Expression of Neurturin, Glial Cell Line–Derived Neurotrophic Factor, and Their Receptor Components in Light-Induced Retinal Degeneration

Abstract

purpose. Dysregulation of neurturin (NTN) expression has been linked to photoreceptor apoptosis in a mouse model of inherited retinal degeneration. To investigate the extent to which any such dysregulation depends on the nature of the apoptotic trigger, the expression of NTN, glial cell line-derived neurotrophic factor (GDNF), and their corresponding receptor components were compared in a rat model of light-induced retinal degeneration.

methods. Retinal expression of NTN, GDNF, their corresponding receptors GFRα-2 and -1, the transmembrane receptor tyrosine kinase (Ret), and cSrc-p60, a member of the cytoplasmic protein-tyrosine kinases family, were analyzed by Western blot analysis and immunocytochemistry in cyclic light- and dark-reared rats in the presence and absence of intense light exposure.

results. All components for NTN-mediated signaling activation are present in rat photoreceptors and retinal pigment epithelium, the cells primarily affected by light-induced damage. The expression levels of GDNF, its receptor components, and NTN, were not affected by light-induced stress. However, GFRα-2 expression strikingly increased with the extent of retinal damage, especially at the photoreceptors, in contrast to decreased levels that were observed previously in an inherited degeneration model.

conclusions. The present study indicates that the expression of receptors of the GDNF family is independently regulated in normal and light-damaged rat retina, and in conjunction with previous work, suggests that the pattern of modulation of these genes during photoreceptor degeneration is determined by the nature of the apoptotic trigger. Such differential responses to different modes of retinal degeneration may reflect influences of the neurotrophic system on photoreceptor survival or in the regulation of neuronal plasticity.

GDNF and neurturin (NTN) are members of the glial cell line–derived neurotrophic factor (GDNF) family ligands (GFL) of neurotrophic factors. GFLs have been shown to influence the development of enteric, sympathetic, parasympathetic, and sensory neurons (for review see Ref. ¹ ). They generally signal through a multicomponent receptor system consisting of the receptor tyrosine kinase Ret and a high-affinity ligand binding glycosyl-phosphatidylinositol (GPI)–linked coreceptor (GFRα). GDNF-mediated bioactivity involves signaling molecules of the src-family of protein-tyrosine kinases; and, in particular, p60 Src has been shown to interact with activated Ret. ² GDNF and NTN are expressed in a wide variety of tissues including the retina, suggesting an implication in diverse biological processes. ^{3 4 5 6 7} In our earlier report, we showed altered expression of NTN and its GFRα-2 receptor component in the rd mouse model of retinal degeneration, suggesting a link between dysregulation of NTN neurotrophic function and apoptotic photoreceptor cell death. ⁶ Upregulation of NTN mRNA expression was associated with progressive retinal neurodegeneration, but GFRα-2 mRNA levels remained lower than in age-matched nondegenerative control retinas. On the assumption that increased NTN expression is a survival-promoting response of the retina to the onset of degeneration, its potential neurotrophic effect on photoreceptors might be constrained by the persistently low GFRα-2 levels in rd retinas. Alternatively, because NTN also signals through the GDNF receptor (GFRα-1) but through a low-affinity interaction, ¹ it is possible that increased NTN is limited in its efficacy by failure to activate sufficient survival-promoting pathways through the GFRα-1 receptors. 

To assess the extent to which such modulations of expression of GFL members and their receptors are dependent on the nature of the apoptotic trigger, we have compared expression patterns of NTN, GDNF, and their receptor components in a model of photoreceptor cell death induced by exposure to intense light. In rats, light-induced retinal damage is rhodopsin-mediated and dependent on light intensity, wave length and duration of the exposure, period of dark adaptation before exposure, and the exposure schedule. ^{8 9 10 11 12} The effects were studied of both the type I (damaging both the photoreceptors and the retinal pigment epithelium) and type II (characterized by the loss of visual cells only) light-induced damage regimens on the expression of two members of the GDNF family. The retinal distributions of NTN, GDNF, and their receptor components were assessed by immunoblot and immunocytochemistry in control and light-stressed rat retinas. 

