[1] 13w Strategies for Neuroprotection Against L-Trans-2,4-Pyrrolidine Dicarboxylate-Induced Neuronal Damage During Energy Impairment In Vitro

[1] 214w Increased levels of extracellular excitatory amino acids and failure of energy metabolism are two conditions associated with brain ischemia. In the present study we have combined the simultaneous inhibition of glutamate uptake and mitochondrial electron transport chain to simulate neuronal damage associated with brain ischemia. Results show that cerebellar granule neurons are not vulnerable to transient glutamate uptake inhibition by L-trans-pyrrolidine-2,4dicarboxylate (PDC) despite the increase in the extracellular concentration of glutamate, unless they are simultaneously exposed to the mitochondrial toxins 3-nitropropionic acid (3-NP) or sodium azide. Cell damage was assessed by light microscopy observation, by reduction of 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), and by the fluorescent markers for live and dead cells, calcein and ethidium homodimer, respectively. The protective effect of alternative energy substrates, such as pyruvate, acetoacetate, and ␤-hydroxybutyrate against PDCinduced neuronal death during 3-NP exposure was studied and compared to the effects of the antioxidant vitamin E, the spin trapper ␣-phenyl-N-tert-butylnitrone (PBN), voltage-dependent calcium channel antagonists, and glutamate receptor antagonists. Results show that neuronal damage can be efficiently prevented in the presence of pyruvate and the N-methyl-D-aspartate (NMDA) receptor antagonist MK-801, whereas the non-NMDA receptor antagonist NBQX, acetoacetate, vitamin E, and PBN showed partial protection. In contrast, ␤-hydroxybutyrate and voltage-dependent calcium channels blockers did not show any protective effect at the concentrations tested. J.

[1] 82w It is accepted that the high-energy brain demands depend mainly on blood supply of external glucose and that ATP production through glycolysis and oxidative phosphorylation is the major source of energy used by brain (Erecinska andSilver, 1989, 1994). When glucose requirements are not satisfied owing to severe hypoglyce-mia or brain ischemia, neuronal damage occurs probably as a result of the combination of energy depletion and increased levels of excitatory amino acids in the extracellular space (Benveniste et al., 1984;Sandberg et al., 1986).

[2] 101w In vitro studies have shown that, during ATPlimiting conditions such as glycolysis inhibition and mitochondrial dysfunction, changes in the intracellular concentrations of Na ϩ and K ϩ occur, resulting in cell membrane depolarization and increased intracellular calcium levels (Silver et al., 1997). Similar changes occur during in vivo cerebral ischemia (Kristia ´n and Siesjo ¨, 1997), leading to the release of excitatory amino acids either by exocytosis or by reverse transport (Longuemare et al., 1999;Rossi et al., 2000) as well as to the activation of N-methyl-Daspartate (NMDA) receptors through the relief of the voltage-dependent Mg 2ϩ blockade (Zeevalk and Nicklas, 1992).

[3] 154w Recent evidence suggests that mitochondrial function is critical for glutamate-mediated neuronal death. Calcium dysregulation associated with the collapse of the mitochondrial membrane gradient and ATP depletion after prolonged exposure to excitotoxic concentrations of glutamate will lead to irreversible neuronal death (Ankarcrona et al., 1995;Budd and Nicholls, 1996;Schinder et al., 1996;Sattler et al., 1998). In addition, mitochondrial dysfunction will enhance the production of reactive oxygen species (ROS), which will also contribute to neuronal damage (Murphy et al., 1999). Substantial evidence indicates that mitochondrial toxins induce neuronal damage through an excitotoxic-mediated mechanism (Storey et al., 1992;Beal et al., 1993;Grenee et al., 1993;Brouillet et al., 1994;Schulz et al., 1994) and potentiate glutamate neurotoxicity both in vivo and in vitro (Novelli et al., 1988;Greene and Greenamyre, 1995;Cebers et al., 1998;Greene et al., 1998;Sa ´nchez-Carbente and Massieu, 1999). All these studies favor the hypothesis of a close and complex interrelation between mitochondrial energy metabolism and glutamate excitotoxicity (Henneberry, 1989).

