Mapping Retinal Degeneration and Loss-of-Function in Rd-FTL Mice

Abstract

Purpose.: Retinitis pigmentosa (RP) is a blinding disease caused by the degeneration of photoreceptors. To further understand the process of degeneration in RP, the authors have traced activation patterns associated with rod and cone photoreceptor degeneration in a mouse model of RP

Methods.: The authors used a double-mutant mouse, Rd-FTL, which contains a natural mutation, rd1, affecting the rod photoreceptors and an axon-targeted β-galactosidase reporter system, which is under the regulation of the promoter of the c-fos gene. These mice allowed the authors to trace degeneration-related activity that corresponded to rod and cone death

Results.: The authors traced cell death-associated activation in both rods and cones, allowing them to accurately determine the time course of cone degeneration in these mice. In the analysis of downstream activation patterns in the inner retina, they found that amacrine and ganglion cells maintain their photopic light-related activation until at least the initiation of cone degeneration. These cell populations then show increased activity during the peak time of cone cell degeneration. The authors also examined light-regulated functional activation of a subclass of amacrine cells, the dopaminergic amacrine cells. These cells showed light-induced functional activation after rod photoreceptor death and until the period of cone photoreceptor death, suggesting that they can be regulated by cone photoreceptors alone

Conclusions.: These findings have helped to accurately trace the periods of photoreceptor degeneration in this model of RP and show that correct light-regulated inner retinal activation is maintained until the time of cone degeneration.

Retinitis pigmentosa (RP) is a blinding disorder characterized by loss of function and gradual degeneration of photoreceptors. Mutations in many genes that are expressed specifically in photoreceptors are a major cause of RP. In mouse models of retinal degeneration, where the initial defect lies in the rod photoreceptor, cone cell loss inevitably occurs and is followed by a myriad of reactive changes, including cell migration, neurite extension, glia cell proliferation, and even cell death. ¹ Consequently, therapeutic methods to maintain remaining cone photoreceptors and inner retinal integrity are needed. To develop such methods, it is important to better understand the scope and nature of retinal degenerative changes and the time course over which they develop. 

In one model, the Rd mouse, a mutation in the PDE6B gene (encoding the beta subunit of the enzyme cGMP-phosphodiesterase) ultimately leads to an elevated level of cyclic guanosine monophosphate (cGMP) in the photoreceptors, causing their death. ² Interestingly, this gene has also been found to be affected in some forms of human RP. ³ Degeneration of photoreceptors in the rd/rd mouse can be detected histologically from postnatal day (P) 10 by the pyknotic appearance of their nuclei. ⁴ Degeneration progresses very rapidly so that by P18, virtually all rod photoreceptors have disappeared. At this stage, it is thought that cone death has already begun. By P90 to P100, however, all cones are lost. ¹ Thus, by this age, all second-order neurons in the retina are in principal deafferented, and the mice become completely blind. 

Our overall aim was to characterize how photoreceptor loss affects individual rod- and cone-mediated inner retinal circuits. Because of the increasing numbers of transgenic mice available, the mouse retina is a valuable tool for understanding both retinal function and disease states. Although the mouse retina is rod dominant, the actual density of cone photoreceptors in the mouse retina is comparable to cone density in areas of the human retina outside the fovea. ⁵ Thus, although there are limits, studies of mouse retinal circuitry can provide insight into human retinal function in normal and disease states. 

We have generated transgenic mice in which an axon-targeted β-galactosidase (β-gal) reporter system is under the regulation of the promoter of the c-fos gene. ⁶ In these fos tau LacZ (FTL) transgenic mice, neurons that express c-fos express β-gal in their cell bodies, axons, and dendrites, permitting direct visualization of their projections. ⁶ In these mice, light stimulation results in the expression of the c-fos gene and thus of lacZ, which can be localized by immunohistochemistry or by using an enzymatic assay. It is simple to unequivocally identify activated cell types in the retina by double staining for β-gal and retinal markers. We have recently used these mice to look at light-activated pathways in the retina and brain. ⁷  

Under normal conditions c-fos is not expressed in photoreceptors; however, c-fos is expressed in many cells preceding the onset of programmed cell death. ⁸ This also appears to pertain to photoreceptors because c-fos is upregulated in mouse rods before their degeneration after light damage. ⁹ Similar observations have been observed in the Rd mouse model. ¹⁰ Nevertheless, this c-fos expression does not appear to be essential for the cell death of photoreceptors because retinal pathology has been described to be unchanged in Rd-c-fos ^{− / −} double mutants. ¹¹  

We crossed Rd transgenic mice with FTL mice to study the degenerative pathology and time course of retinal degeneration. The advantage of the double-mutant mice, Rd-FTL, is twofold. First, it allowed us to accurately follow the scope and time course of photoreceptor degeneration and loss. Although the time course of rod death in Rd mice is known in more detail, the consequences of rod degeneration and the precise onset of cone degeneration remain to be determined. Second, we were able to study the consequences of photoreceptor loss on downstream activation of circuits in the inner retina by comparing the light-induced activity pattern in the retinas of FTL single (control) and Rd-FTL double mutants through the different stages of degeneration. 