Methods

Animals

Weaning male albino Sprague-Dawley rats were obtained from Harlan Inc. (Indianapolis, IN) and kept either in a weak cyclic light environment (1200–1400 lux, 12 h/d) or in darkness for 40 days. At 60 days of age, rats were adapted to the dark overnight and then exposed to intense visible light for up to 24 hours. Light exposures were started at 9 AM and performed in green Plexiglas chambers (Dayton Plastics, Dayton, OH) transmitting 490- to 580-nm light (green light). The illuminance was 1500 to 1750 lux (∼800 μW/cm²). The type of light-induced damage reflects the light history, in which type II occurs in cyclic-light–reared and type I in dark-reared animals. Rats were killed in carbon-dioxide–saturated chambers and the eyes enucleated. After light exposure, some animals were allowed to remain in the dark for 24 hours before killing. For both cyclic light- and dark-reared regimens, control animals were those not exposed to light-induced damage (i.e., 0 hours of light treatment). All animal procedures were performed in compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the Wright State University Laboratory Animal Care and Use Committee. 

Immunohistochemical Analysis

Enucleated eyes were fixed in 4% paraformaldehyde, followed by embedding, freezing, and cryosectioning as previously described. ¹³ Polyclonal antibodies were obtained from Santa Cruz Biotechnology (Santa Cruz, CA) except where stated otherwise. Rabbit anti-human GDNF antibody (2 μg/mL), goat anti-rat GFRα-1 antibody (4 μg/mL), goat anti-human GFRα-2 antibody (4 μg/mL), goat anti-human neurturin (4 μg/mL), rabbit anti-human Ret antibody (4 μg/mL), and rabbit anti-human Src p60 antibody (2 μg/mL), were used on retinal frozen sections (10 μm). Details of the specificities and previous applications of these antibodies are available through the manufacturer’s Web site (www.scbt.com; Santa Cruz Biotechnology). The antibodies were localized using appropriate goat anti-rabbit IgG or rabbit anti-goat IgG conjugated to fluorescein (Sigma-Aldrich Co. Ltd, Poole, UK) at concentrations in accordance with the manufacturer’s recommendations. 

Western Blot Analysis

Retinas were homogenized in sodium dodecyl sulfate (SDS) sample buffer. ¹⁴ After centrifugation (5 minutes, ×200g), the supernatants were used for Western blot analysis. Protein extracts of paired retinas from three animals were separately analyzed in duplicate as a minimum. Proteins (80–90 μg/well) were resolved by 10% or 12% SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene difluoride membranes (Immobilon-P; Millipore, Bedford, MA). Subsequent stripping and detection of actin assessed the transfer. The reagents and polyclonal antibodies used were from Santa Cruz Biotechnology: goat anti-human actin antibody (0.2 μg/mL), rabbit anti-human GDNF antibody (2 μg/mL), goat anti-rat GFRα-1 antibody (1 μg/mL), goat anti-human GFRα-2 antibody (4 μg/mL), goat anti-human neurturin (4 μg/mL), rabbit anti-human Ret antibody (1 μg/mL), and rabbit anti-human Src p60 antibody (1 μg/mL). The immunoreactive proteins were localized with horseradish-peroxidase–linked donkey anti-goat antibody or goat anti-rabbit antibody and enhanced chemiluminescence system (Santa Cruz Biotechnology). The blots were exposed to autoradiograph film (X-omat; Eastman Kodak Company, Rochester, NY). Band intensities were analyzed using a laser densitometer (LKB, Bromma, Sweden) on nonsaturated exposures. Because the optical density (OD) readings were taken in arbitrary units, direct blot-to-blot comparisons were not valid. Therefore, to enable comparisons between different sets of experiments, data were compared by using relative ratios (optic density [OD]/OD_max for a given blot and detected protein) followed by normalization to actin to control for protein loading. For a given data series, incorporating both cyclic light- and dark-reared animals, these normalized values were presented graphically with the maximum ratio arbitrarily set to 1. Analysis was performed using a computer statistical package (Prism; GraphPad, San Diego, CA). 

