Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Glutamate is the major excitatory neurotransmitter in the brain. It is thought to be involved in several neurological conditions by virtue of its ability to stimulate glutamate receptors that can lead to a prolonged increase in intracellular [Ca²⁺] (Garthwaite and Garthwaite, 1986; Choi, 1987). Increased [Ca²⁺]_i can then signal a variety of second messenger systems that could play a role in cell death. For example, elevated cytosolic Ca²⁺ can activate calmodulin-dependent nitric-oxide synthase (Bredt and Snyder, 1992; Alagarsamy et al., 1994) and phospholipase A₂, leading to increased NO and arachidonic acid synthesis, respectively, both of which can ultimately result in production of free reactive oxygen species (ROS) and lipid peroxidation (Lafon-Cazal et al., 1993; Dugan et al., 1995; Reynolds and Hastings, 1995). The use of superoxide dismutase, inhibitors of NO synthesis and nitric-oxide synthase knockouts has implicated NO, O₂, and their reaction product, peroxynitrite, in several forms of glutamate-mediated neurotoxicity (Dawson et al., 1993, 1996; Dugan et al., 1995; Simonian and Coyle, 1996; Wang et al., 2000).

A number of reports suggest that nuclear factor-B (NF-B) and Bax may participate in glutamate-induced neurotoxicity (Grilli et al., 1996; Furukawa and Mattson, 1998; Ko et al., 1998; Xiang et al., 1998; Qin et al.,1999; Djebaili et al., 2000). On the other hand, in certain situations NF-B can be protective (Mayo et al., 1997; Kaltschmidt et al., 2000), perhaps by activating the expression of Bcl-2 family proteins, including Bcl-X_L, which is known to oppose the pro-apoptotic effect of Bax (Tamatani et al., 1999; Chen et al., 2000). Nevertheless, NF-B is an attractive candidate to mediate the effects of ROS because the interaction between the inhibitory protein IB and NF-B proteins (e.g., p50/RelA and p65) is regulated by protein kinases that contain several redox-sensitive cysteine residues in critical kinase domains (for review, see Piette et al., 1997 and Gius et al., 1999; but also see Bowie and O'Neill, 2000). Exactly how NF-B might mediate the neurotoxic effect of glutamate is unknown, but its ability to up-regulate p53 expression (Grilli and Memo, 1999a; Qin et al., 1999), and subsequently Bax (Xiang et al., 1998; Grilli and Memo, 1999b), may play an important role.

Because the relationship between O₂, NF-B, Bax, and Bcl-X_L in excitotoxic cell death is unknown, this study sought to determine whether O₂ production was critical in NMDA-induced cell death and whether it was upstream of either NF-B or Bax in this putative pathway. We also wanted to determine whether NF-B translocation was critical to this process and whether it was involved in the regulation of either Bax or Bcl-X_L. Our results suggest that both O₂ and NF-B are critical to NMDA-induced necrosis and apoptosis. Furthermore, we suggest that activation of the NMDA receptor leads first to the production of O₂ and then subsequently to NF-B translocation, which then in turn signals alterations in Bax and Bcl-X_L expression, which are correlated with the appearance of DNA fragmentation and apoptotic nuclei.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Drugs and Other Materials. N-Methyl-D-aspartic acid (NMDA) was purchased from Tocris Neuramin (Bristol, UK). M40403, a nonpeptidyl superoxide mimic, and its inactive control were synthesized at MetaPhore Pharmaceuticals as previously described (Salvemini et al., 1999). SN50, a peptide inhibitor of NF-B, as well as its inactive control peptide, were purchased from BIOMOL Research Laboratories (Plymouth Meeting, PA). 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) was purchased from Sigma-Aldrich (St. Louis, MO). The medium and fetal bovine serum were purchased from Invitrogen (Carlsbad, CA). All other kits and enzymes were obtained from Roche Applied Science (Indianapolis, IN). Other chemicals were obtained from Sigma-Aldrich.

