Address correspondence and reprint requests to Dr. W.-K. Kim at Department of Pharmacology, College of Medicine, Ewha Women's University, 911-1 Mok-6-dong, Yangchun-Ku, Seoul 158-056, Republic of Korea. E-mail: firstname.lastname@example.org
Abstract: Previously we reported that immunostimulated astrocytes were highly vulnerable to glucose deprivation. The augmented death was mimicked by the peroxynitrite (ONOO--producing reagent 3-morpholinosydnonimine (SIN-1). Here we show that glucose deprivation and ONOO- synergistically deplete intracellular reduced glutathione (GSH) and augment the death of astrocytes via formation of cyclosporin A-sensitive mitochondrial permeability transition (MPT) pore. Astrocytic GSH levels were only slightly decreased by glucose deprivation or SIN-1 (200 μM) alone. In contrast, a rapid and large depletion of GSH was observed in glucose-deprived/SIN-1-treated astrocytes. The depletion of GSH occurred before a significant release of lactate dehydrogenase (a marker of cell death). Superoxide dismutase and ONOO- scavengers completely blocked the augmented death, indicating that the reaction of nitric oxide with superoxide to form ONOO- was implicated. Furthermore, nitrotyrosine immunoreactivity (a marker of ONOO-) was markedly enhanced in glucose-deprived/SIN-1-treated astrocytes. Mitochondrial transmembrane potential (MTP) was synergistically decreased in glucose-deprived/SIN-1-treated astrocytes. The glutathione synthase inhibitor L-buthionine-(S,R)-sulfoximine markedly decreased the MTP and increased lactate dehydrogenase (LDH) releases in SIN-1-treated astrocytes. Cyclosporin A, an MPT pore blocker, completely prevented the MTP depolarization as well as the enhanced LDH releases in glucose-deprived/SIN-1-treated astrocytes.
In various neurodegenerative diseases such as stroke and trauma, the concentrations of cytokines in the CSF are increased, resulting in an activation of glial cells, including astrocytes (Minami et al., 1992; Lees, 1993; Rothwell and Relton, 1993). These immunostimulated glia secrete various bioactive agents, including cytotoxins, that ultimately determine the pattern and degree of functional recovery of the CNS (Giulian, 1990, 1992). One of the glia-derived cytotoxins is nitric oxide (NO), which has been implicated in a large number of pathologies (Bruhwyler et al., 1993). The cytotoxicity of NO is increased by reaction with superoxide anion (O2[UNK]) to form ONOO- (Beckman et al., 1990).
We and other researchers have reported that immunostimulated astrocytes and microglia synergistically enhanced N-methyl-D-aspartate receptor-mediated neuronal death via induction of NO synthase (Hewett et al., 1994; Kim and Ko, 1998). Immunostimulated astrocytes also enhanced death of neuronal cells caused by hypoxia and glucose deprivation (Hewett et al., 1996; Kim et al., 1999a,b). Recently, we also found that immunostimulated astrocytes became highly vulnerable to glucose deprivation (Choi and Kim, 1998a,b). At present, however, the exact mechanism for the autocrine cytotoxicity remains unknown.
The present study was therefore aimed to investigate the mechanism for the augmented cell death. Immunostimulated astrocytes simultaneously produce NO and O2[UNK] via expression and activation of inducible NO synthase (Simmons and Murphy, 1992; Skaper et al., 1995). NO and O2[UNK] rapidly react to form ONOO- in a 1:1 stoichiometry (Beckman et al., 1990). 3-Morpholinosydnonimine (SIN-1) is also known to produce ONOO- by simultaneously releasing NO and O2[UNK] (Hogg et al., 1992). Because immunostimulated astrocytes produce many biologically active molecules, including cytokines, in the present study we used SIN-1 to investigate the synergistic effect of ONOO- on the viability of glucose-deprived astrocytes. The pure NO releasers S-nitroso-N-acetylpenicillamine (SNAP) and S-nitrosoglutathione (GSNO) were also tested for comparison.
MATERIALS AND METHODS
GSH, Triton X-100, SIN-1, GSNO, and L-buthionine-(S,R)-sulfoximine (L-BSO) were purchased from Sigma Chemical Co. (St. Louis, MO, U.S.A.). Dulbecco's modified Eagle's medium (DMEM) and fetal bovine serum (FBS) were purchased from GibcoBRL (Grand Island, NY, U.S.A.). SNAP was obtained from Research Biochemicals International (Natick, MA, U.S.A.). Monochlorobimane and JC-1 were purchased from Molecular Probes (Eugene, OR, U.S.A.). FK-506 was a generous gift from Fujisawa Pharmaceutical Co., Ltd. (Osaka, Japan).
