Abstract
We investigated the role of neuronal (type I) nitric oxide synthase (nNOS) in NMDA-mediated excitotoxicity in wild-type (SV129 and C57BL/6J) and type I NOS knock-out (nNOS−/−) mice and examined its relationship to apoptosis. Excitotoxic lesions were produced by intrastriatal stereotactic NMDA microinjections (10–20 nmol). Lesion size was dose- and time-dependent, completely blocked by MK-801 pretreatment, and smaller in nNOS knock-out mice compared with wild-type littermates (nNOS+/+, 11.7 ± 1.7 mm3; n = 8; nNOS−/−, 6.4 ± 1.8 mm3;n = 7). The density and distribution of striatal NMDA binding sites, determined by NMDA receptor autoradiography, did not differ between strains. Pharmacological inhibition of nNOS by 7-nitroindazole (50 mg/kg, i.p.) decreased NMDA lesion size by 32% in wild-type mice (n = 7). Neurochemical and immunohistochemical measurements of brain nitrotyrosine, a product of peroxynitrite formation, were increased markedly in wild-type but not in the nNOS−/− mice. Moreover, elevations in 2,3- and 2,5-dihydroxybenzoic acid levels were significantly reduced in the mutant striatum, as a measure of hydroxyl radical production.
The importance of apoptosis to NMDA receptor-mediated toxicity was evaluated by DNA laddering and by quantitative histochemistry [terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate–biotin nick end-labeling (TUNEL) staining]. DNA laddering was first detected within lesioned tissue after 12–24 hr. TUNEL-positive cells were first observed at 12 hr, increased in number at 48 hr and 7 d, and were located predominantly in proximity to the lesion border. The density was significantly lower in nNOS−/− mice. Hence, oligonucleosomal DNA breakdown suggesting apoptosis develops as a late consequence of NMDA microinjection and is reduced in nNOS mutants. The mechanism of protection in nNOS−/− mice may relate to decreased oxygen free radical production and related NO reaction products and, in part, involves mechanisms of neuronal death associated with the delayed appearance of apoptosis.
- NMDA
- excitotoxicity
- striatum
- neuronal nitric oxide synthase
- knock-out mice
- nitrotyrosine
- hydroxyl radical
- apoptosis
- DNA laddering
- TUNEL staining
Nitric oxide (NO⋅) synthesized by neuronal (type I) nitric oxide synthase (nNOS) has been implicated in many pathophysiological processes, including cerebral ischemia and excitotoxicity (Bredt and Snyder, 1990; Dawson et al., 1991; Lipton et al., 1993; Moskowitz and Dalkara, 1996). A number of studies have demonstrated that pharmacological inhibition or gene knock-out of nNOS confers resistance to cerebral ischemia (Dalkara et al., 1994; Huang et al., 1994) and excitotoxicity in vivoand in vitro (Dawson et al., 1996). In particular, cortical cells in culture from nNOS−/− mice are resistant to NMDA but not AMPA or kainic acid (KA) receptor-mediated toxicity (Dawson et al., 1996). Pharmacological inhibition of type I NOS protects against intrastriatal NMDA- but not AMPA- or KA-induced excitotoxic lesions (Schulz et al., 1995). Contradictory results have been reported about the importance of nNOS to excitotoxicity (for review, see Pelligrino, 1993; Löschmann et al., 1995), part of which may be explained by the deleterious effects of endothelial (type III) NOS (eNOS) inhibition (Morikawa et al., 1992; Connop et al., 1995;Globus et al., 1995; Huang et al., 1996) and/or insufficient type I NOS inhibition.
Two mechanistically distinct but related forms of neuronal death have been identified and take the form of necrosis or apoptosis. Both NO⋅ and peroxynitrite have been linked to apoptosis in severalin vitro models (Albina et al., 1993; Estevez et al., 1995;Lin et al., 1995). Oxidative DNA damage and strand breaks caused by peroxynitrite are well documented (Inoue and Kawanishi, 1995; Salgo et al., 1995), which might suggest a mechanism for initiating apoptosis after NMDA receptor activation and NO⋅ synthesis. Because decreased neuronal NOS activity is neuroprotective in stroke and excitotoxicity, and because blockade of apoptosis decreases cell death and improves functional recovery (Hara et al., 1997a), it is therapeutically relevant to investigate mechanisms of cell death induced by nitric oxide and its reaction products.
