Expression of Ciliary Neurotrophic Factor Activated by Retinal Müller Cells in Eyes with NMDA- and Kainic Acid–Induced Neuronal Death

Megumi Honjo; Hidenobu Tanihara; Noriaki Kido; Masaru Inatani; Kazushiro Okazaki; Yoshihito Honda

Abstract

purpose. To elucidate the role of retinal Müller cells in N-methyl-d-aspartate (NMDA)- or kainic acid (KA)-induced retinal damage.

methods. In experimental eyes, NMDA or KA was injected into the vitreous of rat eyes. Immunohistochemistry and western blot analysis were conducted to elucidate expression and localization of glial fibrillary acidic protein (GFAP) and ciliary neurotrophic factor (CNTF). In addition, the neuroprotective effects of CNTF were calculated by counting cells in the ganglion cell layer (GCL) and by measuring the thickness of the various retinal layers.

results. Morphometric analysis of retinal damage in NMDA- and KA-injected eyes showed significant cell loss in the GCL and thinning of the inner plexiform layer (IPL) of the retina, but not of other retinal layers. Immunohistochemistry demonstrated disappearance and/or decrease in immunoreactivities of calbindin- and calretinin- positive cells and their neurites and upregulated expression of both GFAP and CNTF in experimental eyes. Western blot analysis showed an increase in protein expression for CNTF in retinas of experimental eyes. Confocal images and sequential localization demonstrated colocalization of CNTF and GFAP in the inner retinal layer and possibly in Müller cells. In addition, pretreatment with CNTF (1 μg) before the intravitreal injection of NMDA (or KA) demonstrated that CNTF has neuroprotective effects against NMDA- or KA-induced neuronal death in the retina.

conclusions. These studies revealed the upregulated expression of CNTF and GFAP in Müller cells in response to NMDA- and KA-induced neuronal death, suggesting that production of CNTF in Müller cells may be a part of the endogenous neuroprotective system in the retina.

Upregulated expression of neuroprotective molecules plays an important role in fending off various insults to neural tissues.¹ ² ³ ⁴ Ciliary neurotrophic factor (CNTF) is a representative neuroprotective molecule that was initially isolated from the eye,⁵ and upregulated expression of CNTF has been reported in eyes subjected to mechanical injury,⁶ ⁷ light exposure,⁸ and ischemia.⁹ Because intravitreal injection of this neurotrophic factor protects retinal neurons from various insults such as axotomy,¹⁰ ischemia,¹¹ inherited retinal degeneration and light induced retinal degeneration,¹² upregulation of CNTF has been hypothesized to play an important role in the endogenous neuroprotective system.⁶ ⁷ ⁸ ⁹ ¹³ On the other hand, changes in the expression and distribution of glial fibrillary acidic protein (GFAP) in retinal glial components have been reported in eyes with various diseases and/or injuries, including glaucoma,¹⁴ retinal detachment,¹⁵ light-induced damage,¹⁶ retinal degeneration,¹⁷ mechanical injury¹⁸ and ischemia.¹⁹ Because upregulation of GFAP occurs in response to a number of conditions, as noted above, this may represent a responsive mechanism elicited by stress or injury in the retina. Because the intraocular origin of CNTF expression is hypothesized to be glial cells, astrocytes, and Müller cells, the number of GFAP-positive cells may reflect the number of CNTF-expressing cells.⁶ ⁷

It is well known that glutamate receptor–related neurotoxicity can be induced by N-methyl-d-aspartate (NMDA) and kainic acid (KA),²⁰ ²¹ ²² and herein, we report upregulated expression of CNTF in retinal Müller cells in response to NMDA- or KA-induced retinal damage.

Materials and Methods

Animals

Experiments were performed on adult male Sprague–Dawley rats (200–300 g). All animals were killed by an intraperitoneal overdose injection of pentobarbital. Animals were given water and food ad libitum. All studies were conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research.

Animal Models of NMDA- and KA-Induced Neuronal Death

NMDA and KA were obtained from Sigma (St. Louis, MO), and rat CNTF was obtained from Genzyme (Cambridge, MA). Animal models were made in a manner similar to that described by Morizane et al.²³ Briefly, rats were anesthetized with an intraperitoneal injection of 50 mg/kg of sodium pentobarbital. The pupil was dilated with phenylephrine hydrochloride and tropicamide drops, and a single dose of 5 μl of 40 mM NMDA (total amount, 200 nmol) or 5 μl of 1 mM KA (total amount, 5 nmol) was injected into the vitreous space. A microsyringe with a 30-gauge needle was inserted 2 mm behind the limbus at the superotemporal quadrant and directed toward the optic nerve. When the tip of the needle reached the midvitreous, the injection was administered in a single, swift action. All procedures were performed under microscopy. In this series, rats that received only an injection of 5 μl of 0.1 M phosphate-buffered saline (PBS) served as controls.

Morphometric Analysis

Morphometric analysis was carried out in a manner as described previously.²³ Briefly, 7 days after NMDA or KA injection, animals were killed by an intraperitoneal overdose injection of pentobarbital, and the eyes were enucleated. Eyes were immersed overnight at 4°C in fixative solution containing 2.5% glutaraldehyde and 2% formalin in 0.1 M phosphate buffer (pH 7.4), followed by fixation at 4°C in 10% formalin in 0.1 M phosphate buffer (pH 7.4) for at least 24 hours, followed by dehydration and paraffin embedding. Transverse sections of the rat retinas, 3-μm thick, were made through the optic disc. The sections then were stained with hematoxylin and eosin and subjected to morphometric analysis. The extent of NMDA- or KA-induced retinal neuronal death was quantified by counts of cells in the ganglion cell layer (GCL) and the thickness of the various retinal layers, such as the inner plexiform layer (IPL), inner nuclear layer (INL), outer plexiform layer (OPL), and outer nuclear layer (ONL) at a distance of 1.0 to 1.5 mm from the optic disc. Units for numerical values were cell number per millimeter for GCL cells and per micrometer for thickness of retinal layers. Data from three sections were averaged for each eye and compiled. The results were expressed as the means ± SE. All data were analyzed by analysis of variance (ANOVA), and appropriate group comparisons between PBS-injected control eyes and experimental eyes were performed.