Astrocytes were cultured from the prefrontal cortices of 2—4-day-old Sprague—Dawley rat pups as previously described (Choi and Kim, 1998b). In brief, cells were dissociated by mild trypsinization (10 min at 37°C with DMEM containing 0.1% trypsin) and passed through sterile nylon sieves (pore size, 80 μm) into DMEM containing 10% heat-inactivated FBS. Cells were then plated (∼50,000 cells/ml) onto poly-L-lysine (20 μg/ml)-coated 75-cm2 culture bottles and maintained for 1 week in DMEM supplemented with 10% FBS. Astrocytes were then trypsinized, washed, and replated in the growth medium onto poly-L-lysine (20 μg/ml)-coated 24-well plates. Cells were used for the experiments 7-9 days later.
Glucose deprivation was achieved by incubation in glucose-free DMEM that was not supplemented with serum because it interfered with the lactate dehydrogenase (LDH) assay.
Assessment of astrocyte injury or death
Astrocyte injury or death was assessed by morphological examination of cells using phase-contrast microscopy and quantified by measuring the amount of LDH released into the bathing medium. As we described before (Kim and Pae, 1998), the activity of LDH was measured using the diagnostic kit from Sigma. In brief, the media to be tested were mixed with NADH and pyruvate (0.1% wt/vol) and warmed at 37°C, and 20 min later coloring reagent was added. Thirty minutes later, the reaction was stopped by adding 0.4 M NaOH. The activity (in units/ml) of LDH in each medium was calculated from the standard curve. Total LDH content, which corresponds to complete astrocyte death, was measured in sister cultures frozen and thawed after the experiments. Cell viability was expressed as a percentage of total LDH.
GSH was quantified as described by Fernández-Checa and Kaplowitz (1990) with some modifications. Astrocytes were washed with phosphate-buffered saline solution (PBS) and then incubated at 37°C for 10 min with 100 μM monochlorobimane. Cells were washed with PBS and then lysed in PBS containing 0.2% Triton X-100. After centrifugation at 13,600 g for 6 min, fluorescence of the monochlorobimane-GSH complex in the supernatant was measured with a Hitachi model 650-60 spectrofluorophotometer (excitation wavelength of 400 nm and emission wavelength of 480 nm). The GSH content of the samples was calculated from the standard curve prepared from GSH standards incubated in the presence of monochlorobimane and glutathione transferase (Sigma). The concentration of GSH was expressed as nanomoles per milligram of protein; protein content was determined using bicinchoninic acid (Smith et al., 1985).
Nitrotyrosylation of bovine serum albumin (BSA)
Nitrotyrosylation of BSA was performed according to the method of Misko et al. (1998). BSA was dissolved at a concentration of 5 mg/ml in PBS (pH 7.4), and SIN-1 was added to BSA in the presence of indicated test reagents. Formation of nitrotyrosine residues on the BSA was monitored by western blot analysis. In brief, samples were solubilized in sodium dodecyl sulfate-containing Laemmli buffer and heated for 10 min before loading onto a 7.5% polyacrylamide gel. The gel was transferred to nitrocellulose and subsequently probed for 18 h at 4°C with a rabbit polyclonal anti-nitrotyrosine antiserum (Upstate Biotechnology, Lake Placid, NY, U.S.A.) at 1 ng/ml containing 10 mM L-tyrosine (to reduce nonspecific staining). Following intensive wash, the blots were probed with a 1:3,000 dilution of a goat anti-rabbit horseradish peroxidase conjugate, and the nitrotyrosine residues were visualized by enhanced chemiluminescence using the Immun-Lite system (Bio-Rad Laboratories, Hercules, CA, U.S.A.).
Cells were lightly fixed in 4% paraformaldehyde in PBS for 10 min at room temperature and then incubated in cold 70% ethanol until use. To inactivate endogenous peroxidase, cells were covered with 3% H2O2 in PBS for 5 min at room temperature. Nonspecific staining was blocked with PBS containing 8% BSA for 30 min at room temperature. All subsequent incubations were carried out in this buffer supplemented with 10 mM L-tyrosine. For detection of nitrotyrosine immunoreactivity, cells were incubated for 2 h at room temperature with PBS containing a rabbit anti-nitrotyrosine antibody at 10 μg/ml and 1% BSA. After incubation with a 1:200 dilution of anti-rabbit horseradish peroxidase-conjugated secondary antibody in 1% BSA for 1 h at room temperature, cells were exposed to a chromogenic mixture of diaminobenzidine.
