in the electroretinogram beyond the second month of life, although the inner retinal response is preserved. ⁹ In accordance, no visual acuity can be measured at advanced stages of degeneration. ¹⁰  

A general question in the field of neurodegeneration has been whether death of a certain cell type or cell layer induces transneuronal alterations in the neighboring layers and cells. ¹¹ Grafstein and coworkers ¹² described ganglion cell axonal transport within the retina in rd mice, where the dystrophy becomes established before retinal maturation. Eisenfeld and coworkers ¹ reported no changes in ganglion cell size, synapse, or axonal transport in retinal ganglion cells (RGCs) of RCS rats, which are characterized by an onset degeneration after maturation of the retina at postnatal day (P) 14. The explanation of this discrepancy between the two animal models was that maturity of the tissue may protect against transneuronal changes. Retinal interneurons like horizontal cells undergo some changes when examined with antibodies to calcium-binding protein. ¹³ To study whether ganglion cells disappear, Villegas–Pérez and coworkers ¹⁴ reported that some perivascular areas were devoid of RGCs at 6 months of age and that the retinas showed axonal abnormalities. A loss of RGCs and axons in the optic fiber layer has also been described at very advanced stages of RP retinas. ^{15 16 17} Stone and coworkers ¹⁶ have suggested that RGC loss might be an effect of transneuronal degeneration. Aspects of intracellular deficit in function like dendritic transport of RGC and ability to regenerate after an injury have not been examined yet, although extensive studies have been performed to assess transneuronal changes after incomplete or complete deafferentation in various areas of the central nervous system (CNS). ¹¹ So far examined, transneuronal degeneration seems, first, to be a rare phenomenon within the CNS and, second, difficult to assess. 

The first goal of this study was to examine whether the intraretinal portion of ganglion cells remains functionally intact, using dendritic transport as the parameter, by injecting a carbocyanine fluorescent dye, which is known to have a high resolution for outlining the finest dendritic branches. ¹⁸ The dye can be retrogradely transported from the superior colliculus to the cell body of the RGCs and from there in an anterograde direction to the dendrites. Second, we were interested in determining whether the propensity of ganglion cells to regenerate their axons, both in vivo ¹⁹ and ex vivo, ²⁰ is influenced by the progression of photoreceptor loss. 

Methods

Determining Ganglion Cell Density with Retrograde Transport

Pigmented non-dystrophic, rdy^+/+ (control group) and congenital dystrophic rdy^−/+ RCS rats ² were bred in our own facilities and were initially obtained from the Institute of Genetics (Nijmegen, the Netherlands). The care and maintenance of the animals conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Three groups of animals at ages P30 (n = 6), P90 (n = 6), and P180 (n = 6) were used to retrogradely label and quantify RGCs. The fluorescent vital dye 4Di-10ASP[ N-4-4-(4-didecylaminostyryl]-N-methylpropidium iodide (No. D291; Molecular Probes, Eugene, OR) was used. ¹⁸ Briefly, the animals were anesthetized intraperitoneally with ketamine (50 mg/ml, 0.2 ml/100 g body weight; Bayer, Leverkusen, Germany). An opening over the superior colliculus (SC) was made in the skull and overlying cortical tissue aspirated, providing access to the left SC. Several (∼7 to 10) solid crystals (∼50 to 100 μm in diameter) of the fluorescent dye were first deposited using a sharp scalpel blade on the top of the pia mater directly over the SC and then inserted with the blade tip into the superficial layer where the retinocollicular terminals are located. This method allowed the entire surface of the SC to be covered with dye and ensured RGCs across the entire retinal surface were labeled. This technique labels most of the RGCs in the contralateral retina ²¹ and a few in the ipsilateral retina. The cavity resulting from aspiration of the cortex was filled with resorbable gelatin sponge (Gelfoam; Upjohn, Aarhus, Denmark), and after suturing the skin, the animals were returned to their cages. The animals were killed 8 or 14 days later, by which time the RGCs were retrogradely labeled. 

