Address correspondence and reprint requests to Simon Pope, Department of Clinical Biochemistry (Neurometabolic Unit), National Hospital of Neurology and Neurosurgery, London WC1N 3BG, UK. E-mail: email@example.com
Glutamate is the major excitatory amino acid of the mammalian brain but can be toxic to neurones if its extracellular levels are not tightly controlled. Astrocytes have a key role in the protection of neurones from glutamate toxicity, through regulation of extracellular glutamate levels via glutamate transporters and metabolic and antioxidant support. In this study, we report that cultures of rat astrocytes incubated with high extracellular glutamate (5 mM) exhibit a twofold increase in the extracellular concentration of the tripeptide antioxidant glutathione (GSH) over 4 h. Incubation with glutamate did not result in an increased release of lactate dehydrogenase, indicating that the rise in GSH was not because of membrane damage and leakage of intracellular pools. Glutamate-induced increase in extracellular GSH was also independent of de novo GSH synthesis, activation of NMDA and non-NMDA glutamate receptors or inhibition of extracellular GSH breakdown. Dose–response curves indicate that GSH release from rat astrocytes is significantly stimulated even at 0.1 mM glutamate. The ability of astrocytes to increase GSH release in the presence of extracellular glutamate could be an important neuroprotective mechanism enabling neurones to maintain levels of the key antioxidant, GSH, under conditions of glutamate toxicity.
total glutathione (amount of GSH + twice the amount of GSSG)
reactive nitrogen/oxygen species
Glutamate is the major excitatory amino acid of the mammalian brain (Danbolt 2001). It acts through a variety of ionotropic and metabotropic receptors: the first exert their effects via ligand-gated ion channels, whereas the second act through coupling to G proteins and activation of intracellular secondary messengers (Greenamyre and Porter 1994; Meldrum 2000). Although glutamate is an important excitatory neurotransmitter it can be toxic if its extracellular levels are not tightly controlled. In conditions where release and/or uptake of glutamate are altered, extracellular glutamate can accumulate causing a persistent or excessive activation of glutamate-gated ion channels (excitotoxicity) (Mark et al. 2001; Coyle and Puttfarcken 1993). A number of pathways have been implicated in glutamate excitotoxicity, namely calcium deregulation, loss of membrane potential, mitochondrial impairment and production of reactive nitrogen/oxygen species (RNOS), which can lead to oxidative/nitrosative stress and ultimately cell death (Coyle and Puttfarcken 1993; Massieu and Garcia 1998; Pitt et al. 2000).
The extracellular levels of glutamate have been measured in various in vivo disease models by microdialysis and have been shown to reach concentrations of > 500 μM following spinal cord injury (McAdoo et al. 1999) and be maintained at concentrations of > 50 μM for 1–2 h during and following ischaemic insult (Orwar et al. 1994; Ritz et al. 2004; Homola et al. 2006). As extracellular glutamate derives from intracellular vesicles (whose glutamate concentrations are between 0.24 and 11 mM; Harris and Sultan 1995), the local concentration of glutamate in these conditions is likely to be even higher. Prolonged exposure to such concentrations of glutamate is likely to result in significant neurotoxicity (Liu et al. 1999). Astrocytes have a fundamental role in the regulation of extracellular glutamate levels and in the protection of neurones from glutamate toxicity (Hertz and Zielke 2004). In normal synaptic transmission, glutamate released into the synaptic cleft by neurones is accumulated in astrocytes (Hertz et al. 1978) by means of glutamate transporters such as glutamate transporter 1 and glutamate aspartate transporter (Gadea and Lopez-Colome 2001), after which it is returned to neurones in the form of glutamine.