[4] 81w In the present study we show that cultured cerebellar granule cells are not vulnerable to the toxic effects of transient glutamate uptake inhibition by L-trans-pyrrolidine-2,4dicarboxylate (PDC) despite the increase in glutamate extracellular levels. However, when cultured cells are simultaneously exposed to PDC and the mitochondrial toxins 3-nitropropionic acid or sodium azide, inhibitors of complexes II and IV of the electron transport chain, respectively, substantial neuronal damage is observed. We have previously reported similar in vivo results (Sa ´nchez-Carbente and Massieu, 1999).

[5] 115w Several in vitro studies have shown the neuroprotective potentiality of energy substrates alternative to glucose, such as pyruvate, lactate, and creatine, against neuronal damage associated with hypoglycemia (Schurr et al., 1988;Izumi et al., 1994Izumi et al., , 1997)), excitotoxicity (Ruiz et al., 1998;Maus et al., 1999;Brewer and Walliman, 2000), zinc neurotoxicity (Sheline et al., 2000), and oxidative stress (Desagher et al., 1997). Similarly, in vivo experiments have shown neuroprotective effects of creatine, CoQ 10 , and nicotinamide against striatal lesions induced by mitochondrial toxins (Schulz et al., 1996;Matthews et al., 1998) and excitotoxins (Malcon et al., 2000), and pyruvate can protect against PDC-induced lesions in an in vivo model of hypoglycemia (Massieu et al., 2000).

[6] 67w In the present study we have extended these observations to an in vitro model of acute neuronal damage that combines two events associated with brain ischemia, glutamate uptake inhibition and mitochondrial failure. The neuroprotective potentiality of pyruvate and the ketone bodies, acetoacetate and ␤-hydroxybutyrate, was tested and compared to that of vitamin E, the free radical scavenger ␣-phenyl-N-tert-butyl-nitrone (PBN), glutamate receptor antagonists, and voltage-dependent calcium channel blockers.

[1] 180w Cerebellar granule cell cultures were exposed during 6 hr to PDC, THA, and DHK (500 M), and 24 hr later cell viability was assessed by the MTT reduction assay. None of the three inhibitors induced statistically significant decreases in MTT reduction relative to control cultures, despite the increase in the extracellular concentration of glutamate induced by PDC and THA (Fig. 1). In control cultures, basal concentration of glutamate was 12.90 Ϯ 2.68 M, and incubation with PDC and THA (500 M) significantly increased this concentration to 31.68 Ϯ 3.87 M and 38.22 Ϯ 4.82 M, respectively. In contrast, incubation with DHK did not alter basal glutamate extracellular levels (Fig. 1). Because similar effects of PDC and THA were found on both glutamate levels and MTT reduction, only PDC was used for the rest of the experiments. In addition, a 4 hr incubation was chosen for all the following experiments, because the changes induced by PDC on both MTT reduction and glutamate extracellular levels were very similar to those observed after 6 hr (cf. Figs. 1 and 4 and Table II).

[2] 64w In contrast to transient treatments (4 -6 hr), prolonged incubation with PDC (500 M) during 24 hr induced a significant decrease in MTT reduction to 64.7% Ϯ 5.7% control (n ϭ 5, P Ͻ 0.05, data not shown), whereas PDC at 100 M had no effect on cell survival when incubated during 24 hr (MTT reduction ϭ 99.8% Ϯ 9.34% control, n ϭ 5).

[3] 135w To test the vulnerability of cerebellar granule cells to glutamate uptake inhibition during impairment of energy metabolism, culture wells were simultaneously exposed to PDC and the mitochondrial toxin 3-NP during 4 hr. When cultures were exposed to 500 M 3-NP during 4 hr, a significant 21.5% decrease in MTT reduction capacity was observed after 24 hr. When 3-NP was simultaneously incubated with PDC (500 M), neuronal damage induced by 3-NP was enhanced as evidenced by a 55.8% decrease in MTT reduction capacity relative to control cultures (see Fig. 4). Neuronal death induced by coincubation with PDC ϩ 3-NP was corroborated by the calcein/ethidium homodimer fluorescent method. In control cultures 12.0% red-fluorescent dead neurons were counted, whereas in cultures coincubated with 3-NP ϩ PDC (500 M during 4 hr) this number increased to 47.0% (Table I).