Results

The expression of NTN, GDNF, and their receptor components was studied by immunoblot and immunocytochemistry analyses of retinas from rats reared either in dim cyclic light or in the dark, with and without exposure to intense visible light. Changes in retinal morphology after light-induced damage, especially at the photoreceptor and retinal pigment epithelium (RPE) cells level, were in accord with those previously described. ⁸ Loss of visual cells was observed in both cyclic light- (type II damage) and dark-reared rats. ^{15 16} In addition, concomitant loss of RPE cells was detected in dark-reared animals (type I damage). ^{15 16}  

NTN Expression

In cyclic-light– and dark-reared rats, no significant difference in the NTN levels was observed by immunoblot analysis between control and light-exposed animals (Fig. 1) . NTN immunoreactivity was detected mainly at the photoreceptor outer and inner segments, the RPE, and very slightly in the inner part of the retina, in both cyclic light– and dark-reared animals (Figs. 2A 2G 2M 2S 3A 3G 3M 3S) . Photoreceptor inner and outer segments and RPE appeared more intensely immunostained after 24 hours of light exposure, with or without a subsequent 24-hour dark period (Fig. 2M 2S) . A weak nonspecific fluorescence of the RPE and choroid was observed in sections without primary antibody (Fig. 4) . 

GFRα-2 Expression

In cyclic-light–reared rats, GFRα-2 expression levels were significantly higher (4.5 times, P < 0.003) after 24 hours of exposure to light, and greatly increased (21 times, P < 0.0001) when animals were kept a further 24 hours in the dark after light treatment (Fig. 5) . 

In dark-reared animals, higher levels of retinal GFRα-2 were also observed in rats exposed to 24 hours of light-induced damage in comparison with unexposed control animals (∼3 times, P < 0.04) and in rats kept 24 hours after light-induced damage (∼3 times, P < 0.03; Fig. 5 ). Moreover, GFRα-2 levels were approximately 20% higher in cyclic-light–reared animals kept 24 hours in the dark after 24 hours of light exposure, compared with their dark-reared counterparts. Protein expression was localized by immunostaining on frozen retinal sections (Figs. 2B 2H 2N 2T 3B 3H 3N 3T) . GFRα-2-immunoreactivity was concentrated at the photoreceptor outer and inner segments and the RPE in both cyclic-light– and dark-reared animals. The signal intensity was strikingly increased at the photoreceptor segments when the rats were kept for a further 24 hours in the dark after light exposure (Figs. 2T 3T) compared with control subjects (Figs. 2B 3B , respectively). In addition, the photoreceptor cell bodies were immunopositive in dark-reared animals subjected to this regimen (Fig. 3T) . 

Ret Expression

Western blot analysis of Ret expression showed no significant difference between cyclic-light– and dark-reared animals. The levels were essentially similar in control and light-exposed rats (Fig. 6) . Immunocytochemistry analysis showed Ret localized mostly to the photoreceptor outer and inner segments, the RPE, the inner nuclear layer, the inner and outer plexiform layers, and, to a lesser extent, the ganglion cell layer, in both cyclic-light– and dark-reared animals (Figs. 2C 2I 2O 2U 3C 3I 3O 3U) . After 24 hours of light-induced damage followed by a 24-hour dark period, the photoreceptor immunostaining was more pronounced in cyclic-light– than in dark-reared animals (Figs. 2U 3U) . 

Src p60 Expression

Evaluation of cSrc-p60 expression by immunoblot analysis showed that levels were approximately four times higher (P = 0.0001) in cyclic-light– than in dark-reared control rats (Fig. 7) . In cyclic-light–reared rats, protein expression was decreased by approximately 50% after 12 hours of light treatment (P < 0.05) and by approximately 80% after 24 hours of exposure, whether followed or not by a 24-hour dark period (P < 0.0005, P = 0.0001 respectively; Fig. 7 ). No significant differences were observed after light exposure in dark-reared animals (Fig. 7) . Immunocytochemistry analysis showed that cSrc-p60 was localized to the photoreceptor segments, the RPE, the inner and outer plexiform layers, the ganglion cell layer, and, to a lesser extent, the inner nuclear layer in control cyclic-light–reared animals (Fig. 2D) . After light exposure, the immunoreactivity signal was reduced in the cyclic-light–reared rats’ retinas, in particular at the photoreceptor segments level, but the ganglion cell layer remained immunostained (Figs. 2J 2P 2V) . In dark-reared rats the immunolocalization was comparable (Figs. 3D 3J 3P 3V) , but the ganglion cell layer staining disappeared in animals kept in the dark after light-induced damage (Fig. 3V) . Immunoreactivity at the photoreceptor segment level was observed mainly after 24 hours of light-induced damage followed by a 24-hour dark period (Fig. 3V) . 

