Retinal Neuronal Death Induced by Intraocular Administration of a Nitric Oxide Donor and Its Rescue by Neurotrophic Factors in Rats

Abstract

purpose. To investigate the neurotoxic outcome in the rat retina exposed to nitric oxide (NO) released from an NO donor and to evaluate the effects of neurotrophic factors on the survival of NO-damaged retinal cells.

methods. An NO releasing compound, N-ethyl-2-(1-ethyl-2-hydroxy-2-nitrosohydrazino) ethanamine (NOC 12), was intravitreously injected into a rat’s right eye. The influences of NOC 12 on retinal neurons and the neuroprotective effects of ciliary neurotrophic factor (CNTF) or brain-derived neurotrophic factor (BDNF) on NOC 12–mediated damage were estimated by counting cells in the ganglion cell layer (GCL) and by measuring the thickness of retinal layers. The exact count of retinal ganglion cells (RGCs) was also confirmed by means of retrograde labeling with a fluorescent tracer.

results. Morphometric analyses of retinal damage in the NOC 12–exposed eyes demonstrated a significant and dose-dependent decrease in cell density in the GCL and a reduction in thickness of the inner plexiform layer and inner nuclear layer, but not of the outer nuclear layer. TdT-dUTP terminal nick-end labeling of retinal sections after intravitreous injection of NOC 12 demonstrated that NO could trigger apoptotic cell death. The counting of the RGCs labeled with a fluorescent tracer suggested that a decrease in GCL cell density induced by NOC 12 reflects a loss in RGCs. Treatment with CNTF (1 μg) or BDNF (1 μg) before the intravitreous injection of NOC 12 (400 nmol) demonstrated that these trophic factors have protective effects against NO-induced neuronal cell death in the retina.

conclusions. Exogenous NO induces retinal neurotoxicity, suggesting that NO plays a pathogenic role in degenerative retinal diseases. BDNF and CNTF protect retinal neurons from NO-mediated neurotoxicity.

In several of the ophthalmic disorders, the mechanisms involved in retinal neuronal cell death are not well understood. However, elevation of intraocular glutamate levels followed by glutamate-receptor–mediated excitotoxicity is regarded as one of the important mechanisms in the pathogenesis of neurodegenerative diseases, such as diabetic retinopathy ^{1 2} and optic neuropathy. ³ N-Methyl-d-aspartate (NMDA) receptors are hypothesized to be a common pathway and a predominant route of glutamate-induced neurotoxicity in many neurodegenerative diseases. In general, activation of NMDA receptors by glutamate leads to an increase in intracellular Ca²⁺, which activates NO synthase (NOS). It has been reported that several neurons of the retina contain NOS ⁴ and expression of NOS is enhanced in the ischemic retina. ⁵ In addition, NOS inhibitors block NMDA-induced retinal damage ⁶ and ischemic injury. ⁷ Therefore, NO appears to play an important role in the retinal neurotoxicity mediated by NMDA receptors. Whether and how NO can induce retinal toxicity is an important question. However, in vivo evidence for the toxic effects of NO donors is quite limited. To our knowledge, there has been only one study that demonstrated in vivo retinal toxicity of an NO donor by intravitreous administration of S-nitro-N-acetyl-dl-penicillamine (SNAP; 200 nmol), assessed by electroretinogram and transmission electron microscopic observations in the rabbit retina. ⁸ Thus, we performed morphometric analysis of quantification of retinal injury by NO released from N-ethyl-2-(1-ethyl-2-hydroxy-2-nitrosohydrazino) ethanamine (NOC 12). 

Several neurotrophic factors such as nerve growth factor, ciliary neurotrophic factor (CNTF), and brain-derived neurotrophic factor (BDNF) have been reported to protect retinal neurons from damage caused by axotomy ^{9 10 11} and ischemia. ^{12 13} It has been hypothesized that upregulation of the expression of neurotrophic factors plays an essential role in the endogenous neuroprotective system. ^{14 15 16 17} In central nervous system neurons in vitro, neurotrophic factors have been shown to exert a protective effect against neurotoxicity caused by NO donors. ^{18 19} In the present study, NOC 12 elicited retinal neurotoxicity, and CNTF and BDNF counteracted retinal damage induced by NOC 12. 

