Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Increased extracellular glutamate and the resulting overstimulation of excitatory amino acid receptors is one of the major consequences of ischemic brain injury (Garcia, 1995; Martin et al., 1998). The resulting excitotoxicity is due to a loss of intracellular ionic homeostasis and the downstream activation of Ca²⁺-dependent cell death cascades (Lipton and Rosenburg, 1994; Koroshetz and Moskowitz, 1996; Nicotera and Lipton, 1999), leading to an apoptotic/necrotic continuum of cellular death (Cheung et al., 1998; Martin et al., 1998). Although several agents targeting glutamate receptors have shown neuroprotective efficacy to reduce brain injury in animal models, none are currently clinically available (Muir and Grosset, 1999).

Seizurogenic activity develops in many patients following a focal ischemic brain lesion and may be involved in the pathophysiological effects of stroke. Indeed, the incidence of poststroke epilepsy has been reported to range from 4 to 43% of human patients depending on the type of insult suffered, including cerebral infarction or cerebral hemorrhage (Kotila and Waltimo, 1992). The role of ischemia-mediated seizure activity in the progression of brain injury is not currently known. However, these seizures can occur prior to, immediately following, or up to several weeks following the onset of stroke (Armon et al., 1991). Although seizure-blocking drugs are prescribed to reduce convulsive seizures that develop following brain injury (Arboix et al., 1997), their effects have not been extensively evaluated either as adjuncts or as novel neuroprotective strategies in brain trauma.

GPI5232 (Fig. 1) is a more lipophilic analog of the NAALADase inhibitor 2-PMPA [2-(phosphonomethyl)pentanedioic acid]. NAALADase is an abundant brain dipeptide responsible for the enzymatic breakdown of N-acetyl-aspartylglutamate (NAAG). Breakdown of NAAG involves the release of glutamate (Vornov et al., 1999). Therefore, the reduction in NAALADase activity may be a homeostatic control mechanism to reduce susceptibility of the injured brain to seizure activity (Meyerhoff et al., 1989). Inhibition of NAALADase with 2-PMPA has been shown to be an effective neuroprotective mechanism in vitro (Tortella et al., 2000) and also to reduce focal ischemic injury in vivo (Slusher et al., 1999; Vornov et al., 1999).

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Chemical structure of GPI5232.

In this study we describe the effect of GPI5232 to inhibit NAALADase activity in vitro and its in vivo neuroprotective profile to reduce brain injury and improve functional recovery from experimental focal ischemia/reperfusion in rats. In addition, we have analyzed quantitative (spectral analysis) and qualitative (visual wave form analysis) measures of brain EEG activity. The most profound results measured were a reduction in brain infarction and attenuation of ischemic seizure activity following GPI5232 treatment. Critically, in normal rats there were no signs of EEG toxicity or behavioral side effects at the highest dose of GPI5232 tested, suggesting a highly favorable therapeutic index for the neuroprotection treatment of an ischemic insult with this compound.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

In Vitro NAALADase Inhibition by GPI5232. Hydrolysis of N-acetyl-aspartyl-[³H]glutamate catalyzed by NAALADase was carried out as previously described (Slusher et al., 1990). Briefly, NAAG (30 nM) was incubated with NAALADase (rat brain synaptosomal membranes ~100 µg of membrane protein/ml) in the presence of CoCl₂ (1 mM) at pH 7.4 (Tris buffer, 20 mM) for 15 min at 37°C in a final volume of 1 ml. The reaction was carried out in the absence and presence of different concentrations of GPI5232. The reaction was terminated with 1 ml of ice-cold sodium phosphate buffer (0.1 M, pH 7.4). The assay mixture (200-µl aliquot) was applied to an anion exchange column (prepared in five 3/4-inch Pasteur capillary pipettes) and [³H]glutamate was eluted with 2 ml of 1 M formate. N-Acetyl-aspartyl-[³H]glutamate remained bound to the column. [³H]Glutamate production was determined by scintillation counting.