Mice with a disruption of the type I NOS gene provide an invaluable tool to study the physiological and pathophysiological functions of NO⋅ (P. L. Huang et al., 1993; Z. Huang et al., 1994; Hara et al., 1996; Ma et al., 1996; Panahian et al., 1996). In this study, we demonstrate that intrastriatal microinjection of NMDA leads to neuronal death, that the absence of NO⋅ synthesized by type I NOS affords 50% protection, and that apoptosis is involved as a late event. The neuroprotection attained in nNOS−/−mice probably relates to a decreased production of OH⋅ radicals and reduced nitration of tyrosine by elimination of the NO⋅ and O2⋅− reaction and subsequent peroxynitrite formation. Our data also suggest that apoptosis is an important delayed mechanism for NO⋅-mediated excitotoxicity after NMDA receptor activation.
MATERIALS AND METHODS
Experimental animals
Wild-type (SV129 and C57BL/6J, 21–27 gm, male; Taconic Farms, Germantown, NY) and type I NOS knock-out mice (20–28 gm, male and female) were housed under diurnal lighting conditions and given food and water ad libitum. In preliminary experiments we observed a significantly larger NMDA (10 and 20 nmol) lesion in SV129 mice compared with C57BL/6J mice; therefore, comparison of the lesion volume, DNA laddering, and terminal deoxynucleotidyl transferase (TdT)-mediated deoxyuridine triphosphate (dUTP)–biotin nick end-labeling (TUNEL)-positive cells were performed in littermate wild-type offspring from the mating of heterozygotes (nNOS+/−) as controls to eliminate differences attributable to background strains.
Intrastriatal NMDA microinjection
Mice were anesthetized with halothane (2.5% for induction, 1–1.5% for maintenance), and body temperature was kept at 36.9 ± 0.1°C with a thermostatic heating pad. The head was fixed to a stereotaxic frame with a mouse head holder (David Kopf, Tujunga, CA). A burr hole was drilled, and an injection needle (26 ga) was lowered into the right striatum (anterior, 0.5 mm; lateral, 2.5 mm; ventral, 2.5 mm, from bregma). Drugs were injected in a volume of 0.3 μl, over 2 min, using a microinjection system (David Kopf), and the needle was left in place for an additional 8 min.
Mice were routinely killed by decapitation after brief halothane anesthesia, at 48 hr. To study lesion development, mice were killed at 3, 6, 12, 24, and 48 hr and 7 d after injection. The brains were frozen, and 20-μm-thick coronal cryostat sections were taken. Every 10th section was stained by hematoxylin and eosin (H&E). The lesion area was identified by the loss of basophilic staining, measured by image analysis (M4, Imaging Research, St. Catharines, Ontario, Canada), and integrated to calculate the volume.
Drug preparation
NMDA was dissolved in 0.1 m PBS, pH 7.4, at 33 and 67 mm. 7-nitroindazole (7-NI) was dissolved in peanut oil (5 mg/cc), administered 30 min before intrastriatal NMDA microinjection (25 mg/kg, i.p.), and repeated 30 min later. MK-801 (4 mg/kg, i.p., in normal saline) was administered 30 min before intrastriatal NMDA microinjection.
Salicylate assay and 3-nitrotyrosine measurement
The salicylate hydroxyl-trapping method (Floyd et al., 1984) was used for measuring OH⋅ radicals in striatal tissues 1, 6, and 24 hr after intrastriatal injection of NMDA (20 nmol). Salicylate (100 mg/kg, i.p.) was administered 1 hr before killing. Mice were killed by decapitation, and the right and left striata were rapidly dissected from a 2-mm-thick slice on a chilled glass plate, and the tissues were placed in 0.5 ml of chilled 0.1 m HClO4. The samples were sonicated, frozen rapidly, thawed, and centrifuged twice. Aliquots of the supernatant were stored at −70°C until assay. Salicylate and its metabolites 2,3- and 2,5-dihydroxybenzoic acid (DHBA) were quantified by HPLC with 16-electrode electrochemical detection. Salicylate, 2,3- and 2,5-DHBA, tyrosine, and 3-nitrotyrosine were measured electrochemically by oxidation at 840, 240, 120, 600, and 840 mV, respectively, with retention times of 20.5, 9.4, 6.3, 10.5, and 18.2 min, respectively. The 3-nitrotyrosine measurements were validated by changing chromatographic conditions, overspiking samples with authentic standards, and demonstrating the correct electrochemical signature across two electrodes. Treatment of both standards and tissue extracts with 1 m sodium hydrosulfite (dithionite) abolished the 840 mV nitrotyrosine peaks by conversion of 3-nitrotyrosine to aminotyrosine. Data were expressed as the ratio of 2,3- and 2,5-DHBA to salicylate and of 3-nitrotyrosine to tyrosine to normalize for varying brain concentrations of salicylate and tyrosine, which could be a consequence of impairment of blood–brain barrier (salicylate), or neuronal loss during treatment (tyrosine).