Primary Cell Culture. Primary forebrain cultures were prepared from newborn rats (Sprague-Dawley) as described by Wang et al. (2000). Briefly, forebrains were dissected and dissociated in cold Hanks' solution without Mg²⁺ and Ca²⁺. Cultures were grown on polylysine-coated coverslips in Dulbecco's modified Eagle's medium supplemented with 10% (v/v) fetal bovine serum. Glial proliferation was stopped with a mitotic inhibitor, cytosine -D-arabinofuranoside (beginning on the 3rd day of culture). After 1 week, these cultures are about 50% neuronal (Wang et al., 2000). Cultures were exposed to NMDA in the presence and absence of M40403 or SN50 (or their inactive control analogs) on day 6 in serum-free defined medium (prepared by adding a supplement mixture consisting of 15 µg/ml insulin, 20 µg/ml transferrin, 20 nM progesterone, 100 µM putrescine, and 30 nM sodium selenite to Dulbecco's modified Eagle's medium). After the addition of NMDA in Mg²⁺-containing medium, neurotoxicity was evaluated at times ranging from 1 to 20 h in four different assays as described below.

Cytotoxicity Detection Assay [Lactate Dehydrogenase (LDH)]. The release of the cytosolic enzyme LDH into the medium was used as a generic index of cell death. One to 20 h after exposure to NMDA or control medium, the medium was collected and assayed for LDH activity using a cytotoxicity detection kit from Roche Applied Science. Briefly, LDH catalyzes the conversion of lactate to pyruvate upon reduction of NAD⁺ to NADH/H⁺; the added tetrazolium salt (yellow) is then reduced to formazan (red). The amount of formazan formed correlates to LDH activity. The formazan product is measured with a microtiter plate reader at an absorption wavelength of 490 nm. In some experiments, the results were presented as the percentage of the total LDH in the culture. Total LDH was estimated following sonication of quadruplicate cultures in distilled water. Expression of the magnitude of the LDH released by a particular treatment as a percentage of the total is thus an estimate of the percentage of the cells killed by this treatment.

MTT Reduction Cell Viability Assay. The dye MTT is taken up and metabolized to a colored product by viable mitochondria. Thus, the measurement of this metabolic reduction reaction was used as a marker of mitochondrial viability. Briefly, after the removal of medium (and detached cells) for use in the LDH assay, 100 µl of MTT (5 mg/10 ml of medium) was added to each well, and the plate was incubated for 4 h at 37°C. The MTT solution was removed, and 100 µl of dimethyl sulfoxide was added to each well; the color intensity was assessed with an enzyme-linked immunosorbent assay (ELISA) plate reader at a wavelength of 590 nm. The magnitude of the decrease in MTT metabolized is a direct estimate of the percentage of cells compromised by any given treatment.

Fragmented DNA Detection by ELISA. Although the LDH release assay and the MTT reduction assay are reliable indices of cell death, neither is specific for apoptotic cell, which is better characterized by DNA fragmentation which is the result of internucleosomal cleavage of DNA by apoptosis-specific activation of endonucleases. The presence of fragmented DNA associated with nucleosomal histone was assessed by a specific two-site ELISA employing an anti-histone primary antibody and a secondary anti-DNA antibody according to the manufacturer's instructions (Roche Applied Science). Briefly, cells were grown in 10-cm tissue culture dishes. At 0, 2, 7, or 20 h following NMDA addition, cells were spun and resuspended in 3 ml of lysis buffer and incubated for 30 min at room temperature. After centrifugation, the supernatants (cytosol containing low-molecular weight fragmented DNA) were diluted 1:2 (v/v) with lysis buffer. Twenty microliters from each sample was transferred to a plate reader well precoated with anti-histone antibody, and 80 µl of immunoreagent mix (including the secondary antibody) was added. After incubation and washes, the wells were treated with the chromogen substrate, and the intensity of the color that developed was assayed at 405/490 nm.