Methods

Mice

All procedures concerning animals were performed in accordance with the University of Melbourne Animal Experimentation Ethics Committee and with guidelines set by the National Health and Medical Research Council and adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. All mice were housed under standard conditions, with food and water provided ad libitum in a 12-hour light/12-hour dark cycle. Ambient light in the animal house was measured using a photometer and was found to be an average of 330 lux (measured 1 m from the floor). Mice exposed to this light regimen were termed light-exposed or photopic light-exposed mice. 

Rd-FTL double-mutant mice were obtained by breeding the heterozygous FTL transgenic mice ⁶ with the Rd1 homozygous mutant mice (obtained from the Jackson Laboratory, Bar Harbor, ME) and subsequent breeding of the F1 progeny of these mice with Rd1 homozygous mutant mice to generate mice heterozygous for the FTL transgene and homozygous for Rd1, termed Rd-FTL. All lines of mice are maintained at the Animal Research Facility of the Faculty of Medicine and Health Sciences, University of Melbourne, Victoria, Australia. These Rd-FTL mice are healthy and viable and display the combined phenotypes of their parental transgenic cohorts. Once generated, they were maintained as a single line via inbreeding. Control mice were heterozygote FTL mice bred on a C57/Bl6 strain background. 

Mice were collected at around noon every second day from P6 to P34. Mouse age was calculated with date of birth designated as P0. Mice were phenotyped by tail assay ¹² for FTL transgene expression. 

Tissue Collection and Fixation

Mice were either taken from their home cage (light exposed) or kept in darkness for at least 48 hours (dark adapted). By this we obtained four different experimental cohorts: a light-exposed Rd-FTL degenerative cohort, a dark-adapted Rd-FTL degenerative cohort, a light-exposed FTL control cohort, and a dark-adapted FTL control cohort. 

All mice between P6 to P20 were killed by cervical dislocation, after which they were decapitated and their eyes were dissected. A small incision was made into the eyecup, and the eye was fixed in 4% paraformaldehyde, 0.1 M phosphate buffer, pH 7.4 (PFA), for 10 minutes. After this, the anterior eyecup and lens were dissected, and the posterior eyecups containing the retinas were further fixed for 30 minutes in the same fixative. Mice older than P20 were deeply anesthetized (dark-adapted mice in darkness) with an intraperitoneal overdose of pentobarbital (Nembutal; Rhone-Merieux, Pinkenba, Australia); then they underwent saline perfusion and were administered PFA. After dissection of the eye, a small incision was made into the eyecup and was postfixed in fresh 4% PFA for 10 minutes. For retinal wholemount stainings, retinas were dissected; for stainings on retinal sections, retinas were left in the eyecup and immersed overnight in 20% sucrose. Next they were equilibrated in a 1:1 solution of 20% sucrose/OCT (Tissue Tek OCT; Sakura Finetek Inc., Torrance, CA) for 30 minutes, then in fresh OCT for 1 hour, before they were snap frozen in an isopentane/liquid nitrogen bath. Eyes were sectioned at 16 μm on a cryostat, and sections we collected onto gelatinized slides. The sections were then stained by immunohistochemistry. 

Immunohistochemistry

For β-gal immunohistochemistry, retinal cryostat sections, attached to gelatinized slides, were blocked in 10% CAS Block (Zymed, San Francisco, CA)/0.1%TritonX/PBS then incubated overnight at 4°C in primary rabbit anti-β-gal antisera (Cappel Laboratories, Cochranville, PA) diluted 1:10,000 in 10% CAS-Block in PBS. After washing in PBS, the sections were incubated for 2 hours in Alexa Fluor 594 goat anti-rabbit IgG (Molecular Probes, Eugene, OR) diluted 1:400 in PBS. After final rinsing stages, sections were coverslipped using fluorescent mounting media (DAKO, Carpinteria, CA). For double-labeling experiments, cryostat sections were incubated overnight in a mixture of both primary antibodies. Depending on the type of primary antisera coincubated, for β-gal immunoreactivity either rabbit anti-β-gal antisera (Cappel Laboratories) or mouse anti-β-gal antibody (1:40; Promega, New South Wales, Australia) was used. The following antibodies were used to recognize certain retinal cell types: mouse anti-tyrosine hydroxylase (TH; Chemicon, Temecula, CA) used at 1:2000; rabbit anti-neuron specific enolase (NSE; DAKO) used at 1:5000, and FITC-conjugated peanut agglutinin (PNA) used at 1:500 (Vector Laboratories, Burlingame, CA). Secondary antisera were goat anti-mouse IgG Alexa Fluor 488 diluted 1:400 and goat anti-rabbit IgG Alexa Fluor 594 diluted 1:800 (both Molecular Probes). 