Materials and Methods

Animals

Male Sprague-Dawley (SD) rats (7 weeks old; Nihon SLC, Shizuoka, Japan) were maintained in a humidity- (55% ± 10%) and temperature- (23 ± 2°C) controlled facility under a 12-hour light–dark cycle (lights on at 8:00 AM) with free access to food (MF chow pellets; Oriental Yeast, Tokyo, Japan) and water. The rats were acclimated for 1 week before experiments. 

All animal procedures were in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

NO Donor-Induced Retinal Cell Death

The following compounds were used: NOC 12 (Dojindo Laboratories, Kumamoto, Japan), recombinant rat CNTF (Genzyme, Cambridge, MA), and human BDNF (Peprotech Ec Ltd., London, UK). 

Each rat was anesthetized by an intraperitoneal injection of pentobarbital sodium (50–80 mg/kg), and then 1% atropine sulfate drops were applied to the right eye to generate a full dilation of the pupil. A single intravitreous injection of NOC 12 (20–1000 nmol in 5 μL of sodium phosphate buffer, containing 0.03 M NaOH) was performed on the right eye, using a 33-gauge needle connected to a 25-μL syringe (Hamilton, Reno, NV). Injection was performed slowly over a period of 1 minute. Histologic sections were prepared from both eyes 7 days after the NOC 12 injection. When we examined the effects of the neurotrophic factors, 5 μL NOC 12 was injected 2 days after an intravitreous injection of 3 μL recombinant rat CNTF or human BDNF (1 μg/3 μL phosphate-buffered saline; PBS). PBS (3 μL) was administered to the right eyes in the control group of rats. 

Morphometric Analysis

Seven days after injection of NOC 12, animals were killed and both eyes were enucleated. Eyes were immediately fixed in phosphate-buffered 4% formalin and 1% glutaraldehyde aqueous solution (pH 7.4), followed by phosphate-buffered 10% formalin solution (pH 7.4). After fixation, the eyes were embedded in paraffin and cut into six horizontal meridional 5-μm-thick sections through the optic disc of each eye. The sections were then stained with hematoxylin and eosin. 

Morphometric analysis was performed to quantify NOC 12–induced injury. Three sections were selected randomly from each eye. A microscopic image of each section within 1 mm of the optic disc was scanned digitally with the aid of an image-analysis system including a 3-charge-coupled device (CCD) camera module (XC-009; Nexus, Tokyo, Japan) and an image-analysis processor (nexusQube; Nexus). For assessment of the degree of injury in each eye, the number of cells in the ganglion cell layer (GCL) was enumerated within a 1-mm range of the optic disc. The thicknesses of the inner plexiform layer (IPL), inner nuclear layer (INL), and outer nuclear layer (ONL) in randomly selected sections were measured at three points with the use of image-analysis software (NIH Image; available by ftp from zippy.nimh.nih.gov/or from http://rsb.info.nih.gov/nih-image; developed by Wayne Rasband, National Institutes of Health, Bethesda, MD). The number of cells in the GCL was calculated by determining linear cell density (cells per millimeter). Finally, the thickness of the IPL, INL, and ONL and the linear cell density in the GCL were each expressed as the mean result of nine measurements. For each animal, these parameters for the right eye were expressed as a ratio (%) to those for the intact left eye. No attempt was made to distinguish the cell types in the GCL for enumeration of the cell number. Data are expressed as the mean ± SEM of results in two to eight animals. Statistical analyses were performed by the Dunnett’s test, and differences were considered significant at P < 0.05. 