[4] 183w Results similar to those observed with 3-NP were obtained when PDC 500 M was coincubated with different concentrations of sodium azide during 4 hr (see Fig. 3). At low concentrations (100 and 300 M) sodium azide had no effect on cell survival, whereas 500 M decreased MTT reduction to 50%. Coincubation of PDC (500 M) and low concentrations of sodium azide (100 and 300 M) clearly potentiated neuronal damage, whereas PDC did not cause damage additional to that induced by 500 M sodium azide (see Fig. 3). The addition of MK-801 (10 M) abolished neuronal death induced by PDC ϩ sodium azide at all concentrations tested. Cultured cells were also vulnerable to glutamate toxicity when added at 500 M during 45 min as evidenced by a decrease in MTT reduction to 55.7% Ϯ 7.2% of control value (n ϭ 7, data not shown). Exposure to 3-NP either alone (0.5, 1.0, and 2.0 mM) or in combination with PDC (500 M) during 4 hr had no effect on cell survival in pure astrocytic cultures, as monitored 24 hr after the experiment (data not shown).

[5] 159w Control cultured cells observed under the microscope show dark and round somata, with an intricate neurite network (Fig. 2A). Neither 3-NP nor PDC (500 M, 4 hr) individual treatments altered the general morphology of cell cultures (Fig. 2B,C). However, 24 hr after coincubation with PDC ϩ 3-NP, most of the cells looked bright and small, and the neurite network had either disappeared or become thinner and fragmented, with numerous varicosities (Fig. 2D). These changes were very similar to those induced by 4 hr of incubation with PDC ϩ sodium azide and those observed after 45 min of exposure to 500 M glutamate (not shown). In some experiments pure astrocytic cultures were treated identically to granule neuron cultures. In agreement with MTT determinations, there were no evident morphological alterations in astrocytic cultures in the presence of 3-NP, PDC, or 3-NP ϩ PDC (not shown). This experiment shows that cell death induced under our experimental conditions is specific to neuronal cells.

[6] 199w The effect of 3-NP on SDH activity was measured immediately after or 24 hr after 30 min of incubation of cell cultures with 3-NP at different concentrations. 3-NP reduced the activity of SDH to 71.30% Ϯ 5.3%, 40.02% Ϯ 7.2%, and 13.87% Ϯ 3.8% of control values when determined immediately after a 30 min incubation period with 3-NP at 0.01, 0.1, and 0.5 mM, respectively (mean Ϯ SEM, n ϭ 4). Enzyme activity remained inhibited to the same extent during the following 24 hr, except in the precence of 0.01 mM 3-NP, which decreased further enzyme activity after 24 hr (53.46% Ϯ 6.07%, 30.28% Ϯ 4.25%, and 10.83% Ϯ 1.60%, respectively, for 3-NP at 0.01, 0.1, and 0.5 mM, mean Ϯ SEM, n ϭ 3). The presence of PDC did not alter 3-NP inhibition of enzyme activity when measured either immediately after (70.19% Ϯ 7.09%, 43.44% Ϯ 9.26%, and 10.93% Ϯ 0.82%, mean Ϯ SEM, n ϭ 4) or 24 hr after 3-NP addition (49.60% Ϯ 6.52%, 30.38% Ϯ 5.59%, and 9.89% Ϯ 1.81%, mean Ϯ SEM, n ϭ 3). Changes in SDH activity induced by 3-NP are very similar to those recently reported by Olsen et al. (1999).

[7] 142w To evaluate the neuroprotective potentiality of some energy substrates alternative to glucose in our experimental model, pyruvate (2 mM), acetoacetate (5 mM), or ␤-hydroxybutyrate (5 mM) was simultaneously added to 3-NP ϩ PDC and incubated during 4 hr. Neuronal survival was assessed 24 hr later. As is shown in Figure 4, pyruvate and acetoacetate improved neuronal survival in the presence of 3-NP ϩ PDC when present during the 4 hr incubation period. The protective effect of these compounds was partial restoring MTT reduction capacity to 59.9% Ϯ 4.0% and 64.4% Ϯ 7.5% of control, respectively. However, notable protection was achieved when pyruvate was present during the 4 hr incubation period and the following 24 hr. Under this condition, MTT reduction capacity was restored to 79.0% Ϯ 4.1% of control. In contrast to pyruvate and acetoacetate, ␤-hydroxybutyrate did not show any protective effect.

