N-Methyl-D-Aspartate (NMDA)–Induced Apoptosis in Rat Retina

Abstract

purpose. The involvement of apoptosis in N-methyl-D-aspartate (NMDA)–induced excitotoxicity in adult rat retinas was examined.

methods. Excitotoxic loss of inner retinal elements was induced by intravitreal injections of various concentrations of neutralized NMDA in adult albino Lewis rats. Tissue responses were quantified by measuring the inner retinal thickness (IRT) in plastic sections of the retinas and cell counts in the retinal ganglion cell layer in flatmount preparations of the whole retinas. Internucleosomal DNA fragmentation, a hallmark of apoptosis, was assayed with agarose DNA gel electrophoresis. The in situ TdT-mediated biotin-dUTP nick end labeling (TUNEL) method was used to locate nicked DNA in paraffin sections of the retinas. Ultrastructural changes of the degenerating cells were examined by electron microscopy. The efficacy of Ac-Tyr-Val-Ala-Asp-CMK (YVAD–CMK), a peptidyl caspase inhibitor, and 3-aminobenzamide (ABA), an inhibitor of poly(ADP-ribose) polymerase (PARP), in ameliorating the loss of inner retinal elements was evaluated using morphometry to examine the apoptotic pathways.

results. Intravitreal injection of NMDA induced a dose-dependent loss of inner retinal elements as evidenced by the measurements of IRT and RGCCs. There were time- and dose-related appearances of internucleosomal fragmentation of retinal DNA and a time-related appearance of TUNEL-positive nuclei in the inner retinas after intravitreal NMDA injection. Ultrastructural features consistent with classic apoptotic changes were noted in degenerating cells in the retinal ganglion cell layer and the inner nuclear layer. Control retinas given vehicle, N-methyl-L-aspartate (the L-isomer of NMDA), or NMDA plus MK-801, a specific antagonist, did not show these changes. Simultaneous administration of NMDA and YVAD–CMK or ABA abolished or attenuated the loss of RGCCs in the posterior retinas.

conclusions. NMDA-induced excitotoxicity involved apoptosis and caspases and PARP may play important roles in the pathways.

Cell death may be conveniently classified into two broad categories: necrosis and apoptosis. ^{1 2 3 4 5 6 7} In necrosis, the cells show early plasma membrane changes, clumping of chromatin, swelling of intracellular organelles, membrane breakdown, and leakage of enzymes and proteins causing extensive inflammatory reactions. ^{1 2} In addition, necrotic cells appear in patches. In contrast, apoptotic cells are characteristically scattered throughout the tissue and initially show condensation of chromatin at the nuclear periphery and reduction of nuclear size and cell volume. ^{1 2 3 4 5 6 7} Genomic DNA is cleaved at internucleosomal regions, resulting in fragments that are multiples of 180 to 200 bp, giving a ladder pattern after gel electrophoresis. ⁷ Subsequently, buttons of nuclear cytoplasm are fragmented into multiple small membrane-bound units known as apoptotic bodies, which may be phagocytosed by neighboring cells. There is minimum inflammatory reaction in the tissue. ^{3 4 5}  

Excitotoxicity, neuronal cell death caused by excitatory neurotransmitters such as glutamate, is linked to stroke, hypoglycemia, trauma, epilepsy, and chronic neurodegenerative diseases such as Huntington’s disease, the acquired immunodeficiency syndrome (AIDS) dementia complex, amyotrophic lateral sclerosis, and Alzheimer’s disease. ^{8 9} In the retina, excitotoxicity is believed to play an important role in retinal ischemia/reperfusion injury ^{10 11 12 13} and more recently in neuronal loss in glaucoma. ¹⁴ However, there are contradicting reports regarding the involvement of apoptosis in excitotoxicity of cerebral neurons. ^{15 16 17 18 19 20 21 22 23}  

In this study, we examined the possible involvement of apoptosis in the excitotoxin N-methyl-D-aspartate (NMDA)–induced neuronal cell death in adult rat retinas. At various times after intravitreal injection of different concentrations of NMDA, internucleosomal cleavage of genomic DNA was examined by agarose gel electrophoresis and the localization of nicked DNA by in situ labeling with the TdT-mediated biotin-dUTP nick end labeling (TUNEL) method. Ultrastructural changes characteristic of apoptosis were noted. In addition, the efficacy of two modulators of the apoptotic pathways was evaluated to probe the pathway or pathways and to explore a new therapeutic approach to excitotoxic retinal cell death using morphometry in the evaluation of retinal changes. 