Photographic Imaging

Fluorescence photographs were obtained using an cooled charge-coupled device (CCD) camera (Imagepoint; Photometrics LTD, Tucson, AZ) and imaging software (V for Windows; Digital Optics, Auckland, New Zealand). 

Cell Counts

Cell counts in the retinas of all experimental cohorts were performed in sections through the optic nerve. With the exception of the dopaminergic amacrine cell population, all cell counts were performed in central retinal segments, defined as the segment of the retina 665 μm in length on both sides of the optic nerve. Measurements of these segments were made using the digital measurement tool in image analysis software (AnalysisD; Soft Imaging System GmbH, Münster, Germany). Counts were made by an experimenter blind to the identity of the sample. Cell counts were then made from three consecutive sections of both eyes giving a total of six segments. The numbers from the six segments were then averaged. For most time points, retinas from two animals were counted; however, in some cases, retinas from a single animal were counted. For the dopaminergic amacrine cell population, the cells were counted through the entire length of the retina; for these cells, four mice were counted for each time point. 

PNA labels the outer segments of cones and cone pedicles in the OPL, but not the cell bodies. To minimize inconsistencies, we counted PNA-positive outer segments in the central retinas (as defined) under high power. We then repeated the same procedure for PNA-positive cone pedicles to compare these numbers to our outer segment counts. If the numbers differed significantly (±5%), we performed recounts with other sections. 

For statistical analysis, data sets were then analyzed using a two-factor analysis of variance (two-way ANOVA, for dopaminergic amacrine cell analysis), or Student's t-test (for analysis of amacrine and ganglion cell numbers). Significance was defined as P < 0.05 (i.e., confidence interval was set at 95%), and statistical analyses were performed (GraphPad Prism 5; GraphPad Software Inc., San Diego, CA). 

Results

FTL Expression during Early Retinal Degeneration

We mapped the expression pattern of the FTL transgene product, β-gal, in the retinas of Rd-FTL mice using immunohistochemistry and compared it with the pattern in control FTL mice. For this, vertical cryostat sections of retinas were stained with an antibody specific to β-gal. We looked at developmental stages before photoreceptor death, from P6 to P8, during rod death, from P10 to P18, and after rod death had ceased, from P19 to P42. 

P6 to P8

At P7, β-gal immunoreactivity was found in the inner retina of both FTL (Fig. 1A) and Rd-FTL mice (Fig. 1B). In both strains of mice, a few amacrine cells in the inner nuclear layer and numerous cell bodies in the ganglion cell layer (GCL) were strongly stained. In the inner plexiform layer (IPL), distinct strata were labeled that most likely represented the processes of β-gal-positive ganglion cells and amacrine cells. In contrast, no immunoreactivity could be detected in the outer retina, indicating that the photoreceptors and bipolar and horizontal cells were negative. This staining pattern remained the same from P6 until P8 in both Rd-FTL and FTL control mice (data not shown). Since the eyes at these ages are still closed, FTL activity must be independent of direct visual processes and thus may reflect activation because of developmental processes, such as waves of ganglion cell activity, synaptic pruning, or both. 

Figure 1.

 

β-Gal expression in retinas of FTL and Rd-FTL mice from P7 to P18. Vertical sections of retinas were stained for β-gal using fluorescence immunohistochemistry. (A, C, F, I) FTL mice. (B, D, E, G, H, J, K) Rd-FTL mice. (A, B) At P7, activity is confined mainly to the inner retina of both strains of mice. (C–E) At P10, there is strong activity in the inner retina, and β-gal staining is present in rods (arrows, D) in the ONL of Rd-FTL mice. (E) High-power view of a β-gal-positive rod. (F–H) At P15, there is very extensive β-gal staining in the ONL of the Rd-FTL mice, and Müller cells become activated (arrows, G). (H) High-power view of a β-gal-positive Müller cell. (I–K) By P18, there is somewhat decreased β-gal activity in the retinas of the Rd-FTL mice compared with FTL controls, though it is still significant. (J) Area in the central retina. (K) High-power view of peripheral retina containing β-gal-positive Müller cells. GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; ONL, outer nuclear layer. Scale bar (shown in A): 50 μm (A–D, F–K); scale bar in (E): 25 μm.

P9 to P18

At P9, several β-gal-positive rods were found in the outer nuclear layer (ONL) of the Rd-FTL mice, especially close to the optic disc in the central retina, which became more prominent at P10 (Fig. 1D; see Fig. 1E for high-power view). The staining pattern in the inner retina closely resembled the pattern at the earlier stages, though staining was stronger and denser. At P11, β-gal-positive rods also became visible in the peripheral retinas of Rd-FTL mice (data not shown), and at P15 there were many more present in the central retina (Fig. 1G). This is consistent with the central-peripheral gradient of rod degeneration described in Rd mice. ¹³ By P15, the ONL started to become thinner (Fig. 1G); by P18, the ONL in the central retina consisted of only a single row of photoreceptors (Fig. 1J; see also Fig. 3B). At this age, there was very little FTL expression in the ONL (Fig. 1J) and only few photoreceptors, mainly in the peripheral retina, were β-gal positive. Thus, by this age, most of the rod photoreceptors ceased to express β-gal most likely because they died. This is consistent with previous observations that rod death is complete in the central retina by P17. ¹³  