After the animals were euthanatized either with an overdose of chloral hydrate or with CO₂, both eyes were removed and the retinas dissected and spread on nitrocellulose filters (Sartorius, Göttingen, Germany) with the ganglion cell layer (GCL) facing upward. After immersion-fixation within 4% paraformaldehyde in 0.1 M phosphate-buffered saline for several hours (usually overnight), the wholemounts were embedded in Mowiol (Hoechst), placed on coverslips, and viewed through the fluorescein filter of a fluorescence microscope (Axiophot; Zeiss, Oberkochen, Germany), with maximum fluorescence emission at 563 nm for 4Di-10ASP. 

In the retina contralateral to the labeled SC, RGC densities were determined by counting 20 fields of 275 × 200 μm, distributed over all four retinal quadrants and over all the different eccentricities, to obtain the optimal average density of RGCs. The counts were averaged for each retina to determine cell density, and within each group the cell densities were averaged and compared across the different groups. Cell density of the different ages and groups were analyzed with the H-test of Kruskal and Wallis, whereas the results between the different groups were compared with the Student’s t-test or the Wilcoxon Mann–Whitney U test. Differences were considered significant at the 95% level of confidence or higher. 

In addition to the retrograde axonal transport from the SC to the cell bodies within the GCL, anterograde dendritic transport occurs within the intraretinal segment of ganglion cells. Thus, the goal of this part of the study was to label the cell bodies from the SC and to determine whether the ganglion cells in the RCS rat can transfer the fluorescent 4Di-10ASP from their cell bodies into the dendritic branches within the inner plexiform layer (IPL). For this, the aforementioned animal groups used for measuring ganglion cell densities within the contralateral retina were analyzed for counting completely labeled dendrites within their ipsilateral retina. Because the number of cells labeled in the ipsilateral retina was much fewer due to the pattern of central projection, ²¹ these cells seemed suitable and were selected for analysis of dendritic transport. A similar analysis was impossible within the contralateral retina because the high degree of overlapping among dendrites in this retina prevents the unequivocal identification of individual dendrites. 

Wholemounts of ipsilateral retinas were prepared as described above. Ipsilateral ganglion cells were quantified within the inferotemporal peripheral retinal crescent, which is known to form this projection. ²¹ In addition, the “dispersed” ipsilateral projection with fewer cells (∼150/retina) ¹⁸ distributed across the entire retina was analyzed. Cells with completely labeled dendrites could be easily identified in wholemounts and were counted at a magnification of 20×. Their number was equal to the total number of retrogradely labeled cell bodies of ipsilateral cells. The percentage of completely stained ganglion cells was compared between RCS rats and age-matched controls at P30, P90, and P180. The averaged counts between the different groups were compared using the Wilcoxon Mann–Whitney U test. 

Microsurgery at the Optic Nerve and Grafting

In this series of experiments the ability of RGCs to regenerate their axons after optic nerve cut was studied. Under anesthesia, the intraorbital segment of the nerve was exposed with the aid of a surgical microscope and cut. Grafting of an autologous piece of the sciatic nerve (SN) at the optic nerve stump was a modification of the original procedure of Vidal–Sanz and colleagues. ¹⁹ In all cases, the distal end of the graft ended blind, that is without connection to central relay structures. ²² Figure 1A shows how the SN piece was exposed and sutured at the ON stump. The experimental animals were divided into three groups (5–10 animals/group) belonging to the ages P30, P90, and P180. Three age-matched groups of the wild-type rdy^+/+ strain constituted the control groups. 

Staining Techniques from the Graft

To label RGCs whose axons had regenerated into the SN graft, the rats were anesthetized again 16 days after the grafting procedure, the time required by RGCs to regenerate their axons in the peripheral nerve graft. The SN graft was microsurgically exposed in its epidural segment and transected at a distance of approximately 12 to 13 mm behind the eye bulb, and solid crystals of 4Di-10ASP were deposited at the cross-sectional area of the nerve. ²² The labeled stump was then surrounded with absorbable sponge (Gelfoam; Upjohn) to prevent displacement of the dye crystal. After suturing the skin wound the animals were kept alive for another 5 days to allow for retrograde transport of the dye and labeling of regenerated ganglion cells. 

Transplanted eyes were enucleated, and the retinas were dissected and prepared for fluorescence microscopy as described above. RGC densities were determined as in the case of normal retinas, and photographic montages were made from representative retinas to document the cell distribution. Data obtained from dystrophic and nondystrophic groups were analyzed with the H-test of Kruskal and Wallis. Comparison between the different groups was done with the Student’s t-test or the Wilcoxon Mann–Whitney U test. In addition to quantification of cell densities, morphometric assessment and characterization of ganglion cells were undertaken to determine whether different types of RGCs contribute differentially to regrowth of axons. 