Surgical Procedures. Male Sprague-Dawley rats (270-330 g; Charles River Labs, Raleigh, VA) were used in all of the following procedures. Anesthesia was induced by 5% halothane and maintained at 2% halothane delivered in oxygen. Indwelling i.v. cannulas (PE-50) were placed into the left jugular vein of all animals for drug delivery. For EEG experiments, epidural electrodes (stainless steel screw electrodes, 0-80 × 1/8 in) were permanently implanted and fixed to the skull using dental acrylate cement (Tortella et al., 1997). Body temperature was maintained normothermic (37 ± 1°C) throughout all surgical procedures by means of a homeothermic heating system (Harvard Apparatus, South Natick, MA). Food and water were provided ad libitum pre- and postsurgery. The animals were individually housed under a 12-h light/dark cycle. For the brain injury studies, 3 to 5 days of recovery was allowed following the surgical procedures described above. Rats were then re-anesthetized and prepared for temporary focal ischemia using the filament method of MCAo and reperfusion as described elsewhere (Britton et al., 1997). Briefly, the right external carotid artery was isolated and its branches coagulated. A 3-0 uncoated monofilament nylon suture with rounded tip was introduced into the internal carotid artery via the external carotid artery and advanced (approximately 22 mm from the carotid bifurcation) until a slight resistance was observed, thus occluding the origin of the middle cerebral artery (MCA). The endovascular suture remained in place for 2 h and was then retracted to allow reperfusion of blood to the MCA. Following MCAo surgery, animals were placed in recovery cages with ambient temperature maintained at 22°C. During the 2-h ischemia period and the initial 6-h postischemia period, 75-W warming lamps were also positioned directly over the tops of each cage in order to maintain body temperature normothermic throughout the experiment. Body temperatures were recorded pre-MCAo, 1, 2, 5, and 24 h postocclusion. At 24 h, the rats were euthanized and their brains removed for quantification of infarction.

MCAo Dose Response and Time Course/No EEG. Intravenous bolus injections of either vehicle (50 mM HEPES in saline) or GPI5232 were administered post-MCAo followed by continuous i.v. infusion. Infusion and dosing protocols are as described in Tables 1 and 2.

MCAo/EEG Experiments. Two electrodes were placed over the right parietal cortex (5 mm right of bregma, and 5 mm right, 4 mm posterior to bregma, respectively), and two electrodes were placed over the right temporal cortex (in the same coronal plane of each parietal electrode but located 1 mm below the right lateral ridge of the skull). A fifth reference electrode was implanted over the occipital cortex. An i.v. bolus injection of vehicle or 30 mg/kg GPI5232 was given at 1 h postocclusion and a continuous i.v. infusion was immediately begun (6 mg/kg/h for 4 h). Rats were housed in EEG recording chambers (as described above), and EEG activity was recorded before the MCAo surgery (baseline recording), continuously during the 2 h of MCAo and the ensuing 3 h following reperfusion, and once again at 24 h postocclusion. Spectral parameters were evaluated 1, 2, 5, and 24 h post-MCAo and compared with baseline, pre-MCAo values. Percentage of recovery in EEG power was calculated as the percentage of drop in power at 24 h minus the percentage drop in power at 1 h as compared with baseline values.

Normal Rats/EEG Experiments. Four recording electrodes were placed bilaterally over the right and left frontal (3 mm anterior, 2 mm lateral to bregma) and parietal (3 mm posterior, 2 mm lateral to bregma) cortices. A fifth reference electrode was implanted over the occipital cortex. Rats were individually housed in Plexiglas recording chambers equipped with custom designed multichannel mercury swivel commutators (Dragonfly Inc., Ridgeley, WV). On the morning of the experiment, the rats were connected to the swivel system by flexible shielded cables providing a noise-free connection from the unrestrained rat to a Grass Instruments model 7D polygraph (West Warwick, RI) and digital analysis system (Neurodata, Inc., Pasadena, CA) while permitting freedom of movement by the animals during all phases of the experiment. Control EEG recordings were collected such that the drug testing was initiated only after each subject exhibited normal behavioral and EEG slow-wave sleep (SWS) patterns. All EEG experiments routinely began between 9:00 AM and 11:00 AM and were considered complete upon the reemergence of SWS. Using this protocol, each animal served as its own control. Throughout these experiments, all rats were drug naïve and used only once. Following the onset of normal SWS rhythms and behavior, the experiment was initiated by giving an i.v. injection of the control vehicle. With the reemergence of EEG SWS after the control injections, the animals (n = 5 per dose) were injected acutely with i.v. GPI5232 (100 and 200 mg/kg) or the same infusion protocol used in the MCAo experiments (i.e., 30 mg/kg i.v. bolus followed by continuous i.v. infusion of 6 mg/kg/h for 4 h). For the duration of the EEG experiments, each animal was observed for signs of abnormal behavioral activity including sedation, ataxia, enhanced locomotion, stereotypic activity (including head weaving, grooming, preening, and scratching), and signs of clonic (convulsant) muscle activity. The behavioral responses for each animal were noted and recorded on the EEG polygraph records as a correlate to their respective changes in EEG activity. EEG samples for spectral analysis were taken at 10, 40, 100, 140 min, and 24 h post-GPI5232 injection and compared with the control, postvehicle injection.