Methods

Administration of NMDA

The procedure by Siliprandi et al. ²⁴ was modified to induce excitotoxic cell loss in rat retinas. Briefly, adult male albino Lewis rats (45–50 days old) were anesthetized with an intraperitoneal injection of 400 mg/kg chloral hydrate. After the application of topical 0.5% proparacaine hydrochloride, the animals were given intravitreal injections of 2 μl of 0.2, 1, 4, 10, or 40 mM (corresponding to 0.8, 2, 8, 20, and 80 nmoles, respectively) of neutralized NMDA (Sigma, St. Louis, MO) in 0.1 M phosphate buffered saline (PBS). For controls, the same volume of 0.1 M PBS or 40 mM neutralized N-methyl-L-aspartate (NMLA), the L-isomer of the excitotoxin, or 40 mM neutralized NMDA plus 1 mg/ml MK-801, was given intravitreally to different groups of animals. There were at least 6 animals per each group of experimental conditions. All solutions for injection had a final pH of 7.4. Animals were allowed to recover and euthanatized with an overdose of pentobarbital at various times after the intravitreal injection. The eyes were enucleated and processed for further analysis. All use and handling of animals adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

DNA Fragmentation Studies by Gel Electrophoresis

Retinal DNA analysis was performed using a previously described procedure. ²⁵ Briefly, after enucleation, the retinas were dissected from the retinal pigment epithelium and choroid, frozen in liquid nitrogen, and stored at –80°C. Retinal DNA was extracted using the standard phenol/chloroform/isoamyl alcohol (25:24:1) method. The concentrations of DNA in each sample were determined by measuring the absorbance at 260 nm. Samples of 10 μg of DNA were loaded in each well of a 2.0% agarose gel. Electrophoresis was performed at 3 to 5 V/cm. DNA was stained with a 1:10,000 dilution of SYBR Green I (Molecular Probes, Eugene, OR) in Tris borate EDTA (TBE) buffer (89 mM Tris base, 89 mM boric acid, 1 mM EDTA, pH 8), visualized by transillumination with UV light, and photographed. 

In Situ Labeling with the TUNEL Method

The retinas were fixed in Davidson’s fixative (32% ethanol, 2.2% neutral buffered formalin, and 11% glacial acetic acid in distilled water) and embedded in paraffin. The deparaffinized sections were washed in double distilled water for 2 minutes four times. Endogenous peroxidase was inactivated by covering the sections with 2% H₂O₂ for 5 minutes at room temperature. TUNEL was performed using the ApopTag kit from Oncor (Gaithersburg, MD). 

Morphometry

Morphometric measurements were performed according to previously published procedures. ²⁶  

Morphology and Inner Retinal Thickness (IRT) Measurements.

The enucleated eyes of the experimental animals were opened and fixed in 2.5% glutaraldehyde in 0.1 M cacodylate buffer overnight. After the anterior segments were removed, four strips of the retina (each from the nasal, inferior, temporal, or superior quadrant) were sampled. The tissue samples were osmicated, dehydrated, and embedded in epoxy resin for sectioning and morphologic and morphometric studies. Injury of the inner retinal layers was evaluated quantitatively by measuring the thickness between the inner limiting membrane and the interface of the outer plexiform layer and the outer nuclear layer in plastic sections. Readings from all quadrants were averaged to obtain the value for one eye. 