Organotypic Retinal Cultures

To assess whether advancing photoreceptor dystrophy (Fig. 1B) or just aging influences the propensity of RGC to regenerate, axonal regrowth was monitored in culture too (Figs. 1C 1D) . The procedure of retinal pretreatment was similar to that described previously. ²⁰ Briefly, the ONs were crushed in vivo to induce a conditioning lesion to the ganglion cells. In contrast to the original protocol of “blind crush,” crush of the ON was performed under visual control after exposing the meningeal sheath (open crush). Five days later the retinas were explanted on polylysine/laminin-coated plates (Petriperm; Hereaus, Rottenburg, Germany) at concentrations determined in previous studies. ²⁰ In the first two groups of RCS animals at ages P30 and P180 we compared the numbers of axons growing out after 1 and 2 days in culture. The same experiments were repeated with control animals of rdy rats of corresponding ages (P30, P180). The absolute numbers of axons per explant (Fig. 1D) were averaged for each retina. These means were averaged within each group to obtain values for intergroup statistical comparison. This was done by means of the Student’s t-test. Statistical significance was set at the 95% level of confidence. 

Results

Age-Related Axonal and Dendritic Transport in RGCs

Fluorescent RGCs were distributed uniformly across the retina contralateral to dye injection in both strains 8 days after labeling (Fig. 2A ). However, because of the high density of cells and massive overlapping of dendrites, individual dendrites could only be discerned within the retina ipsilateral to dye injection (Figs. 2B 2C) . In these cells the dye was transported along the axons in a retrograde manner to label first the cell bodies. Then the dye moved by anterograde transport into the dendrites, located in the IPL (Fig. 2C) . The fluorescent probe labeled by this transport all the RGCs located in the ipsilateral retina and enabled their quantification (Figs. 2C 3B) . 

In the control rats, an average of 2576 ± 140 RGCs/mm² cells could be labeled across the contralateral retina at P30 (Fig. 3A) . This density was similar between retinas analyzed 8 (n = 3) or 14 (n = 3) days after deposition of 4Di-10ASP, indicating that a transportation time of 8 days is sufficient to label the somata of all RGCs uniformly. At P90, the RGC density declined slightly (2110 ± 130 RGCs/mm²) and a further, nonsignificant, decline was observed at P180 (1956 ± 60 RGCs/mm²). Morphologically, the cells showed no differences at all ages investigated and between the two strains used throughout the study. Across the ipsilateral retina, ganglion cell bodies were consistently labeled within the temporoinferior periphery, which is known to form this projection in rodents. ²¹ Within this area the dendrites showed a high degree of overlapping as in the contralateral retina. Outside of this restricted region, an additional dispersed ipsilateral projection that consists of approximately 150 cells spread across the entire retina, ¹⁸ was also labeled. After 8 days of dye transport, almost identical numbers were labeled in the RCS and control rats at P30 (136 ± 17 in RCS, 142 ± 13 in rdy, n = 3 for each group). These numbers remained unchanged both with age (130 ± 12 in RCS and 140 ± 13 in rdy at P180), and with longer time of incubation, which was 14 days after injection of the dye into the SC. 

Dendritic transport occurs in the anterograde direction from the labeled ganglion cell bodies into the dendrites located within the IPL (Figs. 2B 2C) . This transport was evaluated within the retina ipsilateral to the dye injection, because of the sparse distribution of cells. But even within this population, most cells had either only their perikarya or the primary dendrites labeled. In both strains and throughout the stages analyzed, up to 20 of 130 to 140 cells per retina were considered to be completely filled. It appeared that in the control rdy⁺/⁺ rats 17 ± 2 ipsilaterally projecting RGCs had completely labeled dendrites (Fig. 3B) at P30. This picture was slightly changed at P90 (10 ± 2) and at P180 (13 ± 2) RGCs (Fig. 3B) . The same pattern was also obtained in RCS rats at P30 (17 ± 3). However, in older RCS animals of P90, significantly fewer RGCs had completely filled dendrites (2 ± 1, Fig. 3B ). Virtually no complete labeling was obtained at P180 (Fig. 3B) . Dendritic labeling occurred when the time allowed for transport of the dye was increased to 14 days (Fig. 3C) . As Figure 3C also shows, 13 ± 2 control RGCs and only 6 ± 2 RGCs in the RCS retina had labeled dendrites at P180 (Fig. 3C) . No additional cells were labeled when the time of survival after dye injection was greater than 14 days. These data indicate that, although the retrograde axonal transport from the SC to the cell bodies is comparable in speed between control and dystrophic rats, the anterograde intraretinal redistribution of the dye by dendritic transport becomes significantly attenuated as photoreceptor degeneration progresses in the RCS rats. 

