Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

N-Acetylaspartylglutamate (NAAG) is a dipeptide found in the brain in millimolar concentrations (0.5-2.7 mM) (Pouwels and Frahm, 1997) that has been immunohistochemically localized to neurons, particularly those known to be glutamatergic (Williamson and Neale, 1988; Tsai et al., 1990, 1993), and to reactive microglia (Passani et al., 1997). There is evidence that suggests that NAAG could act both as a neurotransmitter or serve as a precursor to synaptic glutamate (Mori-Okamoto et al., 1987; Moffett et al., 1990; Tsai et al., 1990).

NAAG has been shown to be released in a calcium-dependent manner on synaptic stimulation (Tsai et al., 1988, 1990; Zollinger et al., 1988) and thus has been considered to be a candidate neurotransmitter. Although results from early electrophysiological studies of bath or iontophoretically applied NAAG were inconclusive (Bernstein et al., 1985; Westbrook et al., 1986; Joels et al., 1987; Mori-Okamoto et al., 1987; Whittemore and Koerner, 1989; Sekiguchi et al., 1992), more recent studies have demonstrated effects of NAAG as a partial antagonist/agonist at the N-methyl-D-aspartate (NMDA) receptor (Sekiguchi et al., 1989; Puttfarcken et al., 1993; Valivullah et al., 1994) and as an agonist at metabotropic type II glutamate receptors (Wroblewska et al., 1997; Bruno et al., 1998).

NAAG is hydrolyzed to N-acetylaspartate (NAA) and glutamate both in vitro (Robinson et al., 1987) and in vivo (Stauch et al., 1989) by a neurocarboxypeptidase called NAALADase (N-acetylated -linked acidic dipeptidase; glutamate carboxypeptidase II, EC3.4.17.21) found on the external surface of cells (Cassidy and Neale, 1993a), with gene expression predominantly in astrocytes (Berger et al., 1999). The localization of NAALADase suggests the possibility that NAAG could act as a glutamate precursor because the glutamate liberated from NAAG could act directly at glutamate receptors.

Modulation of NAALADase activity could control the effects of NAAG. When NAALADase is active, NAAG would act as a potential glutamate precursor. When NAALADase is inactive, NAAG would act as a potential neuromodulator at both the NMDA and metabotropic glutamate receptor (mGluR) sites. Until recently, the ability to study the role of NAALADase activity in the brain was limited by the lack of specific and potent enzyme inhibitors (Serval et al., 1990; Slusher et al., 1999). However, in 1996, the first potent NAALADase inhibitor termed 2-(phosphonomethyl)-pentanedioic acid (2-PMPA) was described (Jackson et al., 1996). We have previously demonstrated that 2-PMPA is neuroprotective in cell culture and animal models of ischemia through inhibition of NAALADase (Slusher et al., 1999). As predicted by our hypothesis, we observed that 2-PMPA decreased glutamate while simultaneously increasing NAAG extracellularly after middle cerebral artery occlusion (Slusher et al., 1999). Decreased glutamate availability is expected to be neuroprotective by limiting toxic glutamate receptor activation (Rothman and Olney, 1986; Meldrum, 1990), whereas increased NAAG could be neuroprotective via both activation of mGluRs (Bruno et al., 1995, 1998; Buisson and Choi, 1995; Wroblewska et al., 1997) and blockade of NMDA receptors (Sekiguchi et al., 1989; Puttfarcken et al., 1993; Valivullah et al., 1994).

To understand the likely effects of endogenous NAAG in the central nervous system, we have now investigated the activity of NAAG in primary cortical cultures while manipulating NAALADase activity. We report that under hydrolyzing conditions, when NAALADase is active, NAAG was toxic and that this toxicity could be reversed by NAALADase inhibitors and glutamate receptor antagonists. However, under nonhydrolyzing conditions, when NAALADase is inactive, we report that NAAG provided neuroprotection following both NMDA toxicity and metabolic inhibition. This neuroprotective activity was found to be mediated through both NMDA antagonism and mGluR agonism.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Dissociated Cortical Cell Cultures