Astrocytes also protect neurones in other ways such as through metabolic and antioxidant support. One of the most important molecules in this respect is the antioxidant glutathione (GSH) (Schulz et al. 2000). The trafficking of GSH between astrocytes and neurones is particularly important in conditions of oxidative stress (Dringen 2000). Astrocytes are able to increase neuronal GSH levels by secreting GSH into the extracellular environment (Sagara et al. 1996; Dringen et al. 1999; Stewart et al. 2002). Neurones are unable to take up GSH directly but can make use of cysteinyl glycine and cysteine, which are produced from GSH by the consecutive action of γ-glutamyl transferase (γGT) and aminopeptidase N, two enzymes expressed on the surface of astrocytes and neurones respectively (Dringen et al. 1997, 2001). Cysteine is the rate-limiting substrate for GSH synthesis in neurones, so the supply of this substrate by astrocytes is essential for the maintenance of GSH levels in neurones (Dringen et al. 1999). Previous studies have shown that astrocytes increase GSH release in response to increases in RNOS, such as nitric oxide (NO) (Gegg et al. 2003) and hydrogen peroxide (Sagara et al. 1996). This increase in GSH release is hypothesised to be a neuroprotective mechanism which maintains and/or increases neuronal GSH levels to counteract the damaging effects of RNOS. As oxidative stress is considered to be a key component of glutamate toxicity it was the aim of this study to investigate whether high concentrations of extracellular glutamate also had an effect on GSH release from astrocytes.
Materials and methods
Minimum essential medium (l-valine based) and foetal bovine serum were purchased from Gibco-Invitrogen (Paisley, UK). Cell culture flasks were purchased from Nalgene Nunc International (Naperville, IL, USA). Six-well plates were purchased from Corning Costar (High Wycombe, UK). All other chemical reagents were purchased from Sigma Chemical Company (Poole, UK). For the experiments performed on cell cultures on 24-well dishes the following reagents were used: Dulbecco’s modified Eagle’s medium was from Gibco-Invitrogen (Karlsruhe, Germany). Foetal calf serum and penicillin/streptomycin stock solution were from Biochrom (Berlin, Germany). Sulphosalycilic acid and NADPH were from AppliChem (Darmstadt, Germany). Glutathione reductase and GSSG were obtained from Roche Diagnostics (Mannheim, Germany). All other chemicals were obtained from Sigma (Steinheim, Germany), Fluka (Neu-Ulm, Germany) or Merck (Darmstadt, Germany). Sterile 24-well dishes were from Sarstedt (Nümbrecht, Germany).
Primary cultures of astrocytes
Primary cortical astrocytes cultures were prepared from Wistar rat neonates (0–2 days). The cerebral hemispheres were removed from the skull under the dissecting microscope, and cortex and hippocampus were isolated and manipulated separately. Cortical and hippocampal astrocytes were prepared as described previously (Griffin et al. 2005). Astrocytes on day in vitro (DIV) 13 were removed from the flasks with 0.01% trypsin, and seeded on to poly-d-lysine-coated six-well plates at a density of 1 × 106 cells/well. Experiments on these secondary astrocyte cultures were conducted at DIV 14. The experiments shown in Fig. 1b and Table 1 were performed on primary astrocyte cultures that were prepared according to the method described by Hamprecht and Loeffler (1985) by seeding 3 × 105 cells per well of 24-well dishes. These cultures were used at DIV 15–23.
Table 1. Cellular and extracellular GSH contents (nmol/well) of primary astrocyte cultures that were treated with glutamate and/or BSO
0 min Cells
240 min Cells
240 min Media
240 min Cells + Media
Primary astrocyte cultures in wells of 24-well dishes were pre-incubated for 2 h in MM without or with BSO (5 mM) before they were incubated for 4 h in 0.5 mL MM in the presence or absence of glutamate (5 mM) and/or BSO (5 mM). The basal cellular GSH content of untreated primary astrocyte cultures was 23.0 ± 1.8 nmol/mg protein. The 2 h pre-incubation of these cultures without and with BSO (5 mM) lowered the GSH content to 19.7 ± 1.0 nmol/mg and 17.5 ± 0.5 nmol/mg respectively. The data presented are mean ± SEM of experiments performed on 4 independently prepared cultures. The significance of differences to the data obtained for the control condition (no glutamate and no BSO) are indicated as *p <0.05, analysed by anova followed by the Tukey post hoc test. For all conditions, the extracellular activity of LDH was less than 10% of initial cellular LDH and the values did not differ significantly between the individual groups. GSH release rates from cultured astrocytes have previously been reported to be between 2 and 4 nmol/mg/h (Sagara et al. 1996; Hirrlinger et al. 2002; Gegg et al. 2003). In the current study, the GSH release rate was 2.25 nmol/mg/h under control conditions and 3.5 nmol/mg/h after addition of glutamate. LDH, lactate dehydrogenase; BSO, buthionine sulphoxime.