Flat Preparation of the Retina and Cell Counts in the Retinal Ganglion Cell Layer (RGCL).

Flat preparations of the retinas and morphometry of cells in the RGCLs were performed to evaluate cell loss according to published procedures. ²³ Briefly, after fixation of the enucleated eyes, and washing in 0.1 M phosphate buffer (pH 7.4), the retinas were separated and mounted on a slide with the vitreous facing up, using a 1% gelatin solution. Several radial cuts were made at the peripheral retinas, and the retinal surfaces were flattened with a fine brush. The sections were dried, stained with 1% cresyl violet for 3 to 5 minutes, dehydrated, and covered with coverslips. The numbers of nuclei at the RGCL of the posterior pole (approximately 500 μm from the center of the optic disc) and the peripheral retina (approximately 4 mm from the center of the optic disc) per unit area were taken with an eyepiece reticule of a microscope at ×400 magnification. Counts were taken from comparable areas of the nasal, inferior, temporal, and superior quadrants of the retinal flat preparation. Numbers from the four quadrants of an eye were averaged to give the value for one eye. No attempt was made to distinguish the types of retinal ganglion cells (RGCs) and displaced amacrine cells. Morphologically distinguishable glial cells and vascular endothelial cells were excluded from the cell count. 

All numbers represented the mean ± SD, with a minimum n = 6. One-way ANOVA and Tukey’s test were used for statistical analysis for all morphometric measurements. 

Apoptotic Modulators Study

The peptidyl caspase inhibitor Ac-Tyr-Val-Ala-Asp-CMK (YVAD–CMK) (Bachem Bioscience, King of Prussia, PA) or an inhibitor of poly(ADP-ribose) polymerase (PARP), 3-aminobenzamide (ABA; Sigma), was dissolved in 4 mM neutralized NMDA solution. 

Groups of animals were given intravitreal injections of 2 μl 4 mM NMDA solution containing 0.25 mM YVAD–CMK, ²⁵ or 10 mM ABA. ²⁷ One group of animals was injected with 2 μl 4 mM NMDA intravitreally, and another group was used as untreated normal controls. One eye from each animal was used for measurement of IRT, and the other eye was used for retinal ganglion cell counts (RGCCs). 

Results

There were dose-dependent losses of IRT (Fig. 1 A) and RGCCs of the posterior (Fig. 1B) and peripheral (Fig. 1C) retinas when measured at 7 days after intravitreal injection of NMDA. Statistically significant (P < 0.05) losses in all three parameters (IRT and RGCCs at posterior and peripheral retinas) were noted at 8 nmoles NMDA and above. These losses were antagonized by the simultaneous injection of MK-801, a specific antagonist of the NMDA receptor. The vehicle (0.1 M PBS) or NMLA, the L-isomer of the excitotoxin NMDA, did not cause noticeable loss. Hence, the measured losses of inner retinal elements were due to a NMDA receptor–mediated event. Agarose gel electrophoresis of retinal DNA obtained at various times after the injection of 20 nmoles of NMDA (Fig. 2 A) showed a time-dependent appearance of the typical ladder pattern of internucleosomal fragmentation, a characteristic of apoptosis. A faint ladder pattern was barely noticeable at 12 hours, and a maximum band intensity was noted at 24 hours. The band intensities of the ladder also appeared to be dose dependent (Fig. 2B) when retinas were examined at 24 hours after the injection of various doses of NMDA. The appearance of the ladder pattern was abolished with the simultaneous injection of MK-801. NMLA (80 nmoles) did not produce a ladder pattern. Hence, formation of the ladder pattern was also the result of a NMDA receptor–mediated event. 

TUNEL showed a similar time course of appearance of positively labeled nuclei at the RGCL and the inner nuclear layer (INL; Fig. 3 ). Intense labeling of nuclei especially at the inner part of the INL was noted between 12 and 24 hours postinjection, corresponding to the time of maximum band intensity of the ladder pattern in the DNA gel analysis. 