Measurements of mitochondrial transmembrane potential (MTP)
The MTP was measured according to the procedure of Reers et al. (1991) with minor modifications. In brief, astrocytes cultured on 24-well culture plates were loaded for 20 min at 37°C with JC-1 (1.0 μg/ml) in culture medium. Depolarization of MTP was assessed by measuring the fluorescence intensities at 530 and 590 nm of emission wavelengths using a fluorescence microplate reader (FL600; Bio-tek Instruments, Winooski, VT, U.S.A.). During the measurements, cells were maintained at 37°C and protected from light. Fluorescence intensity was measured every 5 min for <2 s to minimize photobleaching. All fluorescent measurements were corrected for autofluorescence; autofluorescence of cells not loaded with JC-1 was constant throughout the experiment. In control experiments, no photobleaching was observed during fluorescence monitoring.
Data are expressed as the mean ± SEM values and analyzed for statistical significance using one-way ANOVA followed by Scheffé's test for multiple comparison. A p value of <0.05 was considered significant.
In agreement with our previous reports (Choi and Kim, 1998a,b), as assessed morphologically using a phase-contrast microscope, the augmented death was observed only in astrocytes deprived of glucose for as short as 3 h in the presence of 200 μM SIN-1 (data not shown). A simultaneous exposure to glucose deprivation and SIN-1 rapidly and synergistically enhanced the LDH release from astrocytes (Fig. 1). However, neither an 8-h glucose deprivation nor an 8-h treatment with SIN-1 (200 μM) alone altered the viability of astrocytes (Fig. 1). Although the pure NO releasers SNAP and GSNO also evoked LDH from glucose-deprived astrocytes, they were much less potent than SIN-1. Thus, 200 μM SNAP and 1 mM GSNO evoked significant LDH release in glucose-deprived astrocytes only after 8 h of exposure, whereas 200 μM SIN-1 evoked this after 3 h (Fig. 1).
The GSH level in control astrocyte cultures was 19.38 ± 0.43 nmol/mg of protein (see the legend of Fig. 2). Assuming a water content of ∼3.5-4.0 μl/mg of protein (Chen et al., 1992), the intracellular concentration of GSH is ∼4.5-5 mM. Intracellular GSH levels were progressively but slightly decreased in astrocytes deprived of glucose or just treated with 200 μM SIN-1 (Fig. 2). In contrast, a rapid and marked decrease in intracellular GSH levels was observed in astrocytes that were both glucose-deprived and treated with SIN-1 (Fig. 2). The decrease of intracellular GSH levels in glucose-deprived/SIN-1-treated astrocytes preceded the release of LDH. Thus, a significant reduction in GSH levels was observed within 1 h, whereas a significant release of LDH was not observed until 3 h (Fig. 1).
To determine whether the decrease in intracellular GSH level was associated with the delayed release of LDH, the GSH synthase inhibitor L-BSO was used (Griffith, 1982). Pretreatment with 100 μM L-BSO almost completely depleted astrocytes of GSH (Fig. 3A). Despite this marked depletion of GSH, no significant death of astrocytes was observed (Fig. 3B). However, L-BSO significantly enhanced the LDH release from astrocytes treated with 200 μM SIN-1 (Fig. 3B).
It has been reported that astrocytes have enzymes for GSH synthesis, such as γ-glutamyl synthetase and GSH synthetase (Marker et al., 1994), and specific transporters for precursors, such as glutamate, cysteine, glycine, cystine, and cysteinylglycine (Dringen and Hamprecht, 1996; Dringen et al., 1998). Astocytes may use extracellular GSH to increase their intracellular GSH content via extracellular breakdown, transport of products, and intracellular synthesis of GSH (Meister, 1988). In good agreement with previous data (Papadopoulos et al., 1997), preincubation of astrocytes with 10 mM GSH for 12 h increased intracellular GSH levels from 19.38 ± 0.43 to 29.07 ± 2.71 nmol/mg of protein. Intracellular GSH levels, however, decreased very rapidly after removal of GSH from the bathing medium. Regardless of whether the astrocytes were preincubated with 10 mM GSH, SIN-1 reduced intracellular GSH levels in glucose-deprived astrocytes to the same extent within 1 h (Fig. 4A). Preincubation with GSH did not block the augmented LDH release in glucose-deprived/SIN-1-treated astrocytes (Fig. 4B).