1.9 ± 0.2
1.0 ± 0.1 (53 ± 5%)
0.9 ± 0.1 (45 ± 4%)
1.9 ± 0.1 (98 ± 3%)
1.9 ± 0.2
1.1 ± 0.1 (57 ± 5%)
1.4 ± 0.1* (70 ± 7%)*
2.5 ± 0.2 (127 ± 10%)
Cont + BSO
1.7 ± 0.2
0.7 ± 0.0* (41 ± 3%)
0.9 ± 0.1 (51 ± 5%)
1.6 ± 0.1 (92 ± 7%)
Glu + BSO
1.7 ± 0.2
0.8 ± 0.1 (48 ± 5%)
1.2 ± 0.1 (70 ± 7%)*
2.0 ± 0.2 (118 ± 11%)
GSH release from astrocytes
The media of six-well plates containing secondary astrocyte cultures at DIV 14 was removed and the cells were washed twice in 1 mL Hank's buffered saline solution; 1 mL minimal medium (MM) (44 mM NaHCO3, 110 mM NaCl, 1.8 mM CaCl2, 5.4 mM MgSO4, 0.92 mM NaH2PO4 and 5 mM glucose, adjusted with CO2 to pH 7.4) was added to each well, supplemented with 5 mM sodium glutamate, 5 mM buthionine sulphoxime (BSO) or both. For BSO experiments, cells were incubated in MM containing 5 mM BSO for 2 h before and during supplementation with glutamate. After stimulation for 15, 45, 120 and 240 min, 500 μL of medium was removed and centrifuged at 3000 g for 5 min to remove cell debris (NB: Different wells were used for each time point). A total of 250 μL of supernatant was added to the same volume of 30 mM o-phosphoric acid and kept at −80°C for up to 3 weeks until HPLC determination of GSH. For experiments on primary cultures on 24-well dishes, cells were washed with 0.5 mL of pre-warmed (37°C) MM, pre-incubated for 2 h in 0.5 mL MM with 100 μM of the γ-glutamyl transpeptidase (γGT)-inhibitor acivicin (Dringen et al. 1997) in the absence or the presence of BSO (5 mM), and incubated in the cell incubator with 0.5 mL incubation medium (MM with 100 μM acivicin) in the absence or presence of glutamate (5 mM) and/or BSO (5 mM). Extracts of cells and media in 1% (w/v) of sulphosalicylic acid were used to determine the total glutathione content (GSx = amount of GSH plus twice the amount of GSSG). For determination of the content of GSSG in lysates or media the GSH present was derivatised with 2-vinylpyridine as described previously (Minich et al. 2006). For all conditions investigated the GSSG values were in the range of the detection limit of the assay used (< 5% of GSx). Therefore, the GSx amounts determined are considered and addressed here as GSH amounts.
GSH levels were determined electrochemically following extraction of GSH into 15 mM o-phosphoric acid (final concentration) and separation by reverse-phase HPLC (Riederer et al. 1989). The levels of GSx and of GSSG in cells and media of primary astrocyte cultures in wells of 24-well dishes were determined as previously described (Minich et al. 2006) by a modification of the colorimetric Tietze assay.
Lactate dehydrogenase release
Lactate dehydrogenase (LSH) activity was determined by measurement of NADH oxidation at 340 nm in the presence of pyruvate. The assay was performed in 96-well plates as described (Dringen et al. 1998). The percentage of LDH released into medium was calculated for three separate preparations (mean ± SEM) by the following: (LDH activity in medium/Total LDH in medium after cell lysis with Triton X-100) × 100.