Nitrotyrosine immunohistochemistry
nNOS−/− (n = 2) and wild-type (SV129, n = 5) mice were processed for immunohistopathological examination. Mice were deeply anesthetized and transcardially perfused with cold (4°C) saline, followed by 0.1m sodium phosphate buffer, pH 7.4, paraformaldehyde solution. Brains were removed directly after perfusion, post-fixed for 2 hr, washed in 0.1 m sodium phosphate buffer, and cryoprotected in increasing concentrations of 10 and 20% glycerol-2% DMSO solution. Frozen serial coronal sections of the entire brain were made at 50 μm intervals. Sections were subsequently stained for Nissl substance using cresyl violet to identify the lesioned locus and for immunohistochemical localization of 3-nitrotyrosine (monoclonal antibody marker for peroxynitrite-mediated nitration; Upstate Biotechnology, Lake Placid, NY; 1:500 dilution) using a previously reported conjugated second antibody method (Ferrante et al., 1993). Tissue sections were preincubated in absolute methanol-0.3% hydrogen peroxide solution for 30 min, washed (three times) in PBS, pH 7.4, 10 min each, placed in 10% normal goat serum (Life Technologies, Gaithersburg, MD) for 1 hr, incubated free floating in primary antiserum at room temperature for 12–18 hr (all dilutions of primary antisera above included 0.3% Triton X-100 and 10% normal goat serum), washed (three times) in PBS for 10 min each, placed in peroxidase-conjugated goat anti-rabbit IgG (1:300 in PBS; Boehringer Mannheim, Mannheim, Germany), washed (three times) in PBS 10 min each, and reacted with 3,3′-diaminobenzidine-HCl (1 mg/ml) in Tris·HCl buffer with 0.005% hydrogen peroxide.
Specificity for the antisera used in this study was examined in each immunohistochemical experiment to assist with interpretation of the results. This was accomplished by preabsorption with excess target proteins and by omission of the primary antibody to determine the amount of background generated from the detection assay. Tissue sections for 3-nitrotyrosine immunocytochemistry were preincubated for 6 hr at room temperature with either 20 mm nitrotyrosine or 1 mg/ml nitrated BSA containing ∼30 μm nitrotyrosine to establish the specificity of antibody binding.
TUNEL
TUNEL staining was performed according to the method of Gavrieli et al. (1992) with minor modifications. Frozen tissue sections were fixed in 4% paraformaldehyde (Sigma, St. Louis, MO) and then washed in PBS (0.1 m, pH 7.4; Poly Scientific). The sections were treated with 10 μg/ml proteinase K (Boehringer Mannheim) at room temperature (RT) for 5 min and then washed in PBS. Sections were post-fixed in 4% paraformaldehyde for 15 min and then washed in PBS. Sections were immersed in TdT buffer (in mm: 30 Tris base, pH 7.2, 140 sodium cacodylate, and 1 cobalt chloride) for 5 min three times and then incubated with TdT buffer (974 μl) containing TdT enzyme (8 μl, Boehringer Mannheim) and biotinylated dUTP (10 μl, Boehringer Mannheim) for 60 min at 37°C. The reaction was terminated by NaCl (300 mm) and sodium citrate (30 mm) for 15 min at 25°C. Sections were then washed in PBS, followed by immersion in 1% H2O2 (Sigma) in PBS for 10 min at RT, and then rinsed in PBS for 5 min three times. The reaction product was visualized with a peroxidase standard Vectastain ABC kit (Vector Laboratories, Burlingame, CA) and diaminobenzidine (Sigma). For negative controls, sections were incubated without enzyme or biotinylated dUTP. Positive controls were immersed in DNase I (5 μl/5 ml of distilled water, Boehringer Mannheim) for 10 min at RT, before equilibration in TdT buffer.
TUNEL-positive cell counting
Terminal transferase labels 3′-DNA nicks, which are not specific for oligonucleosomal damage. Moreover, internucleosomal DNA breakdown can occur in necrotic cells (Grasl-Kraupp et al., 1995; vanLookeren-Campagne and Gill, 1996). Therefore, we used both morphological and histochemical (TUNEL positivity) criteria to classify cells as apoptotic (Li et al., 1995a; Charriaut-Marlangue et al., 1996) and excluded from counting those cells showing features of necrosis. The cells classified as apoptotic were TUNEL-positive and showed condensed, occasionally fragmented nuclei with clumped chromatin but without cytoplasmic staining. Necrotic cells showed weak, diffuse DAB-positive cytoplasmic staining with or without a condensed and fragmented nucleus. Apoptotic cells were counted in a single tissue section at the largest lesion diameter within a preselected 100-μm-wide band (∼2.5 mm; see Fig. 3A).