Regeneration of Axons within the Grafts

Regeneration occurred in all animals grafted at all ages of observation. The numbers and the morphology of RGCs whose axons grew into a SN graft were determined with retrograde transport of the fluorescent dye 4Di-ASP, which completely fills the cells. ²² The total number of regenerating RGCs was determined from retinal wholemounts by counting all cells across the total retinal surface. Brightly fluorescent RGCs were distributed across the retinas of all animals in both strains and were not restricted to a certain area. Figure 4A illustrates the peripapillar area of the RCS retina at P30, indicating that an evenly distributed population of RGCs was retrogradely labeled from their regenerated axons within the SN graft. By examining these cells at higher magnifications, different types of RGCs could be identified and classified according to morphometric criteria. ²² As Figures 4B 4C and 4D show, the major classes of RGCs are represented among the regenerated cells, thus indicating that the different types of RGCs possess the ability for regeneration. 

Photoreceptor loss is directly proportional to increasing ages of RCS rats (Fig. 1B) . However, the ability of RGCs to regenerate their axons remained intact in all ages. As Figure 5 shows, labeled cells were found both in the central and peripheral retinas at P30 (Fig. 5B) , P90 (Fig. 5C) , and P180 (Fig. 5D) . Background fluorescence increased with age, because photoreceptor death leads to accumulation of lipofuscin, which is ingested by microglial cells and appears brownish (Fig. 6A ), in contrast to green fluorescence of the RGC (Fig. 6A) . Because of this autofluorescence, most of the fluorescent spots seen in Figures 5C (P90) and 5D (P180) represent microglial cells containing lipofuscin. RGCs of control rats retained the ability to regenerate at all ages, but the microglial cells lacked lipofuscin and, therefore, remained nonvisible with the fluorescence microscope. 

In the nondystrophic rats a total of 3844 ± 733 RGCs (n = 5) were retrogradely labeled from their axons, which were extended into the SN graft at P30. At the same age, an average of 2868 ± 516 RGCs (n = 5) was counted in the dystrophic RCS retina (Fig. 7) . This difference in the numbers of regenerating RGCs is not likely due to the fact that these two strains have a different RGC density, because their numbers are almost identical, as determined by retrograde labeling (Fig. 3A) . More likely, this difference may be caused by the nonganglionic retinal cells. Indeed, we observed numerous lipofuscin-filled microglia within the GCL (Fig. 8A ) and in the deeper layers (Fig. 8B) in the retinas of RCS rats at P30. Such microglial cells, which are not found in the control strain, may have a negative influence on the process of regeneration. At P30, the difference in the number of regenerating cells was not significant between the two strains at the 95% level of confidence. At P90, 3962 ± 594 cells were counted within the group of control rats (n = 5) as opposed to 2572 ± 494 cells in the RCS rats (n = 8, Fig. 7 ). The difference was highly significant (P < 0.01). At P180, 3964 ± 281 cells/retina (n = 5) were labeled in the control rats and only 1453 ± 439 cells/retina (n = 5) in the RCS rats. This difference was also highly significant (P < 0.01). These data indicate that progression of photoreceptor cell death negatively influences the propensity for axonal regeneration in RCS rats. 

