INTRODUCTION

Exposure to stress is acknowledged to be involved in a wide range of physiological and psychopathological processes. In the brain, most of the deleterious consequences after stress (molecular, cellular and morphological changes, and even mood disorders) are attributed to supraphysiological effects of glucocorticoids (reviewed in Sapolsky (2000)) as a consequence of a compromised metabolic capacity of the brain induced by overexposure to these stress hormones (reviewed in McEwen (2003), Raison and Miller (2003)).

It is well known that energy consumption in the mammalian brain is supplied mainly by the oxidation of glucose. During the physiological activation of the brain areas there is a correlative increase in local glucose uptake (Sokoloff, 1999). However, several studies have found that in stressful situations, supraphysiological levels of corticosterone have inhibitory effects on glucose transport and metabolism in several brain regions in rats (Landgraf et al, 1978; Virgin et al, 1991) and also in humans (De León et al, 1997). In peripheral tissues, glucocorticoids released by stress act by decreasing glucose transport and metabolism and increasing serum glucose levels (Munck, 1971).

This decrease in brain glucose transport by the excess of glucocorticoids during stress exposure accounts for the decrease in ATP production observed in hippocampal neurons and glia (Tombaugh and Sapolsky, 1992; Lawrence and Sapolsky, 1994). The consequence is an increased vulnerability to excitotoxicity (ie, calcium overload, reduction in the reuptake of glutamate capacity of neurons) (Novelli et al, 1988; Cheng and Mattson, 1992; Joëls and Vreugdenhil, 1998) and less ability to afford the costly task of managing the consequences of an excitotoxic or metabolic insult (Yusim et al, 2000). This impairment in the brain glucose uptake, ATP loss and excitotoxicity has been seen not only after stress or glucocorticoids, but also as an early step before neuronal degeneration takes place, such as in Alzheimer disease (Hoyer et al, 1988).

Together with these deleterious effects of stress on energy production and excitotoxic processes in the brain, the oxidative/nitrosative and neuroinflammatory consequences of stress in the central nervous system are now receiving great attention (Elenkov and Chrousos, 2002). Indeed, it is now well accepted that inflammatory responses in the brain contribute to the cell death and damage during neurological and neuropsychiatric diseases related to stress exposure (neurodegenerative diseases, depression, post-traumatic stress disorder and schizophrenia) (reviewed in McLeod et al (2001)).

Recently, it has been shown that the peroxisome proliferator-activated receptors (PPARs) are involved in the regulation of inflammatory responses (Murphy and Holder, 2000; Chawla et al, 2001). The activation of one of their three major subtypes—PPARγ—contributes by reducing the secretion of proinflammatory cytokines and neurotoxic substances in brain cells (Petrova et al, 1999), that is downregulating inducible nitric-oxide synthase expression, reducing cell death in in vitro (Heneka et al, 1999) and in vivo (Heneka et al, 2000) brain experiments. The possible clinical use of PPARγ agonists in neurological diseases has been demonstrated in experimental models, and this has led to the possibility that PPARγ agonists provide protection in neurodegenerative diseases, such as Alzheimer's disease (Heneka et al, 2001), stroke (Uryu et al, 2002), Parkinson's disease (Breidert et al, 2002), and multiple sclerosis (Niino et al, 2001).

Previous studies from our laboratory have demonstrated that both synthetic and natural PPARγ ligands prevent inflammatory and oxidative/nitrosative consequences of stress exposure in the central nervous system of rats subjected to immobilization stress (García-Bueno, 2005a, 2005b). The mechanisms by which these compounds prevent these consequences include inhibition of stress-induced increase in inducible nitric oxide synthase activity, nuclear factor κB blockade (by preventing stress-induced IκBα decrease) and inhibition of TNFα release in stressed animals. Therefore, we are particularly interested in the search of other mechanisms of protection due to PPARγ activation in the stressed brain.