GFRα-1 Expression

Expression of GDNF receptor GFRα-1 was examined by immunoblot analysis and by immunocytochemistry in control and light-exposed animals. No significant difference was observed between cyclic-light– and dark-reared animals (Fig. 8) . Immunostaining was mainly localized to the inner nuclear layer, inner plexiform layer, and ganglion cell layer (Figs. 2E 2K 2Q 2W 3E 3K 3Q 3W) . The photoreceptor outer and inner segments and RPE were immunopositive in cyclic-light–reared animals after 24-hours of light exposure, with or without a subsequent dark period (Figs. 2Q 2W) . The localization was similar in dark-reared rats (Figs. 3K 3Q 3W) . 

GDNF Expression

The levels of GDNF were essentially similar in control and in light exposed cyclic-light– and dark-reared animals (Fig. 9) . GDNF was localized mainly to the outer and inner plexiform layers (Figs. 2F 2L 2R 2X 3F 3L 3R 3X) . Close examination by light microscopy revealed that immunoreactivity was associated with the RPE in cyclic-light–reared animals (Figs. 2F 2L 2R 2X) and in dark-reared rats (Figs. 3F 3L 3R) . 

Discussion

In our earlier study, we observed a discrepant expression between NTN and its GFRα-2 receptor component in the rd mouse model of retinal degeneration. ⁶ Upregulation of NTN mRNA expression suggested that NTN could have a survival function modulated by stress before the onset of photoreceptor apoptosis in this model, ^{17 18 19 20} but its function may have been constrained by downregulation of the GFRα-2 receptor. To investigate the extent to which such modulations depend on the nature of the apoptotic trigger, we analyzed the expression of NTN and its receptor components in a rat model of light-induced retinal degeneration. Exposure to intense light triggers photoreceptor cell death, ^{21 22} and light-induced damage can be classified based on the primary and secondary manifestation of the damage. ^{15 16} Type I is characterized by damage to both the retina and the retinal pigment epithelium (RPE), ¹⁵ whereas visual cells are primarily affected with little or no damage to the RPE in type II. ^{15 16} Despite the different types of retinal damage known to occur in rats and the variables that influence the extent of damage, a hallmark of light-induced damage is the morphologic preservation of the inner retinal neurons. ⁸  

NTN and GFRα-2 are coexpressed in the outer and inner segments of the photoreceptors and RPE of control rat retinas, the target cells of both light-induced damage types. This localization correlates with the mRNA expression detected at the photoreceptor outer segments in control adult mouse retina, suggesting that neurturin plays a physiological role in these photosensitive cells in both species. ⁶ Similarly, the transmembrane receptor tyrosine kinase Ret, which forms, with GPI, the functional GFL receptor complex, ^{23 24 25 26 27} is localized to the RPE and photoreceptor outer segments, as previously reported in adult mouse retinas. ⁶ Ret was found to be colocalized with cSrc-p60, a member of the cytoplasmic protein-tyrosine kinase family that interacts with activated Ret. cSrc-p60 activity is necessary for differentiation and survival events elicited by GDNF family ligands. ² Taken together, these data indicate that the components for NTN-mediated signaling activation were detected in the photoreceptor and RPE cells, and that NTN could most probably act in an autocrine or paracrine manner, as previously suggested for the central nervous system, ²⁸ or similarly for GDNF in the retina. ⁷  