Dissociated cortical cultures were prepared from 17-day-gestation Harlan Sprague-Dawley rat fetuses as previously described (Vornov, 1995). In brief, the fetuses were removed from timed pregnant rats, and the cortices were rapidly isolated in Dulbecco's phosphate-buffered saline. The cortices were then mildly triturated and incubated in Earle's balanced salt solution containing papain, at 37°C, for 15 min. At the end of the incubation period the cortices were pelleted at 2000g. The pellet was then resuspended in Earle's balanced salt solution containing DNase (0.05 mg/ml), BSA (1 mg/ml), and trypsin inhibitor (1 mg/ml) and repelleted at the same centrifugal force. The pellet was then suspended in plating medium containing 10% fetal bovine serum, 5% heat-inactivated equine serum, and 85% modified minimum essential medium (MEM) with high glucose (4.5 g/l). The cultures were subsequently plated at 2.5 × 10⁵ cells/ml on 24-well plates that had been pretreated with poly(D-lysine) (10 µg/ml) for 1 h. Every 3 to 4 days, half the medium was exchanged for new growth medium containing 5% heat-inactivated horse serum, 2 mM glutamine, and 95% modified MEM. Cultures were maintained for 18 to 20 days, at 37°C, in a 5% CO₂ incubator before they were used in the experiments.

NAAG Exposure under Hydrolyzing Conditions

NAALADase activity is inhibited by the millimolar phosphate concentrations (Robinson et al., 1987; Stauch et al., 1989) generally used in tissue culture experiments. To study the effects of NAAG exposure under hydrolyzing conditions, it was necessary to expose cells and allow for their subsequent recovery in solutions that were phosphate free. All NAAG exposures were performed at 37°C, for 20 min in phosphate-free HEPES buffered saline solution (HBSS) containing 143 mM NaCl, 5.4 mM KCl, 1.2 mM MgSO₄, 5.0 mM HEPES, 2.0 mM CaCl₂, 11.1 mM D-glucose, and 0.005% phenol red. After exposure, cultures were washed once with phosphate-free HBSS and allowed to recover for 24 h in phosphate-free Dulbecco's modified Eagle's medium. Unless otherwise specified, protective drugs were present during both exposure and throughout recovery.

NAAG Exposure under Nonhydrolyzing Conditions

To examine the effects of NAAG under nonhydrolyzing conditions, cultures were exposed to NAAG at 37°C for 20 min in standard HBSS containing 1.2 mM NaH₂PO₄, washed with the HBSS, and allowed to recover for 24 h in growth medium.

NMDA Toxicity. The effects of intact NAAG on NMDA toxicity were examined by exposing cultures to NAAG and 300 µM NMDA for 5 min in phosphate containing HBSS. We have previously reported that this concentration of NMDA causes complete morphological destruction of neurons (Thomas et al., 1999). Cultures were then washed with the HBSS and allowed to recover for 1 h in growth medium containing NAAG, followed by a further 24-h recovery in fresh growth medium without NAAG. The NAAG was removed after 1 h because in preliminary experiments we observed toxic effects from prolonged exposures to NAAG.

Metabolic Inhibition. As previously described (Vornov et al., 1996), ischemic conditions were simulated by a 20-min exposure to 10 mM 2-deoxyglucose (2-DG) to inhibit glycolysis and 5 mM KCN to inhibit mitochondrial function in glucose-free HBSS. We have previously reported that this exposure to metabolic inhibitors causes complete morphological destruction that is equivalent to that caused by the NMDA exposure, as described above (Vornov, 1995; Vornov et al., 1996; Thomas et al., 1999). As in NMDA toxicity experiments, the effects of NAAG were examined by adding NAAG during metabolic inhibition and during the 1st h of recovery.

Lactate Dehydrogenase (LDH) Measurements

Cellular injury was assessed by measuring the accumulation of LDH (Koh and Choi, 1987) in the culture medium at the end of the 24-h recovery period. Each experiment contained an independent control in which no metabolic inhibitors or NMDA was added and an independent measurement of maximal injury. In most experiments maximal injury was defined as that caused by a 20-min exposure to 5 mM KCN and 10 mM 2-DG. In the NMDA toxicity experiments maximal injury was defined by a 5-min exposure to 300 µM NMDA. Data were then normalized within each experiment as a percentage of maximal injury and control ((condition  control)/(maximum injury  control) × 100). All data are presented as mean ± S.E. of three determinations in three or more independent experiments. The values were compared by a two-population t test.