Ultrastructural studies showed a shrunken nucleus with condensed chromatin and densified cytoplasm in the RGCL (Fig. 4 A) at 18 hours after injection with 8 nmoles of NMDA. Similarly, in the INL, cells with clumping of chromatin at the nuclear membrane and dense cytoplasm were noted (Fig. 4B) . At a more advance stage, dense and shrunken nuclei, invaginated nuclear membrane, condensed chromatin, and dense cytoplasm were seen (Figs. 4C 4D) . These morphologic changes are consistent with apoptosis. However, adjacent to these typical apoptotic cells, swollen cells could be seen (Fig. 4D) . 

Simultaneous administration of NMDA with YVAC–CMK, the peptidyl caspase inhibitor, completely abolished the reduction in RGCCs in the posterior (Fig. 5 A) and peripheral (Fig. 5B) retinas. However, ABA only attenuated the reduction in posterior RGCCs (Fig. 5A) , with no significant effect seen in the peripheral RGCCs (Fig. 5B) . 

Discussion

In the present study, we demonstrated a dose-related loss of inner retinal elements with intravitreal injections of NMDA, a time- and dose-related appearance of internucleosomal fragmentation of retinal DNA after NMDA injection, a time-related appearance of TUNEL-positive nuclei in the inner retina, and ultrastructural changes of inner retinal cells consistent with apoptotic changes after NMDA injection. These changes are NMDA receptor–mediated and suggest a prominent role for apoptosis in NMDA-induced loss of retinal elements. The caspase inhibitor YVAD–CMK abolished the loss in posterior and peripheral RGCCs when given simultaneously with NMDA, suggesting important roles for the caspases in the apoptotic pathways of NMDA-induced neuronal death. The partial effect of the PARP inhibitor ABA also suggests a possible involvement of PARP in NMDA-induced apoptotic cell death. 

Our findings in the dose-dependent losses of IRT and RGCCs were consistent with those of Sabel et al., ²⁸ who reported a significant loss of cells in the RGCL, and of Siliprandi et al., ²⁴ who illustrated a dose-dependent reduction of inner plexiform layer at 8 days after administration of 20 to 200 nmoles NMDA. Sabel et al. ²⁸ observed a significant loss of horseradish peroxidase–positive cells, presumed to be retinal ganglion cells (RGCs), after the administration of NMDA. Siliprandi et al. ²⁴ showed that the loss of RGCs with a size bigger than 8 μm in diameter was more severe than that of RGCs of smaller size and suggested that cholinergic amacrine cells might be vulnerable to NMDA excitotoxicity in a dose-dependent manner. However, Dreyer et al. ²⁹ also revealed that the larger size RGCs were more susceptible to NMDA. In the present study, we did not distinguish the type or size of the remaining cells in our study. Hence, whether apoptosis is restricted to certain types of the inner retinal neurones is not clear. However, it seems that neurones at the inner part of the INL showed more TUNEL-positive nuclei, suggestive of the involvement of amacrine cells of the INL. 

Kure et al. ¹⁵ reported that endonuclease was activated in glutamate excitotoxicity and that it was responsible for the internucleosomal DNA cleavage both in vivo and in vitro. Activation of endonuclease with subsequent internucleosomal DNA fragmentation is believed to be the hallmark of apoptosis. Consistent with the observation of Kure et al, we also noted the time- and dose-dependent appearance of internucleosomal DNA fragmentation after the administration of NMDA. The lack of the ladder pattern in NMLA-treated retinas, as well as in retinas with simultaneous administration of NMDA and MK-801, suggests that the apoptotic change is a NMDA receptor–mediated event. 

The TUNEL technique provided in situ visualization of nicked DNA in a single cell. ³⁰ In the present study, the nicked DNA was noted in the RGCL and INL. The locations of these labeled cells were similar to those of the excitotoxic dying cells (densified nuclei in the RGCL and INL) found in this study and previous reports. ^{24 30 31 32 33} These cells were suggested by previous reports to be the RGCs and displaced amacrine cells in the RGCL and amacrine cells in the INL, especially the cholinergic neurones. Our study also illustrated that there was no evidence of TUNEL in the photoreceptor cell. Thus, excitotoxicity appeared to be confined to the inner retina. 