Infarct Analysis. From each rat brain, analysis of ischemic cerebral damage included infarct volume (cortical and subcortical) and hemispheric infarct size (calculated as percentage of infarcted tissue referenced to the corresponding contralateral uninjured cerebral hemisphere, to exclude the possible contributing effect of hemispheric edema to infarct volume). Ischemic regions were defined as those areas completely lacking 2,3,5-triphenyl tetrazolium chloride (TTC) staining in seven coronal sections (2 mm thick). Brain sections were taken from the region beginning 1 mm from the frontal pole and ending just rostral to the cortico-cerebellar junction. Computer-assisted image analysis (Loats Associates, Westminster, MD) was used to calculate infarct volumes by sequential integration of the respective areas to yield infarct volume (mm³), as described in detail elsewhere (Tortella et al., 1999). Similarly, ipsilateral and contralateral hemispheric volumes were measured where hemispheric swelling (edema) was expressed as the percentage of increase in size of the ipsilateral (occluded) hemisphere over the contralateral (uninjured) hemisphere.

EEG Data Analysis. Bipolar EEG was recorded from each pair of recording electrodes. Off-line quantitative spectral analysis of several spectral parameters including power, relative power, mean frequency, complexity, mobility, and zero crossing was done on 60-s sample epics of cortical EEG activity using QND signal analysis software (Neurodata Inc.). For each animal, a control EEG sample was recorded before the experiment began and was used as a baseline value to compare all subsequent EEG recordings for that animal. Analog signals were also visually studied for the presence of abnormal rhythmic or polymorphic delta activity as well as seizure spikes, sharp waves, polyspikes, or spike/slow-wave complexes.

Statistical Analysis. Data are presented as the mean ± standard error of the mean. Unless otherwise noted, statistical analysis of neuroprotective recovery was done by analysis of variance followed by Dunnett's post hoc test to compare individual treatment doses to the vehicle, control group. Spectral parameters were evaluated by planned comparisons using independent t tests with the Bonferroni correction for multiple tests. Statistical analysis was done using the Minitab Statistical Analysis software program (Minitab, State College, PA).

Compound. GPI5232 and 2-PMPA were received from Guilford Pharmaceuticals (Baltimore, MD). The compounds were dissolved in 50 mM HEPES in physiological saline immediately prior to testing and administered as an i.v. bolus followed by continuous i.v. injection or as a single acute i.v. injection (100 and 200 mg/kg).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

NAALADase Activity. Using a purified NAALADase enzyme isolated from rat brain synaptosomal membranes, the IC₅₀ for GPI5232 to inhibit NAALADase activity was determined to be 80 ± 14 nM (mean ± standard deviation). By comparison, the IC₅₀ measured for 2-PMPA was 0.029 ± 0.05 nM.