Antibodies

The following specific antibodies were used: rabbit polyclonal antibody to human GFAP (Dako Japan, Kyoto, Japan); mouse monoclonal antibody to rat CNTF (Boehringer Mannheim, Mannheim, Germany); mouse monoclonal anti-calbindin D antibody (Sigma Chemical Co.); rabbit polyclonal antibody to calretinin (Chemicon International, Temecula, CA), biotinylated anti-rat CNTF antibody (R&D Systems Inc., Minneapolis, MN). Appropriate secondary antibodies labeled with dichlorotriazinylamino fluorescein or indocarbocyanin dyes Cy3 were obtained from Chemicon International.

Immunohistochemical Procedures

Adult rats were perfusion-fixed with 4% paraformaldehyde/PBS before enucleation. Subsequently, the enucleated eyes were further fixed for 2 hours at 4°C in 4% paraformaldehyde/0.1 M PBS, washed for 5 minutes in PBS, then gently shaken overnight at 4°C in 15% sucrose/0.1 M PBS, embedded in Tissue-Tek (Miles, Inc., Elkhart, IN), and frozen in liquid nitrogen. Sections (10 μm) were cut on a cryostat and collected onto silanized slides (Dako Japan), and air-dried.

The samples were incubated successively in methanol at −20°C for 20 minutes, in 5% skim milk in PBS (“blocking solution”) for 30 minutes, and in a solution of antibodies to CNTF (1:50 dilution in blocking solution) (Boehringer Mannheim), GFAP (1:100 dilution in blocking solution), calbindin-D (1:500 dilution in blocking solution), calretinin (1:500 dilution in blocking solution), for 60 minutes at room temperature. They were then treated for 30 minutes with secondary antibodies (Cy3-conjugated anti-mouse IgG or anti-rabbit IgG) diluted 1:200. For double-label immunostaining, the same procedures were repeated. Sections were mounted in 90% glycerol/10% PBS. The secondary antibodies used for double-staining experiments were labeled with Cy3 for CNTF and fluorescein isothiocyanate (FITC) for GFAP. As for immunohistochemistry for CNTF and GFAP, sequential sections were stained with each antibody as well as double-stained sections. Fluorescence was visualized under an epifluorescence microscope (Zeiss Axioplan, Oberkochen, Germany) and with a confocal laser scanning microscope (Bio-Rad Laboratories, Hercules, CA). To display results from double-labeled sections simultaneously, confocal images were color-coded and superimposed. After immunostaining of the 10-μm-thick transverse sections of rat retinas, immunoreactive cells at a determined layer were counted at a distance of 1.0 to 1.5 mm from the optic disc. Units for numerical values were cell number per millimeter. Data were obtained from at least six eyes, and the results are shown as the means ± SE. P values and statistical significance were calculated by ANOVA between PBS-injected control eyes and experimental eyes.

Western Blot Analysis

For western blot analysis, 7 days after NMDA or KA injection, animals were killed by an intraperitoneal overdose injection of pentobarbital, and the eyes were enucleated. The animals that received PBS injection served as controls. The retinal tissues were isolated with fine forceps and Vannas scissors under a dissecting microscope. Then samples were lysed in sample buffer containing 2% sodium dodecylsulfate (SDS) and were subjected to SDS-polyacrylamide gel electrophoresis (PAGE) using a polyacrylamide gradient gel (15%–25%) (Daiichi Pure Chemicals, Tokyo, Japan). Each sample was containing 50μ g protein. The concentration of each sample was measured by Bio-Rad DC protein assay (Bio-Rad Laboratories). After electrophoresis, the proteins were electrophoretically transferred onto polyvinylidene difluoride membranes (Millipore Co., Bedford, MA). After blocking the membrane with 5% skim milk and 0.002% Tween 20 in Tris-buffered saline (TBS; pH 7.4) at 4°C for 16 hours, the membrane was incubated in a 1:500 dilution of biotinylated anti-rat CNTF antibody (R&D Systems Inc.) for 2 hours at room temperature and then washed with TBS five times for 5 minutes. The membrane was then incubated with ABC solution (ABC Elite kit; Vector, Burlingame, CA). After washing the membrane with TBS for 5 minutes five times, the blotted protein bands were stained with dimethylaminoazobenzene (DAB) solution. Optical densities of the labeled bands were measured by a Power Macintosh G3 computer (Apple Computer, Cupertino, CA) and NIH Image 1.59.

Pretreatment with CNTF

A single injection of 1 μl CNTF at two concentrations (0.1 or 1μ g/μl) was administered into the vitreous space using a 30-gauge needle in the same manner as described above. Two days later, a single injection of 5 μl of 40 mM NMDA (200 nmol) or 5 μl of 1 mM KA (5 nmol) was administered into the vitreous space in a single dose. One week later the animals were killed, and the eyes were prepared for histologic examination. Morphometric analyses and immunohistochemical studies were conducted as described above. The control animals received injection of 1 μl of vehicle (PBS) 2 days before NMDA or KA injection and were killed 1 week later in the same manner.