NAALADase Activity in Cortical Neurons under Both Hydrolyzing and Nonhydrolyzing Conditions

To determine NAALADase activity in the cortical cultures, under hydrolyzing and nonhydrolyzing conditions, cultures were first washed with HBSS containing zero and 1.2 mM NaH₂PO₄, respectively. The cells were then scraped off the 24-well plate, centrifuged at 1000g for 10 min, and resuspended in the respective buffers (1 ml), without and with phosphate. The enzymatic activity in these cultures was determined using the NAAG hydrolysis assay as described elsewhere (Robinson et al., 1987). Briefly, approximately 100 µg each of tissue protein, suspended in phosphate and phosphate-free buffers, were preincubated in 50 mM Tris-Cl (pH 7.4) and 1 mM CoCl₂ for 10 min at 37°C. Subsequently, 50 µl of 0.6 µM [³H]NAAG, radiolabeled at the terminal glutamate, was added and the incubation continued. After 15 min, the enzymatic reaction was stopped with ice-cold 0.1 M phosphate buffer, and the reaction products were separated using an ion-exchange chromatographic column. NAALADase activity in these cortical cells, under both hydrolyzing and nonhydrolyzing conditions, was determined as a measure of the liberated radiolabeled glutamate using liquid scintillation spectrometry.

Materials

NAAG was obtained from either Tocris-Cookson (Ballwin, MO) or Bachem (Torrance, CA); radiolabeled NAAG from NEN Life Science Products (Boston, MA); (+)-MK801 hydrogen maleate (MK-801) and 6-cyano-2,3-dihydroxy-7-nitroquinoxaline (CNQX) from Research Biochemicals International (Natick, MA); and (+)--methyl-4-carboxyphenylglycine (MCPG), (2S)--ethylglutamic acid (EGLU), (R,S)-1-aminoindan-1,5-dicarboxylic acid (AIDA), and (R,S)--methylserine-O-phosphate (MSOP) from Tocris-Cookson. 2-PMPA was synthesized by SRI International (Menlo Park, CA). All other drugs and chemicals were obtained from Sigma (St. Louis, MO). Regular and phosphate-free HBSS were custom made by Paragon Biotech Inc. (Baltimore, MD). Dulbecco's modified Eagle's medium and MEM were purchased from Life Technologies, Inc. (Rockville, MD). The serum used in making up the tissue culture medium was obtained from HyClone Laboratories (Logan, UT).

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Under Hydrolyzing Conditions, NAAG Caused Glutamate Toxicity

NAAG Is Toxic in Phosphate-Free Medium. To determine whether NAAG would cause glutamate toxicity via hydrolysis by NAALADase, we compared the effects of NAAG in phosphate-containing and phosphate-free medium because phosphate is an inhibitor of NAALADase (Robinson et al., 1987; Stauch et al., 1989). As shown in Fig. 1, NAAG caused dose-dependent toxicity to dissociated rat cortical cultures in phosphate-free medium. Under these hydrolyzing conditions, the injury caused by 10 mM NAAG was maximal, defined herein as that caused by the maximal toxic concentrations of metabolic inhibitors, 10 mM 2-DG and 5 mM KCN. We confirmed that NAALADase was active (specific activity = 0.2 pmol/mg/min) in the phosphate-free medium using the previously described "[³H]NAAG radioenzymatic assay" (Robinson et al., 1987). In contrast, when the phosphate concentration in the medium was raised to 1.2 mM, a concentration 10-fold higher than phosphate's IC₅₀ for NAALADase inhibition, NAAG no longer showed significant toxicity. Concurrently, we determined that at this millimolar phosphate concentration, NAALADase was inactive (specific activity = 0 pmol/mg/min).

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Neurotoxic effects of NAAG when NAALADase is active. Rat cortical cultures were exposed to NAAG for 20 min under both hydrolyzing (without phosphate) and nonhydrolyzing (1.2 mM phosphate) conditions. Cellular injury was quantified by measuring LDH released into the medium during a 24-h recovery period. Data were normalized within each experiment as a percentage of maximal injury and control, as detailed under Experimental Procedures, and presented as mean ± S.E. The measured activity of the enzyme under each condition is shown. Without phosphate, NAALADase is active, and NAAG is dose dependently neurotoxic. With 1.2 mM phosphate, NAALADase is inactive, and NAAG is not neurotoxic. At each concentration, toxicities were compared using Student's two-population t test (*P < .05).

2-PMPA, a NAALADase Inhibitor, Inhibited NAAG Toxicity. We next examined whether pharmacological blockade of NAALADase with 2-PMPA would also limit NAAG toxicity. 2-PMPA (10 µM), present during exposure and throughout the 24-h recovery, significantly reduced the toxicity of 3 mM NAAG (Fig. 2; 76% reduction, P < .001).