[8] 155w The effects on cell survival of energy substrates were compared to those of the antioxidant vitamin E (250 M) and the free radical scavenger PBN (500 M), which showed identical partial protective effects, restoring MTT reduction capacity to 61.55% Ϯ 7.7% and 61.04% Ϯ 8.6% of control, respectively. The NMDA receptor antagonist MK-801 (10 M) induced almost complete protection (88.79% Ϯ 4.9% of control MTT reduction), whereas the non-NMDA antagonist NBQX (100 M), showed only partial protective effects (57.39% Ϯ 7.7% of control MTT reduction). In contrast to these compounds, the calcium channel blockers nitrendipine (2 M) and nimodipine (2 M) were not neuroprotective at all. None of the drugs tested increased MTT reduction in control cultures or in cultures treated with 3-NP alone, except for pyruvate and MK-801 (4 hr incubation), which restored MTT reduction capacity to 95.11% Ϯ 7.1% and 93.41% Ϯ 3.5% of control, respectively, when added simultaneously with 3-NP (Table III).

[9] 72w The protective effect of MK-801 (4 hr incubation) and pyruvate (24 hr treatment) was corroborated by counting the percentage of red-fluorescent dead cells relative to the total of cells present in the cultures (Table I). Morphological observations corroborated as well the neuroprotective effect of the different drugs tested. As can be seen in Figure 5 cultures were better preserved in the presence of pyruvate (24 hr incubation), acetoacetate, vitamin E, or MK-801.

[1] 275w In a previous study we have shown that glutamate uptake inhibition leads to neuronal death in cortical cultured neurons, which is directly correlated to increases in the extracellular levels of glutamate from 8.0 M to up to 24 M (Velasco et al., 1996). In the present study we show that extracellular glutamate concentration increased from 18.1 to 35.0 M after 4 hr of exposure to PDC. Insofar as PDC is a substrate inhibitor of glutamate transporters (Bridges et al., 1991), glutamate elevation induced by PDC might result from carrier-mediated heteroexchange as has been suggested for cortical cultures (Blitzblau et al., 1996;Volterra et al., 1996). Previous studies have shown that low glutamate concentrations in the range of 10 -30 M are partially toxic to cerebellar granule cultures when exposed for short (15 min) periods in ionic Mg 2ϩ -free medium, a condition that favors the activation of NMDA receptors, whereas larger concentrations, up to 100 -500 M, incubated for 15 min or longer periods dramatically reduce cell survival (Manev et al., 1991;Milani et al., 1991). In the present study we show that transient (4 hr) PDC-induced increases in glutamate concentrations up to 35 M are not toxic to cerebellar granule neurons. This is probably because PDC was not incubated in ionic medium but was directly added to the culture medium. Furthermore, prolonged exposure to PDC during 24 hr under the same conditions did induce a 35% decrease in MTT reduction relative to control cultures. Nonetheless, transient increases in glutamate extracellular levels are neurotoxic when mitochondrial metabolism is compromised. These observations are in agreement with previous in vivo studies (Massieu et al., 1995;Sa ´nchez-Carbente and Massieu, 1999).

[2] 164w According to our results, 3-NP rapidly and severely inhibited SDH activity. After 30 min of incubation with 500 M 3-NP, SDH activity was reduced to close to 10.0% of control activity and had not recovered 24 hr after toxin withdrawal. Despite this dramatic reduction in SDH activity, 3-NP caused only a 20% decrease in cell survival. Similarly, sodium azide at low concentrations did not induce cell death. This result suggests that granule neurons can tolerate well mild mitochondrial failure and that electron transfer through the active mitochondrial complexes as well as glycolytic activity can fulfill cell energy demands at least for 24 hr. However, increases in excitatory amino acids levels up to 35 M were highly toxic to these partially metabolically inhibited neurons. PDC toxicity in the presence of 3-NP might not be explained by a further increase in excitatory amino acid concentrations, in that 3-NP did not increase glutamate or aspartate extracellular levels either in the presence or in the absence of PDC.