As GSH has been reported to scavenge ONOO- (Barker et al., 1996), the rapid decrease in intracellular GSH levels in glucose-deprived/SIN-1-treated astrocytes might result in an increased level of ONOO-. Because ONOO- can nitrate tyrosine residues, the detection of nitrotyrosine, which is chemically stable, has become a reliable biochemical marker for the presence of ONOO- in pathophysiological processes. Nitrotyrosine immuno-reactivity was the highest in glucose-deprived/SIN-1-treated astrocytes (Fig. 5A). Nitrotyrosylation of BSA by the ONOO--producing reagent SIN-1 was attenuated by GSH but not by glucose (Fig. 5B). Furthermore, the increased death of glucose-deprived/SIN-1-treated astrocytes was prevented by superoxide dismutase (SOD), which blocks the formation of ONOO- from SIN-1, and also by cysteine and GSH, which are known as ONOO- scavengers (Table 1). Mn(III)tetrakis(1-methyl-4-pyridyl)porphyrin (MnTMPyP), described as a cell-permeable SOD mimetic and ONOO- scavenger (Pfeiffer et al., 1998), also completely blocked the augmented death, whereas the structurally related inactive compound without metals, 5,10,15,20-tetrakis(N-methyl-4′-pyridyl)porphyrin (H2TMPyP), did not (Table 1). In contrast, the lipid-soluble vitamin E analogue Trolox, a well-known antioxidant that protects against lipid peroxidation, did not alter the augmented death (Table 1). These data indicate that ONOO- is involved in the augmented death of glucose-deprived/SIN-1-treated astrocytes.
Table 1. LDH release by astrocytes exposed to various conditions
During glucose deprivation (GD) for 3 h in the absence and presence of SIN-1 (200 μM), astrocytes were treated with various drugs. Data are mean ± SEM values (n = 4-12), expressed as a percentage of total LDH as defined in Materials and Methods.
ap < 0.001, significantly different from untreated astrocytes.
Because apoptotic and necrotic cell death may involve a disruption in MTP (Zamzami et al., 1995, 1997; Kroemer et al., 1997; Petit et al., 1997), alterations in the astrocytic MTP were investigated. The MTP was not altered by glucose deprivation alone but was depolarized by 200 μM SIN-1 (Fig. 6). However, a severe depolarization of the MTP, as indicated by the loss of the red fluorescence (J-aggregates), was observed in glucose-deprived/SIN-1-treated astrocytes. Conversely, green fluorescence (J-monomer) was increased in glucose-deprived/SIN-1-treated astrocytes. After 5-7 h, both red (J-aggregates) and green (J-monomer) fluorescences disappeared, indicating that the mitochondrial membrane was permeabilized (data not shown). Although pretreatment with L-BSO did not alter the MTP in control or glucose-deprived astrocytes, GSH depletion by L-BSO synergistically enhanced SIN-1-induced depolarization (Fig. 6).
Additional experiments were undertaken to determine whether the mitochondrial permeability transition (MPT) pore was contributing to the loss of viability of glucose-deprived/SIN-1-treated astrocytes. Addition of 2 μM cyclosporin A (CsA), a well-known blocker of the MPT pore, completely prevented the enhanced LDH release (Fig. 7) and the decrease in MTP (Fig. 8) in glucose-deprived/SIN-1-treated astrocytes. As CsA is also an inhibitor of calcineurin activity, its effects on astrocytic death and mitochondrial membrane depolarization may not be solely related to a decrease in permeability of the MTP pore. The possible involvement of calcineurin in astrocytic death and mitochondrial membrane depolarization was investigated using FK-506, a more selective calcineurin inhibitor. FK-506 (1 and 10 nM) attenuated neither the loss of viability (Fig. 7) nor the decrease in MTP (data not shown) in glucose-deprived/SIN-1-treated astrocytes. At higher concentrations (≥50 nM), FK-506 was itself cytotoxic (data not shown). Lipid peroxidation also has been reported to increase membrane proton leakage (Radi et al., 1991; Gadelha et al., 1997; Brookes et al., 1998). However, the lipid-soluble antioxidant Trolox did not block depolarization of the MTP in glucose-deprived/SIN-1-treated astrocytes (Fig. 9).