Axonal Regeneration In Vitro

The in vivo grafting model of regeneration is best suited to test the response of RGCs to the positive environment of Schwann cells found in the graft. To examine the intrinsic ability of RGCs to regenerate axons on extracellular substrates, retinal strips were explanted in petri dishes previously coated with polylysine and laminin. This experiment also aimed to compare whether the aging retinas of dystrophic and nondystrophic rats show differences in their ability to regenerate axons in culture. At P30, the number of axons growing out from control and dystrophic tissue was different (Fig. 8C) . Although an average of 176 ± 8 axons grew out from control tissue (n = 40), only 52 ± 8 axons were counted in explants (n = 40) from RCS rat tissue (Fig. 8C) on the second day in culture. This difference was significant at the 99% level of confidence. Morphologic examination of growth cones and observation of movement of the axonal tip revealed that growth patterns were identical in both strains (data not shown). At P180, an identical number of axons grew out of retinal explants in both strains (Fig. 8C) . 

Discussion

In this study we used an animal model of photoreceptor dystrophy to assess possible transneuronal effects of this defect on RGCs. The principal new findings of the present study are as follows: Increasing photoreceptor degradation is associated with a decreased anterograde dendritic transport in RGCs; the regenerative ability of RGC axons through an autologous SN piece diminishes significantly in older RCS rats; and RGCs retained their intrinsic ability to extend axons in an in vitro environment, at all ages studied. These results suggest that some RGC functions are adversely influenced by the lack of sensory input from photoreceptors. Alternatively, RGCs may be abnormal in the RCS rat, although this remains to be verified by examining additional parameters. 

Loss of Photoreceptors and Dendritic Transport

The retina of the RCS rat is completely differentiated and mature before any signs of degeneration become evident at P19 to P20. ^{4 5} It was not possible to study younger animals because of maternal rejection after surgery. Hence, the earliest age studied was P30. As the retrograde transport from the SC showed, there was no difference between ages P30 and P180. This result is in accordance with a previous study, which showed that both the retrograde and the anterograde axonal transport of radioactive amino acids remain unaffected in ganglion cell axons until 515 days of life. ¹ The obvious discrepancy in the reduced axonal transport in rd mice ¹² may be simply explained by the earlier onset of photoreceptor dystrophy that, in contrast to the RCS rat, affects the rd mouse retina at immature stages. 

The dystrophy-dependent reduction of anterograde transport of the dye within the RGC dendrites is a new finding. This observation becomes possible because of the bidirectional tracer characteristic of 4Di-ASP in the retinocollicular projection. ¹⁸ As also documented in the present study, complete delineation of the finest branches within the IPL can be achieved. Although difficult to assess quantitatively, the transport of 4Di-ASP achieved in RCS rats 14 days after injection into the SC resulted in dendritic labeling that was still lower than that observed after 8 days in age-matched control rats. Within the RCS strain, the number of completely filled dendrites decreased as a function of age. Although the mechanisms of dendritic transport have not been investigated in detail, it is assumed that they are similar to those of axonal transport. The fact that dendritic transport is selectively reduced with ongoing dystrophy of photoreceptors points to either a reduction in metabolic activity of deafferented RGCs, or to the influence of the neighboring glial cells. Whereas glial cells of peripheral nerves like Schwann cells are conducive for axonal growth, ²³ Müller cells, ²⁴ macrophages, ²⁵ neuroglia, ²⁶ or microglial cells that become activated and migrate throughout the RCS retina ^{7 27} may adversely affect axonal regeneration. The numerous microglia cells observed in the axotomized retina of the present study are additional evidence for activation and proliferation of these cells within the RCS retina. A possible mechanism of the microglia-mediated influence may be the production of neurotoxic substances as has been analyzed in neuronal/microglial cocultures. ^{28 29} Roque and colleagues ³⁰ recently found that cultured microglial cells from the RCS rat exert neurotoxic influences on cocultured photoreceptors. On the other hand, Banerjee and Lund ³¹ observed that microglial cells play a role in the maintenance of photoreceptors in transplanted retinal tissue lacking retinal pigment epithelium. Among the neurotoxic products, nitric oxide may play a crucial role. ^{32 33} In addition, free radicals or even neurotrophins like nerve growth factor ³⁴ are produced by retinal microglial cells and are able to regulate neuronal cell survival. 