Analysis for Retinal Ganglion Cells Loss

Retrograde labeling was performed in a manner similar to that described previously. ²⁰ Three days after an intravitreous injection of NOC 12, the animal was anesthetized by an intraperitoneal injection of pentobarbital sodium (50–70 mg/kg), and the head was placed firmly in a stereotaxic apparatus. A solution of 4% aminostilbamidine methanesulfonate (Molecular Probes, Eugene, OR) was microinjected bilaterally into the superior colliculus (−6.3 mm posterior to bregma, ±1.5 mm lateral to the midline, −3.6 mm from the surface of the skull). Four days after aminostilbamidine injection, the animals were killed, and both eyes were enucleated. The eyes were fixed for 45 minutes in 4% paraformaldehyde-PBS. The retina was divided into six radial cuts and then removed from the sclera and mounted on slides. The regions to be subjected to counting of retinal ganglion cells (RGCs) were selected from four fields of 376 × 505 μm of the central area (1–2 mm from the optic disc) and four fields of the peripheral area (4 mm from the optic disc) for each retina. The aminostilbamidine-labeled RGCs were counted with the aid of the image analysis software (NIH Image). 

TdT-dUTP Terminal Nick-End Labeling

Staining was performed according to a method described previously. ²¹ Twenty-four hours after NOC 12 injection, the animals were killed and their eyes were enucleated. Eyes were immediately fixed in 4% paraformaldehyde-PBS. After fixation, eyes were embedded in paraffin and cut into several horizontal meridional 5-μm-thick sections through the optic disc of each eye. TdT-dUTP terminal nick-end labeling (TUNEL) was performed with an apoptosis kit (Mebstain; MBL, Nagoya, Japan). 

Results

Dose-Dependent Neurotoxicity of NOC 12 in Rat Retina

NOC 12 was injected into the vitreous humor of rats to assess neurotoxicity of NO released by an NO donor in the rat retina. Figure 1 shows histopathologic images of representative retinas in the vehicle- or NOC 12 (400 and 1000 nmol)–treated eyes at 7 days after intravitreous injection. The NOC 12-induced retinal damage was confined to the inner area of the retina. The intravitreous injection of NOC 12 (200 nmol) resulted in significant cell loss in the GCL and thinning of IPL (Figs. 2A 2B) . Cells in INL exhibited relatively low sensitivity to NOC 12–induced cytotoxicity, and 600 nmol of NOC 12 was necessary to reduce the density of INL cells significantly (Fig. 2C) . In contrast, NOC 12 did not affect the thickness of ONL at doses up to 1000 nmol. No histologic alterations were observed after intravitreous injection of the vehicle solution. The selective vulnerability of the inner retinal layers in ischemic injury has been extensively investigated, ^{7 22} and retinal damage induced by an intravitreous injection of NMDA (200 nmol) gave rise to selective cell loss in the GCL and thinning of the IPL. ^{6 15 23} Thus, we used 400 nmol NOC 12 to elicit retinal damage (in the GCL and the IPL) in the following experiments. 

To determine whether the residue of NOC 12, which released NO, affects retinal cell survival, intravitreous injection was performed with 400 nmol NOC 12 solution that had undergone incubation for 7 days at 37°C, a procedure that permits the release of essentially all the NO from the compound (expended NOC 12). The expended NOC 12 exerted no effects either on cell counts in the GCL or on the thicknesses of the IPL, INL, and ONL (Fig. 3) . 

Characterization of NOC 12–Induced Retinal Cell Damage

In the rat retina, the GCL contains two neuronal populations: ganglion cells and displaced amacrine cells. ²⁴ To evaluate NOC 12-induced loss of RGCs in the GCL specifically, we analyzed the retinas isolated from animals that had received an aminostilbamidine injection to the superior colliculus for a retrograde labeling of RGCs. The number of aminostilbamidine-labeled RGCs in the retina treated with NOC 12 (200 nmol: 1288 ± 271 and 400 nmol: 801 ± 182 cells/mm², centrally) were significantly decreased in a dose-dependent manner, compared with the number in the vehicle-treated retina (1836 ± 101 cells/mm², centrally; (Fig. 4) . The cell density in the vehicle-treated retina was consistent with the findings of previous studies in which the mean density of the labeled RGCs in the normal retina was 2342.3 ± 103.4 cells/mm^{2 23} or 1746 ± 45 cells/mm². ¹³ The degree of retinal damage assessed by tracer labeling was in good agreement with the results obtained by histologic analysis of retinal cross-sections (Fig. 2) . 