[3] 238w In the present study we tested the protective effect of different agents against cell death of cerebellar granule neurons under experimental conditions that may mimic ischemia. Pyruvate and acetoacetate showed partial pro- tection when incubated during 4 hr with 3-NP and PDC. The protective effect of these compounds may be related to their capacity to restore the energy deficit induced by 3-NP by acting as substrates of the tricarboxylic acid cycle through their conversion to acetyl-CoA. This hypothesis is consistent with the observation that pyruvate can prevent the decrease in MTT reduction induced by 3-NP (Table III). In contrast, acetoacetate and ␤-hydroxybutyrate did not restore mitochondrial activity in the presence of 3-NP. However, acetoacetate did improve neuronal survival in the presence of 3-NP ϩ PDC. These results may indicate that acetoacetate might be used as a putative substrate of energy metabolism during severe injury, whereas pyruvate can also be utilized during mild energy depletion. No protective effect of ␤-hydroxybutyrate was observed in the present study. As a precursor of acetoacetate, ␤-hydroxybutyrate has to be me-tabolized to acetoacetate before entering the TCA. Therefore, higher concentrations or prolonged exposure to this ketone body might be needed to achieve neuroprotection. Consistently with this suggestion, it was recently reported that ␤-hydroxybutyrate can protect dopaminergic cultured neurons against MPP ϩ neurotoxicity after chronic exposure to a concentration of this ketone body higher than that used in the present study (Kashiwaya et al., 2000).

[4] 137w Substantial evidence supports the protective role of pyruvate in different experimental paradigms. Pyruvate can preserve synaptic function and morphological cell integrity of hippocampal slices in the absence of glucose and during glycolysis inhibition (Izumi et al., 1997) and protects against glutamate (Ruiz et al., 1998) and NMDA neurotoxicity in the absence of glucose, by a mechanism probably involving ATP production, increased mitochondrial respiration rate, and decreased free cytosolic calcium concentration (Villalba et al., 1994;Eimerl and Schramm 1995;Maus et al., 1999). Additionally, pyruvate protective effects might derive from its nonenzymatic H 2 O 2 scavenging capacity (Desagher et al., 1997). The improved protective efficacy of pyruvate when incubated during 24 hr may be related to this effect. Alternatively, it has recently been shown that pyruvate significantly increases the formation of kynurenic acid, an endogenous antagonist of NMDA recep-

[5] 231w TABLE II. Extracellular Levels of Amino Acids (M) and Their Changes Induced by Exposure to 3-NP, PDC, and 3NP ؉ PDC in Cerebellar Granule Neuron Cultures † Glutamate Aspartate Glutamine Glycine Taurine Alanine Control 18.10 Ϯ 4.2 2.54 Ϯ 0.71 648.0 Ϯ 79.1 147.9 Ϯ 6.3 30.62 Ϯ 4.2 218.0 Ϯ 18.3 3-NP 23.76 Ϯ 4.9 2.40 Ϯ 0.62 605.1 Ϯ 75.2 145.7 Ϯ 9.3 30.80 Ϯ 3.3 211.7 Ϯ 11.7 PDC 35.09 Ϯ 9.9* 4.80 Ϯ 1.0* 521.1 Ϯ 72.6 160.8 Ϯ 12.0 32.02 Ϯ 3.9 212.6 Ϯ 14.5 3-NP ϩ PDC 29.57 Ϯ 6.4 3.70 Ϯ 0.88 526.6 Ϯ 50.0 140.9 Ϯ 6.7 39.92 Ϯ 2. TABLE III. Effects of 3-NP on Cell Survival and Protection by Different Compounds in Cerebellar Granule Cell Cultures † Condition MTT reduction (% of control) 3-NP (500 M) 78.5 Ϯ 1.6* ϩ Pyruvate (2 mM) 95.11 Ϯ 7.1** ϩ Acetoacetate (5 mM) 72.0 Ϯ 10.0 ϩ ␤-Hydroxybutyrate (5 mM) 70.72 Ϯ 8.7 ϩ Vitamin E (250 M) 85.12 Ϯ 8.1 ϩ PBN (500 M) 85.61 Ϯ 4.0 ϩ MK801 (10 M) 93.41 Ϯ 3.5** ϩ NBQX (100 M) 78.23 Ϯ 6.2 tors (Hodgkins and Schwarcz, 1998a). A similar effect has been found for the ketone bodies acetoacetate and ␤-hydroxybutyrate (Hodgkins and Schwarcz, 1998b). In addition, acetoacetate and ␤-hydroxybutyrate can also increase brain ␥-aminobutyric acid (GABA) levels after intraperitoneal injection (Daikhin and Yudkoff, 1998).