Results are expressed as a mean ± SEM values for the number of preparations indicated. Statistical significance for the comparison of two groups was assessed using Student’s t-test. Multiple comparisons were made by one-way anova followed by the Bonferroni test unless otherwise stated. A value of p <0.05 was considered significant. Data expressed as ratios were transformed as previously described (Gegg et al. 2003), prior to statistical analysis.
Glutamate induces an increase in extracellular GSH in cultures of rat cortical astrocytes
To assess the effect of extracellular glutamate on GSH release, rat cortical astrocytes were treated with glutamate and extracellular GSH was measured at various time points by HPLC (Fig. 1a). In these initial experiments 5 mM glutamate was used. Although this could be thought of as a comparatively high glutamate concentration to use, similar glutamate concentrations are thought to be reached in the synaptic cleft following release of a single synaptic vesicle (hypothesised to be between 0.24 and 11 mM) (Harris and Sultan 1995) and millimolar glutamate has been used before to model glutamate excitotoxicity in astrocytes (Chen et al. 2000).
In the absence of glutamate, extracellular GSH increased to 0.52 ± 0.05 μM after 120 min and 1.23 ± 0.18 μM after 240 min (Fig. 1a). In the presence of 5 mM glutamate, the concentration of extracellular GSH was significantly higher after 120 and 240 min compared with control astrocytes, reaching 1.22 ± 0.08 and 2.33 ± 0.17 μM respectively (Fig. 1a□, p <0.05). Similar results were obtained for primary astrocyte cultures on 24-well dishes using a different assay to determine GSH (Fig. 1b). These results indicate that glutamate, at this concentration, induces a strong increase in extracellular GSH in rat astrocyte cultures. However, an apparent increase in extracellular GSH could be the result of increased GSH synthesis following incubation with glutamate, increased leakage of intracellular contents because of glutamate toxicity or inhibition of extracellular GSH breakdown. These were investigated in turn.
Increased extracellular GSH is not a result of de novo synthesis
Glutamate can be used by cells for GSH synthesis, provided other precursors are not limited (Dringen and Hamprecht 1998), and constitutive GSH release from astrocytes correlates with intracellular GSH concentration (Sagara et al. 1993). The increase in extracellular GSH observed in the presence of high extracellular glutamate could therefore result from increased GSH synthesis. To determine whether this was the case, glutamate-induced GSH release was measured in the presence and absence of the GSH synthesis inhibitor BSO (Fig. 1). Astrocytes were incubated with or without 5 mM BSO (a concentration that has previously been shown to inhibit de novo GSH synthesis (Gegg et al. 2002)) in MM for 2 h prior to and throughout experiments. In the absence of glutamate, extracellular GSH levels for BSO-treated astrocytes were not significantly different from control astrocytes, reaching 0.99 ± 0.09 μM after 240 min (Fig. 1a•). When glutamate was added to BSO-treated astrocytes a significant increase in extracellular GSH was detected, reaching 2.28 ± 0.17 μM after 240 min (Fig. 1a○, Cont + BSO vs. Glu + BSO, p <0.05), similar to what was observed in glutamate-treated astrocytes in the absence of BSO. These results were confirmed for primary astrocyte cultures on 24-well dishes (Fig. 1b).
Glutamate does not induce LDH release from cortical astrocytes
In order to determine if the increase in extracellular GSH was because of glutamate-induced cellular damage, LDH levels were measured in media and cells as an indicator of membrane disruption. As determined for the 240 min time point, LDH levels were not significantly different between control (1.6 ± 0.3%) and glutamate-treated astrocytes (2.0 ± 1.2%), suggesting that the increase in extracellular GSH was not a consequence of leakage of intracellular content. LDH release levels were also not significantly different between glutamate-treated and control astrocytes in the presence of BSO (2.2 ± 0.3% vs. 2.4 ± 0.6% respectively).