Results

Morphometric Analysis of NMDA- and KA-Induced Retinal Damage

In eyes that underwent intravitreal injection of NMDA (200 nmol), cell loss in the GCL and thinning of the IPL were observed. Statistical analysis of results obtained from the morphometric studies showed significant differences between experimental and control eyes in cell counts in the GCL and thickness of the IPL (P < 0.0001 and P < 0.0001, ANOVA), but not for other layers; that is, INL, OPL, and ONL (Fig. 1A ). Furthermore, similar analyses of eyes that received intravitreal injections of KA (5 nmol) showed statistically significant differences between experimental and control eyes in cell counts in the GCL and thickness of the IPL (P < 0.0001 and P < 0.0001, ANOVA), but not for other layers (Fig. 1B) .

Immunohistochemical Studies for Calcium-Binding Proteins

In PBS-injected (control) eyes, immunohistochemical studies demonstrated that calretinin immunoreactivities were seen in cells located in the GCL and INL as well as in axons of the nerve fiber layer (NFL) and IPL. Furthermore, three calretinin-positive layers were observed within the IPL (Fig. 2A ). This result seemed to be in agreement with previously reported data.²⁴ In NMDA-injected eyes, there was a significant loss of GCL cells and partial cell components in the INL, and more intense staining of cells in the INL and small granulelike immunostaining in the IPL were seen (Fig. 2B) . In experimental eyes using intravitreal injection of KA, no calretinin-positive bands were observed, although small granulelike immunostaining was seen in the IPL (Fig. 2D) . In control eyes, immunoreactivities of calbindin, another calcium-binding protein and cell type marker for amacrine and horizontal cells,²⁴ ²⁵ ²⁶ ²⁷ were faint in cells at the vitreous side of the INL and much more conspicuous in cells of the outer layer of the INL (Fig. 2F) . This result seemed to be in good accordance with previous reports.²⁴ ²⁵ ²⁶ ²⁷ One week after intravitreal injection of NMDA, numerous calbindin-positive cells in the inner side of the INL (possibly amacrine cells) showed disappearance and/or decrease in immunoreactivities although quantification of the cells was difficult because of the faint immunoreactivities (Fig. 2G) . In contrast, the population of calbindin-positive cells in the outer side of the INL (possibly horizontal cells) in experimental eyes seemed to be same as in control eyes. A similar observation was confirmed by experiments using intravitreal injection of KA (Fig. 2I) .

Immunohistochemical Studies for GFAP and CNTF

In PBS-injected (control) eyes, immunohistochemical studies using polyclonal antibodies against GFAP showed immunoreactivity in some of the longitudinal cells embedded in the nerve fiber layer, which might be astrocytes, but no immunoreactivity was found in the outer layer of the retina (Fig. 3A ). In contrast, in the experimental eyes that had received an intravitreal injection of NMDA, faint radial staining was seen at early stage (3 days after NMDA injection) (Fig. 3B) , and strong immunoreactivity against GFAP was found in the inner layer (NFL and GCL), and radial staining was seen throughout the retina at later stage (7 days after NMDA injection) (Fig. 3C) . This was confirmed in other experimental eyes that had undergone intravitreal injection of KA (Figs. 3D 3E) .

Additional immunohistochemical studies were conducted in an attempt to elucidate the expression of CNTF in rat retinas. In control eyes, no CNTF immunoreactivity was detected, save for faint immunostaining in the NFL (Fig. 3F) . However, as early as 3 days after treatment with NMDA, CNTF immunoreactivity became detectable in the NFL and individual cells in the INL. Faint labeling also was detectable in radial processes of Müller cells (Fig. 3G) . At a later stage as late as 7 days after NMDA injection, more conspicuous CNTF immunoreactivities were observed in the NFL and cells in the INL; there also was less conspicuous radial staining throughout the retina (Fig. 3H) . A greater intensity of immunostaining was observed in the region of the cell nucleus compared with the cytoplasm. This was confirmed in other experimental eyes that had undergone intravitreal injection of KA (Figs. 3I 3J) . Alteration in immunoreactivities for CNTF was more drastic in KA-injected eyes than in NMDA-injected eyes. Immunohistochemical study of sequential sections as well as confocal microscopic observation of the double-stained sections by GFAP and CNTF revealed colocalization of the two proteins in the radial staining pattern (Fig. 4) . We quantitated cell counts of CNTF-positive cells associated with immunostaining for both GFAP and CNTF in radial staining. The mean cell count (±SE) of CNTF-positive cells was 3.3 ± 1.8, 48.3 ± 4.5, and 44.1 ± 3.5 (number/mm), respectively, in PBS-injected control eyes, NMDA-injected eyes, and KA-injected eyes at 7 days after NMDA or KA injection (Table 1) . Statistical analysis showed a significant increase in the mean cell count (± SE) of CNTF-positive cells during the observation period (up to 1 week) in both NMDA- and KA-injected eyes (P < 0.0001 for NMDA treatment and P < 0.0001 for KA treatment, ANOVA) (Table 1) .

Western Blot Analysis

To identify and quantify protein expression of CNTF, we carried out a western blot analysis. Western blot analysis showed a band of approximately 23 kDa, which corresponds to the reported molecular size of rat CNTF,⁹ ²⁸ in experiments using anti-rat CNTF antibody (Fig. 5) . Densitometric analysis of the positive bands of expected length showed that, in experimental eyes with NMDA injection (Fig. 5A) , the optical density was 2.13 ± 0.38-fold that of the control eyes (1.00 ± 0.13) (P = 0.0082, ANOVA). Also, similar experiments showed that, in experimental eyes with KA injection (Fig. 5B) , the optical density was 7.51 ± 0.98-fold that of the PBS-injected (control) eyes (1.00 ± 0.24) (P = 0.0004, ANOVA).