Further, Müller cells, which under normal conditions are FTL negative, became β-gal positive from P15 (Figs. 1G, H, K). This retinal glia cell is typically activated after degenerative changes in the retina ¹⁴  

P19 to P34

Beginning at P18 or P19, faint β-gal staining began to reappear in the cells in the ONL, but only in the central region of the retina. This β-gal staining became more prominent at P20, when a significant number of β-gal-positive cells were found close to the optic disc (Figs. 2B, F). This staining must have reflected expression in cones, given that all the rod cells had died by this time. To verify that FTL expression was in cones, we performed double-labeling immunohistochemistry for β-gal and neuron-specific enolase (NSE). In the development of the retina, NSE is initially expressed on both rods and cones (from P8 to P10), but after this it is only expressed on cones ¹⁰ (Fig. 3A). Double-labeling studies were performed in vertical cryostat sections of Rd-FTL retinas from P14 until P34. From P14 through to P17; the β-gal-positive cells in the ONL were not labeled with antibodies for NSE (data not shown), indicating that these cells were most likely rods. However, from P19, β-gal-positive cells were found in the ONL and were double labeled for NSE (data not shown), establishing that these were cone cells. Further, the parallel arrangement of NSE-labeled cones at this age was severely disrupted in the Rd-FTL mice (compare Figs. 3A and B). 

Figure 2.

 

β-Gal expression in retinas of FTL and Rd-FTL mice from P20 to P42. (A–E) Vertical sections of retina. (A) FTL control mice at P42. (B–E) Rd-FTL mice at ages indicated. There is no β-gal staining in the ONL in FTL retina, whereas in Rd-FTL retinas, extensive β-gal staining is seen in the ONL from P20 to P24. By P42, only isolated areas of the ONL have β-gal-positive cells in retinas from Rd-FTL mice. (F–H) Images of wholemount retina from Rd-FTL mice at ages indicated. At P20, there is some β-gal staining restricted to areas directly surrounding the optic disc (F). By P22, there is extensive β-gal staining in the central retina and some staining in some regions of the peripheral retina (G). By P28, β-gal staining is restricted to regions of the peripheral retina. OD, optic disc. Scale bar (shown in F): 25 μm (A–E); 250 μm (F); 500 μm (G, H).

Figure 3.

 

Cone photoreceptors and expression of β-gal in Rd-FTL retinas at P19 and P22. (A) At P19, NSE-labeled cones are present in parallel arrays in retinas from FTL control mice, (B) but their spatial organization is disrupted in Rd-FTL retinas. (C) At P22, extensive expression of β-gal was evident in the ONL of Rd-FTL mice. (D) Same section of Rd-FTL retina stained for NSE. (E) Merged images (C and D) with extensive colocalization of β-gal and NSE labeling. (F) Another retinal section of β-gal labeling in the ONL of Rd-FTL mice. (G) Same section stained for PNA. (H) Merged images (F and G) with some overlap of β-gal expression expressed mainly on the cell bodies and PNA labeling expressed on the cone outer segments. Scale bars, 10 μm.

The number of these NSE/β-gal double-labeled cells in the ONL increased strongly at P22 (Figs. 3C-E), indicating that extensive degeneration of cones was occurring at this time. Given that rod cell death was complete by this time, it also followed that the only cells left in the ONL were the cones. Thus, to determine the time course of cone degeneration in Rd-FTL retinas, it was only necessary to determine numbers of β-gal-positive cells in the ONL. The number and intensity of these β-gal-positive cells peaked at P22 to P26, with expression starting first in the central retina and then extending into the periphery over several days (Fig. 4). The number of β-gal-positive cells decreased thereafter, but at later ages (at least until P42) there were still patches of β-gal-positive cells distributed randomly over the retina that became less frequent with increasing age (Fig. 2E). 

Figure 4.

 

Time course of expression of β-gal on cones and loss of PNA-labeled cones in Rd-FTL retinas. Dotted line: number of β-gal-labeled cones per millimeter of ONL in retina. Solid line: number of PNA-labeled cone outer segments in the ONL in retinas of Rd-FTL mice expressed as a percentage of PNA-labeled cone outer segments at the same age in the ONL of FTL control mice.

In addition, β-gal staining appeared in the peripheral region of the retina, but at a later stage than seen in the central retina, and it could be observed at P22 to P23 in the peripheral retina (Fig. 2G) and was present until at least P28. At the same time, β-gal staining in the central retina was decreasing and was essentially gone from this area of the retina by P28 (Fig. 2H). 