Increased extracellular GSH is not a result of inhibition of γGT by glutamate
γ-Glutamyl transferase, expressed on the surface of astrocytes, breaks down extracellular GSH by catalysing the transfer of the glutamyl residue of GSH to a variety of amino acid and dipeptide acceptors (Dringen et al. 1997). Inhibition of γGT by acivicin has been shown to result in an increase in extracellular GSH (Dringen et al. 1997). To investigate the possibility that glutamate was increasing extracellular GSH levels by inhibiting γGT, the effect of acivicin with or without glutamate on the release of extracellular GSH by primary rat astrocytes was tested (Fig. 2). Treatment with 100 μM acivicin resulted in a slight but not significant increase in extracellular GSH in control astrocytes after 240 min (1.42 ± 0.01 μM for Cont vs. 1.70 ± 0.03 μM for Cont + Aciv), suggesting that γGT was not particularly active in our cultures to metabolise the GSH released from the cells. However, a combination of acivicin and glutamate did result in a significant increase in extracellular GSH after 240 min compared with astrocytes treated with glutamate alone (3.17 ± 0.13 μM for Glu alone vs. 3.79 ± 0.13 μM for Glu + Aciv, p <0.05). As acivicin was used at a concentration which has previously been reported to maximally inhibit γGT (Dringen et al. 1997) and glutamate increased extracellular GSH even in the presence of acivicin, this data suggests that glutamate does not act by inhibiting γGT.
Determination of cellular GSH
In order to better understand the effects of glutamate and BSO on GSH metabolism and GSH release in astrocytes, intra- and extracellular GSH was measured before and after glutamate stimulation of primary astrocytes on 24-well dishes. In the absence of glutamate, approximately 50% of the initial cellular GSH was found in the medium after 240 min incubation. This amount was increased to approximately 70%, if glutamate was present during the incubation. In contrast, the presence of BSO did not alter the extracellular GSH content compared with the respective controls without BSO. The differences found for the sum of cellular plus extracellular GSH after 240 min of incubation were not significant (p >0.05). For all conditions shown in Table 1, GSSG accounted for less than 5% of the GSx contents in cells or media (data not shown), indicating that GSH and not GSSG is released from astrocytes and that the presence of glutamate does not significantly affect the extracellular GSH/GSSG ratio.
Effect of glutamate receptor agonists on GSH release
To study whether GSH release was dependent on activation of glutamate receptors, agonists were added to astrocyte cultures. Incubation of astrocytes with agonists to the NMDA (50 μM) or non-NMDA ionotropic glutamate receptors [50 μM α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA)] had no significant effect on GSH release at the 240 min time point compared with control astrocytes (Table 2).
Table 2. Effect of glutamate receptor agonists on GSH release from astrocytes
Extracellular GSH (μM, mean ± SEM)
Percentage control (%)
Cortical astrocyte cultures were incubated for 4 h with or without glutamate or glutamate receptor agonists, as indicated. Only glutamate had a significant effect on extracellular GSH compared with controls (p <0.01). Percentage control is the extracellular GSH concentration after 4 h compared with the control for that experiment. n numbers are as indicated. AMPA, α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid.
1.4 ± 0.2
100 ± 13.9
5 mM glutamate
2.8 ± 0.3
208.8 ± 24.1
50 μM NMDA
1.3 ± 0.3
104.1 ± 14.8
50 μM AMPA
1.2 ± 0.2
93 ± 7.1
Glutamate-induced GSH release from hippocampal astrocytes
In order to investigate whether glutamate-induced increase in extracellular GSH could be observed in astrocytes from other brain regions, hippocampal astrocytes at DIV 14 were compared with cortical astrocytes. In the absence of glutamate, extracellular GSH increased to 1.14 ± 0.23 μM after 240 min in hippocampal cultures compared with 1.22 ± 0.08 μM for cortical cultures (Fig. 3). In the presence of 5 mM glutamate, the concentration of extracellular GSH in hippocampal cultures was significantly increased compared with controls (2.73 ± 0.42 μM vs. 1.14 ± 0.23 μM, respectively, at 240 min; p <0.05). This increase in extracellular GSH is of the same order of magnitude to that observed in cortical astrocyte cultures. As observed for cortical astrocytes, no significant difference could be observed between control and glutamate-treated hippocampal cells in terms of LDH release (1.7 ± 0.5% vs. 1.5 ± 0.3% respectively).