To examine whether NOC 12-induced retinal neuronal cell death is apoptotic, we conducted TUNEL staining of the retina 24 hours after intravitreous injection of 400 nmol NOC 12. As shown in Figure 5 , several TUNEL-positive cells were observed in the GCL and in the inner side of the INL (probably amacrine cells), but not in the ONL. This result suggests that NO released from NOC 12 triggers apoptotic cell death in these layers. No TUNEL-positive cells were found in the retinas of the vehicle-treated control group. 

Neuroprotective Effects of CNTF and BDNF against NOC 12–Induced Retinal Damage

It has been reported that pretreatment with CNTF (1 μg) and BDNF (1 μg) protects the inner retinal cells from NMDA-mediated neuronal death. ^{14 22} To examine the potential neuroprotective effects of these neurotrophic factors against NO-mediated retinal damage, animals were given an intravitreous injection of CNTF or BDNF. At 2 days after the intravitreous injection of CNTF (1 μg) or BDNF (1 μg), 400 nmol of NOC 12 was injected into the eyes of the same animals. PBS was injected instead of a neurotrophic factor–containing solution in several rats as a control. 

Pretreatment with intravitreous CNTF (1 μg) resulted in a significant amelioration of cell loss in the GCL and tended to improve the thinning of the IPL induced by NOC 12 (Fig. 6A) . In addition, pretreatment with BDNF (1 μg) ameliorated the decrease in cell density in the GCL and thinning of the IPL to statistically significant levels (Fig. 6B) . 

Discussion

Elevation of vitreous glutamate levels has been reported in several ophthalmic disorders, including optic neuropathy and ischemia. ^{1 2 3 25} Abnormal levels of glutamate are likely to result in extensive neurodegeneration mediated by NMDA receptors, followed by a series of insults including activation of NOS. In fact, administration of NOS inhibitors or deficiency of neuronal NOS diminishes NMDA-induced retinal damage and RGCs degeneration after axotomy in addition to ischemic injury. ^{6 7 26 27} Although NO released by NO donors leads to the death of cultured retinal neurons, ²⁸ little is known about the retinal neurotoxicity of NO donors in vivo. 

The present study demonstrated that an intravitreous injection of NOC 12, an NO donor, resulted in a significant and dose-dependent cell loss in GCL and thinning of IPL and INL, without any effect on ONL. Additional experiments by the method of retrograde labeling of RGCs with fluorescent dye further confirmed that NOC 12 causes a marked loss of RGCs in GCL. Some studies have shown that NO leads to axonal degeneration. ²⁷ The decrease in retrograde labeling of RGCs due to NO-induced axonal degeneration after intravitreous injection of NOC 12 may partially involve RGC death induced by NOC 12 observed in the present study. Expended NOC 12 did not damage the retinal structure, suggesting that NOC 12–induced damage is attributable to the action of NO released from the compound. The cell loss in the GCL and thinning of IPL were reported in rat eyes that had undergone ischemic injury or received an intravitreous injection of NMDA. ^{6 7 15} Some neuronal (ganglion and amacrine) cells in the retina contain NOS and express different NOS isoforms on ischemic damage. ^{4 5 29} The inner retinal layers injured by NOC 12 were consistent with the area susceptible to glutamate toxicity and with the area of NOS localization. 