View larger version (15K): [in this window] [in a new window]   Fig. 2.   NAALADase inhibition by 2-PMPA significantly reduces NAAG toxicity. Neurotoxicity of NAAG under hydrolyzing conditions was measured with and without the specific NAALADase inhibitor 2-PMPA (10 µM), as described in Fig. 1. 2-PMPA was present during exposure and throughout recovery. At each concentration, toxicities were compared using Student's two-population t test (*P < .05, **P < .01, ***P < .001).

Glutamate-Receptor Antagonists Protected against NAAG Toxicity. We also investigated the effects of glutamate receptor antagonists on NAAG toxicity to determine whether the toxicity was caused by the glutamate liberated from NAAG via NAALADase. As shown in Fig. 3, both MK-801 (50 µM), a noncompetitive NMDA receptor antagonist, and CNQX (100 µM), a competitive non-NMDA receptor antagonist, protected against the toxicity of 3 mM NAAG (83%, P < .001 and 63%, P < .01, respectively). In contrast, the mGluR group II antagonist MCPG (100 µM) did not provide statistically significant protection (24%, P > .05).

View larger version (10K): [in this window] [in a new window]   Fig. 3.   Effects of glutamate receptor antagonists on NAAG toxicity. Neurotoxicity of 3 mM NAAG was measured as in Figs. 1 and 2. Both MK-801 (50 µM), a noncompetitive NMDA receptor antagonist, and CNQX (100 µM), a competitive non-NMDA receptor antagonist, provide significant neuroprotection. The protection afforded by the mGluR group II antagonist MCPG (100 µM) is not statistically significant. All antagonists were present during NAAG exposure and throughout recovery (**P < .01, ***P < .001).

Under Nonhydrolyzing Conditions, NAAG Protected against NMDA Toxicity and Metabolic Inhibition

NAAG Protection against NMDA Toxicity. To investigate whether intact NAAG would act as a partial NMDA receptor antagonist/agonist as previously suggested (Sekiguchi et al., 1989; Puttfarcken et al., 1993; Valivullah et al., 1994) we examined the effect of NAAG on NMDA toxicity. As shown in Fig. 4, under nonhydrolyzing conditions, NAAG protected against NMDA toxicity (300 µM) when present at concentrations up to 100 µM (80%, P < .001). However, at higher concentrations, the protective effect of NAAG was lost, consistent with previous reports of NMDA agonist effects at concentrations greater than 100 µM (Bruno et al., 1998). Additionally, we cotreated cultures with AIDA (100 µM; group I antagonist), MCPG (10, 30, and 100 µM) and EGLU (100 µM; group II antagonists), and MSOP (100 µM; group III antagonist) to determine whether the protection afforded by NAAG against NMDA toxicity was mediated through the various mGluR receptors. As shown in Fig. 4, B and C, MCPG, AIDA, EGLU, and MSOP had no significant effect.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Neuroprotective effects of NAAG against NMDA toxicity when NAALADase is inactive. Rat cortical cultures were exposed for 5 min to NMDA (300 µM) with various concentrations of NAAG. Cultures were then washed and allowed to recover for 1 h in growth medium containing the same concentrations of NAAG, followed by a further 24-h recovery in fresh growth medium without NAAG. Cellular injury was quantified as in previous figures except that NMDA (300 µM) was used as maximal injury to normalize the data. A, dose-dependent neuroprotection is provided by NAAG against NMDA toxicity. B, neuroprotection provided by 100 µM NAAG, against NMDA toxicity, cannot be reversed by MCPG (10, 30, and 100 µM). C, neuroprotection provided by 100 µM NAAG, against NMDA toxicity, cannot be reversed by other mGluR antagonists, AIDA (100 µM), EGLU (100 µM), or MSOP (100 µM). AIDA, EGLU, MCPG, and MSOP were present during exposure and throughout recovery (*P < .05, ***P < .001 versus NMDA; P < .05, P < .001 versus NMDA + NAAG).

NAAG Protection against Metabolic Inhibition. At concentrations up to 100 µM, intact NAAG also protected against neuronal injury caused by metabolic inhibition (82%, P < .001; Fig. 5A). At higher concentrations, the protective effect of NAAG was lost. The U-shaped dose-response curve was similar to that observed in the NMDA toxicity experiments. However, the protection afforded by NAAG against metabolic inhibition was reversed by both group II mGluR antagonists, MCPG (100 µM; Fig. 5B) and EGLU (100 µM; Fig. 5C). However, as shown in Fig. 5C, neither the group I antagonist AIDA nor the group III antagonist MSOP had any significant effect.