[6] 145w Notable protection was observed in the presence of the NMDA receptor antagonist MK-801. This result favors the proposed secondary excitotoxic hypothesis, suggesting that energy-limiting conditions owing to partial mitochondrial impairment or glycolysis inhibition lead to excitotoxically mediated neuronal death. According to this hypothesis, Mg 2ϩ ions that block the NMDA receptor channel during resting conditions may be extruded during impaired energy production owing to depolarization resulting from the loss of membrane ionic gradients as a consequence of Na ϩ /K ϩ ATPase inactivation. This would lead to Ca 2ϩ influx and excitotoxicity even in the absence of substantial release of excitatory amino acids (Novelli et al., 1988;Zeevalk andNicklas, 1990, 1992). This hypothesis is supported by the observation that MK-801 restored mitochondrial MTT reduction capacity after exposure of cultured neurons to 3-NP alone, a condition that did not enhance the extracellular levels of excitatory amino acids.

[7] 195w NBQX, a non-NMDA antagonist, showed partial protective effects against 3-NP ϩ PDC-induced neuronal death and had no effect on 3-NP toxicity. This result is interesting in that it suggests the participation of AMPA receptors in ischemic neuronal death. Previous experiments have shown that cultured cerebellar granule neurons are not sensitive to AMPA neurotoxicity unless AMPA receptors desensitization is blocked by cyclothiazide (Hack and Bala ´zs, 1995;Cebers et al., 1997). Interestingly, it has been recently found that AMPA neurotoxicity is highly potentiated during ATP-limiting conditions induced by sodium cyanide in cerebellar granule neurons by a Ca 2ϩ -mediated mechanism, suggesting that regulation of AMPA receptor desensitization might be influenced by the energy state of the cell (Cebers et al., 1998). Alternatively, the NBQX protective effect might reflect the contribution of kainate receptors to PDC-mediated neurotoxicity; at the concentration used in this study (100 M), it can also block kainate receptors (Sheardown et al., 1990). In contrast, antagonists of the voltagesensitive calcium channels did not protect at all against neuronal damage induced under our experimental conditions. This result suggests that Ca 2ϩ influx through these channels might not be involved in excitotoxic calcium entry in our model.

[8] 169w Numerous studies favor the role of free radicals or oxidative stress in glutamate-induced neuronal death (Lafon-Cazal et al., 1993;Dugan et al., 1995;Gunasekar et al., 1995), and it was recently reported that mitochondrial toxins such as 3-NP and malonate increase reactive oxygen species both in vitro and in vivo (Kim et al., 2000;Zeevalk et al., 2000). Therefore, important increases in reactive oxygen species might be predicted in the presence of 3-NP ϩ PDC. Results suggest that this is indeed the case; vitamin E and PBN partially protected neurons against the toxic effects of 3-NP ϩ PDC coincubation. Interestingly, neither vitamin E nor PBN prevented decreases in MTT reduction induced by exposure to 3-NP alone probably, indicating that these compounds may be more efficient during severe rather than mild oxidative stress. These results are in accordance with previous studies showing the protective potentiality of PBN and the related spin trapper S-PBN against cell death induced by mitochondrial toxins both in vitro and in vivo (Schulz et al., 1996;Olsen et al., 1999).

[9] 58w In the present report we show the relevance of the use of energy substrates such as pyruvate and acetoacetate as neuroprotective agents in a model that may mimic ischemic damage. Experiments are now in progress to clarify the cellular events leading to neurotoxicity in this model and to help in understanding the protective mechanisms of the compounds tested.

[1] 61w Fetal bovine serum, soybean trypsin inhibitor, and penicillin/streptomycin were purchased from Gibco (Grand Island, NY). PDC, DL-threo-␤-hydroxyaspartate (THA), and dihydrokainate (DHK) were obtained from Tocris (Bristol, United Kingdom). MK-801, NBQX, nitrendipine, and nimodipine were purchased from RBI (Natick, MA), and all other chemicals were from Sigma (St. Louis, MO). Calcein/ethidium homodimer reagents (LIVE/DEAD kit) were purchased from Molecular Probes (Eugene, OR).