To obtain another measure of the progression of cone cell survival and death in the ONL, we labeled retinas with fluorescein-labeled lectin peanut agglutinin (PNA), which specifically labels the outer segments of cones (Fig. 3G). Double labeling with β-gal showed some overlap of expression (Figs. 3F-H), though this was limited because β-gal predominantly labels cell bodies and PNA labels the outer segments. Nevertheless, it was simple to quantitate these outer segments in retinal sections (each of which corresponds to a single cone). Thus, in the central retina, these counts showed that there were essentially identical numbers of PNA-positive cones in the Rd-FTL mice compared with control FTL mice until P17 (Fig. 4). After this time, the number of PNA-labeled cones began to decrease compared with controls, indicating death of these cells began at this time. This was consistent with our data with β-gal expression (Fig. 4). Indeed, comparison of the curves of β-gal-expressing cells in the ONL with that for numbers of PNA-labeled cones shows a close correspondence; as the number of β-gal cells increased, there was a corresponding decrease in the number of PNA-labeled cones (Fig. 4). Both these curves indicated that cone cell degeneration and death began at P18 and proceeded for a considerable period after, but with most cell death over by P30. These data are largely consistent with cone survival counts ^13,15,16 that indicate that 60% of cones have died by P30. 

Changes in Activation in Inner Retina

We also examined β-gal expression in the inner retina to determine the consequences of rod and cone degeneration on downstream signaling processes within the retina. We concentrated on amacrine cells and cells in the GCL (representing ganglion cells and displaced amacrine cells) because they were easily visible, showed clear β-gal labeling, and were simple to count. Several of the β-gal-positive cells displayed an axon in the nerve fiber layer that was clearly labeled, suggesting that they represented ganglion cells, and our previous studies showed that there was significant β-gal expression in ganglion cells in FTL mice. ⁷ We were thus confident that we were analyzing expression in both amacrine cells and ganglion cells. Cell counts of both amacrine cells and cells in the GCL were taken every second day from P8 to P30 to produce a temporal profile. Figure 5A shows the graph for amacrine cells in Rd-FTL mice compared with FTL controls. In FTL control mice, there was a brief spike in numbers of β-gal-positive cells just before eye opening (P15), and then a constant number of cells showed β-gal expression until P30. In comparison, there was a clear increase in the number of β-gal-positive amacrine cells in Rd-FTL retinas between P16 and P23. However, at P24, the number of β-gal-positive amacrine cells decreased dramatically. 

Figure 5.

 

Time course of β-gal expression in cell populations in inner retinas from P8 to P30. Shown are numbers of amacrine cells (A) and cells in the ganglion cell layer (B) from retinas of FTL (dotted lines) and Rd-FTL mice (solid lines) exposed to photopic light conditions (gray) and after dark adaptation (black).

It was unknown whether this expression was related to normal light-induced functional activation in the retina or was associated only with degenerative processes in the Rd-FTL mice. To determine this, we dark adapted a cohort of FTL and Rd-FTL mice and then analyzed their retinas for β-gal expression. As expected, dark-adapted control FTL mice showed lower counts of β-gal-positive amacrine cells than light-exposed mice, though a number of cells still expressed β-gal immunoreactivity (Fig. 5A). Again, by comparison, the pattern observed in the dark-adapted Rd-FTL mice was initially similar to that seen in FTL control mice until P15. Thereafter, there was again a clear increase in the number of β-gal-positive cells until P23, followed by a sharp decrease to almost no β-gal-positive cells from P24 onward. These differences in β-gal-positive cell numbers between FTL and Rd-FTL were highly significant between P18 and P22 (P < 0.001 for light-exposed and dark-adapted conditions; statistical analysis was performed using cell counts pooled from P18-P22). 

Overall, these data indicate that the large increase in the number of β-gal-positive amacrine cells in the Rd-FTL retinas between P18 and P22 is independent of light and almost entirely associated with degenerative processes occurring in these retinal regions. This upregulation coincides precisely with the onset of cone degeneration. This conclusion is also supported by the finding that this upregulation of β-gal in the inner retina occurs only in the same retinal regions in which there is upregulation of β-gal in cones, as shown in Figures 6A and B. After P22, there is a dramatic reduction in the number of β-gal-positive amacrine cells, which coincides with peak upregulation of β-gal in the cones and, at the same time, with significant loss of this population. 

Figure 6.

 

Expression of β-gal in inner retinas of Rd-FTL mice during cone photoreceptor degeneration. (A) Low-power view of vertical section of retina from mouse labeled for β-gal at P22. Only areas with strong labeling in the ONL also show labeling in the inner retina (arrows). (B) Higher power view of an area of Rd-FTL retina with strong β-gal labeling in the ONL bordering an area with no β-gal labeling in the ONL. There is stronger β-gal activity within the inner retina where the ONL is labeled. A short distance from this area, the β-gal activity in the inner retina is greatly diminished. Scale bar (shown in B): 250 μm (A); 50 μm (B).