SIN-1 has been shown to produce ONOO- by simultaneously releasing NO and superoxide (Hogg et al., 1992). At least four observations indicate that the augmented death by SIN-1 in glucose-deprived astrocytes is mediated by ONOO- and not by NO. First, compared with SIN-1, SNAP and GSNO, which cause release of NO, not ONOO-, were much less effective in increasing glucose-deprived astrocytic death. Second, SOD and the SOD-mimetic MnTMPyP, which remove superoxide, blocked the augmented death. Third, the ONOO- scavengers, cysteine and GSH, blocked the SIN-1-induced augmented death. Fourth, nitrotyrosine immunoreactivity (a marker of ONOO-) markedly increased in glucose-deprived/SIN-1-treated astrocytes.
Astrocytes, a central component of the brain's antioxidant defense, appear to be particularly resistant to the actions of ONOO-. This resistance may arise from an ability to sustain cellular energy demands by glycolysis and by a superior capacity to handle oxidizing species such as ONOO- (Bolaños et al., 1995; Lizasoain et al., 1996). In view of the latter suggestion, the intracellular GSH concentration may be particularly important as ONOO- readily reacts with thiol-containing compounds (Bolaños et al., 1995; Lizasoain et al., 1996). Astrocytes contain 4.5-5 mM glutathione (Juurlink et al., 1996; present study). Glucose deprivation caused only a slight decrease of astrocytic GSH levels, possibly because the ATP required for GSH synthesis might be generated from amino acid metabolism even in the absence of glucose. In addition, the availability of GSH precursors was not restricted during incubation because the bathing medium (DMEM) contains amino acids and was only deficient for glucose. In the present study astrocytic GSH levels were not much decreased by the ONOO--releasing reagent SIN-1 (200 μM) alone. The GSH concentration in rat astrocytes was also previously reported to be little affected by 1-2 mM ONOO- (Bolaños et al., 1995). Thus, the sustained levels of GSH might protect the astrocytes from glucose deprivation- or ONOO- (released from 200 μM SIN-1)-induced cytotoxicity. It is interesting, however, that in the presence of ONOO-, glucose-deprived astrocytes rapidly lost the capability to maintain normal levels of intracellular GSH. Astrocytic GSH levels rapidly decreased even after prior elevation by adding GSH in the bathing medium. Consequently, GSH-depleted astrocytes became much more susceptible to ONOO--induced cytotoxicity. Also, 200 μM SIN-1 caused significant deterioration in the astrocytes depleted of GSH by L-BSO.
At present, the exact mechanism whereby astrocytic GSH levels are synergistically decreased by ONOO- and glucose deprivation remains unknown. As in other cells and tissues, the pentose phosphate pathway appears to be the predominant source in brain cells for regeneration of NADPH (Hotta, 1962; Hotta and Seventko, 1968; Baquer et al., 1988). Glucose deprivation interferes with NADPH production by the pentose phosphate pathway in astrocytes (Kussmaul et al., 1999). Because recycling of GSH from oxidized glutathione (GSSG) requires NADPH (Shan et al., 1990), in the absence of glucose the regeneration of GSH from GSSG produced during reduction of ONOO- might be compromised owing to insufficient regeneration of NADPH by the pentose phosphate pathway. Synthesis of GSH from glutamate, cysteine, and glycine also requires the intracellular free energy source ATP. Therefore, it is likely that glucose deprivation and ONOO- synergistically decrease the synthesis of GSH by inhibiting the reduction of GSSG to GSH, by lowering NADPH levels, and by reducing availability of precursors and ATP.