Morphometric Analysis for Neuroprotective Effects of CNTF

In an effort to elucidate neuroprotective effects of CNTF against NMDA- or KA-induced retinal damage, we conducted morphometric analyses after intravitreal injection of NMDA and KA (Fig. 6) . Two days after injection of 1 μl (0.1 or 1 μg/μl) recombinant rat CNTF (or 1 μl of PBS as a negative control) into the vitreous, 5μ l of NMDA (200 nmol) or KA (5 nmol) was injected. With CNTF or PBS pretreatment, no inflammatory findings were observed on histologic examination. Also, in both NMDA- and KA-injected eyes, cell loss in the GCL and thinning of the IPL appeared to be inhibited.

In NMDA-injected eyes, the mean number (±SE) of GCL cells was 25.9 ± 0.9, 25.8 ± 1.6, and 40.1 ± 3.9, respectively, in PBS-treated eyes (control), CNTF (0.1 μg)-treated eyes, and CNTF (1 μg)-treated eyes (PBS versus CNTF 0.1 μg, P = 0.9681; PBS versus CNTF 1 μg, P < 0.0001, ANOVA) (Table 2) . The mean thickness (±SE) of the IPL was 18.8 ± 1.4, 18.7 ± 1.2, and 25.1 ± 1.6, respectively, in PBS-treated eyes, CNTF (0.1 μg)-treated eyes, and CNTF (1 μg)-treated eyes (PBS versus CNTF 0.1 μg, P = 0.9622; PBS versus CNTF 1 μg, P = 0.0073, ANOVA) (Table 2) . Also, in KA-injected eyes, the mean number (±SE) of GCL cells was 26.4 ± 1.6, 28.5 ± 5.9, and 30.6 ± 3.5, respectively, in PBS-treated eyes (control), CNTF (0.1 μg)-treated eyes, and CNTF (1 μg)-treated eyes (PBS versus CNTF 0.1 μg, P = 0.6191 and PBS versus CNTF 1 μg, P = 0.3720, ANOVA) (Table 2) . The mean thickness (±SE) of IPL was 9.0 ± 1.2, 9.8 ± 2.3, and 16.9 ± 2.2, respectively, in PBS-treated eyes (control), CNTF (0.1 μg)-treated eyes and CNTF (1 μg)-treated eyes (PBS versus CNTF 0.1 μg, P = 0.9084 and PBS versus CNTF 1 μg, P = 0.0485, ANOVA) (Table 2) .

To show a neuroprotective effect of CNTF pretreatment, immunohistochemical studies for calbindin/calretinin were performed in NMDA (200 nmol)- or KA (5nmol)-injected eyes after CNTF (1 μg) pretreatment (Figs. 2C 2E 2H 2J) . As for calretinin immunoreactivity, the same staining pattern was observed in eyes with PBS pretreatment before NMDA or KA injection (negative controls) as shown in NMDA- or KA-injected eyes without any pretreatment. In eyes with CNTF pretreatment, the small granulelike immunostaining against calretinin seen in the IPL was preserved compared to the control (Fig. 2C) . As for calbindin immunoreactivity, decrease of the number of calbindin-positive cells in the inner side of the INL (possibly amacrine cells) appeared to be inhibited in the experimental eyes, although quantification of the cells was difficult because of the faint immunoreactivities (Fig. 2H) . In contrast, the population of calbindin-positive cells in the outer side of the INL (possibly horizontal cells) in experimental eyes seemed to be same as in control eyes. A similar observation was confirmed by experiments using intravitreal injection of KA (Figs. 2E 2J) .

Discussion

In several common ocular diseases, including glaucoma, ischemia, and optic nerve damage, high levels of glutamate in the vitreous have been reported.²⁹ ³⁰ ³¹ Because glutamate is not only a neurotransmitter in retina, but also a neurotoxic excitatory amino acid against neuronal cells, prolonged high levels of glutamate have been hypothesized to result in serious damage to the retina.³² ³³ Glutamate receptor–related neuronal death can be explained by an influx of calcium ion, nitric oxide synthesis and subsequent free radical formation, depletion of ATP and various enzymatic reactions of calcium-dependent enzymes.³⁴ It is well known that glutamate receptor–related neurotoxicity can be induced by NMDA and KA.²⁰ ²¹ ²² In the study described herein, we investigated alterations in the expression of CNTF in NMDA- or KA-injected eyes and also the efficacy of CNTF in experimental eyes.

Our study showed upregulation of GFAP in NMDA- and KA-injected rat eyes, suggesting that glial components throughout the retina may be part of the retinal response against NMDA- and KA-induced neuronal cell death. Because responsive upregulation of GFAP as well as CNTF in retinal glial cells was elicited within a short incubation period (as early as 3–7 days) in our study, NMDA or KA treatments may have their effect directly on retinal glial cells. However, we were unable to draw any conclusion on this point because we cannot deny the possibility that these events may follow the early events of neurodegenerative processes even before cell loss of retinal neuronal cells. Astrocytes are ubiquitous in the NFL and in the GCL, and Müller cells extend all the way from the external to the internal limiting membrane.³⁵ ³⁶ Our immunohistochemical studies on control sections demonstrated faint expression of GFAP in the NFL and GCL, but none in outer layers, which is in agreement with previous reports.¹⁴ ¹⁵ ¹⁶ ¹⁷ ¹⁸ ¹⁹ After the intravitreal injection of NMDA and KA, conspicuous immunolabeling for GFAP was observed in a radial pattern between the NFL and external limiting membrane in addition to longitudinal immunoreactivity in the NFL and GCL. Thus, the major cell types responsible for the upregulation of GFAP in eyes with experimental eyes are thought to be retinal Müller cells. Changes in the expression and distribution of GFAP in retinal glial components have been reported in eyes with various diseases and/or injuries, such as retinal detachment,¹⁵ laser-induced injury,¹⁶ retinal degeneration,¹⁷ and mechanical injury.¹⁸ Because upregulation of GFAP occurs in response to a number of conditions, as described above, it may represent a nonspecific neuroprotective mechanism in response to stress or injury in the retina. In our previous report,¹⁴ upregulation of GFAP was identified in monkey eyes with experimental glaucoma and our immunohistochemical studies demonstrated that the cells responsible are most probably both astrocytes and Müller cells. Because glaucoma is one of the most common ocular diseases related to elevated levels of glutamate in the vitreous body,³⁷ it is possible that it may be caused by abnormal release and/or uptake of glutamate, secondary to intraocular pressure–related primary neural damage. This also may be the case in eyes with ischemia and optic nerve damage, both of which result in elevation in intravitreal glutamate concentrations.²⁹ ³⁰ ³¹ ³⁸ Thus, our results show that the treatments with NMDA or KA elicit activation of retinal glial components, whether by a direct or an indirect mechanism.