Light-induced damage did not affect retinal NTN levels, but GFRα-2 expression was strikingly increased. Similarly discrepant changes in NTN and GFRα-2 expression have been reported after brain insults in the rat, and it has been postulated that GFRα-2 may associate with other signaling receptors or ligands. ²⁸ The fact that lower levels of expression of cSrc-p60 were detected after light-induced damage in the cyclic-light–reared animals is consistent with this hypothesis. Alternatively, increased GFRα-2 expression could reflect a neurotrophic activity induced by GDNF, as the latter has been shown to improve outer segment survival times of isolated photoreceptors in vitro. ²⁹ Both GFRα-1 and -2 have been shown to mediate responses to GDNF, even if GFRα-1 is more efficient than GFRα-2. ^{27 30 31} However, no variations in GDNF or GFRα-1 levels were observed after light exposure suggesting that GDNF family receptor components are regulated independent of each other. At most, the partial redistribution of the GFRα-1 immunostaining to the photoreceptor inner and outer segments and RPE could reflect a requirement for GDNF- or NTN-mediated trophic effect in these cells after light-induced damage. 

Intense light exposure can trigger photoreceptor apoptosis, ^{6 21} and the concomitant upregulation of GFRα-2 and increased expression of NTN at the photoreceptor outer segments observed in this model together raise the question as to whether NTN may have a direct or indirect involvement in counteracting the apoptotic processes in these dying neurons. Indeed, NTN has been shown to enhance the survival of axotomized retinal ganglion cells and has been suggested to have both a direct and indirect neuroprotective role on these retinal neurons. ⁷ It is possible therefore that increased expression of GFRα-2 reflects a direct survival action of NTN on the light-damaged photoreceptor cells. In addition, the coexpression of both NTN and GDNF in the photoreceptor outer segments could indicate that NTN is acting synergistically with GDNF in a neuroprotective role in the photoreceptors themselves. Such synergistic effects could also occur through combinations of NTN and/or GDNF with other neurotrophic factors, such as BDNF, CNTF, or FGF, which are synthesized by other retinal cells ³² and have been shown to rescue photoreceptors from light-induced degeneration. ³³ Studying the effect of light-induced damage on mice lacking GFRα-2 ³⁴ may provide insight into the role of NTN receptors in photoreceptor-induced apoptosis. Of note, NTN but not GFRα-2 mRNA expression was found to be increased at the onset of photoreceptor cell death in the rd model of retinal degeneration, ⁶ suggesting that different triggers of apoptosis can differentially modulate expression of NTN and its receptor complex, perhaps to promote either neuronal survival or remodeling. 

As in the rd mouse model, the level of Ret expression was not affected by induction of apoptosis, indicating that there is not a simple coordinate effect relating the regulation of the expression of NTN to that of the components of its receptor complex. Indeed, insult-specific regulation of GDNF components was also observed in the rat brain, ²⁸ indicating an extremely complex control of the GDNF family activity. Neuronal survival elicited by GFLs is dependent on the activity of members of the Src family of kinases, in particular cSrc-p60. ² It is interesting to note therefore that in dark-reared animals, which are more susceptible to light-induced damage (for review see Ref. ¹² ) and exhibit a higher density of apoptotic photoreceptor nuclei, ⁶ we observed persistently low levels of cSrc-p60 expression. An assessment of the phosphorylation-dependent activation of cSrc-p60 will ascertain the extent to which GFL survival signaling may be impaired in this model. 

In summary, the present study has demonstrated independent regulation of expression of receptors of the GDNF family in the normal and light-damaged rat retina and in conjunction with previous work suggests that the pattern of modulation of these genes during photoreceptor degeneration is determined by the nature of the apoptotic trigger. Such differential responses to different modes of retinal degeneration may reflect influences of the neurotrophic system on photoreceptor survival or in the regulation of neuronal plasticity. 

 Supported by the Iris Fund for Prevention of Blindness, the British Retinitis Pigmentosa Society, and National Eye Institute Grant EY01959 (DTO).

 Submitted for publication October 10, 2003; revised December 5, 2003; accepted December 17, 2003.