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Neuroprotection by NAAG against metabolic inhibition is blocked by MCPG. To simulate ischemic injury, cultures were exposed for 20 min to 10 mM 2-DG/5 mM KCN with various concentrations of NAAG. Cultures were then washed, and allowed to recover for 1 h in growth medium containing the same concentrations of NAAG, followed by a further 24-h recovery in fresh growth medium without NAAG. A, dose-dependent neuroprotection provided by NAAG against metabolic inhibition. B, neuroprotection provided by 100 µM NAAG, against metabolic inhibition, is dose dependently reversed by MCPG (10, 30, and 100 µM). C, neuroprotection provided by 100 µM NAAG, against metabolic inhibition, is reversed by EGLU (100 µM) and not by the other mGluR antagonists, AIDA (100 µM) and MSOP (100 µM). AIDA, EGLU, MCPG, and MSOP were present during exposure and throughout recovery (*P < .05, **P < .01, ***P < .001 versus 2-DG/KCN; P < .05, P < .01, P < .001 versus 2-DG/KCN + NAAG).

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

These results provide evidence that the effects of NAAG depend on the activity of NAALADase. Under hydrolyzing conditions, when NAALADase was active, NAAG had toxic effects, presumably due to the liberation of glutamate from NAAG and subsequent activation of NMDA and -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)/kainate receptors. Under nonhydrolyzing conditions, when NAALADase was inhibited, NAAG demonstrated neuroprotective effects against both NMDA toxicity and metabolic inhibition. In NMDA-induced toxicity, NAAG appeared to act through its partial antagonist activity at the NMDA receptor. In metabolic inhibition, NAAG's primary effect appeared to be mediated through its agonist properties at the mGluR.

Under hydrolyzing conditions, the observation that glutamate receptor antagonists block NAAG toxicity is entirely consistent with the simple hypothesis that the exogenous NAAG is converted to extracellular glutamate. However, NAAG was less toxic than equimolar glutamate. Although our previous reports show 1 mM glutamate to cause toxicity about equal to that caused by metabolic inhibition (Vornov, 1995), these experiments show 1 mM NAAG to cause only 50% of the toxicity caused by metabolic inhibition. This suggests that, under our culture conditions, the conversion of NAAG to glutamate may not be rapid and quantitative. In addition, glutamate transporters could limit the toxic potency of NAAG if the rate of glutamate liberation from NAAG is not significantly greater than the rate of glutamate transport. This hypothesis is consistent with what is known about glutamate toxicity. Glutamate is a less potent toxin, both in vitro and in vivo, than would be predicted by its receptor affinities because glutamate transporters remove glutamate from the extracellular space before it has an opportunity to act at the receptors (Robinson et al., 1993).

Our results are consistent with the hypothesis that NAALADase is the enzyme responsible for the hydrolysis of NAAG. NAAG toxicity could be prevented by two distinct manipulations known to inhibit NAALADase. Both 2-PMPA, a potent and selective NAALADase inhibitor, and millimolar phosphate completely blocked NAAG toxicity. It is clear that millimolar phosphate does not directly block glutamate toxicity because phosphate does not affect the toxicity of NMDA (Fig. 4) or glutamate (Vornov, 1995). In addition, we have previously reported that 2-PMPA has no significant affinity for glutamate receptors (Slusher et al., 1999).

Under nonhydrolyzing conditions, the effects of NAAG were predominantly neuroprotective. NAAG as a neuroprotectant, under nonhydrolyzing conditions, was about 30 times more potent than NAAG as a neurotoxin under hydrolyzing conditions. Although NAAG transport has been demonstrated (Williamson and Neale, 1992; Cassidy and Neale, 1993b), the transporters themselves must not have sufficient capacity or affinity to limit NAAG availability in the way glutamate is limited by glutamate transport, resulting in the greater potency under nonhydrolyzing conditions.

The blockade of NMDA toxicity by NAAG can be attributed simply to NAAG's action as an NMDA receptor antagonist at micromolar concentrations. Since NMDA toxicity is a direct result of ionic fluxes through the receptor-gated channels, NAAG's blockade of the NMDA receptor would directly limit these ionic fluxes. The U-shaped dose-response curve is most likely explained by the partial antagonist properties of NAAG. As observed by Bruno et al. (1998), at higher concentrations, the antagonism is lost as NAAG gains agonist properties. We note, however, that under nonhydrolyzing conditions we saw no evidence of NMDA receptor-mediated toxicity by NAAG even at concentrations as high as 10 mM. This is consistent with our previous observations that NAALADase inhibitors do not cause NMDA receptor-mediated toxicity through the accumulation of NAAG (Slusher et al., 1999).