Immunohistochemical studies against calcium-binding proteins demonstrated that, in experimental eyes, GCL cells and partial cell components in the INL showed disappearance and/or decrease in immunoreactivities. This result seems to be in good accordance with the cell loss of these layers, as shown in morphometric analysis. On the other hand, we were unable to detect a significant change in the population of the calbindin-positive cells in the outer side of the INL (possibly horizontal cells) in experimental eyes in comparison with control eyes. Although some investigation suggested that horizontal cells possess sensitivity to KA,²¹ ²² ³⁸ ³⁹ ⁴⁰ Morgan et al.²² ⁴¹ have reported that, with low (6 nmol) and intermediate (60 nmol) doses of KA, the retinal lesions were confined to the IPL and amacrine cells, and horizontal cells were substantially intact. Our study was conducted under experimental conditions using an experimental condition with 5 nmol of KA in which only inner retinal layers were affected. The used dose of KA in this study was comparable to “low dose” of the previous report by Morgan et al., and the result of our study also seemed to be in good accordance with their result.

Our immunohistochemical studies showed upregulated expression of CNTF in retinal Müller cells, as well as in astrocytes of the NFL and GCL. Confocal images of double staining and sequential localization demonstrated colocalization of GFAP and CNTF in radial processes, indicating that activated glial components express this neurotrophic factor throughout the structure of the retina. Similar upregulation of CNTF has been reported in eyes with mechanical injury, light-induced retinal damage, and ischemia.⁶ ⁷ ⁸ ⁹ Also, we observed that CNTF might have a preferential disposition to the nucleus rather than the cytoplasm. This result is in good accordance with CNTF immunoreactivity in adult rodent central nervous system and rat retina after pressure-induced ischemia.⁹ ²⁸ The nuclear localization of CNTF may be important in defining the function of these CNTF-positive cells and suggests the possibility that CNTF might interact with cellular components directly to affect gene transcription. Also, faint to moderate expression of CNTF was observed in cells (possibly astrocytes) in the NFL and GCL, which corresponds to GFAP expression. Responsive upregulated expression of CNTF and GFAP in Müller cells may take part in a series of events of NMDA- and KA-induced neurodegeneration. Because CNTF activates transcriptional expression of GFAP in glial cells,⁴² the secondary upregulation of GFAP expression in Müller cells may be caused by the primary release of the CNTF. Also, it is possible that alteration in distribution of the expression may result in neuroprotective effects in a larger area.

In conclusion, our studies have shown that upregulated expression of CNTF and GFAP in retinal Müller cells occurs in response to NMDA- and KA-induced neuronal death. Thus, in addition to astrocytes, activation of retinal Müller cells may play a role in the intrinsic neuroprotective system of retina.

Supported in part by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture, Japan, and from the Ministry of Health and Welfare, Japan.

Submitted for publication March 12, 1999; revised August 26, 1999; accepted September 23, 1999.

Commercial relationships policy: N.

Corresponding author: Hidenobu Tanihara, Department of Ophthalmology, Tenri Hospital, Mishima-cho 200, Tenri, 632-0015, Japan. [email protected]

Figure 1.

View Original Download Slide

Morphometric analysis of glutamate receptor–mediated retinal damage on the seventh posttreatment day. In NMDA (200 nmol)-injected eyes (A) and kainic acid (5 nmol)-injected eyes (B), the degree of NMDA-induced retinal neuronal death was quantified by cell counts (cell number/mm) of cells present in the ganglion cell layer (GCL) and the thickness (μm) of retinal layers such as the inner plexiform layer (IPL), inner nuclear layer (INL), outer plexiform layer (OPL) and outer nuclear layer (ONL) at a distance of 1.0 to 1.5 mm from the optic disc. Units for numerical values were cell number per millimeter for GCL cells and per micrometer for thickness of retinal layers. Data from three sections were averaged for each eye. The results are shown as the means ± SE (n = 10). P values and statistical significance were calculated by ANOVA between PBS-injected control eyes and experimental eyes. Morphometric analysis showed statistically significant cell loss in GCL cells (P < 0.0001) and thinning of IPL thickness (P < 0.0001). NS, not significant.

Figure 2.