 Disclosure: C. Jomary, None; R.M. Darrow, None; P. Wong, None; D.T. Organisciak, None; S.E. Jones, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Western blot analysis of NTN in the retinas of cyclic-light– or dark-reared animals subjected to light-induced damage of 12-, 24-, or 24-hour exposure followed by 24 hours in darkness. Densitometric analyses of the normalized signals are presented as the mean ratio ± SEM of NTN to actin from three independent experiments. Representative exposures of enhanced chemiluminescence detection of the immunoreactive proteins are shown in the bottom panels.

Figure 2.

 

Immunofluorescent analysis using antibodies to NTN, GFRα-2, Ret, cSrc-p60, GFRα-1, and GDNF, in cyclic-reared animals: control (A–F), and exposed to intense visible light for 12 (G–L) or 24 (M–R) hours and for 24 hours followed by a 24-hour dark period (S–X). LE, light exposure; dk, dark period; PR, photoreceptors; ONL, outer nuclear layer; INL, inner nuclear layer; GCL, ganglion cell layer. Arrows: photoreceptor immunostaining. Bar, 100 μm.

Figure 3.

 

Immunofluorescent analysis using antibodies to NTN, GFRα-2, Ret, cSrc-p60, GFRα-1, and GDNF in dark-reared animals: control (A–F), and exposed to intense visible light for 12 (G–L) or 24 (M–R) hours and for 24 hours followed by a 24-hour dark period (S–X). LE, light exposure; dk, dark period; PR, photoreceptors; ONL, outer nuclear layer; INL, inner nuclear layer; GCL, ganglion cell layer; OS, photoreceptor outer segments; IS, photoreceptor inner segments; RPE retinal pigment epithelium. Arrowheads: photoreceptor nuclear immunostaining; arrows: photoreceptor immunostaining. Bar, 100 μm.

Figure 4.

 

Control immunostaining on retinal sections lacking primary antibody (A) incubated with goat anti-rabbit IgG conjugated to fluorescein and (B) incubated with rabbit anti-goat IgG conjugated to fluorescein, indicating a weak nonspecific fluorescence of the RPE and choroid. Bar, 100 μm.

Figure 5.

 

Western blot analysis of GFRα-2 in the retina of cyclic-light– or dark-reared animals exposed to various periods of light-induced damage. Densitometric analyses of the normalized signals are presented as the ratio ± SEM of GFRα-2 to actin from three independent experiments. Versus control in cyclic-reared rats: *P < 0.003, 24-hour light exposure; *P < 0.0001, 24-hour light exposure followed by 24-hour dark period; unpaired t-test. Versus control in dark-reared rats: *P < 0.04, 24-hour light exposure; *P < 0.03, 24-hour light exposure followed by 24-hour dark period; unpaired t-test. Representative exposures of enhanced chemiluminescence detection of the immunoreactive proteins are shown in the bottom panels.

Figure 6.

 

Western blot analysis of Ret in the retina of cyclic-light– or dark-reared animals exposed to various light-induced damage treatments. Densitometric analyses of the normalized signals are presented as the mean ratio ± SEM of Ret to actin in three independent experiments.

Figure 7.

 

Western blot analysis of cSrc-p60 in the retina of cyclic-light– or dark-reared animals exposed to various light-induced damage treatments. Densitometric analyses of the normalized signals are presented as the mean ratio ± SEM of cSrc-p60 to actin from three independent experiments. Versus control in cyclic-reared rats: *P < 0.05, 12-hour light exposure; *P < 0.0005, 24-hour light exposure; *P = 0.0001, 24-hour light exposure followed by a 24-hour dark period; comparison between the two controls (cyclic-light– and dark-reared): ▴P = 0.0001; unpaired t-test). Representative exposures of enhanced chemiluminescence detection of the immunoreactive proteins are shown in the bottom panels.

Figure 8.

 

Western blot analysis of GFRα-1 in the retina of cyclic-light– or dark-reared animals exposed to various light-induced damage treatments. Densitometric analyses of the normalized signals are presented as the ratio ±SEM of GFRα-1 to actin from three independent experiments.

Figure 9.

 

Western blot analysis of GDNF in the retina of cyclic-light– or dark-reared animals exposed to various light-induced damage treatments. Densitometric analyses of the normalized signals are presented as the ratio ±SEM of GDNF to actin from three independent experiments.

The authors thank John Grist and Hannah Stewart for technical assistance.