DNA fragmentation
Because TUNEL staining may not differentiate internucleosomal versus random DNA breakdown, we studied DNA fragmentation on agarose gels. Nonspecific DNA damage as well as internucleosomal DNA breaks were studied at 12, 24, and 48 hr and 7 d after NMDA injection (20 nmol) in SV129 mice (n = 4 at each time point) and 48 hr after NMDA in nNOS+/+ and nNOS−/− littermates (n = 3 each). Brains were rapidly removed, and 2-mm-thick coronal slices were cut. The second slice containing both cortex and striatum was used for analysis after separating the hemispheres.
DNA was isolated (Puregene Systems, Minneapolis, MN), treated with DNase-free RNase (Boehringer Mannheim, Indianapolis, IN), and extracted by phenol/chloroform. The DNA was precipitated in 0.1 × volume of 3 m sodium acetate and 2.5 × volume of ice-cold 100% ethanol followed by incubation at −70°C for at least 60 min. The DNA was pelleted at 15,000 × g for 30 min at 4°C, washed with 80% ethanol, and air-dried for 30 min with tubes inverted. The pellets were resuspended in 25 ml of sterile water, and DNA concentration was determined by absorbance using the Christian and Warburg coefficient.
DNA damage was assessed by a radioactive end-labeling method by terminal transferase (Tilly and Hsueh, 1993) with minor modifications (Hara et al., 1997b). The DNA samples were labeled together with [α-32P]dideoxy-ATP (3000 μCi mmol−1; Amersham, Oakville, Ontario, Canada) and 25 U of terminal transferase (Boehringer Mannheim) in a final volume of 50 μl. The reaction was stopped by addition of 5 μl of 0.25m EDTA, pH 8.0. Labeled DNA was separated from unincorporated radionucleotide by adding 0.2 volume 10 mammonium acetate and 3 volume ice-cold 100% ethanol and incubating at −70°C for 60 min together with 50 μg of yeast tRNA (2 μl of a 25 mg/ml stock) to reduce background. The DNA was pelleted by centrifugation at 15,000 × g at 4°C for 30 min, washed with 80% ethanol, and allowed to air dry for 30 min with tubes inverted. The pellets were resuspended in 20 μl Tris-EDTA buffer. The labeled DNA samples were electrophoresed in a 2% agarose gel (agarose 3:1; Amresco, Solon, OH) at 50 V for 3.5 hr. The gels were placed on several sheets of Whatman (Maidstone, UK) 3M chromatography paper and dried in a slab gel dryer (model 224; Bio-Rad, Rockville Centre, NY) for 3–4 hr without heat. Dried gels were sealed in a plastic bag and exposed to Kodak (Rochester, NY) X-Omat films.
In vitro receptor autoradiography
We measured the density and distribution of glutamate receptors to determine whether differences in NMDA-induced toxicity were related to changes in glutamate receptor density. We assessed all three ionotropic glutamate receptor subtypes, because NMDA causes glutamate release and secondary AMPA and kainate receptor activation.
Mice were anesthetized with halothane and decapitated. The brains were quickly removed, frozen in liquid N2, and kept at −80°C until used. Ten-micrometer sections were cut with a cryostat–microtome and thaw-mounted onto gelatin-coated slides. The slides were brought from −80°C to room temperature 30 min before the autoradiographic experiments.
NMDA receptors.Slides were preincubated three times for 15 min each at RT in 50 mm Tris·HCl, pH 7.5, and incubated for 60 min at 4°C in buffer containing 5 nm[3H]d,l-(E)-2-amino-4-propyl-5-phosphono-3-pentanoic acid (CGP-39653) (44.5 Ci/mmol; DuPont NEN, Boston, MA). Nonspecific binding was assessed by the addition of 100 μm glutamate. Slides were washed four times for 15 sec each in ice-cold buffer, dipped in ice-cold distilled water, dried under a stream of cold air, and exposed to 3H-Hyperfilms (Amersham) for 1 month.
AMPA receptors.Slides were preincubated three times for 15 min each at RT in (in mm) 50 Tris·acetate, 100 KSCN, and 2.5 CaCl2, pH 7.3, and incubated for 45 min at 4°C in buffer containing 9 nm [3H]AMPA (45.3 Ci/mmol, DuPont NEN). Nonspecific binding was assessed by the addition of 1 mm glutamate. Slides were washed three times for 10 sec each in ice-cold buffer, dipped in ice-cold distilled water, dried under a stream of cold air, and exposed to3H-Hyperfilms (Amersham) for 1 month.