View Original Download Slide

Immunohistochemistry for calcium-binding proteins in retina. (A through E) Light microscopic photographs show representative immunohistochemistry for calretinin. In PBS-injected control eyes (A), calretinin immunoreactivities were seen in cells in the GCL and INL, in addition to their neurites. Three calretinin-positive layers were observed in the IPL of control eyes. In NMDA (200 nmol)-injected eyes (B), at the first posttreatment week, significant loss of GCL cells and partial cell components in the INL were found, and calretinin-immunoreactive small granulelike immunostaining was observed. On the other hand, in KA (5 nmol)-injected eyes (D), no calretinin-positive bands were observed although small granulelike immnunostaining was found in the IPL (D). With pretreatment of CNTF, loss of small granulelike immunostaining was recovered (C). This observation was confirmed by experiments using intravitreal injection of KA (D, E). (F through J) Light microscopic photographs showed representative immunohistochemistry for calbindin. In PBS-injected control eyes (F), immunoreactivities of calbindin were found in cells of the INL. In NMDA (200 nmol)-injected eyes (G), at the first posttreatment week, the number of calbindin-positive cells in the inner side of the INL decreased, but the population of calbindin-positive cells in the outer side of the INL (possibly horizontal cells) in experimental eyes seemed to be same as in control eyes. This observation was confirmed by experiments using intravitreal injection of KA (I). Pretreatment with CNTF showed no significant protective effect against decreasing the number of calbindin-positive cells (H, J). NFL, nerve fiber layer; GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; OPL, outer plexiform layer; ONL, outer nuclear layer. Bar, 50μ m.

Figure 3.

View Original Download Slide

Immunohistochemistry for glial fibrillary acidic protein (GFAP) and ciliary neurotrophic factor (CNTF). Upregulated immunoreactivities of GFAP were shown in both NMDA- and KA-injected eyes. Arrowheads indicate the NFL. (A through E) Light microscopic photographs showed representative immunohistochemistry for GFAP. (A) PBS-injected control eyes; (B) 3 days after NMDA (200 nmol) injection; (C) 7 days after NMDA (200 nmol) injection; (D) 3 days after KA (5 nmol) injection; (E) 7 days after KA (5 nmol) injection. (F through J) Light microscopic photographs showed representative immunohistochemistry for CNTF. Upregulated immunoreactivities of CNTF were shown in both NMDA- and KA-injected eyes. (F) PBS-injected control eyes; (G) 3 days after NMDA (200 nmol) injection; (H) 7 days after NMDA (200 nmol) injection; (I) 3 days after KA (5 nmol) injection; (J) 7 days after KA (5 nmol) injection. Bar, 50 μm.

Figure 4.

View Original Download Slide

Sequential localization and colocalization of GFAP and CNTF. Immunohistochemistry of sequential sections of KA-injected eyes showed sequential localization of CNTF (A) and GFAP (B). Also, confocal images were shown in the right-most panel. The CNTF signals (red; Cy3) coincide with the GFAP ones (green; FITC) (C). (A) Arrowheads indicate CNTF-positive radial staining throughout the retina. (B) Arrow indicates GFAP-positive Müller cell and radial staining. (C) Large arrowheads point the colocalization of CNTF and GFAP in radial staining. Bar, 50 μm.

Table 1.

View Table

Cell Counts of CNTF- and GFAP-Positive Cells

Table 1.

Cell Counts of CNTF- and GFAP-Positive Cells

	CNTF-Positive Cells (cells/mm)	P Values
Control	3.3 ± 1.8
NMDA
1 day	13.0 ± 2.3	<0.0001
3 days	32.2 ± 5.2	<0.0001
7 days	48.3 ± 4.5	<0.0001
KA
1 day	15.3 ± 3.0	<0.0001
3 days	32.0 ± 3.7	<0.0001
7 days	44.1 ± 3.5	<0.0001

Cell counts (cell number/mm) of CNTF-positive cells associated with immunostaining for both GFAP and CNTF in radial staining during the observation period (up to 1 week) in both NMDA- and KA-injected eyes were performed. Immunoreactive cells at a determined layer (INL) were counted at a distance of 1.0 to 1.5 mm from the optic disc. Statistical analysis showed a significant increase in the mean cell count (±SE) of CNTF-positive cells during the observation period (up to 1 week) in both NMDA- and KA-injected eyes (P < 0.0001 for NMDA treatment and P < 0.0001 for KA treatment, ANOVA).

Figure 5.

View Original Download Slide

Western blot analysis for ciliary neurotrophic factor (CNTF). Western blot analysis for CNTF showed upregulated expression of CNTF proteins in NMDA (200 nmol)-injected eyes (n = 3) (A) and KA (5 nmol)-injected eyes (n = 3) (B). Lanes 1 to 3: PBS-injected eyes as controls; lanes 4 to 6: NMDA- or KA-injected eyes. A single band of approximately 23 kDa was detected (arrowhead).

Figure 6.

View Original Download Slide

Light micrographs of retinal sections of eyes with glutamate receptor-mediated retinal damage and those pretreated with CNTF. Light microscopic photographs show representative retinal tissues in control and experimental eyes after intravitreal injection of NMDA (200 nmol) and KA (5 nmol). (A) PBS-injected eyes as controls; (B) NMDA-injected eyes without pretreatment; (C) NMDA-injected eyes after pretreatment with CNTF (1 μl of 0.1 μg/μl); (D) NMDA-injected eyes after pretreatment with CNTF (1 μl of 1 μg/μl); (E) PBS-injected eyes as controls; (F) KA-injected eyes without pretreatment; (G) KA-injected eyes after pretreatment with CNTF (1 μl of 0.1 μg/μl); (H) KA-injected eyes after pretreatment with CNTF (1 μl of 1 μg/μl). Bar, 50 μm.

Table 2.

View Table

Morphometric Analysis of Protective Effects of CNTF in NMDA-Injected Eyes

Table 2.