The reduction of dendritic transport is in accord with the observations of Valverde, ³⁵ who showed loss of dendritic spines in cortical cells after dark-rearing and, therefore, reduction of sensory input. In addition, Ruiz–Marcos and Valverde ³⁶ observed a reduction of basal branches of dendrites in the visual cortex of mice after enucleation. Biochemical or molecular changes within the dendrites have not been reported to date. However, dendrites have been reported to be vulnerable to various environmental changes such as aging, ³⁷ Alzheimer’s disease, ³⁸ and ganglioside storage disease. ³⁹ Changes in other retinal cells within the inner nuclear layer were not studied here, because this was previously analyzed using electron microscopy and quantitative measurements. ¹ In agreement with this study, RGCs show normal morphology and sizes. Also, the present finding of normal retrograde axonal transport is in agreement with previously published data ¹ showing that the axonal transport of radioactive substances from the eye to the SC is normal. 

Regeneration of Ganglion Cell Axons

The fact that transected axons within long intraspinal tracts ⁴⁰ and of RGCs can regenerate over appreciable distances has been well documented in different rat strains ^{19 22} and in hamsters. ⁴¹ Moreover, as reported in the albino Sprague–Dawley and pigmented rat strains, ²² various types of RGCs contribute proportionally to axonal regeneration. The aim of the present study was to examine whether RGC axons, which become disconnected from photoreceptor sensory input, sustain the intrinsic ability to regrow within an autologous SN graft and in organ culture. The reduction in the number of regenerating cells with increasing age can be attributed to external influences like microglial activation or even to the lack of functional input. The basic observation that axonal growth does not depend on afferent input is new and underlines the fact that the ability for growth of cut axons is an indigenous feature of adult neurons that can be supported by favored surrounds. 

Some outer retinal degenerative diseases (such as rod-cone dystrophies, juvenile macular degeneration, and others) are associated with alterations in the retinal fiber layer. ¹⁵ However, most of these diseases are established at immature retinal stages and correspond more to the transneuronal changes observed in the rd mouse ¹² than in the RCS rat. Stone and coworkers ¹⁶ investigated ganglion cell changes in 41 patients with different genetic forms of RP, a disease similar to but not identical to that seen in the RCS rat. ² Approximately 50% to 75% of the ganglion cells survived after death of photoreceptors in RP retinas. ¹⁶ The present data support the theory that ganglion cells in the RCS rat survive, and these data analyzed three major functional aspects, those of normal axonal transport, reduced dendritic transport, and, finally, their regenerative ability. The vessel-induced decrease of ganglion cells in the RCS rat ¹⁴ occurred at later stages of life than those analyzed in the present study and showed that prolonged deafferentation is accompanied by progressive changes in the inner retina. This result is in agreement with our data. The present findings are new and could be vital for attempts at restoring vision of diseased retinas by replacing the photoreceptors or stimulating with implanted devices. Such attempts have been undertaken recently. For example, complete loss of the photoreceptor cell layer can be delayed by transplantation of wild-type retinal pigment epithelial cells, ^{42 43} by intravitreal injection of basic fibroblast growth factor ⁴⁴ into the eye, or by lesioning the eye. ⁴⁵ Another promising approach is the grafting of cryopreserved embryonic retinal tissue. ⁴⁶ However, the cellular targets of neurotrophic factors and the interactions between grafted and host cells remain to be elucidated. Although the dendrites of RGCs undergo certain functional deficits, they do not degenerate at advanced stages of the disease, which could be a prerequisite for restoring vision in this model. 

In conclusion, we analyzed two basic features of RGCs in the mutant RCS rat, which becomes blind due to photoreceptor dystrophy. The disease reduces dendritic transport in RGCs at advanced stages of dystrophy and the ability to regenerate axons both in vivo and in vitro. These data may be of relevance in approaches to replace photoreceptors with transplanted tissue and recover visual function. 

 Supported by the Interdisciplinary Center for Clinical Research (IZKF, Project E3) and the Deutsche Forschungsgemeinschaft (Grants DFG Th 386/8-1 and 10-1).

 Submitted for publication November 12, 1999; revised February 2, 2000; accepted February 10, 2000.

 Commercial relationships policy: N.

 

Figure 1.