Kainate receptors.Slides were preincubated three times for 15 min each at RT in 50 mm Tris·acetate, pH 7.4, and incubated for 45 min at 4°C in buffer containing 12 nm[3H]kainate (58.0 Ci/mmol, Dupont NEN). Nonspecific binding was assessed by the addition of 1 μmkainate. Slides were washed three times for 10 sec each in ice-cold buffer, dipped in ice-cold distilled water, dried under a stream of cold air, and exposed to 3H-Hyperfilms (Amersham) for 6 weeks.
Autoradiograms were analyzed by comparing the optical density of the film over specific brain regions with that over tritiated standards (Amersham) using a computerized system (M4, Imaging Research). Data are given in nanocuries of bound radioligand per milligram of tissue.
Data analysis. The data are expressed as mean ± SEM. Two-way ANOVA for repeated measures or one-way ANOVA followed by Tukey’s multiple comparison test were used to compare three or more groups. Student’s t test was used to compare two groups.
RESULTS
NMDA lesions in wild-type and nNOS-deficient mice
Intrastriatal microinjections of NMDA consistently produced well delineated lesions. The lesions were usually located approximately at the level of bregma and confined to the striatum except at the highest dose (20 nmol) when the adjacent cortex was affected. In preliminary experiments using SV129 mice, we determined at the light microscopic level that the loss of striatal neurons started as early as 3 hr, without significant tissue edema, and became grossly recognizable 6 hr after NMDA injection. At 12–24 hr, there was marked edema and pallor, with a much lower cell density and the beginning of glial proliferation. At 24 and 48 hr, the lesion contained many reactive glia and necrotic as well as shrunken cells with pyknotic, densely stained nuclei, particularly at the margins. Remnants of fragmented cells (apoptotic bodies), mainly distributed along the lesion boundary, were observed only occasionally at 48 hr but frequently at 7 d. The lesions were characterized by marked gliosis at 7 d. Maximum lesion volume was reached at 6–48 hr. For all subsequent experiments, animals were killed at 48 hr after NMDA injection.
The NMDA receptor antagonist MK-801 (4 mg/kg, i.p.) completely prevented lesion development (19.5 ± 3.7 and 0 ± 0 mm3 in vehicle and MK-801 groups, respectively, after 10 nmol of NMDA) when tested in SV129 mice (n = 5).
NMDA-induced lesions were 45% smaller in nNOS−/−mice (nNOS+/+, 11.7 ± 1.7 mm3; n = 8; nNOS−/−, 6.4 ± 1.8 mm3;n = 7; p < 0.05). In addition to smaller lesions, nNOS−/− mice showed more sparing of neurons within the lesion bed compared with nNOS+/+ mice.
Effects of 7-NI on NMDA toxicity
7-NI (50 mg/kg, i.p.) was used to evaluate the importance of neuronal NOS in SV129 mice. 7-NI decreased lesion volume by ∼32% (30.0 ± 2.9 mm3; n = 7) compared with vehicle injection (43.8 ± 5.4 mm3; p < 0.01; n = 6).
3-Nitrotyrosine production after NMDA microinjection
Brain levels of 3-nitrotyrosine were measured as a reflection of peroxynitrite formation. 3-nitrotyrosine levels increased by ∼100% in wild-type mice 1 hr after NMDA injection (Fig.1A). The levels were at baseline 6 (n = 6) and 24 (n = 6) hr later. 3-Nitrotyrosine levels did not change in nNOS−/− mice.
We further evaluated the 3-nitrotyrosine production by immunohistochemistry. Twelve hours after NMDA injection, striatal lesions were observed within the NMDA injection site in both nNOS−/− and wild-type mice (Fig.2A,B); the lesion size was smaller in the nNOS−/− mice and was primarily restricted to the injection site. There was a marked increase in immunohistochemical expression of 3-nitrotyrosine within the lesioned striatum of wild-type mice (Fig. 2C). Immunostaining was present in both neurons and the surrounding neuropil (Fig.2E). 3-Nitrotyrosine immunoreactivity was not observed in nonlesioned brain areas. In contrast, 3-nitrotyrosine immunolabel was not present in the lesioned striatal areas of nNOS−/− mice (Fig.2D,F). 3-Nitrotyrosine immunoreactivity was completely eliminated from the tissue sections by preincubation using nitrotyrosine or nitrotyrosine bovine serum albumin.
Hydroxyl radical (OH⋅) production after NMDA
We examined striatal levels of 2,3- and 2,5-DHBA as a measure of OH⋅ radical formation. NMDA injection significantly increased OH⋅ radical production in the ipsilateral striatum after 1 hr in both C57BL/6J and SV129 mice. 2,3-DHBA levels returned to baseline 6 (n = 6) and 24 (n = 6) hr after NMDA in SV129 mice, whereas 2,5-DHBA levels remained elevated for 24 hr (data not shown). One hour after NMDA injection, 2,3- and 2,5-DHBA levels did not increase in nNOS−/− mice (Fig.1A,B).