Along with the interference with expression and release of mediators of insulin resistance, the improvement in glucose uptake by cells in patients with uncomplicated type 2 diabetes has been demonstrated as the main antidiabetic mechanism of action of some of the clinically available PPARγ agonists thiazolidinediones (TZDs) (reviewed in Stumvoll and Haring (2002)). Indeed, studies carried out in adipose tissue, liver, and muscle (Chawla et al, 1994; Norris et al, 2003) showed that in periphery, PPARγ regulates glucose metabolism by increasing glucose uptake through facilitative glucose transporter proteins (GLUTs). More recently, it has been demonstrated that some synthetic PPARγ agonists increase glucose utilization in rat cortical slices as well as astroglial glucose metabolism (Dello Russo et al, 2003), adding further support for considering these drugs as potential neuroprotective agents.

As brain glucose metabolism could be adversely affected in some neurodegenerative disorders, including stress-related neuropsychopathologies (reviewed in Peppard et al (1990), Hoyer (2000), Weinstock and Shoham (2004), Peters et al (2004)), we hypothesized that one mechanism of PPARγ agonists' protection during stress exposure could be related to an increase in the cerebral metabolism of glucose. The delivery of glucose within the brain is mediated primarily by GLUT-1, present in the blood–brain barrier and in astrocytes and by GLUT-3 present in neurons (Vannucci et al, 1997; Pellerin, 2005). As 85% of the energy expenditure of the brain occurs in neurons (Attwell and Laughlin, 2001), and glucose is almost the only fuel used by the organ, we tested the possibility that the preventive effects of PPARγ agonists are through glucose transporters, possibly through the neuronal component, GLUT-3. Furthermore, we examined the possibility that PPARγ agonists prevent stress-induced ATP loss and excitotoxicity in an animal model that has been reported to cause accumulation of oxidative mediators (reviewed in Madrigal et al (2004)) to explore the mechanisms implicated. To test this hypothesis, we examined the effects of two PPARγ agonists: the natural ligand 15-deoxy-Δ-12,14-prostaglandin J2 (PGJ2, Kliewer et al, 2001) and one of the synthetic TZDs, rosiglitazone (RG, Houseknecht et al, 2002).

MATERIALS AND METHODS

Animals

Adult male Wistar rats (Han:Wist, ANUC) weighing 225–250 g were used. All experimental protocols adhered to the guidelines of the Animal Welfare Committee of the Universidad Complutense following European legislation (2003/65/EC). The rats were housed individually under standard conditions of temperature and humidity and a 12-h light/dark cycle (lights on at 0800) with free access to food and water. All animals were maintained under constant conditions for 4 days prior to stress.

Restraint Stress

Rats were exposed to 6 h stress between 0900 and 1500 in the animal homeroom. The restraint was performed using a plastic rodent restrainer that allowed for a close fit to rats for 7 or 14 days (Leza et al, 1998) in their home cages. Control animals were not subjected to stress, but were handled at 0900 for a few seconds. Animals were killed immediately after restraint (still in the restrainer) using sodium pentobarbital. Blood for plasma determinations was collected by cardiac puncture and anti-coagulated in the presence of tri-sodium citrate (3.15% w:v, 1 vol citrate per 9 vol blood). After decapitation, the brain was removed from the skull and one cortical area was excised from the brain and frozen at −80°C until assayed. The other cortex was processed to obtain synaptosomes (as described below).

Plasma Corticosterone

Plasma was obtained from blood samples by centrifuging the sample at 1000g for 15 min immediately after stress on day 7 or 14. All plasma samples were stored at −20°C before assay by using a commercially available kit by RIA of 125I-labeled rat corticosterone (DPC, Los Angeles, CA, USA). A gamma counter was used to measure radioactivity of the samples. The values obtained in control animals (208.43±15.09 ng/ml) match with the kit manufacturer's expected values in adult male Wistar rats at the time of blood extraction (≈1500 h).