TUNEL Assay. This assay is widely used to assess apoptosis in situ. It relies on the detection of fragmented DNA strands, but because fragmentation can occur via nonapoptotic mechanisms, it is not absolutely specific for apoptosis. Following 20 h of treatment with NMDA, the cells were rinsed with PBS, fixed by ice-cold (4°C) 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.2, and processed for evaluation of nuclei containing fragmented DNA in situ. Terminal deoxynucleotidyl transferase, a template-independent polymerase, was used to incorporate biotinylated nucleotides at sites of DNA breaks as previously described (Johnson et al., 1998). After processing each slide as described, the slides were treated with Hoechst 33258 (bis-benzimide, 0.1 µg/ml) to stain all nuclei of cells that were not TUNEL-positive (Wang et al., 2000). The TUNEL- and Hoechst 33258-positive cells were then photographed with the use of an Olympus light microscope equipped with epifluorescence (excitation wavelength of 365 nm for Hoechst 33258), and the percentage of TUNEL-positive cells was estimated in five 0.24-mm² fields in each of three dishes in each treatment condition. Each condition was assessed at least in triplicate, and experiments were repeated three times independently. Data are presented as the means ± S.E.M. A probability of P < 0.05 was considered significant (one-way ANOVA).

Western Blot Analysis. Following treatment with NMDA for either 0, 2, 7, or 20 h, the medium was removed and the attached cells were washed with PBS. Protein extraction was accomplished by cell lysis with SDS. Protein samples were measured for protein concentration with BCA protein reagent (Pierce Chemical, Rockford, IL). Equal amounts of total protein (10 µg) were loaded on each lane and run on SDS-polyacrylamide gel with a Tris/glycine running buffer system and then transferred to a polyvinylidene difluoride membrane (0.2 µm) in a mini electrotransfer unit (Bio-Rad, Hercules, CA). The blots were probed with an anti-Bcl-X_L (1:1000, polyclonal; Santa Cruz Biotechnology, Inc., Santa Cruz, CA) antibody, anti-Bax (1:1000, polyclonal; Santa Cruz Biotechnology, Inc.) antibody, and anti-actin (1:3000, monoclonal, housekeeping protein; Amersham Biosciences, Piscataway, NJ). Immunoblot analysis was performed with horseradish peroxidase-conjugated anti-rabbit and anti-mouse IgG using the enhanced chemiluminescence Western blotting detection reagents (Amersham Biosciences). The Bcl-X_L/Bax ratio was analyzed by an automatic image analysis system (Alpha Innotech Corporation, San Leandro, CA).

Electrophoretic Mobility Shift Assay Following treatment with NMDA for 0, 2, 7, or 20 h, the medium was removed, and the attached cells were washed with PBS; nuclear extracts were prepared according to published methods (Dignam et al., 1983; Osborn et al., 1989) with some modifications. Briefly, cells were homogenized in buffer A [10 mM HEPES, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM dithiothrietol, 1 mM phenylmethylsulfonyl fluoride, 2 µg/ml antipain, 2 µg/ml chymostatin, 2 µg/ml pepstatin, and 2 µg/ml leupeptin (pH 7.8)] with approximately 15 strokes of a 1-ml manual Wheaton Tenbroeck tissue grinder (Fisher Scientific, Houston, TX). The lysate was microcentrifuged (8000 rpm for 2 min) to collect nuclei. Nuclear protein was extracted by suspending the nuclei in extraction buffer B [20 mM HEPES, 400 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 2 µg/ml antipain, 2 µg/ml chymostatin, 2 µg/ml pepstatin, and 2 µg/ml leupeptin (pH 7.9)] for 20 min (4°C). The nuclei were subjected to centrifugation, and the supernatant was divided into aliquots.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

In pilot studies as well as previously published studies, NMDA produced a concentration-dependent cell death that is characterized by increases in the number of TUNEL-positive cells, DNA fragmentation, LDH release, and a decrease in mitochondrial uptake and metabolism of MTT (Wang et al., 2000). Using a 20-h treatment period, the standard concentration chosen for this study, 300 µM, is maximally effective.