The expression pattern observed in the GCL from P8 to P30 was different from that seen in the amacrine cells, but there were clear similarities (Fig. 5B). In the GCL, the patterns of expression could most easily be described by subdividing the period into three subperiods: until eye opening (P15); between P15 and P23; and after P23. Thus, in the first period, until eye opening, similar numbers of cells in the GCL showed β-gal expression in both FTL and Rd-FTL mice and in both light-exposed conditions and after dark adaptation. This finding suggests that ganglion cell activity during this period is mainly independent of light stimulation and is not affected by rod degeneration. 

In the second period, until P23, the expression patterns diverged between the two mice strains and between light-exposed and dark-adapted conditions. In the light-exposed conditions, the FTL control mice showed fairly constant and high numbers of β-gal-positive cells in the GCL. In the Rd-FTL mice, the pattern was similar; however, from P18, the number of β-gal-positive cells decreased compared with that seen in the FTL mice. After dark adaptation, in the FTL control mice, there was a progressive decrease in the number of β-gal-positive cells in the GCL through the entire period. The pattern in the Rd-FTL mice was similar but, under these conditions, there was an increase in the numbers of β-gal-positive cells compared with FTL mice (P < 0.03; statistical analysis performed as described). These findings suggested that a number of changes occurred during this period. First, the ganglion cells in control mice were losing their light-independent activation over this period. Second, there was still considerable light-regulated activation of ganglion cells in the Rd-FTL mice during and after the period of rod degeneration. Third, there was some light-independent stimulation of the ganglion cells in the Rd-FTL mice, similar to that observed with the amacrine cells during this period. 

In the third period, after P23, the patterns of β-gal expression in the GCL became simpler. In the FTL mice, β-gal expression was strongly maintained in light-exposed mice, and little expression was observed after dark adaptation. In Rd-FTL mice, few cells in the GCL showed β-gal expression, either in light-exposed or dark-adapted conditions, which again was similar to what was seen with the amacrine cells. This is most simply explained by the very significant loss of cone photoreceptors by this time. 

β-Gal Expression in Dopaminergic Amacrine Cells

We also undertook a study of a subpopulation of amacrine cells, the dopaminergic amacrine cells. Dopamine is involved in adapting retinal circuitry for photopic light, ¹⁷ but how the dopamine amacrine cells are regulated is still a matter of debate. We have previously shown that these cells strongly express β-gal in the light-stimulated retinas of FTL control mice. ⁷ We therefore further investigated this cell type in the Rd-FTL mice to gain insight into how they might be regulated. We stained retinal cryostat sections of cohorts of light-exposed and dark-adapted Rd-FTL and FTL control mice at different ages for tyrosine hydroxylase to identify the dopaminergic amacrine cells as well as β-gal. We then quantified amacrine cells, which were double labeled for tyrosine hydroxylase and β-gal, and compared the numbers between light-exposed and dark-adapted conditions. 

At all ages investigated (P16-P26), after dark adaptation, few (if any) TH-positive amacrine cells were double labeled for β-gal (Figs. 7, 8). After light exposure in the FTL mice, there was clear β-gal upregulation in these cells at all ages (Figs. 7A, B, I, J; Fig. 8). However, in the Rd-FTL mice, we found a different pattern that was dependent on age (degeneration stage). At P16 to P18, almost all TH-positive amacrine cells also expressed β-gal. After this age, there was a progressive decrease in the percentage of double-labeled cells. At P20 to P22, approximately 40% of the dopaminergic amacrine cells were double labeled, but by P24, few of these cells were labeled for β-gal. These differences were highly significant (Fig. 8). These results indicate that dopaminergic amacrine cells of the retina in Rd-FTL mice are activated by light and maintain their activity until P18 to P20, even after rod photoreceptors have completely disappeared in the retina. After this, from P22 onward, they lose their ability to be activated by light, which occurs in parallel to the increasing degeneration of cones. 

Figure 7.

 

Light-activated expression of β-gal in dopaminergic amacrine cells of Rd-FTL mice before cone photoreceptor degeneration and after cone degeneration has occurred. (A, B) Vertical section of retina from FTL mice exposed to photopic light at P18. (A) β-gal labeling. (B) Labeling for tyrosine hydroxylase. (C, D) Analogous images from Rd-FTL mice at the same age. Arrow: dopaminergic amacrine cell also labeled with β-gal. (E, F) Images of a vertical section of retina from FTL mice dark adapted from P16 to P18. (E) β-gal labeling. (F) Labeling for tyrosine hydroxylase. (G, H) Analogous images from dark-adapted Rd-FTL mice at the same age. Arrow: dopaminergic amacrine cell not labeled with β-gal. (I–P) Same series as in (A–H) except at P24. Arrow: dopaminergic amacrine cell also labeled with β-gal (I, J), or not double labeled with β-gal (L, N, P). Scale bar (shown in B), 25 μm.

Figure 8.

 

Quantification of dopaminergic amacrine cells showing light-regulated β-gal expression in Rd-FTL mice. Graph shows mean ± SD percentage of dopaminergic amacrine cells that express β-gal in retinas of Rd-FTL under photopic light conditions (gray bars) or after dark adaptation (black bars) at postnatal ages before and during the process of cone degeneration, as indicated. Percentage of double-labeled cells in Rd-FTL retinas at P20 to 22 and at P24 to 26, are significantly different than FTL controls and Rd-FTL at P16 to 18 (**^,***P < 0.001).