DNA fragmentation
To assess the importance of apoptosis to lesion development, we examined oligonucleosomal DNA breakdown (laddering) after NMDA injection. Oligonucleosomal DNA breakdown as well as total DNA fragmentation was observed beginning at 12–24 hr and persisted at 48 hr (Fig. 3A) and 7 d in wild-type mice. nNOS−/− mice had lower total as well as oligonucleosomal DNA breakdown compared with wild-type littermates 48 hr after NMDA injection (Fig. 3B).
We counted the number of TUNEL-positive cells after NMDA microinjections (n = 4 per time point at 3, 6, 12, 24, and 48 hr and 7 d). There were no TUNEL-positive cells 3 or 6 hr after NMDA injection, even though histopathological evidence for necrosis started as early as 3 hr. Positively stained cells first appeared at 12 hr and were distributed relatively homogenously throughout the lesion. Over the subsequent 36 hr, TUNEL-positive cells were noted primarily at the margins of the lesion, and the cells acquired a more characteristic apoptotic morphology (more fragmented nuclei and apoptotic bodies; Fig.4D,E). The total number of TUNEL-positive cells did not increase significantly beyond 48 hr. However, their distribution shifted further toward the periphery, so that at 7 d, TUNEL-positive cells formed a dense ring at the outer rim of the lesion (Fig. 4B) and were also observed at the outside of the lesion margins when compared with adjacent H&E-stained sections. The density of TUNEL-positive cells in general was higher in the medial compared with the lateral striatum (Fig. 4B,C).
At 48 hr, ∼30% of TUNEL-positive cells showed diffuse cytoplasmic staining and no clear nuclear or cytoplasmic border, more consistent with the appearance of necrotic cells. Approximately 10% of TUNEL-positive cells possessed nuclear fragmentation and chromatin condensation superimposed on a diffuse cytoplasmic staining, thus exhibiting both apoptotic and necrotic characteristics.
The morphology of TUNEL-positive cells in nNOS−/−mice (n = 6) did not differ from that in the wild-type strain (n = 7), although the density of apoptotic neurons in the lesion was less in nNOS−/− mice (406 ± 28 and 280 ± 8 neurons/mm2 in nNOS+/+ and nNOS−/− mice, respectively; p < 0.01). The differences were pronounced in the lateral half of the lesion only (Fig.4C).
Glutamate receptor autoradiography
Because glutamate receptor density and/or distribution could contribute to the observed resistance to NMDA in nNOS−/− mice, we performed NMDA, AMPA, and kainate receptor autoradiography in coronal brain sections from three strains (Table 1). The data from knock-out mice lacking type III NOS are also listed for comparison. The highest density binding for all three ligands was in the hippocampus. Intense labeling was also observed in the striatum and neocortex. The density of NMDA receptors was higher in the lateral striatum compared with the medial striatum in all strains. Strain differences were not observed for any of the three ligands, except that the density of AMPA receptors in the CA1 region of SV129 mice was higher than in C57BL/6J mice. It should be noted that the listed values were obtained at nonsaturating ligand concentrations; the validity of the comparisons thus relies on the assumption that the affinity of the receptors for their respective ligands does not differ among strains.
DISCUSSION
We demonstrated that type I NOS gene knock-out or pharmacological inhibition by 7-nitroindazole confers neuroprotection against NMDA receptor-mediated toxicity in striatum. We also showed that the mechanism was probably related to a deficiency of NO⋅and its products, as reflected by lower levels of free 3-nitrotyrosine as well as 2,3- and 2,5-DHBA and nitrotyrosine immunostaining. Our findings suggest that apoptosis may be an important delayed mechanism of cell death after NMDA receptor activation, observed at 12 hr after injection but not before. The findings are consistent with results of previous in vivo studies showing that nNOS deficiency confers resistance to injury after permanent (Huang et al., 1994) or transient focal (Hara et al., 1996) or global cerebral ischemia (Panahian et al., 1996) or after striatal malonate injection (Schulz et al., 1996). The findings also agree with results showing that cortical cells cultured from nNOS−/− mice are resistant to NMDA toxicity (Dawson et al., 1996), or that inhibition of nNOS by 7-nitroindazole protected against focal ischemia (Yoshida et al., 1994) or NMDA-induced toxicity in rats (Schulz et al., 1995). The extent of protection in nNOS−/− mice (∼45%) was greater than that achieved by 7-NI (32%), probably because the degree of NOS inhibition obtained using knock-out technology is greater and longer-lasting than after drug administration.