In time course studies, untreated cultures released about 13 to 15% of total available LDH over the 20-h incubation period (Table 1). Addition of 300 µM NMDA released an additional 4.5% of total available LDH into the medium after 1 h (131% of control), 9.5% at 2 h (165% of control), and 15.5 to 18.5% from 4 to 20 h (~220% of control). On the other hand, assessment of the effects of NMDA treatment on mitochondrial viability on the cells remaining in the well after removal of detached cells (for the LDH assay) revealed a much different time course. There was no significant difference between control and NMDA treatment after 1 or 2 h and about a 25% decrease after 4 h. With time, this loss of viability increased gradually to about 80% at 20 h. Thus, after 20 h, NMDA kills about 18.5% of the cells in culture by a mechanism consistent with necrosis (compromised plasma membrane and LDH release). Of the remaining cells, NMDA kills or dramatically damages about 80% (or 65% of the total) (0.80 × [100% - 18.5%]) via a mechanism that has similarities to necrosis and apoptosis (see below). The time course for NMDA induction of apoptosis was assessed using an ELISA specific for histone-associated fragmented DNA. Figure 1 shows that this marker is not changed at 2 h but is approximately doubled at 7 h and then further increased at 20 h.

View larger version (24K): [in this window] [in a new window]   Fig. 1.   Time course analysis for NMDA-induced DNA fragmentation as assessed by an ELISA that measures nucleosomes (DNA associated with histone proteins). , P < 0.05 compared with 300 µM NMDA treatment (one-way ANOVA with Dunnett's post hoc test). O.D., optical density.

The possible temporal relationship between NMDA-induced cell death and potential intracellular mediators was assessed in experiments that measured NF-B nuclear translocation and Bax expression. Figure 2 shows a typical result, and these data are quantitated and summarized in Table 2. NF-B translocation is increased by about 57% after 2 h of NMDA treatment, and this remains more or less stable throughout the 20-h period. Increased Bax expression, however, lags behind NF-B with no significant change at 2 h but with a 2.7-fold increase at 7 h and a 3.4-fold increase at 20 h. This pattern mirrors the increase in NMDA-induced DNA fragmentation shown in Fig. 1.

View larger version (58K): [in this window] [in a new window]   Fig. 2.   Time course analysis for 300 µM NMDA-induced NF-B translocation (top, gel shift assay) and Bax expression (bottom, Western blot). These data are representative of four independent experiments, which are summarized in Table 2.

View this table: [in this window] [in a new window]   TABLE 2 Time course analysis of NF-B nuclear translocation (electrophoretic mobility shift assay) and Bax expression (Western blot analysis)

To begin to investigate the role of superoxide and NF-B, we turned to the pharmacological "antagonists", M40403 and SN50, a peptide inhibitor of NF-B transport (Lin et al., 1995). M40403 produced a concentration-dependent protection against NMDA-induced cell death as assessed by LDH release after a 20-h treatment (Fig. 3, top). This effect was maximal (about 77% protection) at 2.5 µM, because higher concentrations, beginning with 10 µM, were actually neurotoxic when administered alone (data not shown). This latter effect is most likely a nonspecific rather than a mechanism-based effect, in that similar effects of the inactive analog of M40403, M40404 (Salvemini et al., 1999), have been observed in a variety of other cell systems (D. Salvemini, unpublished observations). In this study, 2.5 µM M40404 had no effect on NMDA-induced LDH release (data not shown). SN50 also dose dependently protected against the neurotoxic effect of NMDA, reaching a maximum of about 67% at 2.5 µM (Fig. 3, bottom). Similar to M40403, higher concentrations of SN50, including those used commonly by many laboratories (e.g., 50 µg/ml or 18 µM; Lin et al., 1995), were neurotoxic when administered alone (data not shown). Finally, the inactive control peptide for SN50 was inactive at 2.5 µM.

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Protective effect of M40403 and SN50 in NMDA-induced cell death. M40403 (top panel) and SN50 (bottom panel) protect, in a dose-dependent manner, against cell death induced by 300 µM NMDA treatment. LDH released into the culture medium was assayed and used to measure cell death. Results are the means ± S.E. of three independent experiments each performed in triplicate. , P < 0.05 compared with 300 µM NMDA treatment (one-way ANOVA with Dunnett's post hoc test). O.D., optical density.