NOC 12, used in the present study, is known to release large amounts of NO spontaneously, compared with typical NO donors such as sodium nitroprusside (SNP). ^{30 31} Some typical NO donors are known to have disadvantages including cofactor requirements for NO release and biological activity in themselves. For example, SNP has to be metabolized in the presence of cofactor and its effects involve the biological activity of cyanoferrate. ³² In contrast, NOC 12 is a relative slow NO-releasing compound with a half-life of 327 minutes (in PBS, pH 7.4, 22°C), which requires no cofactors, and is easy to handle, because it remains stable in a solid state at room temperature and cannot release NO in alkaline solutions. ³³ Therefore, the present study was performed to assess neurotoxicity of NO in the rat retina by using NOC 12. 

In one study, the investigators reported that intravitreous injection of SNAP (200 nmol), an NO donor, resulted in a decrease in cell density in the GCL and IPL in albino rabbit retina, when assessed by electroretinogram and transmission electron microscopic observations. ⁸ However, several other studies failed to demonstrate retinal toxicity of NO donors. ^{34 35} Similar experiments involving SNAP (1000 nmol) treatment in pigmented rabbit eye and SNP (5 pmol) in the rat eye did not demonstrate any retinal damage. ^{34 35} These discrepant results may be due to the differences in the amounts and time course of the release of NO by NO donors. The large amounts of the radical form of NO, adequate to interact with oxygen radicals, may be required to induce retinal damage, because NO per se has some pharmacologic functions, such as lowering the intraocular pressure, and increasing retinal blood flow. ³⁴  

Several TUNEL-positive cells were detected in NOC 12–treated retina, in the GCL and in the inner side of the INL (probably amacrine cells), but not in the ONL. Although NO can cause cell death through either necrosis or apoptosis, ³⁶ retinal cell death in cases of ischemia, axotomy, and injection of NMDA have been shown to take the course of apoptosis. ^{37 38 39} The present results support the idea that NO produced by stimulation of NOS triggers apoptotic cell death in various types of ocular diseases and that NO-induced retinal injury may be prevented by procedures interfering with the signaling pathway of apoptosis. Because the TUNEL method also detects necrotic cells, the possible involvement of necrosis induced by NO should be taken into consideration. Morphometric analysis of retinal damage in the INL was determined by measuring the thickness, because accurate counting of cell numbers in the INL was difficult. Therefore, it is likely that the loss of TUNEL-positive cells in the inner side of the INL (probably amacrine cells) did not cause the apparent reduction in thickness of the INL. 

In the current study, CNTF and BDNF prevented retinal cytotoxicity induced by NOC 12. The present findings are in agreement with those of several studies in which a similar treatment with CNTF and BDNF (dose; 1 μg/eye, pretreatment 2 days before injury insult) significantly protected the inner retinal cells from neuronal death induced by NMDA or elevated intraocular pressure. ^{12 15 23} Some studies have shown that BDNF has a limited neuroprotective action with a trend toward a bell-shaped dose–response curve in axotomized RGCs. ^{40 41} This phenomenon is interpreted as a consequence of upregulation of retinal NOS activity. ^{41 42} These findings suggest that increased production of NO due to BDNF treatment exacerbates excitotoxic RGC death, thereby limiting the neuroprotective effect of BDNF. The pathologic consequences of BDNF may also be relevant to the modest protective effect of BDNF observed in the present study. The protective effects of these neurotrophic factors against NO-induced neurotoxicity were investigated in primary cultures of rat hippocampal neurons ⁴³ and of cortical neurons. ¹⁸ Although there is no in vitro evidence showing protective effects of CNTF or BDNF on NO-induced toxicity in the retina, CNTF, BDNF, and their receptors (CNTF receptor [R]α, TrkB, and p75) have been reported to be present in cells in the normal retina of adult rats and also in the retina exposed to neurodegenerative insults such as ischemia and elevation of intraocular pressure. ^{14 15 16 44 45 46 47 48} The protective actions of CNTF and BDNF may involve their binding to CNTFRα and TrkB or p75. 