The precise mechanism(s) by which NO⋅ mediates neurotoxicity is not clear, and many mechanisms have been proposed, including DNA damage (Wink et al., 1991), energy depletion attributable to poly(ADP-ribose) polymerase activation (Berger, 1985; Zhang, 1994), and inhibition of mitochondrial respiration (Stadler et al., 1991; Radi et al., 1994). One of the most attractive mechanisms involves peroxynitrite formation, which is initiated via NMDA receptor activation, intracellular Ca2+ increase, and augmented NO⋅ synthesis (Garthwaite et al., 1988). Peroxynitrite (ONOO−) is formed by the reaction of NO⋅with O2⋅− (Lafon-Cazal et al., 1993;Dykens, 1994), and this complex rapidly decomposes into NO2⋅ and hydroxyl radical (OH⋅) or a reactive intermediate with OH⋅−-like activity (Crow et al., 1994). Hydroxyl radical is a highly reactive species that leads to oxidation of sulfhydryl groups, lipids, DNA, and proteins (Beckman et al., 1996). Peroxynitrite can directly inhibit glutamate transporters (Trotti et al., 1996) or produce nitronium ions, causing irreversible nitration of tyrosine residues in proteins (Beckman et al., 1996). Protein tyrosine nitration may contribute to NO⋅ toxicity by reducing phosphorylation by tyrosine kinases (Beckman et al., 1996) or targeting nitrated proteins for degradation (Gow et al., 1996).
We studied brain levels of free 3-nitrotyrosine as an indirect measure of protein tyrosine nitration and as a footprint of peroxynitrite formation during NMDA receptor activation. Compared with robust increases in wild-type mice, brain levels of 3-nitrotyrosine and OH⋅ radical did not rise in nNOS−/− brains, suggesting that the reaction of NO⋅ with O2⋅− is a major pathway that generates peroxynitrite and OH⋅ radical-like activity in brain after NMDA receptor activation. Nevertheless, O2⋅−can induce neurotoxicity independent of its reaction with NO⋅(Chan, 1996), because superoxide dismutase can still protect cortical cell cultures from nNOS−/− mice against NMDA toxicity (Dawson et al., 1996). In the same study, sodium nitroprusside and 3-(4-morpholinyl)-sydnone imine hydrochloride were equally toxic to cultured cells from wild-type and nNOS−/− mice, confirming the importance of a deficiency of NO⋅ synthesis as a mechanism of resistance to cytotoxicity. Despite greatly reduced NO⋅ and OH⋅ radical production, the protection was ∼50% in nNOS−/− mice, suggesting a role for additional mechanisms, such as acute neuronal swelling and lysis, and O2⋅− toxicity (Chan, 1996; Kamii et al., 1996).
Basal 3-nitrotyrosine levels in wild-type and NOS−/− mice were measurable in the nonlesioned striatum as well, suggesting alternative pathways for nitrotyrosine formation in both normal wild-type and nNOS−/−brains that may be independent of NMDA-induced NO formation. Of interest, alternative splice variants of NOS have been reported that generate NO in vitro but lack the PDZ-containing domain and potential coupling to the NMDA receptor (Brenman et al., 1996; Eliasson et al., 1997). However, NOS activity in nNOS−/−brains was very low, measured by the conversion of [3H]l-arginine to [3H]l-citrulline in vitro(Hara et al., 1996). The time course of 3-nitrotyrosine and 2,3-DHBA production was similar, suggesting that NO⋅ and hydroxyl radical synthesis occurs predominantly within the first hour. The elevation in 3-nitrotyrosine was no longer detectable 6 hr after NMDA injection, suggesting that the removal of free 3-nitrotyrosine can be fairly rapid. In contrast, robust staining for nitrotyrosine was observed immunohistochemically 12 hr after NMDA injection in wild-type striatum, suggesting a much longer half-life for nitrotyrosine found in proteins (Beckman et al., 1996). 2,5-DHBA levels may be more indicative of cytochrome P-450 activity (Halliwell et al., 1991) and remained elevated in wild-type brain up to 24 hr.
NMDA-induced striatal cell loss and pallor were visible at 3 hr, and the lesion reached its maximum volume at 6 hr. Accompanying this early cell loss, 2,3-DHBA and 3-nitrotyrosine levels increased at 1 hr, completely recovered by 6 hr, and stayed at baseline levels until 24 hr, suggesting that OH⋅ radical and peroxynitrite formation takes place early after NMDA injection. On the other hand, in situ labeling of 3′-OH DNA breaks (TUNEL staining) and oligonucleosomal DNA breakdown (DNA laddering) were observed only after 12–24 hr and increased at 48 hr. The delayed onset of oligonucleosomal DNA breakdown, nuclear fragmentation, and chromatin condensation suggest that NMDA induces an initial cell loss (3–6 hr), followed many hours later by a form of death resembling apoptosis. The delayed onset probably indicates that one or more intervening steps initiate apoptosis after NMDA microinjection and receptor activation. In a recent study, inhibition of interleukin-1β-converting-enzyme decreased NMDA lesion size by 25% 48 hr after microinjection (Hara et al., 1997), further establishing the linkage between the two.