After intravitreal injection of NMDA, we noted some pyknotic nuclei and nuclei with clumping of the nuclear chromatin at the nuclear periphery, invagination of nuclear membrane, and nuclear fragments in the retina. These nuclear changes may be accompanied by rapid cell volume reduction together with densified cytoplasm, convolution of cell membrane, and compaction of cytoplasmic organelles. These morphologic features are characteristic of apoptosis. ^{34 35 36 37} However, as shown in Figure 4D , some morphologic features such as swollen cytoplasm with vacuolations and granular nuclei in some degenerating cells were different from the classic apoptotic features and were classified as necrotic features by Olney ³² and Sisk and Kuwabara. ³³ However, these morphologic features may be a variant of apoptotic changes. ^{38 39} In Clarke’s classification of apoptotic neurones during development, degenerating cells of type 3A show initial swelling of intracellular organelles, formation of empty spaces in the cytoplasm and further fusion of these spaces to extracellular cavity, and disintegration of cellular structures into smaller pieces without autophagic or heterophagic activity, whereas type 3B, known as the “cytoplasmic” type of degeneration, resembles type 3A in that there are dilated organelles and vacuoles in the cytoplasm,but differs in that the cell membrane retracts and “rounds up” and the nucleus becomes karyolytic or edematous. ³⁸ Zakeri et al. ³⁹ also observed vacuolization of the cytoplasm and delayed collapse of the nucleus with endonucleolytic cleavage of DNA in rat mammary gland. More recent studies by Portera–Cailliau et al. ⁴⁰ also have suggested an apoptosis–necrosis continuum in morphologic changes in excitotoxicity. Hence, it is possible that the observed nonclassic apoptotic changes may be a variant of apoptotic features but may also reflect a mixture of necrosis and apoptosis in excitotoxic cell death. 

According to Choi, ⁴¹ it is important to apply multiple criteria to define apoptotic cell death, including the use of inhibitors of apoptosis. In this study, we examined the effects of a caspase inhibitor and a PARP inhibitor on the loss of retinal elements after NMDA injection. In vitro studies ^{25 42 43 44 45} have implicated a pivotal role of ICE-like enzymes or caspases in apoptosis and suggested that caspases may be targeted for therapy. However, there are few in vivo reports on the use of caspase inhibitors in apoptosis. Milligan et al. ⁴⁶ reported that caspase inhibitors prevented apoptosis during the development of interdigital (nonneuronal) cells of the limbs in vivo. In the present study, a protective effect of YVAD–CMK, a broad-spectrum peptidyl caspase inhibitor, on RGCL cells in NMDA-induced excitotoxic damage was observed. This is supportive of a major role for apoptosis in NMDA-induced excitotoxicity and of an important role for caspases in the apoptotic cell death after intravitreal injection of NMDA in vivo. There are at least 12 known caspases, and many of them are involved in apoptosis. ⁴⁷ Because of the broad-spectrum activity of YVAD–CMK, it is not clear which of them is involved. 

We also demonstrated that ABA, a PARP inhibitor, provided a partial protection of RGCs against NMDA excitotoxicity in vivo. This finding was consistent with earlier studies showing that inhibition of PARP was beneficial in preventing cell death as reported by Zhang et al. ⁴⁸ or when interleukin-1β-converting (ICE)-like enzymes were activated to cleave PARP into fragments, enhancing DNA degradation in apoptotic cell death. ⁴⁹ Durkacz et al. ⁵⁰ also demonstrated that ABA is a specific inhibitor of PARP by preventing the nicotinamide adenine dinucleotide (NAD⁺) decrease completely. This decrease in NAD⁺ is believed to cause cell death. ⁵¹ However, our study only showed a partial effect in the posterior retina and no effect in the peripheral retina. This may be due to differences in the pathways of cell death at different locations of the retina or may be a pharmacodynamic or pharmacokinetic problem, or both. 