[2] 203w Primary cultures of cerebellar granule neurons were prepared from 7-or 8-day-old Wistar rats as described by Moran and Patel (1989), with slight modifications. Briefly, after dissec-tion, seven cerebelli were chopped into 300 m cubes, incubated in 0.25% trypsin solution, and dispersed by trituration in a DNAase and soybean trypsin inhibitor containing solution (0.08% and 0.52%, respectively). Cells were suspended in basal Eagle's medium supplemented with 10% fetal bovine serum, 2 mM L-glutamine, 50 U/ml penicillin, 50 g/ml streptomycin, 50 g/ml gentamicin, and 25 mM KCl and plated at a density of 260 -290 ϫ 10 3 /cm 2 (1.5 ϫ 10 6 cells/ml/well) in Costar 24-well plates (Cambridge, MA) precoated with poly-L-lysine (5 g/ml). Cells were cultured for 9 days in vitro (DIV) at 37°C in a humidified 5% CO 2 /95% air atmosphere. Glucose (5 mM) and cytosine arabinoside (10 M) were added to cultures 4 days after plating. Neuronal population in these cultures was 93% as determined by immunocytochemistry using antibodies against microtubule-associated protein (MAP)-2 and glial fibrillary acidic protein (GFAP), neuronal and glial cells markers, respectively (not shown). All animals were handled according to the Rules for Research in Health Matters (Mexico) with approval of the local animal care committee.

[3] 162w To test the differential effects of glutamate uptake inhibitors on cell survival and extracellular amino acids levels, cerebellar neuronal cultures were exposed during 4 or 6 hr to 500 M of either one of three inhibitors of glutamate uptake: PDC, THA, and DHK. Before the experiment, 0.5 ml of the conditioned medium was removed, and inhibitors dissolved in 10 mM phosphate buffer were added to the remaining 0.5 ml medium. At the end of the incubation period, the drugcontaining medium was substituted by the conditioned medium filter-sterilized and maintained at 37°C, and cultures were kept for additional 24 hr, when cell viability was assessed. Control cultures were identically treated with the exception that vehicle solution (phosphate buffer) was added to the remaining 0.5 ml medium instead of drugs. Because similar results were obtained after 4 or 6 hr of incubation with PDC and THA, only PDC was used in the following series of experiments, and a 4 hr incubation period was chosen.

[4] 142w The effect of PDC (500 M during 4 hr) on cell survival was tested in the presence of the mitochondrial toxin 3-NP (500 M), and putative neuroprotective drugs were added simultaneously (at concentration indicated in the figures), except for pyruvate, which was also incubated during 24 hr in some experiments. To corroborate that inhibition of mitochondrial metabolism was indeed the trigger of PDC toxicity, another mitochondrial toxin was used. In this case PDC was coincubated during 4 hr with sodium azide (0.1, 0.3, and 0.5 mM), an inhibitor of complex IV of the electron transport chain. In all the experiments, cell viability was assessed 24 hr after the onset of drug exposure. All of the neuroprotective agents tested were dissolved in 10 mM phosphate buffer, except for vitamin E, which was diluted in ethanol (final concentration of ethanol in the medium 0.5%).

[5] 73w Pure cultures of cerebellar astrocytes were obtained from the dissociated cell suspension plated at the same density in the same culture medium in Costar 24-well plates without added poly-L-lysine as previously described (Velasco et al., 1995). All experiments were performed in duplicate or tetraplicate. Control cultures were included in each experiment and were han-dled in parallel. Cell morphology and general appearance of the cultures were monitored before treatment. Inadequately preserved cultures were discarded.

[6] 67w Succinate dehydrogenase (SDH) activity was determined immediately after and 24 hr after 30 min of incubation of cell cultures in the presence of 3-NP (0.01, 0.1, and 0.5 mM) as described by Olsen et al. (1999). SDH activity was also determined in the presence of 3-NP (0.01, 0.1, and 0.5 mM) ϩ PDC (500 M). Enzyme activity is expressed as percentage SDH activity relative to control cultures.

[7] 27w All data are expressed as means Ϯ SEM values. Statistical analysis was carried out by one-way ANOVA, followed by a Fisher's least significant difference multiple comparison test.