Preparation of Synaptosomes

After decapitation, half the forebrain was dissected on ice. All subsequent steps were performed at 4°C. Cortical tissue was immediately homogenized in 25 vol of 0.32 M sucrose in a glass homogenizer fitted with a Teflon pestle. The homogenate was centrifuged at 200g for 10 min, and the supernatant was then collected and centrifuged at 20 000g for 20 min. The pellet was resuspended in 0.32 M sucrose and centrifuged at 20 000g for 20 min. The crude synaptosomal pellet was finally resuspended in Locke solution (NaCl 154 mM, KCl 5.6 mM, CaCl2 2.3 mM, MgCl2 1 mM, NaHCO3 3.6 mM, glucose 5 mM, HEPES 5 mM, pH 7.2) without glucose for the 2-deoxy [3H] glucose transport assay and in 1 ml of 0.32 M sucrose for the glutamate transport assay. Synaptosomal preparations, consisting of pre- and post-synaptic elements of neurons, and associated astrocytic end feet, have provided insights into regulation of a variety of synaptic functions including neurotransmitter release and reuptake, energy metabolism and ion transport systems in physiological, neurological, and neurodegenerative disorders (reviewed in Begley et al (1998)).

2-Deoxy [3H] Glucose Uptake by Synaptosomes

2-Deoxy [3H] glucose uptake in synaptosomes was measured according to the procedure described previously (Keller et al, 1997) with some modifications. Briefly, synaptosomes (200 μg/tube) were washed twice in glucose-free Locke's solution, and the assay was started by the addition of 1 μCi of 2-deoxy [3H] glucose. The incubation lasted 7 min at 37°C in a shaking bath and was stopped by pelleting the synaptosomes washing twice with cold glucose-free Locke's solution and lysing the synaptosomes in 200 μl of 1% sodium dodecyl sulfate (SDS)/PBS solution. The 3H-bound radioactivity was measured using a liquid scintillation counter (Beckman LS-6500).

Brain ATP Levels

ATP levels were determined with a firefly luciferin-luciferase assay commercial kit (ATP Bioluminescence Assay Kit HSII, Roche, Barcelona, Spain) using a Fluoroskan Ascent FL microplate reader (Labsystems, Helsinki, Finland). Data are expressed as percentage of control levels. Basal ATP levels are expressed in μmol/g (modified from Hurtado et al, 2003).

[3H]Glutamate Uptake by Synaptosomes

Sodium-dependent glutamate uptake in synaptosomes was measured according to the procedure previously described by Robinson et al (1991) with some modifications. In brief, 25-μl aliquots of synaptosomes were added to 250 μl of incubation buffer (5 mM Tris, 10 mM HEPES, 2.5 mM KCl, 1.4 M NaCl, 1.2 mM CaCl2, 1.2 mM MgCl2, 1.2 mM KH2PO4, and 10 mM dextrose, pH 7.4) containing L-[3H]glutamic acid 0.125 μM (1 mCi/ml; Amersham Biosciences Europe GmbH, Friburg, Germany) and incubated for 3 min at 37°C in a shaking bath. The reaction was terminated using 1 ml of ice-cold choline buffer (incubation buffer in which an equimolar concentration of choline chloride was substituted for NaCl), and the samples were centrifuged at 10 000g for 2 min to recover synaptosomes. The 3H-bound radioactivity was measured using a liquid scintillation counter.