MCAo Injury. Control, vehicle-treated rats exhibited striatal and cortical brain infarction in the right hemisphere lobe from approximately 3 to 13 mm from the frontal pole. The mean total infarct volume in control rats ranged from 176 to 237 mm³ (Tables 1 and 2). These values correspond to a predominately cortical injury (81-89%) as compared with subcortical injury (11-19%). MCAo resulted in significant hemispheric edema of the injured hemisphere, representing approximately an 8 to 19% increase in cerebral volume compared with the contralateral, uninjured hemisphere. Both vehicle and GPI5232-treated animals lost approximately 15 to 20% body weight over the 24-h recovery period, regardless of treatment group, with no significant differences in body weight loss between groups. In a representative subset of vehicle and GPI5232-treated animals (n = 7 per group), MCAo caused a transient, mild hyperthermia from baseline values of 36.5 ± 0.2°C (preocclusion) to 38.2 ± 0.4°C at 2 h and 37.6 ± 0.2°C at 5 h (postocclusion). The rectal temperatures then returned to normal by 24 h (36.3 ± 0.2°C). Temperature measurements of GPI5232-treated animals were not significantly different from vehicle-treated animals at corresponding time points.

Neuroprotection/Dose Response. GPI5232 administration reduced both cortical and subcortical infarct volume following 2 h of MCAo and 22 h of reperfusion (Fig. 2). A dose-dependent reduction of infarct volume was established using an initial i.v. bolus injection at 60 min postocclusion followed by continuous i.v. infusion of drug or vehicle as summarized in Table 1. Significant reductions in infarct volume were measured using bolus injections as low as 10 mg/kg, but not at 3 mg/kg, with optimal reductions in infarct volume of nearly 50% in both cortical and subcortical regions using a 30-mg/kg bolus (as well as continuous infusion of drug as summarized in Table 1).

View larger version (47K): [in this window] [in a new window]   Fig. 2.   Representative forebrain images of vehicle versus GPI5232 (30-mg/kg bolus, followed by 6 mg/kg/h for 4 h) administration following 2 h of MCAo and 22 h of reperfusion. TTC-stained brain sections show completely white regions representing the core infarcted tissue. Arrows, locations of parietal and temporal electrodes.

Neuroprotection/Therapeutic Window. The therapeutic window of GPI5232 was assessed using the 30-mg/kg injection protocol but delaying the initial bolus injection time to 120 and 180 min postocclusion (Table 2). A significant reduction in infarction approaching 50% was again measured following a delay of 120 min, but not 180 min, postocclusion.

MCAo/EEG Analysis. EEG was continuously monitored for 5 h following MCAo and again at 24 h in a group of rats treated with either vehicle or GPI5232 (30-mg/kg i.v. bolus injection at 60 min post-MCAo and 4 h of continuous i.v. infusion as defined above). In this experiment, individual animals were excluded from the EEG study if they showed signs of subarachnoid or intracranial hemorrhage upon pathological examination of the brains or if the EEG amplitude did not drop upon insertion of the filament following MCAo surgery (Fig. 3). Consistent with results from our dose-response neuroprotection experiments, vehicle-treated animals (n = 7) had core infarct volumes of 229 ± 17 mm³ (22% hemispheric infarction as compared with contralateral, uninjured hemisphere). Post-treatment with GPI5232 (n = 7) significantly reduced ischemic core infarction to 108 ± 17 mm³ (53% neuroprotection) with no significant change in cerebral edema.

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Cortical EEG recorded from parietal brain regions ipsilateral to the injured hemisphere (peri-infarct zone). Shown are representative samples from a vehicle (A)- and GPI5232 (B)-treated animal before and after initiation of MCAo.

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Cortical EEG recorded from temporal brain regions ipsilateral to the injured hemisphere (core infarct zone). Shown are representative samples from a vehicle (A)- and GPI5232 (B)-treated animal before and after initiation of MCAo.

View larger version (24K): [in this window] [in a new window]   Fig. 5.   Effect of GPI5232 treatment on EEG power scores in both the parietal (A and B) and temporal (C and D) brain regions following MCAo as seen in each frequency band. Data are presented as mean ± S.E.M. *p < 0.05, power values at 24 h as compared with values at 1 h post-MCAo for each frequency range using independent t tests with Bonferroni's correction for multiple tests.

Brain Seizure Activity. EEG analysis also revealed the presence of brain seizure activity both during the 2 h of MCAo and after reperfusion. In the vehicle-treated group, five of seven animals exhibited seizure activity (Fig. 6A) in both parietal and temporal regions, which continued following vehicle injection at 1 h and ceased in all animals except one following reperfusion. Upon reperfusion, spike/slow wave bursting (Fig. 6B) appeared in the parietal regions in three of the seven animals during the 24-h recovery period. In the GPI5232-treated group, similar brain seizures developed in all seven animals during the MCAo but ceased in three of the animals following injection of GPI5232 and completely ceased in all animals except one upon reperfusion (Fig. 7). Furthermore, none of the GPI5232-treated animals exhibited spike/slow wave complexes during the 22-h reperfusion period.