NAAG is a type II mGluR agonist with high degree of specificity for mGluR3 receptors (Wroblewska et al., 1997). Agonists at the mGluRs have been reported to protect against NMDA toxicity (Bruno et al., 1995, 1998; Buisson and Choi, 1995). We therefore reasoned that NAAG could potentially protect against NMDA toxicity also via the mGluR. However, in our experiments, the blockade of NMDA toxicity by NAAG was unaffected by any of the mGluR antagonists, AIDA (group I), MCPG and EGLU (group II), or MSOP (group III). The partial antagonist effects of NAAG at the NMDA receptor appear to be sufficient, and activation of mGluRs not necessary to protect cells against NMDA toxicity (Bruno et al., 1998).

Under nonhydrolyzing conditions, NAAG also protected against metabolic inhibition. The U-shaped dose-response curve of neuroprotection against metabolic inhibition was similar to that observed against NMDA toxicity and is probably due to similar mechanisms of action, partial antagonism of NAAG at the NMDA receptor at lower concentrations and agonism of NAAG at the NMDA receptor at higher concentrations. This same U-shaped dose-response curve was described by Bruno et al. (1998) in their demonstration of NAAG's metabotropic protective effects, which they also explained as being due to gain of NMDA receptor activation at high NAAG concentrations. We did not observe this toxicity in normal cells at similar NAAG concentrations. This might be due to increased sensitivity to NMDA toxicity in metabolically inhibited cells as has previously been reported by our laboratory (Vornov, 1995) and by other investigators (Novelli et al., 1988).

Although MCPG or EGLU did not reverse the protective effects of NAAG against NMDA exposure, both MCPG and EGLU did reverse the protective effects of NAAG against metabolic inhibition. In the case of NMDA toxicity, exogenous application of NMDA stimulates postsynaptic NMDA receptors. Coapplication of NAAG could directly antagonize the activation of the NMDA receptors as has been previously reported (Bruno et al., 1998). In the case of metabolic inhibition, excess glutamate is released during and after the insult (Goldberg et al., 1987; Gill et al., 1988; Vornov, 1995). The reversal of the protective effects of NAAG against metabolic inhibition by the mGluR antagonists MCPG and EGLU is consistent with the hypothesis that NAAG's protective effects are primarily due to its presynaptic activation of mGluR3 receptors, which can limit glutamate release (Hayashi et al., 1993; Sanchez-Prieto et al., 1996).

However, if NAAG acts as an NMDA receptor antagonist, as discussed above, it is perhaps surprising that we did not observe any protective effects, in the presence of either MCPG or EGLU, against metabolic inhibition. One possibility is that NAAG, as a partial antagonist at the NMDA receptor, was not able to compete with the glutamate endogenously released during recovery from metabolic inhibition. In addition, the glutamate released during recovery from metabolic inhibition would be expected to activate both NMDA and non-NMDA receptors, whereas the toxic effects of NMDA are mediated solely through NMDA receptors. Thus, partial NMDA receptor antagonism could be less effective in metabolic inhibition. Regardless of the precise mechanisms involved, it may be significant that the protective effects of NAAG against injury caused by metabolic inhibition are mediated primarily through metabotropic receptor activation and not through NMDA receptor blockade.

In these experiments examining exogenous NAAG, we have not investigated what role endogenous NAAG plays in the central nervous system. Clearly, this will depend on the activity of NAALADase because the effect of NAAG that we observed was dependent on whether the enzyme was active. We have previously shown that NAALADase inhibition can limit the accumulation of extracellular glutamate during middle cerebral artery occlusion and reperfusion in rats, suggesting that NAALADase is active during pathological events such as ischemia. There is hydrolysis of labeled NAAG microinjected into the brain, suggesting that NAALADase has a normal basal activity and NAAG will generate glutamate (Stauch et al., 1989). However, NAALADase inhibitors have no acute behavioral effects in normal animals (Slusher et al., 1999), suggesting that normal synaptic activity is not dependent on glutamate generated from NAAG. NAAG and the glutamate liberated from NAAG could play a role in metabolism or modulatory signaling events. Regardless of the normal physiological role of NAAG and NAALADase, NAALADase inhibitors are a promising approach to neuroprotection because these drugs may both increase NAAG, thus providing protection via metabotropic receptor activation and NMDA receptor blockade, and decrease glutamate availability.