Morphometric Analysis of Protective Effects of CNTF in NMDA-Injected Eyes

	Counts of GCL Cells (cells/mm)	Thickness of IPL (μm)
Control	59.6 ± 1.6	39.2 ± 1.0
NMDA
PBS+ NMDA (200 nmol)	25.9 ± 0.9	18.8 ± 1.4
CNTF (0.1 μg)+ NMDA (200 nmol)	25.8 ± 1.6	18.7 ± 1.2
CNTF (1 μg)+ NMDA (200 nmol)	40.1 ± 3.9^*	25.1 ± 1.6^{, †}
KA
PBS + KA (5 nmol)	26.4 ± 1.6	9.0 ± 1.2
CNTF (0.1 μg)+ KA (5 nmol)	28.5 ± 5.9	9.8 ± 2.3
CNTF (1 μg)+ KA (5 nmol)	30.6 ± 3.5	16.9 ± 2.2^{, ‡}

In NMDA (200 nmol)- and KA (5 nmol)-injected eyes, counts (cell number/mm) of cells present in the GCL and the thickness (μm) of the IPL were evaluated. Data from three sections were averaged for each eye. The results are shown as the means ± SE. The number of experiments was at least four [n = 31 for control experiments; n = 16 for NMDA-injected eyes; n = 4 for CNTF (0.1 μg)-treated eyes; n = 11 for CNTF (1 μg)-treated eyes] for NMDA-injected eyes, and at least three [n = 27 for control experiments; n = 9 for NMDA-injected eyes; n = 4 for CNTF (0.1 μg)-treated eyes; n = 3 for CNTF (1 μg)-treated eyes] for KA-injected eyes. Morphometric analysis showed statistically significant protective effects of CNTF (1 μg) to the decrease in GCL cells (

* P < 0.0001) and the thinning in IPL thickness (

† P = 0.0073). P values and statistical significance were calculated. Morphometric analysis showed statistically significant protective effects of CNTF (1 μg) against thinning of the IPL thickness (

‡ P = 0.0485).

Calza L, Giardino L, Ceccatelli S, Hokfelt T. Neurotrophins and their receptors in the adult hypo- and hyperthyroid rat after kainic acid injection: an in situ hybridization study. Eur J Neurosci. 1996;8:1873–1881. [CrossRef] [PubMed]

Ip FC, Fu AK, Tsim KW, Ip NY. Differential expression of ciliary neurotrophic factor receptor in skeletal muscle of chick and rat after nerve injury. J Neurochem. 1996;67:1607–1612. [PubMed]

Clatterbuck RE, Price DL, Koliatsos VE. Further characterization of the effects of brain-derived neurotrophic factor and ciliary neurotrophic factor on axotomized neonatal and adult mammalian motor neurons. J Comp Neurol. 1994;342:45–56. [CrossRef] [PubMed]

Lee MY, Naumann T, Kirsch M, Frotscher M, Hofmann HD. Transient up-regulation of ciliary neurotrophic factor receptor-alpha mRNA in axotomized rat septal neurons. Eur J Neurosci. 1997;9:622–626. [CrossRef] [PubMed]

Barbin G, Manthorpe M, Varon S. Purification of the chick eye ciliary neurotrophic factor (CNTF). J Neurochem. 1984;43:1468–1478. [CrossRef] [PubMed]

Cao W, Wen R, Li F, Lavail MM, Steinberg RH. Mechanical injury increases bFGF and CNTF mRNA expression in the mouse retina. Exp Eye Res. 1997;65:241–248. [CrossRef] [PubMed]

Wen R, Song Y, Cheng T, et al. Injury-induced upregulation of bFGF and CNTF mRNAs in the rat retina. J Neurosci. 1995;15:7377–7385. [PubMed]

Wen R, Cheng T, Song Y, et al. Continuous exposure to bright light upregulates bFGF and CNTF expression in the rat retina. Curr Eye Res. 1998;17:494–500. [CrossRef] [PubMed]

Ju WK, Lee MY, Hofmann KD, Kirsch M, Chun MH. Expression of CNTF in Müller cells of the rat retina after pressure-induced ischemia. Neuroreport. 1999;10:419–422. [CrossRef] [PubMed]

Mey J, Thanos S. Intravitreal injections of neurotrophic factors support the survival of axotomized retinal ganglion cells in adult rats in vivo. Brain Res. 1993;602:304–317. [CrossRef] [PubMed]

Unoki K, LaVail MM. Protection of the rat retina from ischemic injury by brain-derived neurotrophic factor, ciliary neurotrophic factor, and basic fibroblast growth factor. Invest Ophthalmol Vis Sci. 1994;35:907–915. [PubMed]

LaVail MM, Yasumura D, Matthes MT, et al. Protection of mouse photoreceptors by survival factors in retinal degenerations. Invest Ophthalmol Vis Sci. 1998;39:592–602. [PubMed]

Liu C, Peng M, Laties AM, Wen R. Preconditioning with bright light evokes a protective response against light damage in the rat retina. J Neurosci. 1998;18:1337–1344. [PubMed]

Tanihara H, Hangai M, Sawaguchi S, et al. Up-regulation of glial fibrillary acidic protein in the retina of eyes with experimental glaucoma. Arch Ophthalmol. 1997;115:752–756. [CrossRef] [PubMed]

Erickson PA, Fisher SK, Guerin CJ, Anderson DH, Kaska DD. Glial fibrillary acidic protein increases in Müller cells after retinal detachment. Exp Eye Res. 1987;44:37–48. [CrossRef] [PubMed]

Grosche J, Hartig W, Reichenbach A. Expression of glial fibrillary acidic protein (GFAP), glutamine synthetase (GS), and Bcl-2 protooncogene protein by Müller (glial) cells in retinal light damage of rats. Neurosci Lett. 1995;185:119–122. [CrossRef] [PubMed]