View larger version (12K): [in this window] [in a new window]   Fig. 6.   Representative parietal EEG recordings (A) of seizure bursts during MCAo (B) and spike/slow-wave complexes following reperfusion (C).

View larger version (30K): [in this window] [in a new window]   Fig. 7.   Percentage of incidence of seizure spikes occurring during MCAo and spike/slow-wave activity following reperfusion. *p < 0.05, **p < 0.01, -square test as compared with predrug seizure activity.

EEG and Behavioral Toxicity Analysis. Single acute injections of GPI5232 (100 and 200 mg/kg) in normal (non-MCAo) rats produced no significant changes in EEG either qualitatively (Fig. 8) or quantitatively (spectral parameters including power, relative power, mean frequency, complexity, mobility, and zero crossing; data not shown). This included the lack of any drug-induced convulsant or brain (EEG) seizure activity. Furthermore, no abnormal behavioral changes, as detailed under Materials and Methods, were observed following drug injection. There was also no change in the time required for the animals to enter SWS following injections of either vehicle (10.6 ± 3.2 min) or GPI5232 (13.6± 3.2 min) at the highest dose tested (200 mg/kg).

View larger version (18K): [in this window] [in a new window]   Fig. 8.   Representative EEG tracings following acute injection of GPI5232 (200 mg/kg, i.v.) in normal rats.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

These data confirm a relatively potent (i.e., IC₅₀ <100 nM) inhibitory action of GPI5232 on brain NAALADase activity and describe a neuroprotection profile of postinjury treatment with GPI5232 to reduce brain injury and improve functional recovery from experimental transient focal ischemia and reperfusion in the rat. Brain ischemia represents a complex pathological process activating many signaling mechanisms promoting cell death. Using TTC staining, we have defined regions of injured brain that appear to be either completely infarcted (those areas completely lacking TTC stain) or pathologically compromised (regions of light pink staining). Since TTC can only be reduced to a red-formazan product in the presence of active oxidative mitochondrial enzymes, it represents a metabolic marker of cell viability and has been shown to correlate to other histological markers of brain injury such as hematoxylin and eosin (Bederson et al., 1986; Park et al., 1988). The core of the injury, which completely lacks TTC stain, may be undergoing a rapid form of necrotic cell death leading to infarction (Lipton, 1999). Coincidentally, the outer (i.e., peri-infarct) regions of the injury, which retain at least some active mitochondrial enzyme function, may also be undergoing energy-dependent controlled cell death processes of apoptosis and autophagocytosis (Martin et al., 1998; Lipton, 1999). In the present study, EEG from these two regions have been shown to display distinctive patterns following ischemic injury, patterns which we are now using to evaluate functional recovery following neuroprotective drug treatment.

As mentioned above, several mechanisms initiating cell death following ischemia have been studied, including apoptotic cascades, free radical formation, inflammation, and the excitotoxic involvement of glutamate release (Muir and Lees, 1995; Lipton, 1999). Related to the latter mechanism, i.e., glutamate release, is the role of NAAG in brain physiology and the inhibition of NAALADase (Wroblewska et al., 1993; Burlina et al., 1994). Since NAAG may also act as a storage form of glutamate that is in turn released by NAALADase activity, the inhibition of NAALADase may prove to be an effective target for neuroprotection. In primary cerebellar neuronal cultures, both NAAG and the NAALADase inhibitor 2-PMPA have been shown to be 100% neuroprotective against hypoxia/hypoglycemia (Tortella et al., 2000), whereas in rodent models of focal ischemia the NAALADase inhibitor 2-PMPA has been described as neuroprotective (Slusher et al., 1999; Vornov et al., 1999). In this study, calculation of inhibitory constants for rat brain NAALADase kinetics showed GPI5232 to possess inhibitory actions on NAALADase activity which, while less potent than 2-PMPA, was still measurable in the low nanomolar range (IC₅₀ = 80 nM). Furthermore, the selectivity of GPI5232 for the NAALADase enzyme has been confirmed in independent experiments (B. Slusher, unpublished data, Guilford Pharmaceuticals) where, with the exception of approximately 50% inhibition measured at guinea pig sigma-1 binding sites at high (i.e., 10 µM) concentrations of GPI5232, inhibition constants for GPI5232 targeting 43 other receptor systems and 70 enzymes have revealed no significant binding properties of GPI5232 at any sites other than NAALADase.