The above experiments were all performed using 5 mM glutamate, a relatively high concentration that is only likely to be present transiently under physiological conditions. Therefore, the above GSH release experiments were repeated using lower concentrations of glutamate. The dose–response curves generated for both cortical and hippocampal astrocytes indicate that GSH release is increased after 240 min even at relatively low glutamate concentrations (0.1 mM) and maximal GSH release is already achieved with 0.5 mM glutamate (Fig. 4). Half-maximal GSH release was achieved at approximately 250 μM glutamate for both hippocampal and cortical cultures.
In the present report, we demonstrate that prolonged exposure to glutamate induces an increase in the concentration of extracellular GSH in three different types of cultured astrocytes. These cells are known to release GSH (Sagara et al. 1996), and when cultured with 5 mM glutamate we observed a significant increase in the amount of extracellular GSH over 240 min (Fig. 1), without evidence of cellular damage. At least a twofold increase in extracellular GSH was observed in both cortical and hippocampal astrocytes after 240 min treatment with glutamate, suggesting this to be a feature common to astrocytes from different brain regions. Dose–response curves also indicated that glutamate induces GSH release from astrocytes at concentrations as low as 0.1 mM (Fig. 4). A number of possible causes for this increase in extracellular GSH have been investigated in this study and are discussed in more detail below.
Glutamate is one of the precursors of GSH (Kranich et al. 1996), and an increase in the synthesis of GSH could result in its increased release into the media. However, under our experimental conditions, glutamate did not cause a significant increase in intracellular GSH (Table 1). This is not surprising as it has been shown previously that addition of 1 mM glutamate to astrocytes only results in an increase in intracellular GSH concentration if cystine/cysteine and glycine are also added (Dringen and Hamprecht 1996). The absence of these substrates in our media suggests that de novo GSH synthesis does not explain the increase in extracellular levels. Support for this argument also come from our experiments with BSO, a potent and specific inhibitor of glutamate-cysteine ligase (the rate-limiting step in GSH synthesis) (Griffith and Meister, 1979). Presence of BSO had no significant effect on GSH release in the time frame of the experiment (Fig. 1). A longer BSO incubation would be expected to lower intracellular GSH to a larger extent, and possibly have an effect on glutamate-induced release if critical intracellular GSH levels were reached. Altogether, these results are in agreement with reports showing that astrocytes rely on stored GSH to resist otherwise harmful conditions, failing to survive only when these pools are depleted (Chen et al. 2000), and emphasise the capacity of astrocytes to release GSH when exposed to glutamate.
High concentrations of glutamate can be toxic to some cell types, leading to necrotic cell death with membrane rupture and leakage of intracellular content (Coyle and Puttfarcken 1993). As intracellular GSH concentrations are about 1000-times extracellular concentrations (mM vs. μM respectively; Dringen 2000), an increase in membrane leakage could explain the significant increase in extracellular GSH in the current study. However, no significant differences could be detected between control and glutamate-treated cells in terms of LDH release, suggesting that increased extracellular GSH detection was not a result of membrane rupture induced by glutamate. Our results are consistent with those of others in terms of the gliotoxic action of glutamate. Chen et al. (2000) demonstrated that 10 mM L-glutamate leads to LDH release only after a very prolonged incubation period (16 h), during which changes in cell morphology and oxidative stress occurs. These changes could be terminated by removal of glutamate before the onset of cell damage (estimated to occur at 4–6 h), indicating that the glutamate effect was reversible and that continuous exposure was required for astrocyte death. As glutamate did not appear to cause release of GSH through non-specific cell leakage other mechanisms were investigated.