Ekstrom P, Sanyal S, Narfstrom K, Chader GJ, van Veen T. Accumulation of glial fibrillary acidic protein in Müller radial glia during retinal degeneration. Invest Ophthalmol Vis Sci. 1988;29:1363–1371. [PubMed]

Bignami A, Dahl D. The radial glia of Müller in the rat retina and their response to injury. An immunofluorescence study with antibodies to the glial fibrillary acidic (GFA) protein. Exp Eye Res. 1979;28:63–69. [CrossRef] [PubMed]

Osborne NN, Block F, Sontag KH. Reduction of ocular blood flow results in glial fibrillary acidic protein (GFAP) expression in rat retinal Müller cells. Vis Neurosci. 1991;7:637–639. [CrossRef] [PubMed]

Siliprandi R, Canella R, Carmignoto G, et al. N-methyl-d-aspartate-induced neurotoxicity in the adult rat retina. Vis Neurosci. 1992;8:567–573. [CrossRef] [PubMed]

Ehrlich D, Morgan IG. Kainic acid destroys displaced amacrine cells in post-hatch chicken retina. Neurosci Lett. 1980;17:43–48. [CrossRef] [PubMed]

Morgan IG, Ingham CA. Kainic acid affects both plexiform layers of chicken retina. Neurosci Lett. 1981;21:275–280. [CrossRef] [PubMed]

Morizane C, Adachi K, Furutani I, et al. N^ω-nitro-l-arginine methyl ester protects retinal neurons against N-methyl-d-aspartate-induced neurotoxicity in vivo. Eur J Pharmacol. 1997;328:45–49. [CrossRef] [PubMed]

Pasteels B, Rogers J, Blachier F, Pochet R. Calbindin and calretinin localization in retina from different species. Vis Neurosci. 1990;5:1–16. [CrossRef] [PubMed]

Hamano K, Kiyama H, Emson PC, Manabe R, Nakauchi M, Tohyama M. Localization of two calcium binding proteins, calbindin (28 kD) and parvalbumin (12 kD), in the vertebrate retina. J Comp Neurol. 1990;302:417–424. [CrossRef] [PubMed]

Peichl L, Gonzalez–Soriano J. Unexpected presence of neurofilaments in axon-bearing horizontal cells of the mammalian retina. J Neurosci. 1993;13:4091–4100. [PubMed]

Chun MH, Kim IB, Ju WK, et al. Horizontal cells of the rat retina are resistant to degenerative processes induced by ischemia-reperfusion. Neurosci Lett. 1999;260:125–128. [CrossRef] [PubMed]

Henderson JT, Seniuk NA, Roder JC. Localization of CNTF immunoreactivity to neurons and astroglia in the CNS. Brain Res Mol Brain Res. 1994;22:151–165. [CrossRef] [PubMed]

Delbarre G, Delbarre B, Calinon F, Ferger A. Accumulation of amino acids and hydroxyl free radicals in brain and retina of gerbil after transient ischemia. J Ocul Pharmacol. 1991;7:147–155. [CrossRef] [PubMed]

Neal MJ, Cunningham JR, Hutson PH, Hogg J. Effects of ischemia on neurotransmitter release from the isolated retina. J Neurochem. 1994;62:1025–1033. [PubMed]

Perlman JI, McCole SM, Pulluru P, Chang CJ, Lam TT, Tso MO. Disturbances in the distribution of neurotransmitters in the rat retina after ischemia. Curr Eye Res. 1996;15:589–596. [CrossRef] [PubMed]

Olney JW. Brain lesion, obesity and other disturbances in mice treated with monosodium glutamate. Science. 1969;164:719–721. [CrossRef] [PubMed]

Lucas DR, Neuhouse JP. The toxic effect of sodium l-glutamate on the inner layers of the retina. Arch Ophthalmol. 1957;58:193. [CrossRef]

Kristian T, Siesjo BK. Calcium in ischemic cell death. Stroke. 1998;29:705–718. [CrossRef] [PubMed]

Singelman J, Ozanics V. Retina. Jakobiec FA eds. Ocular Anatomy, Embryology, and Teratology. 1982;485. Harper & Row Philadelphia.

Bjorklund H, Bignami A, Dahl D. Immunohistochemical demonstration of glial fibrillary acidic protein in normal rat Müller glia and retinal astrocytes. Neurosci Lett. 1985;54:363–368. [CrossRef] [PubMed]

Dreyer EB, Zurakowski D, Schumer RA, Podos SM, Lipton SA. Elevated glutamate levels in the vitreous body of humans and monkeys with glaucoma. Arch Ophthalmol. 1996;114:299–305. [CrossRef] [PubMed]

Yoles E, Schwartz M. Elevation of intraocular glutamate levels in rats with partial lesion of the optic nerve. Arch Ophthalmol. 1998;116:906–910. [CrossRef] [PubMed]

Hampton CK, Garcia C, Redburn DA. Localization of kainic acid-sensitive cells in mammalian retina. J Neurosci Res. 1981;6:99–111. [CrossRef] [PubMed]

Brandstatter JH, Hartveit E, Sassoe–Pognetto M, Wassle H. Expression of NMDA and high-affinity kainate receptor subunit mRNAs in the adult rat retina. Eur J Neurosci. 1994;6:1100–1112. [CrossRef] [PubMed]

Yazulla S, Kleinschmidt J. The effects of intraocular injection of kainic acid on the synaptic organization of the goldfish retina. Brain Res. 1980;182:287–301. [CrossRef] [PubMed]

Kahn MA, Huang CJ, Caruso A, et al. Ciliary neurotrophic factor activates JAK/Stat signal transduction cascade and induces transcriptional expression of glial fibrillary acidic protein in glial cells. J Neurochem. 1997;68:1413–1423. [PubMed]

Figure 1.

View Original Download Slide