In this study we have used a model of temporary MCAo and reperfusion to produce brain injury in the rat. Although reperfusion may aggravate the injury process due to production of reactive oxygen species and an increased inflammatory response (Kuroiwa et al., 1988), it also restores metabolic activity to the compromised tissue (Halsey et al., 1991), closely resembling clinical stroke pathology (Ringelstein et al., 1992). The MCAo injury per se has been widely used as an experimental model of ischemic brain injury and has proven to be highly sensitive to neuroprotective drug intervention with the use of a variety of glutamate receptor antagonists and voltage-gated ion channel modulators (Margaill et al., 1996; Kawasaki-Yatsugi et al., 1998; Tortella et al., 1999; Williams et al., 2000), as well as modulation of inflammation, hypothermia, or free radical formation (Callaway et al., 2000; Phillips et al., 2000; Zausinger et al., 2000).

The NAALADase inhibitor 2-PMPA (100-mg/kg bolus injection followed by 10 mg/kg for 4 h) has been shown to provide a significant reduction of brain infarction when administered 60 min but not 120 min following 2 h of transient MCAo in the same injury model used in our study (Slusher et al., 1999; Vornov et al., 1999). This neuroprotective effect of 2-PMPA correlated to nearly complete attenuation of excitotoxic glutamate levels in injured brain tissue as well as to an increase in NAAG (Slusher et al., 1999; Vornov et al., 1999). Using a similar drug administration protocol, in the present study we measured a significant 50% neuroprotective reduction in brain infarction with GPI5232. GPI5232 neuroprotection also remained significant even when the initial drug injection was delayed for 120 min post-MCAo. Importantly, these effects were not likely due to drug-induced brain hypothermia since rectal temperatures, which have been shown to correlate to brain temperatures (Xue et al., 1992; Zhang et al., 1994), never fell below normal levels following MCAo and treatment with GPI5232. Equally important, pharmacokinetic analysis of the brain concentration of GPI5232 measured at 1 and 6 h post i.v. injection was 4792 and 1004 ng/g (Drs. M. Lu and B. Slusher, personal communication).

We have recently described the spatio-temporal pathophysiology of brain ischemia in our MCAo model using high-resolution 10-electrode topographic mapping in the rat (Williams and Tortella, 2000; Lu et al., 2001). These studies revealed the presence of ischemia-induced brain seizures as well as significant disruption of the EEG activity throughout the injured brain. Although the direct relation between these types of seizures and their possible involvement in promoting the pathology of the brain injury has not been established, this seizure activity is an inherent part of the overall pathophysiology of ischemic brain injury and is probably detrimental to recovery.

Although this study did not use high-resolution topographic EEG mapping of brain function, by focusing on the bipolar EEG derived from two brain regions directly overlying the injured hemisphere, i.e., the ipsilateral temporal cortex (core infarct zone) and the ipsilateral parietal cortex (peri-infarct zone), we were able to identify related changes, including brain seizures. Quantitatively, power scores in the temporal and parietal regions revealed a significant and profound drop in EEG activity at 1 h post-MCAo (prior to the i.v. bolus injection) from both groups, as would be expected from ischemic tissue. This marked reduction in EEG power, which provides an indirect, physiological assessment of a successful occlusion of the MCA, was seen in all frequency bands and was not significantly different between treatment groups. This decrease in power was still evident 2 h post-MCAo (immediately prior to reperfusion) in both vehicle- and GPI5232-treated rats, but following reperfusion the GPI5232-treated group exhibited a faster recovery in EEG function in both the parietal and temporal regions. Importantly, visual and spectral analysis of the parietal and temporal regions revealed distinct EEG patterns inherent upon the pathological processes taking place, i.e., peri-infarct and core infarct, respectively.