To separate the mechanisms underlying necrosis and apoptosis, we first determined the ability of 2.5 µM M40403 and SN50 to prevent NMDA-induced alterations in LDH release and MTT metabolism early in the cell death process. As shown in Table 3, M40403 and SN50 completely blocked the decrease in MTT metabolism caused by NMDA treatment for 4 h. Similarly, these agents blocked LDH release by about 80% when assessed at 4 h, whereas the inactive controls for these agents had no effect on either measure. On the other hand, these agents were less effective in preventing the NMDA-induced increase in LDH observed at the 2-h time point (only about 40% protection). The role of superoxide and NF-B translocation were then determined in NMDA-induced apoptosis observed after 20 h of NMDA treatment. First, NMDA-induced cell death was assessed by an ELISA that is thought to be specific for DNA fragments associated with histone proteins. Similar protective effects of 2.5 µM M40403 and SN50 were again observed (about 75%; Fig. 4). Next, the protective effects of M40403 and SN50 were assessed in situ where the apoptotic effect of NMDA can be easily seen with the TUNEL assay. Figures 5 and 6 show that 300 µM NMDA causes a massive increase in TUNEL-positive cells characterized by fragmentation and nuclear condensation. Figure 5 shows that 2.5 µM M40403 had no significant toxic effect alone but almost completely prevented the apoptotic effect of 300 µM NMDA. Figure 6 shows a similar effect for SN50. That is, at its optimal concentration (2.5 µM), SN50 had no effect alone but substantially diminished the apoptotic effect of NMDA. Quantitation of the TUNEL data is depicted in Fig. 7. Here it can be seen that 300 µM NMDA resulted in TUNEL-positive staining of about 75% of the cells in culture. Both M40403 and SN50 dramatically inhibited the effect of NMDA, although SN50 was marginally less effective in this regard than M40403 (76 versus 93% protection).

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Protection against NMDA-induced DNA fragmentation by 2.5 µM M40403 or 2.5 µM SN50. An ELISA was used to measure nucleosomes (DNA associated with histone proteins). , P < 0.05 compared with 300 µM NMDA treatment (one-way ANOVA with Dunnett's post hoc test). O.D., optical density.

View larger version (63K): [in this window] [in a new window]   Fig. 5.   Representative micrographs depicting the protective effects of M40403 on NMDA-induced TUNEL staining. The addition of 300 µM NMDA produced a marked increase in the number of cells containing dark brown, condensed, and/or fragmented nuclei (panel B) relative to control cells (panel A). Panel D shows that cotreatment with 2.5 µM M40403 prevents the effect of NMDA on TUNEL staining while having no effect of its own (panel C).

View larger version (61K): [in this window] [in a new window]   Fig. 6.   Representative micrographs depicting the protective effects of SN50, an inhibitor of NF-B transport, on NMDA-induced TUNEL staining. The addition of 300 µM NMDA produced a marked increase in the number of cells containing dark brown, condensed, and/or fragmented nuclei (panel B) relative to control cells (panel A). Panel D shows that cotreatment with 2.5 µM SN50 prevents the effect of NMDA on TUNEL staining while having no effect of its own (panel C).

View larger version (11K): [in this window] [in a new window]   Fig. 7.   Quantitative analysis of the effects of M40403 and SN50 in preventing the ability of NMDA to increase the number of TUNEL-positive cells. The total number of cells in a 0.24 mm2 field was estimated by adding the number of Hoescht counterstained cells as counted by three independent observers. Each experimental condition was assessed in five different fields and the mean ± S.E. were calculated from six cultures for each condition. , P < 0.05 compared with 300 µM NMDA treatment (one-way ANOVA with Dunnett's post hoc test).

To determine the possible relationship between the protective effects of M40403 and the NF-B signaling pathway, this same paradigm was repeated in experiments in which nuclear protein extracts from control and treated cultures were tested in an electrophoretic mobility shift assay using a ³²P-labeled oligonucleotide containing the classical NF-B consensus sequence. Figure 8A shows a representative assay in which nuclear protein from control cultures retards the migration of the labeled NF-B binding sequence (lane 1). The two bands that can be seen represent unidentified proteins (probably dimers consisting of either p50, p52, p65, or others) that bind to this sequence. NMDA (lane 2) caused an increase of about 40% in the density of these bands relative to control. This effect of NMDA is not seen in the presence of either M40403 (lane 4) or, as expected, SN50 (lane 6). These two drugs had no effect alone (lanes 3 and 5, respectively). Figure 8B shows the combined results of four identical experiments in which the density of these bands is presented relative to control. Thus, the ability of M404003 to protect against NMDA-induced apoptosis is consistent with its ability to prevent NMDA-induced NF-B translocation to the nucleus.