In summary, we provided morphologic, biochemical, and pharmacological evidence of NMDA receptor–mediated apoptosis in the inner retinal cells in adult rat retinas. These findings showed similarity to those of our retinal–reperfusion study, ²⁵ suggesting a pivotal role for NMDA-mediated excitotoxicity in retinal ischemia–reperfusion injury. In addition, these studies demonstrated that it was feasible to protect neurons from excitotoxic death by modulating the apoptotic mediators. This may lead to new strategies on neuroprotection. Whether these findings will lead to therapeutic agents for the glaucomatous loss of retinal neurons due to excitotoxicity is not clear, but further studies are warranted. 

 Supported in part by a Campus Research Grant, University of Illinois at Chicago; and Direct Grant P204-0554 from the Chinese University of Hong Kong.

 Submitted for publication March 3, 1999; accepted April 12, 1999.

 Proprietary interest category: N.

 Corresponding author: Tim T. Lam, Department of Ophthalmology and Visual Sciences, Hong Kong Eye Hospital, 147K Argyle Street 3/F, Kowloon, Hong Kong. E-mail: [email protected]

 

Figure 1.

 

Dose-dependent loss of inner retinal elements at 7 days postinjection. IRT (A) and retinal ganglion cell count (RGCC) at the posterior (B) and peripheral (C) retinas. Noticeable losses at 2 nmoles and significant losses (P < 0.05, Tukey’s test) at 8 to 80 nmoles. The losses induced by 80 nmoles NMDA were totally abolished by the presence of 1 mg/ml MK-801. Vehicle and 80 nmoles of NMLA, the L-isomer of NMDA, showed no effect.

Figure 2.

 

Agarose gel electrophoresis of retinal DNA. (A) Time-dependent appearance of the ladder pattern. Recognizable ladder patterns were noted at 18 hours and maximum band intensity at 24 hours. (B) Dose-dependent appearance of the ladder pattern. At 24 hours after intravitreal injection of various amounts of NMDA, a dose-dependent appearance of the ladder pattern was noted. The ladder pattern was noticeable with 8 nmoles, and its intensity appeared to increase at 20 and 80 nmoles. However, with 80 nmoles NMLA (lane L) or simultaneous injection of 20 nmoles NMDA and 1 mg/ml MK-801 (lane MK), no ladder pattern was noted.

Figure 3.

 

TUNEL of retinas at various times after 20 nmoles NMDA injection. (A) Control normal retina and retinas at 4 (B), 8 (C), 12 (D), and 18 (E) hours. Scattered and intensely labeled nuclei are seen in the RGCL and the INL. (F) Retinas at 24 hours. Few positive nuclei are seen in the INL. Retinas at 36 (G) and 48 (H) hours. No noticeable labeled nuclei are seen in the RGCL and the INL. Scale bar, 25 μm.

Figure 4.

 

Ultrastructural changes of degenerating cells at the RGCL (A) and the INL (B, C, and D). (A) RGCL: Note dense chromatin (arrow), condensed nucleus, and dense but vacuolated cytoplasm (arrowhead). (B) INL: Note clumping of chromatin at the nuclear membrane (arrow) and dense cytoplasm (arrowhead). (C) INL: A shrunken cell with dense nucleus (arrow), cytoplasm (arrowhead), and vacuolation. (D) INL: Cells showing typical invaginated nuclear membrane (arrowhead) or dense nuclei (arrow) were present, with an adjacent cell showing swollen cytoplasm and nucleus (∗).

Figure 5.

 

Effects of YVAD–CMK (YVAD) and ABA on RGCCs after NMDA administration. YVAD–CMK antagonized the effects of NMDA on the posterior (A) and peripheral (B) retinas, whereas ABA only partially antagonized its effect on the posterior retina and had no effect on the peripheral retina.