[1] 181w Cell viability was evaluated 24 hr after the onset of drug exposure by direct examination of cultures under an inverted light microscope and by the 3-(4,5-dimethylthiazol-2-yl)-2,5diphenyltetrazolium bromide (MTT) reduction assay. The MTT assay is based on the findings that the mitochondrial enzymes are able to transform MTT salt into formazan and thus indicate the integrity of mitochondrial enzymes in viable neurons (Mosmann, 1983;Berridge and Tan, 1993). Although it has been reported that MTT reduction does not exclusively occur in mitochondria, it is still considered a measure of cell viability because it is incorporated into cells by endocytosis, an ATPdependent processes characteristic of living cells, owing to its lack of permeability through plasma membrane (Liu et al., 1997). In brief, 24 hr after the experiment, cell cultures were incubated with MTT (150 M) during 1 hr at 37°C in 5% CO 2 /95% O 2 -containing atmosphere. The medium was aspirated, and the precipitated formazan was solubilized with 0.8 ml isopropanol and quantified spectrophotometrically at a wavelength of 570 nm. Cell viability is expressed as percentage MTT reduction relative to control cultures.

[2] 167w Cell survival was also monitored by the fluorescent markers calcein (C-AM; green-fluorescent viable cells) and ethidium homodimer (HE; red-fluorescent dead cells) according to the live/dead viability assay kit (Molecular Probes). The assay was performed according to the instructions of the supplier. For these experiments, cells were grown on coverslips, and 24 hr after the experiment they were washed twice in D-PBS for 5 min and exposed to C-AM (2 M) and HE (4 M) in D-PBS for an additional 5 min. They were washed in D-PBS and fixed during 30 min in cold paraformaldehyde. Immediately afterward, they were observed under an epifluorescent microscope, and the number of live and dead neurons was counted with the aid of an image analyzer (NIH Image 1.6 for Macintosh). The number of dead neurons is expressed as percentage of redfluorescent dead cells relative to the total (red-and greenfluorescent) number of neurons counted in 10 different fields/ coverslip from three independent experiments. The approximate number of cells present per field was 213.

[3] 54w Amino acid concentrations present in the media were measured by high-performance liquid chromatography (HPLC). In brief, at the end of the incubation with the drugs (4 or 6 hr), an aliquot was taken from the medium, deproteinized, and processed for amino acid analysis after derivatization with o-phthaldialdehyde as previously described (Velasco et al., 1996).

[4] 109w To elucidate whether potentiation of PDC-induced neuronal damage by 3-NP was due to an additional accumulation of excitatory amino acids in the culture medium, amino acid concentrations were determined in the medium after 4 hr incubation either with PDC alone or in the presence of 3-NP. As is shown in Table II, 4 hr of incubation with PDC caused substantial elevation in the extracellular concentrations of glutamate and aspartate; however, in the presence of 3-NP no further increases in the extracellular levels of these amino acids were observed. No changes were observed in the extracellular levels of glutamine, glycine, taurine, or alanine with any of the treatments (Table II).

[5] 181w TABLE I. Assessment of Cell Viability by the Calcein/Ethidium Homodimer Fluorescent Markers and MTT Reduction † Control 3-NP ϩ PDC ϩMK 801 ϩ Pyruvate Neuronal death (% dead cells) 12.12 Ϯ 1.5 47.16 Ϯ 2.0* 20.14 Ϯ 1.6** 16.7 Ϯ 2.6** MTT reduction (% of control) 100 44.16 Ϯ 2.9* 88.79 Ϯ 4.9** 79.0 Ϯ 4.1** † Granule neuron cultures were incubated with 3-NP ϩ PDC (500 M, 4 hr) in the absence or presence of MK-801 (10 M, 4 hr) and pyruvate (2 mM, 24 hr); 24 hr after drug exposure, cell viability was assessed by MTT reduction and by the fluorescent markers ethidium homodimer (4 M) and calcein-AM (2 M) as described in Materials and Methods. Cell death is expresesed as percentage of red-fluorescent dead cells relative to the total cells present in cultures (red-and green-fluoresncent) and as percentage of MTT reduction relative to control cultures. Similar results are obtained by both methods. Values are mean Ϯ SEM of four to six independent experiments. *P Ͻ 0.05 relative to control values. **P Ͻ 0.05 relative to 3-NP ϩ PDC.