View larger version (36K): [in this window] [in a new window]   Fig. 8.   Inhibition of NMDA-induced nuclear translocation of NF-B by M40403 and SN50. The top panel (A) shows a representative electrophoretic mobility shift assay, where each lane corresponds to the conditions shown in the bottom panel (B). This panel shows the data as the mean ± S.E. from four identical experiments in which the amount of NF-B protein is expressed as the percentage of control in each individual experiment. Treatment with 300 µM NMDA was significantly different from treatment with either SN50 or M40403, either alone or in combination with NMDA (one-way ANOVA with Dunnett's post hoc test).

In this culture system, we have previously observed that NMDA produced both an increase in Bax and a decrease in Bcl-X_L (Wang et al., 2000). In Fig. 9A, a representative Western blot confirms this previous observation. In four independent experiments, a 20-h NMDA treatment increased Bax to 216 ± 14% of control and decreased Bcl-X_L to 40 ± 2.2% of control (Fig. 9B). This experiment also demonstrates that both M40403 and SN50, although having no significant effect by themselves on these proteins at 2.5 µM, were able to completely prevent the effect of NMDA. This suggests that both superoxide formation and NF-B play a role in NMDA-induced Bax expression and is consistent with their role in NMDA-induced apoptosis.

View larger version (39K): [in this window] [in a new window]   Fig. 9.   A, representative Western blot demonstrating the effect of M40403 and SN50 on NMDA-induced changes in BCL-XL (top) and BAX (bottom) expression. Control (lane 1), 300 µM NMDA (lane 2), 2.5 µM M40403 (lane 3), M40403 plus NMDA (lane 4), 2.5 µM SN50 (lane 5), SN50 plus NMDA (lane 6). B, the data from three independent experiments were quantified using densitometry and expressed as the ratio of Bcl-XL to Bax. , P < 0.05 compared with 300 µM NMDA treatment (one-way ANOVA with Dunnett's post hoc test).

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

For the purpose of discussion, a cartoon depicting a working model of NMDA-induced apoptosis is provided in Fig. 10, although these data should be interpreted in the context of earlier observations that NMDA-induced cell death has the characteristics of both apoptosis and necrosis including DNA fragmentation and loss of membrane integrity (Sohn et al., 1997; Cheung et al., 1998).

View larger version (53K): [in this window] [in a new window]   Fig. 10.   This cartoon depicts a working model of NMDA-induced neuronal apoptosis. Prolonged activation of NMDA receptors results in an overload of intracellular Ca2+ that exceeds the buffering capacity of the mitochondrion and interferes with electron transport in a manner that results in the production of superoxide anion (O2). The increase in O2 turns on I-B kinases (IKK), resulting in the phosphorylation of I-B serine residues, the dissociation of NF-B proteins and their transport into the nucleus. In the nucleus, these transcription factors bind to several DNA sequences of several known genes including p53 and Bcl-XL. The consequence of this binding is not completely certain, but the transcription of p53 and a subsequent increase in Bax is enhanced in several systems. How increases in nuclear NF-B down-regulate Bcl-XL is unknown, but this decrease, in combination with an increase in Bax, diminishes the formation of anti-apoptotic Bax/Bcl-XL heterodimers in favor of pro-apoptotic Bax/Bax homodimers. These homodimers are thought to create mitochondrial membrane pores through which cytochrome c can leak into the cytoplasm where it can activate caspases that play a critical role in the ultimate demise of the cell. M40403 can prevent this cascade by converting O2 into O2 and hydrogen peroxide, which is then converted to O2 and water by endogenous catalase.

The excitotoxic effects of glutamate are largely mediated by increased Ca²⁺ influx through activated NMDA receptors (Garthwaite and Garthwaite, 1986; Choi, 1987). Associated with Ca²⁺ influx is an increase in ROS that appears to originate in the mitochondria (Malis and Bonventre, 1985). It is thought that Ca²⁺ loading by the mitochondria beyond its buffering capacity reduces the membrane potential and disrupts electron transport, resulting in the increased production of the reactive free radical superoxide anion (O₂) (Luetjens et al., 2000). Several experiments have shown that inhibition of the electron transport chain with rotenone (complex I), antimycin A (complex III), and oligomycin (complex V) prevents ROS formation and in some cases can be neuroprotective (Luetjens et al., 2000).