The stimulus for apoptotic cell death is unknown but probably depends in part on NMDA receptor-mediated Ca2+ influx and free radical formation (e.g., NO⋅, O2⋅−, peroxynitrite, and OH⋅). Apoptosis can be observed in cortical or cerebellar granule cell cultures (Ankarcrona et al., 1995) after the addition of NMDA, especially at low concentrations (Bonfoco et al., 1995), or after intrastriatal injection (Ferrer et al., 1995; vanLookeren-Campagne et al., 1995; Qin et al., 1996). Direct application of NO⋅ or peroxynitrite can also trigger apoptotic cell death (Albina et al., 1993; Ratan et al., 1994; Bonfoco et al., 1995; Estevez et al., 1995;Lin et al., 1995; for review, see Nicotera et al., 1995; Palluy and Rigaud, 1996) possibly related to DNA damage (Inoue and Kawanishi, 1995; Salgo et al., 1995; Szabo, 1996; Tamir et al., 1996), an increase in p53 gene expression (Messmer et al., 1994; Messmer and Brüne, 1996), or inhibition of mitochondrial respiration (Stadler et al., 1991; Radi et al., 1994; Wolvetang et al., 1994; Behrens et al., 1995; Cassina and Radi, 1996; Poderoso et al., 1996). Consistent with these findings, blockade of superoxide dismutase-1 by antisense oligonucleotides causes apoptosis, which is decreased by also blocking NO⋅ (and probably peroxynitrite) formation (Troy et al., 1996).
TUNEL-positive cells were evenly distributed within the lesion at 1 d. Time-dependent studies showed that the density of TUNEL-positive cells gradually increased in the periphery and decreased in the center of the lesion. Hence, TUNEL-positive cells were distributed predominantly within the periphery at 7 d and resembled the pattern reported in focal cerebral ischemia (Li et al., 1995a). Relatively lower NMDA concentrations in the periphery versus core might explain the distribution of positive cells at the lesion margin, as observed in the penumbra of an ischemic lesion (Li et al., 1995a). In fact, damage caused by mild ischemia can be successfully prevented by drugs that block apoptosis, such as cycloheximide (Du et al., 1996) or cysteine protease inhibitors (Endres et al., 1997). Unlike NMDA lesions, however, TUNEL-positive cells and DNA laddering appear within a few hours after 2 hr middle cerebral artery occlusion (Li et al., 1995b), suggesting additional mechanisms that differentiate ischemia and excitotoxicity.
Regions of high NMDA receptor density may be linked more closely to necrotic than apoptotic cell death, because the density of NMDA receptor binding sites, higher in lateral than medial striatum, contrasts with the higher density of TUNEL-positive cells found within medial striatum. The data from nNOS−/− mice suggest that apoptotic cell death may be linked to NO⋅ in regions with a high density of NMDA receptors. These apparent discrepancies point to mechanisms in addition to NOS activation linking NMDA receptor occupancy and development of delayed cell death, at least within medial striatum. Measuring levels of NO⋅ within medial and lateral striatum may help clarify this issue, because the ability of NO⋅ to promote apoptosis or necrosis may well be concentration-dependent (Bonfoco et al., 1995).
In conclusion, our study demonstrates that NO⋅ synthesized by the neuronal NOS isoform is an important mediator in NMDA receptor-mediated toxicity and promotes apoptosis as a delayed and probably secondary mechanism of cell death.
Footnotes
This work was supported by the Massachusetts General Hospital Interdepartmental Stroke Program Project NS10828 (M.A.M.), an unrestricted award in neuroscience by Bristol-Myers Squibb (M.A.M.), the Huntington’s Disease Society of America (R.J.F.), the Department of Veterans Affairs (R.J.F.), National Institutes of Health Grants AG12922 (R.J.F. and M.F.B.), 1P30AG13846 (R.J.F.), NS35255 (R.J.F.), NS33335 (P.L.H.), and AG08487 (B.T.H.), and Deutsche Forschungsgemeinschaft Grant En 343/1-1 (M.E.). P.L.H. is an Established Investigator of the American Heart Association. We thank Karen Smith and Tom Kilgallen for their technical assistance.
Correspondence should be addressed to Michael A. Moskowitz, Stroke and Neurovascular Regulation Laboratory, CNY 149-6403, 13th Street, Charlestown, MA 02129.