Discussion

A major aim of this study was to trace cone photoreceptor degeneration in a mouse model of RP and to assess the overall effects of photoreceptor degeneration on inner retinal activation. We used a double-mutant mouse, Rd-FTL, which carries one gene for RP and FTL, a neuronal activity marker. In the FTL mice, β-gal expression reflects the expression of c-fos, which is activated by a number of distinct stimuli within the retina. We compared staining patterns at different stages and quantified cellular expression of the β-gal product of the FTL transgene in photoreceptors, amacrine cells, and ganglion cells. 

In our studies, we have identified four distinct conditions in which β-gal (and thus c-fos) is upregulated. First, we found in early postnatal ages that β-gal was expressed independently of light activation and was most likely associated with developmental processes in the retina. Second, β-gal expression in amacrine and ganglion cells in normal FTL mice is a good marker of light-regulated activation of these cells, as we have found previously. ⁷ For example, in the inner retina of Rd-FTL mice, we found that amacrine cells maintain their photopic light-related activation until at least P20, which is after all the rods have died. Third, we found that β-gal was a useful marker for dying photoreceptors; for the first time, it allowed the tracing of degeneration-related activity in cones. Fourth, β-gal expression is present in some cells in the retina as a consequence of the upstream degenerative process. Thus, the amacrine and ganglion cell populations respond to cone degeneration by upregulating β-gal for a short period, followed by dramatic downregulation to very low levels by P30, when a large proportion of cones had died. We also have obtained insight into the functional regulation of a subclass of amacrine cells, the dopaminergic amacrine cells. Overall, our findings have helped to accurately trace the period of cone degeneration in this model of RP and show that light-regulated inner retinal activation is maintained until the time of cone degeneration. 

Rod Degeneration and Death

We first analyzed β-gal expression at early postnatal periods, from before eye opening. We did not examine rod death in detail because previous studies have analyzed this process in the Rd mice. Nevertheless, our results show β-gal was first detected in rods from P9 and became very extensive until P15. By P18, little β-gal expression was left in the ONL, and indeed the entire ONL was reduced in thickness to a single layer of cells. These findings indicate that most rod degeneration and death occurred over the period P9 to P17. These findings are broadly in line with published findings for FOS quantification in the ONL from P8 to P16 by Rich et al., ¹⁰ which suggested negligible staining after P15 after a peak at P12. 

Cone Degeneration and Death

We detected large numbers of cone photoreceptors expressing β-gal at times after all rods had died after P18. Double labeling with NSE established that these β-gal-positive cells were cones. We found some double labeling of NSE and FTL at P20 and extensive double labeling at P22, substantiating our argument that the degeneration of cones begins by P20 and becomes significant at P22. These times have not been examined for FOS expression, ¹⁰ and markers of cell death such as TUNEL have also not been detected from P21 in the Rd mouse model. ^11,18  

We also carried out cell counting with PNA to ascertain what proportion of cones was present through this whole period. Our graph of PNA cone cell loss suggests that cone death starts as early as P18 and that there is a large amount of cone death between this age and P25. After this period, the number of β-gal-positive cells decreases, and there is also a decrease in the rate of loss of PNA-positive cones. Approximately 40% of the cones still remain. This is consistent with other studies of cone survival ^16,19 that indicate that at P30, approximately 60% of cones have degenerated. The rate of cone degeneration seems to slow so that approximately 1.5% of cones persist at P600. ¹⁵  

Effect of Photoreceptor Degeneration on Inner Retinal Activation

Clearly, the loss of photoreceptors should have profound effects on inner retinal function, but it is unclear what these changes are and over what time the changes occur. Our data and subsequent statistical analyses both indicate photoreceptor degeneration affects FTL expression in amacrine cells and ganglion cells at specific times. 

Both amacrine and ganglion cells respond to heightened photoreceptor death by upregulating β-gal, but only during the period of cone death. The fact that this occurs for both light-exposed and dark-adapted circuits indicates a process independent of light stimuli. This raises the possibility of cone photoreceptor death triggering increased activity in these populations of cells. The context within which this upregulation occurs should also be examined. In this age group, major second-order neuronal alterations are starting to take place. ^20–22 Even though no morphologic or functional changes have been reported in amacrine cells in this age group, our data seem to indicate a response of amacrine cells to either the start of this restructuring or a loss of photoreceptors. It is conceivable that the loss of rods and the near-complete loss of cones could potentially unmask intrinsic expressions or other extrinsic triggers for β-gal that have not been discovered. Alternatively, it is possible that the degenerating photoreceptors release neuroactive substances such as glutamate or other, yet undiscovered, factors to which amacrine cells respond with β-gal upregulation. Nevertheless, without more investigation, it cannot be determined what has caused the increased β-gal expression within the inner retina. It is still possible that this expression reflects the very early stages of degeneration of the amacrine and ganglion cells. 