Recently, an increasing body of evidence suggests that the mitochondria may be a key regulator of cell death (Bernardi et al., 1999). In general, intact mitochondria maintain a large (up to 180 mV) negative membrane potential across the mitochondrial inner membrane. A decrease in MTP followed by an intense production of reactive oxygen species and a reduction of mitochondrial mass has been shown to occur in various models of cell death (Zamzami et al., 1995, 1997; Kroemer et al., 1997; Petit et al., 1997). Our electron microscopic studies showed that astrocytic mitochondria became markedly swollen within 3 h after starting glucose deprivation in the presence of SIN-1 (authors' unpublished data). Previous reports showed that the concentration of GSH in the brain was lowered in cerebral ischemia (Cooper et al., 1980; Lyrer et al., 1991) and that depletion of GSH caused a striking enlargement of mitochondria in the brain (Jain et al., 1991). ONOO- can alter the activity of the mitochondrial enzymes in the respiratory chain and cause opening of MPT pores (Packer and Murphy, 1995; Lizasoain et al., 1996; Gow et al., 1998). A recent report showed that dysfunction of the mitochondrial respiratory chain led to an opening of the MPT pores (Chavez et al., 1997). At low concentrations of ONOO- produced by 200 μM SIN-1, it did not exert cytotoxicity and only changed the MTP to limited extent. Therefore, it is concluded that glucose deprivation and ONOO- synergistically deplete intracellular GSH and depolarize the MTP in astrocytes.
Previously, the opening of MPT pores in isolated mitochondria was reported to be modulated by the oxidation-reduction equilibrium of glutathione (Reed and Savage, 1995; Chernyak and Bernardi, 1996; Costantini et al., 1996). However, the present study demonstrates that in rat astrocytes the MTP depolarization is not induced simply by the depletion of GSH. Thus, GSH depletion caused by L-BSO did not alter the MTP or the viability of control astrocytes. This discrepancy may be explained by the different experimental systems (isolated mitochondria vs. primary astrocytes). As demonstrated in Fig. 5, GSH depletion in astrocytes would lead to the intracellular accumulation of ONOO-, i.e., increased nitrotyrosine immunoreactivity, which in turn stimulates the opening of the MTP pore. However, it is possible that GSH depletion can facilitate the ONOO--evoked opening of the MTP pore. The present observations also imply that in control astrocytes GSH depletion does not generate sufficient oxidative stress to cause significant cellular damage (Barker et al., 1996). Furthermore, GSH depletion did not alter the viability and the MTP in glucose-deprived astrocytes, indicating that glucose deprivation also did not cause a significant production of reactive oxygen species in astrocytes. Alternatively, however, other antioxidant systems may compensate for the GSH depletion. With regard to this latter suggestion, the α-tocopherol (vitamin E) concentration of astrocytes was reported to be relatively high in chick brain (Makar et al., 1994).
The MTP can be changed by the formation of CsA-sensitive MPT pores (Minamikawa et al., 1999). Prevention of cell death and MTP depolarization by CsA may indicate that the MPT pore is involved in the increased death of glucose-deprived/SIN-1-treated astrocytes. A more selective calcineurin inhibitor, FK-506, did not attenuate the augmented death, implying that the MPT pore and not calcineurin is involved in the increased death evoked by SIN-1 in glucose-deprived astrocytes. Lipid peroxidation also has been reported to increase membrane proton leakage (Radi et al., 1991; Gadelha et al., 1997; Brookes et al., 1998). Thus, mitochondrial oxidative damage induced by ONOO- could cause lipid peroxidation, which by increasing proton leakage could disrupt glial mitochondria. However, lipid peroxidation, a proposed mechanism involved in ONOO- cytotoxicity, may not be a major cause of the depolarization of the MTP as loss of viability and depolarization of MTP in glucose-deprived/SIN-1-treated astrocytes were not prevented by the lipid-soluble vitamin E analogue Trolox.
Classically astrocytes are considered less vulnerable than neurons to the ischemic injury initiated by large artery occlusion (Obrenovitch et al., 1988). However, several reports showed that astrocytes were also vulnerable to ischemia. In the cerebral ischemic penumbra, progressive metabolic deterioration leads to pan-necrosis of both glial and neuronal cells (Brierly and Graham, 1984; Pantoni et al., 1996). Hypoxia, acidosis, and elevated extracellular K+ concentration have been considered as major causes of astrocytic injury or death during incomplete ischemia (Obrenovitch et al., 1988; Swanson et al., 1997). In the present study we provide further evidence that ONOO-, but not NO, can augment the death of energy-depleted astrocytes via a synergistic reduction in intracellular GSH levels. This reduction of GSH content in turn results in a buildup of the intracellular level of ONOO-, which then induces MTP depolarization, resulting in inevitable cell death. The enhanced level of ONOO- that is secondary to the reduction of GSH levels may, at least in part, explain the death of astrocytes in the ischemic penumbra. In view of the importance of antioxidants, in particular GSH, agents that are capable of increasing or maintaining the cellular concentration of this molecule may prove to be of therapeutic value.