Parietal EEG power analysis at 24 h post-MCAo revealed the presence of slow-wave activity in both vehicle- and GPI5232-treated groups, which was correlated to a 7- to 9-fold increase in power in the delta band due to the presence of polymorphic delta activity (PDA). PDA is a common form of slow-wave activity associated with cerebrovascular disorders in humans, which involve large cortical and subcortical brain infarcts (Duffy et al., 1989). Parietal EEG recordings are a measure of predominately peri-infarct, possibly penumbral, tissue, and it is apparent in both groups that there is a great amount of activity in this region that may characterize peri-infarct neuropathology. Following treatment with GPI5232, as opposed to vehicle treatment, there were significant increases in EEG power in the higher-frequency ranges of the alpha and beta bands toward baseline values. Temporal EEG power analysis at 24 h post-MCAo also revealed an increase in power in the delta band of 2- to 3-fold in either of the treatment groups, but with significant recovery of power in the alpha band as well in the GPI5232-treated group. This recovery of high-frequency power toward pre-MCAo, baseline values may in fact be due to the presence of recovering neuronal activity in the parietal and temporal regions of the injury. GPI5232 did not significantly alter the slow-wave (PDA) activity in either the parietal or temporal regions, which may correlate to its lack of effect to alter slow wave sleep in normal rats (latency to SWS).

In kindled and suprathreshold stimulation-induced seizure models in rats, studies have shown a decrease in NAALADase activity in several brain regions following brain seizure activity (Meyerhoff et al., 1989), which they postulated to be an inherent control mechanism involved in reducing brain seizure activity. In our study, the attenuation of seizures both during the MCAo and upon reperfusion occurred following administration of GPI5232, an inhibitor of NAALADase activity. This is further support of the action of NAALADase inhibition as a possible endogenous antiseizure mechanism. Although other neuroprotective mechanisms may be involved with GPI5232 treatment, the measured reduction in brain seizure activity is certainly important. Clinically, the therapeutic use of anticonvulsants following brain injury is not widely practiced except in those patients who develop convulsive seizure activity (Arboix et al., 1997).

Importantly, in normal rats treated with acute high doses of GPI5232, there were no signs of neurotoxicity as defined by the absence of behavioral changes, the absence of drug-induced brain seizures, or lack of changes in spectral parameters of recorded EEG. Additionally, there was no delay in the latency to SWS as compared with the control group.

On the basis of the 10-mg/kg i.v. bolus injection of GPI5232, which provided a significantly effective neuroprotection (i.e., a 41% reduction of total brain infarct volume), and comparing this to the lack of neurotoxicity measured at doses at least as high as 200 mg/kg, we can estimate a protective index (PI) for GPI5232 equal to, or greater than, 20. This PI compares favorably to that described for other drugs, including the low-affinity N-methyl-D-aspartate antagonist dextromethorphan, which also provided 41% neuroprotection in MCAo-injured rats but was somewhat limited by seizure-producing effects at higher doses, yielding a PI equal to 15 (Tortella et al., 1999). Furthermore, the PI for GPI5232 is much greater than that described for MK801, which has been shown to provide as much as a 30% reduction in infarction in experimental brain injury models (Dezsi et al., 1992; Margaill et al., 1996), but which possesses serious brain seizure toxicity, yielding a calculated PI of less than 1 (Tortella et al., 1999). GPI5232 has also shown a more potent neuroprotective reduction of brain infarct volume and longer therapeutic window than its parent compound 2-PMPA (Slusher et al., 1999).

In conclusion, a comprehensive dose-response study has established the neuroprotective efficacy and therapeutic window of GPI5232 to effectively reduce brain infarct volumes in a focal cerebral brain injury model of stroke. Furthermore, this neuroprotective effect was also associated with an improvement in EEG recovery to baseline values and attenuation of ischemia-induced seizure activity. GPI5232 also possesses a favorable neuroprotective to neurotoxic index as compared with other experimental neuroprotective agents. Collectively, these results support the possible clinical efficacy of GPI5232 as a postinjury treatment of ischemic central nervous system insults.