 

Diagrammatic representation of the experimental strategy to assess axonal growth. (A) Schematic drawing of the rat eye showing the intraocular compartments and cut ON. Depiction of the in vivo methodology of grafting a SN piece to facilitate transected axons to regenerate. (B) Schematic drawing of the retina of RCS rats between P30 and P180 indicating increasing degeneration. (C, D) Diagram of an ex vivo retina preparation (C) and photomicrograph after massive axonal growth at P30 (D). Scale bars (D and E), 100 μm. RPE, retinal pigment epithelium; PRL, photoreceptor layer; ONL, outer nuclear layer; OPL, outer plexiform layer; INL, inner nuclear layer; OFL, optic fiber layer; mg, microglia.

Figure 2.

 

Efficacy of retrograde staining of RGCs of RCS rats at P30, 8 days after injection of 4Di-10ASP into the SC. (A) RGCs in the retina contralateral to dye injection. (B, C) Small group of RGCs within the retina ipsilateral to the dye injection. The same cells are focused within the GCL (B) and the IPL (C), making visualization of individual dendrites (arrows) possible. Scale bar, 50 μm.

Figure 3.

 

(A) Quantification of RGCs after retrograde labeling in the retina contralateral to dye injection in RCS and rdy rats at P30, P90, and P180. Note a slight, but not significant (P > 0.1), decrease of the cell numbers in both strains. (B) Number of completely labeled ipsilateral RGCs in both strains 8 days after dye injection. Note the drastic decrease of completely labeled cell dendrites in the RCS rats (P < 0.001 between P30 and P90 rats and P90 to P180 rats). No ganglion cell dendrites were labeled in RCS rats after 8 days of transport at P180. (C) Retrograde transport of 14 days at P180 of RCS rats showed more labeled dendrites compared with 8 day transport (P < 0.01), but still fewer cells than in control rats of the same age (P < 0.05).

Figure 4.

 

(A) Montage of the central RCS retinal surface after grafting at P30 and labeling of regenerating RGCs with 4Di-10ASP. Each spot represents one cell. Note the uniform distribution in all quadrants. OD, optic disc. (B through D) Higher magnifications showing individual cells of type RI (B), RII (C), and RIII (D) according to a previous categorization. ²² Scale bars, (A) 100 μm; (B through D) 25 μm.

Figure 5.

 

Age-related regeneration of RGCs in the RCS rat. (A) Section of the normal RCS retina after retrograde labeling of RGCs from the SC at P90. (B) Regenerating RGCs at P30 shown at higher magnification (OD, optic disc). (C) At P90, there is still regeneration of RGCs (arrows), but increasing background fluorescence appeared either “cloud-shaped” (arrowheads) or particulate, in particular in the peripheral retina (right edge). (D) At P180 fewer RGCs regenerated (arrows), whereas the “cloudy” or particulate background of brownish fluorescence increased all over the retinal eccentricity. Discrimination of RGCs and microglial cells occurred on the basis of the wavelength of fluorescence (Fig. 6) . Scale bar, 100 μm.

Figure 6.

 

Color photographs of intersection of the same retina shown in Figure 5D . (A) Surrounding microglial cells (thin arrows) are located slightly deeper (IPL) and have brownish fluorescence due to ingested lipofuscin. (B) Same intersection shows a regenerating RGCs (arrow) with green-yellowish fluorescence. It is the same cell, which is out of focus in (A). Scale bar, 25μ m.

Figure 7.

 

Quantification of axonal regeneration in control rdy and RCS retinas in vivo. The total number of retrogradely filled cells is plotted versus the age of both strains. The numbers remain remarkably stable in controls, but show a significant decrease in the RCS retina. P < 0.05 between P30 and P90, and between P90 and P180; P < 0.01 between P30 and P180.

Figure 8.

 

(A) Retrogradely filled RGCs in the RCS retina at P30 seen in the wholemount. Individual lipofuscin-containing microglial cells (arrows) are scattered between the RGCs. (B) Same intersection focused in the deeper layers with massive appearance of microglial cells (arrows). Scale bar, 50 μm. (C) Regeneration of ganglion cell axons in culture. The histogram corroborates the number of axons measured after 1 and 2 days in culture and are plotted versus the age of the animals and the strains of rats. Regeneration occurs throughout examined ages, but was reduced only in RCS rats at P30 to become identical with control retinas at P180.

The authors thank Mechthild Langkamp–Flock and Ilka Romann for technical assistance, Evan Dreyer and Rita Naskar for linguistic advice, and Lambros Panagis for helpful comments on the manuscript.