The data in Table 1 show that glutamate increases the proportion of GSH that is extracellular in astrocyte cultures. Two possible explanations for this rise in extracellular GSH have been discounted in this study – namely glutamate inhibition of extracellular processing of GSH by γGT (Fig. 2) and glutamate affecting the extracellular GSH/GSSG ratio (Determination of cellular GSH). Therefore, the most likely explanation for the increase in extracellular GSH in astrocyte cultures upon exposure to glutamate is stimulation of GSH release (Fig. 5). This increased release of GSH from rat astrocytes could result from the activation of glutamate receptors and/or activation of downstream signalling pathways by glutamate. Glutamate receptors are considered to be expressed mainly on neurones but are also present on astrocytes (Porter and McCarthy 1996, 1997), where they have been increasingly implicated in a number of important pathways, including e.g. regulation of intracellular Ca2+ levels, stimulation of protein kinase C and inhibition of adenylate cyclase (Porter and McCarthy 1996, 1997; Winder and Conn 1996). Neurotransmitter(s) released from pre-synaptic terminals could therefore activate receptors located on astrocytes, leading to GSH release. However, data from experiments using agonists for ionotropic glutamate receptors suggests that neither NMDA nor AMPA/kainate receptors are involved in GSH release, as we were unable to detect elevated extracellular levels of GSH after incubation with NMDA or AMPA. Preliminary experiments have also so far failed to show a role of metabotropic receptors in GSH release (data not shown). Glutamate is also able to induce various changes in astrocytes which are not mediated via glutamate receptors. These changes include a switch of astrocytic metabolism from glycolytic to oxidative, via decreased glucose utilisation and increased mitochondrial activity (Liao and Chen 2003). Such changes to astrocyte energy metabolism may also affect GSH metabolism and export, although this remains to be elucidated.
Reactive nitrogen/oxygen species such as hydrogen peroxide and NO have also been implicated in the increase of GSH in cultured astrocytes (Sagara et al. 1996; Gegg et al. 2003), and oxidative stress was shown to result in the over-expression of Nrf2, a transcription factor implicated in GSH use, production and efflux pathways in astrocytes, via antioxidant-response element activation (Shih et al. 2003). Hypothetically, such transcription factor regulated changes could also be induced by glutamate and increase GSH efflux. However, significant changes to gene expression are likely to take hours rather than minutes and are therefore unlikely to contribute to the initial glutamate-induced GSH release observed in this study. Several of the transporters reported to transport GSH are expressed in cultured astrocytes (Minich et al. 2006). However, so far only multidrug resistance protein 1 has been identified in astrocytes to participate in GSH transport under basal conditions (Minich et al. 2006). Whether this transporter, other multidrug resistance proteins, organic anion transporters or the cystic fibrosis transmembrane conductance regulator protein contribute to the elevated GSH release from astrocytes in the presence of glutamate remains to be elucidated.
This increased release of GSH in response to high extracellular glutamate can be regarded as a candidate antioxidant defence mechanism preventing neuronal damage (Drukarch et al. 1997, 1998; Gegg et al. 2005), but GSH may also be implicated in other regulatory events. GSH has been described as candidate modulator of CNS excitability, through binding to the NMDA receptor complex as either an agonist or antagonist in particular circumstances (Ogita et al. 1995; Oja et al. 2000); has been shown to limit cell sensitivity to NO-mediated mitochondrial injury (Bolanos et al. 1996; Gegg et al. 2005); and GSH and other reductants have also been demonstrated to increase the glutamate uptake current of glutamate transporters, an event that could be reversed by the oxidative agent 5,5′-dithio-bis(2-nitrobenzoic) acid (Trotti et al. 1997). In light of this, the ability of astrocytes to release GSH may prove to be important in protecting neurons from glutamate toxicity in distinct brain structures by means other than its role as an antioxidant.
In conclusion, our experimental strategy models conditions where extracellular glutamate levels are raised for prolonged periods such as during ischaemia. Considering the range of glutamate-mediated mechanisms leading to neuronal death, including nitrosative and oxidative stress, the increased availability of GSH, an endogenous low-molecular weight antioxidant, may constitute an important protective mechanism in response to excitotoxic insults.
We are grateful to Dr Mary Hughes and Dr Richard Foxton for all their help with astrocyte cultures. This work was supported by FCT, Portugal (POCTI/BCI/42365/2001 and SFRH/BD/5356/2001) and the Hospitals Savings Association, UK.