Others ^23–25 have reported functional abnormalities in Rd mice from early degenerational stages that point to dysfunction of bipolar cells. We do see β-gal expression in bipolar cells in FTL mice ⁷ (and see Fig. 2A), but not at the relatively early stages examined here; thus, β-gal expression is not useful in examining bipolar cell activation in this study. 

Effect of Degeneration on Light-Regulated β-Gal Expression

Our analysis on light-stimulated β-gal expression in amacrine cells showed no change in response to photoreceptor death through the period of rod cell death. However, we did not conduct any studies under scotopic light, where the rods are active. What these results do show is that complete loss of rods had little apparent effect on gross signaling of cones to amacrine cells, at least as measured using the c-fos-related expression of β-gal in the FTL mice. Recent studies show that neurochemical remodeling takes place in the inner retina of Rd mice at these early ages (in the form of misexpression of glutamate receptors), which should result in dysfunction. ²³ However, there appears to be some decrease in light-related β-gal expression from around P22, which is in the middle of cone degeneration. The light-related β-gal expression decreases further and, by P30, there appears to be little light response left in the amacrine cells. This is consistent with the loss of cone photoreceptors, leading to loss of light-regulated activation in the amacrine cell population. 

We also found light-stimulated β-gal expression in ganglion cells decreased markedly from P24 in the Rd-FTL mice. These results indicate that ganglion cells lose light-stimulated function as a result of cone photoreceptor degeneration. As stated, there was no effect during rod degeneration, but we did not assess the effects of scotopic light, when rods are active. The finding that alterations in ganglion cell function occur as a result of photoreceptor death in early stages of degeneration agrees with similar studies conducted in rats. ²⁶ Furthermore, electrophysiological studies report decreased response amplitudes in retinal ganglion cells in Rd mice after P14, ²⁷ though function continued for many weeks. 

β-Gal expression in the inner retina of the Rd-FTL mouse reverts to baseline levels from P24 on, when approximately 40% of the cones remain. It is possible that β-gal expression in the FTL mice is relatively insensitive and does not reflect all activity in the inner retina. After all, small and delayed light responses can be recorded from rd mice after this age. ²⁰ It is also possible that the remaining cones are defective by this age. Certainly our findings show that the entire spatial organization of the cones is disrupted from as early as P19 and becomes progressively worse after this time. It is thus possible that signaling from the cones also becomes more and more disrupted over this period. 

Our studies with dopaminergic amacrine cells suggest that they maintain light-regulated activation until cone cell degeneration and death. Three different regulatory pathways have been postulated for these cells: a photopic input initiated in the cone photoreceptors and mediated by ON bipolar cells; an inhibitory input via GABAergic and glycinergic, which are in turn regulated by cone OFF and rod ON bipolar cells; and a photopic input originated in the melanopsin class of photosensitive ganglion cells. ^17,28,29 In a recent study, using dopamine measurements and c-fos expression, Cameron et al. ²⁸ studied mice lacking melanopsin, rod phototransduction, or rods and cones. They found that expression in the dopaminergic amacrine cells was decreased in the rod mutants and absent in the mutants lacking both rods and cones and concluded that light regulation of dopamine was dependent mainly on rods and cones. Overall, our results are in agreement with these findings; however, we see very little loss of β-gal activation in the Rd-FTL mice after rod loss and before cone degeneration. Thus, our results are more consistent with the regulation of this class of amacrine cells occurring predominantly by cone pathways. Measurements of dopamine release during rod and cone degeneration in Rd mice would further help to clarify this issue. 

Given that our experiments were conducted under photopic light conditions, we could not rule out the possibility that dopaminergic amacrine cells are regulated by rods as well as cones. We have recently undertaken experiments with FTL mice under scotopic light to examine this possibility. Our preliminary data suggest that there is no β-gal expression in the dopaminergic amacrine cells under these conditions and, thus, that rod activation does not regulate the activation of these cells. 

Several experimental treatments for photoreceptor degeneration, such as stem cell technology, ³⁰ tissue transplantation, ³¹ retinal prosthetic devices, ^32–36 administration of growth factors, ^37,38 and gene therapy ^39,40 are nearing or are in clinical trials in humans. These therapies, designed to replace or treat diseased photoreceptors, rest on the assumption that inner retinal layers are healthy and would function normally if photoreceptor activity were restored or replaced. Our studies provide evidence that inner retinal function is maintained in this model of RP, at least until the stage of cone photoreceptor degeneration. We thus suggest that such treatments, as described, would not be compromised by loss of inner retinal function if given early enough. 

Footnotes

 Supported by a Charitable Trust grant from ANZ Trustees (Melbourne, Australia).

Footnotes

 Disclosure: U. Greferath, None; H.C. Goh, None; P.Y. Chua, None; E. Åstrand, None; E.E. O'Brien, None; E.L. Fletcher, None; M. Murphy, None

Footnotes

 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.