In conclusion, in the present study NO released by NOC 12 triggered the death of inner retinal neurons in rats and CNTF and BDNF protected retinal neurons from NO-mediated neuronal death. These results suggest that NO plays a pathogenic role in degenerative retinal diseases and that the supply of neurotrophic factors represents an important strategy for neuroprotective treatment against NO-mediated neuronal injury. 

 Supported in part by a Grant-in-Aid for Scientific Research from Japan Promotion of Science.

 Submitted for publication May 16, 2002; revised September 10 and November 26, 2002; accepted December 6, 2002.

 Commercial relationships policy: E (KT, SM, FY); N (all others).

 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.

 

Light micrographs of a representative vehicle and NOC 12–treated eyes at 7 days after intravitreous injection. Photomicrographs of hematoxylin-eosin–stained, representative retinal sections from eyes that received an injection of vehicle (A) or 400 (B) or 1000 (C) nmol NOC 12. The intravitreous injection of NOC 12 led to cell loss in the inner layers of the retina. Scale bar, 20 μm.

Figure 2.

 

Morphometric analysis of NOC 12–mediated retinal damage on day 7 after treatment. The degree of NOC 12–induced retinal damage was quantified by a linear count (per millimeter) of cells present in the GCL (A) and by measuring the thicknesses (in micrometers) of the IPL (B), INL (C), and ONL (D) at a distance of 1 mm from the optic disc. For each animal, these parameters for the experimental right eye were expressed as a ratio (%) of those for the intact left eye. The intravitreous injection of NOC 12 (20–1000 nmol) caused cell loss in the GCL and thinning of IPL and INL in a dose-dependent manner. Data represent the average results in one group of two to eight rats (number on each bar) ± SEM; *P < 0.05; **P < 0.01 versus the vehicle-treated control group.

Figure 3.

 

Influence of expended NOC 12 in experimental eyes 7 days after intravitreous injection. To prepare expended NOC 12, 400 nmol NOC 12 was incubated for 7 days at 37°C. The expended NOC 12 did not damage the retinal structure. Data represent an average ± SEM of results in one group of six rats.

Figure 4.

 

Loss of aminostilbamidine-labeled RGCs at 7 days after intravitreous injection of NOC 12 (200–400 nmol). (A, B) Fluorescence micrographs of representative retinal sections from eyes that received an injection of vehicle (A) or 400 nmol (B) NOC 12. Scale bar, 200 μm. (C) Number of RGCs per square millimeter in the retina treated with vehicle or NOC 12 (200–400 nmol). The intravitreous injection of NOC 12 caused RGC loss in a dose-dependent manner. Data represent the average ± SEM of results in one group of four rats; **P < 0.01 versus the central area of the vehicle-treated control group; #P < 0.05, ##P < 0.01 versus the peripheral area of the vehicle-treated control group.

Figure 5.

 

TUNEL staining of a representative retina from the experimental eyes at 24 hours after intravitreous injection of vehicle (A, B) and NOC 12 (400 nmol; C, D). Micrographs show bright-field (A, C) and TUNEL staining (B, D). In the NOC 12–treated retina (D), several TUNEL-positive cells were detected in the GCL and in the inner side of the INL (probably amacrine cells). Scale bar, 100 μm.

Figure 6.

 

Effects of CNTF and BDNF on NOC 12–induced retinal damage. CNTF (1 μg) or BDNF (1 μg) was administered 2 days before an intravitreous injection of 400 nmol NOC 12. The protective effect of CNTF (A) and BDNF (B) against NOC 12 toxicity was quantified by linear counting of cells per millimeter in the GCL and by measuring the thickness (in micrometers) of IPL and INL at a distance of 1 mm from the optic disc. For each animal, these parameters for the experimental right eye were expressed as a ratio (%) to those for the intact left eye. Both neurotrophic factors ameliorated a decrease in the cell density of the GCL and the thinning of the IPL. Data represent the average ± SEM in one group of three to five rats; *P < 0.05, **P < 0.01 versus the PBS+NOC 12–treated control group.