Although apoptosis can be the final result of an excitotoxic insult, the pathways leading from mitochondrial dysfunction and ROS generation are not completely understood. The use of metalloporphyrins such as manganese tetrakis(4-benzoyl acid) porphyrin has implicated O₂ in glutamate excitotoxicity previously (Patel et al., 1996; Luetjens et al., 2000). However, relative to M40403, these reagents lack specificity for O₂ (Patel et al., 1996; Salvemini et al., 1999) and are much less potent. For example, the current study demonstrates that this nonpeptidyl superoxide dismutase mimic significantly blunts NMDA-induced cell death over a range of about 10-fold (0.3-2.5 µM), whereas the maximal effect of manganese tetrakis(4-benzoyl acid) porphryin was observed at 150 to 200 µM (Patel et al., 1996). Unlike manganese tetrakis(4-benzoyl acid) porphryin, M40403 lacks affinity for other ROS such as peroxynitrite (Patel et al., 1996; Salvemini et al., 1999). Thus, the neuroprotective role of M40403 against NMDA-induced neurotoxicity confirms the hypothesis that O₂ plays a crucial role in NMDA receptor-mediated excitotoxicity. Furthermore, the use of the TUNEL assay and the nucleosome-specific ELISA in this study allows the conclusion that O₂ is involved in NMDA-induced apoptosis. Thus, the use of M40403 appears to be a valuable tool for exploring the specific role of O₂ in NMDA-induced apoptosis. Interestingly, M40403 also appears to be able to protect against NMDA-induced neurotoxicity early in the process before classic markers of apoptosis such as DNA fragmentation appear.

One consequence of increased oxidative stress is the activation and inactivation of redox-sensitive proteins (Kamata and Hirata, 1999). The transcription factor NF-B is known to respond to changes in the redox state of the cytoplasm and has been shown to translocate in response to NMDA-induced cellular stress (Ko et al., 1998). NF-B is normally sequestered in the cytoplasm, bound to the regulatory protein IB. In response to a wide range of stimuli including oxidative stress, infection, hypoxia, extracellular signals, and inflammation, IB is phosphorylated on serine residues Ser-32 and Ser-36 by the enzyme IB kinase. This targets the IB protein for ubiquination and subsequent degradation by the 26 S proteasome (Bowie and O'Neill, 2000). The net result is the release of the NF-B dimer, which is then free to translocate into the nucleus. The ability of M40403 to prevent NMDA-induced neurotoxicity and nuclear translocation of NF-B strongly suggests that O₂ is a key reactive oxygen species in this pathway.

Previous research has shown a very complex and often contradictory role for NF-B in neuronal apoptosis. Studies examining hypoxia/reoxygenation, serum withdrawal and extracellular signaling proteins have generally found a protective role for NF-B (Tamatani et al., 1999). On the other hand, NF-B translocation appears to be a necessary step in apoptosis induced by cyanide and excitotoxic stimuli (Shou et al., 2000). The mechanism by which NF-B translocation induces apoptosis is not completely clear, but it is assumed that this mechanism involves the regulation of one or more genes known to play a role in apoptosis. However, because NF-B is known to regulate both anti-apoptotic and pro-apoptotic proteins depending on the cell type and the nature of the stimulus (Bowie and O'Neill, 2000; Grilli and Memo, 1999b; Tamatani et al., 1999), it is possible that the ultimate effect will be the sum of several downstream regulators. In neurons, astrocytes, and glia these regulators include the anti-apoptotic Bcl family members, Bcl-X and Bcl-2 (Tamatani et al., 1999; Chen et al., 2000), the important antioxidant, manganese superoxide dismutase, and the potentially detrimental proteins p53, inducible nitric-oxide synthase, and cyclooxygenase-2 (Mattson et al., 2000).

The prevention of NMDA-induced cell death in this preparation by SN50 clearly suggests that the overall effect of NF-B translocation is pro-apoptotic. Among the potential downstream regulators, we measured the effect of NMDA on Bax and Bcl-X_L, pro- and anti-apoptotic members of the Bcl family that have been reported to be up- and down-regulated, respectively, in various models of apoptosis (Miller et al., 1997; Niwa et al., 1997; Reed, 1998; Gonzalez de Aguilar et al., 2000; Ravishankar et al., 2001). Bax is activated by p53 in neurons (Morrison and Kinoshita, 2000). In hippocampal neurons cultured from p53 / and +/+ mice, it was recently demonstrated that an NMDA-induced increase in Bax was p53-dependent (Djebaili et al., 2000). This study also demonstrated that NMDA-induced DNA fragmentation and TUNEL staining were found only in cultures from p53 +/+ mice. The ability of SN50 to prevent NMDA-induced apoptosis and increased expression of Bax in the present study demonstrates that NF-B is crucial to this process and, when considered in light of the study on hippocampal neurons, suggests that p53 may be an intermediate in this pathway. Furthermore, the antagonistic effect of M40403 on NMDA-induced increases in NF-B translocation and Bax strengthens the argument that increased NF-B is critical in Bax up-regulation.

Recently, it has been demonstrated in several lymphoid cell lines that Bcl-X_L is selectively up-regulated by the NF-B proteins c-Rel and RelA (p65), but not by p50 (Chen et al., 2000). Functional analysis of the bcl-x promoter suggested that it is under the direct control of c-Rel. Furthermore, it was recently determined that there are two functional NF-B DNA binding sequences clustered upstream of the brain-specific transcription start site in the upstream region of the murine bcl-x promoter (Glasgow et al., 2000). These sequences have an affinity for p50/p50 and p50/p65 heterodimers that is similar to the consensus IgG-B binding sequence (Glasgow et al., 2000). Both dimers have been demonstrated to act in a positive sense, i.e., if NF-B proteins bind these promoter regions, Bcl-X synthesis is increased (Chen et al., 2000; Glasgow et al., 2000). Thus, despite its inhibition by SN50, the mechanism by which NMDA decreases Bcl-X_L is not a straightforward result of NF-B up-regulation. There is evidence in the literature that suggests that the transcriptional regulation of target genes by NF-B is tissue-specific and possibly gene-specific within a given cell type. For example, c-Rel/p50 and p50/p50 bind to the CS4 promoter site (GGGGGTCTCC) in hippocampus, but only the latter dimer binds to this sequence in basal forebrain (Qui et al., 2001). Also, following hypoxia, the temporal pattern of c-Rel/p50 and p50/p50 binding in the hippocampus is quite different depending on whether the CS4 or IgG-B sequence (GGGACTTTCC) is used to assess binding. In fact, in that study, p50/p65 binding to the IgG-B sequence was significantly decreased at several intervals after hypoxia whereas binding of this dimer to the CS4 sequence was undetectable (Qui et al., 2001). Thus, it is possible that the classical consensus sequence used in this study may not actually reflect the binding of NF-B to the bcl-x promoter and that the observed NMDA-induced decrease in Bcl-X_L could be indirectly mediated by another NF-B gene target. Interestingly, in preliminary experiments, we have found that NMDA (at 20 h) induces a change in the makeup of NF-B proteins from p50 and p65 containing dimers to p52 containing dimers. If p50/p65 is protective as others have suggested, perhaps apoptosis is the result of loss of this function. Resolution of this question will require additional experiments.

In summary, we have attempted to clarify the role of superoxide anion in NMDA receptor-mediated apoptosis and to delineate the downstream signaling pathways. It is hypothesized that superoxide generation via mitochondrial dysfunction and perhaps other Ca²⁺-sensitive enzymatic pathways occurs upstream of NF-B activation and acts as an intracellular signal by changing the redox state of the cell. The resulting translocation of NF-B is proposed to be pro-apoptotic either directly or indirectly leading to a relative change in the expression of the Bcl family proteins Bax (pro-apoptotic) and Bcl-X_L (anti-apoptotic). The ability of M40403 to prevent apoptosis at a relatively early stage in this cascade suggests that it may be possible to intervene therapeutically in clinical conditions such as stroke with drugs that target the production of superoxide anion.