Western Blot Analysis

Succesful synaptosomal preparations were verified by determining expression of membrane associated protein synaptophysin (Sigma, 1:5000) and the absence of the cytoplasmatic protein B-tubulin (Sigma, 1:1000) (Reagan et al, 2000). Synaptophysin was used as loading control. To determine the levels of these proteins and the facilitative glucose transporter proteins GLUT-1, GLUT-3, and the high-affinity sodium-dependent glutamate transporters EAAT-2 and EAAT-3, proteins present in synaptosomes were loaded (20 μg) and size-separated in 10% SDS-polyacrylamide gel electrophoresis (90 V). After blotting onto a polyvinylidene difluoride membrane (Millipore, Bedford, MA, USA), were incubated with specifics rabbit polyclonal GLUT-1 (Chemicon international, 1:3000), GLUT-3 (Chemicon international, 1:5000) EAAT-3 (Santa Cruz, 1:500) and mouse monoclonal EAAT-2 (Transduction Labs, 1:1000) antibodies. Proteins recognized by the antibody were visualized on X-ray film by chemiluminiscence (ECL) following the manufacturer's instructions (Amersham lbérica). Autorradiographs were quantified by densitometry (Scion Image, Scion Corp., Frederick, MD, USA), and several time expositions were analyzed to ensure the linearity of the band intensities. Data are presented as arbitrary units (AU) of optical density.

Pharmacological Tools

Various groups of animals were i.p. injected with rosiglitazone (RG) maleate (3 mg/kg; Alexis, San Diego, CA), a high-affinity synthetic PPARγ receptor ligand (reviewed in Berger and Moller 2002), and 120 μg/kg of 15d-PGJ2 (Cayman Chem., Ann Arbor, MI, USA). RG was dissolved in saline and 15d-PGJ2 in DMSO (10%). None of the parameters studied were modified in vehicle-treated rats when compared with noninjected animals. The drugs were injected at the onset of each daily stress session. Animals receiving vehicles at the onset of stress were used as the control group in the Results section and in figures. Each experimental group contained at least eight animals.

Protein Assay

Proteins were measured using bicinchoninic acid (Hill and Straka, 1988).

Chemicals and Statistical Analyses

Unless otherwise stated, the chemicals were from Sigma Spain, Madrid. Data in text and figures are expressed as mean±SEM of the indicated number of experiments. Multiple comparisons were analyzed by the ANOVA and Newman–Keuls test, and P<0.05 was considered as statistically significant.

RESULTS

Effects of Rosiglitazone and PGJ2 on Stress-Induced Decrease in Synaptosomal Glucose Uptake and Expression of Glucose Transporters

We tested the effect of rosiglitazone (RG) and PGJ2 on synaptosomal glucose uptake. Control, nonstressed rats were injected to either RG or PGJ2 for 7 or 14 days (3 mg/kg/day and 120 μg/kg/day, respectively). Cerebral glucose uptake was not modified by the compounds in control rats at the tested doses (not shown, n=4). Exposure to repeated stress (S7 and S14) leads to a marked decrease in glucose uptake (S7:60±6%; S14:68±8% vs control, 5.9±0.5 cpm/min/mg protein, both P<0.05 Figure 1). Treatment with RG and PGJ2 prevented this stress-induced impairment in glucose uptake function at both times studied (S7+RG:107±11%; S7+PGJ2:135±17%, both P<0.05 vs control; S14+RG:94±4%; S14+PGJ2:96±6%, both P<0.05 vs control) (Figure 1).

Figure 1
figure 1

Effect of repeated restraint exposure on glucose uptake in synaptosomes of control (nonstressed) and stressed rats during 6 h for 7 days (S7) or 14 days (S14) receiving vehicle (SS, see Materials and methods for details), rosiglitazone (RG; 3 mg/kg) or 15d-PGJ2 (PGJ2 120 μg/kg), respectively at the onset of restraint. The data represent the mean SEM of eight rats. *P<0.05 vs control; #P<0.05 vs correspondent SS (S7 or S14) (Newman–Keuls test).

Full size image

We also assessed the possibility that these effects might be due to expressional changes in glucose transporters. Successful isolation of cerebral membranes fraction was verified by immunoblot analysis showing the presence of the membrane protein associated to synaptic vesicles synaptophysin (Syn, MW 38 kDa), as well as the absence of the cytoplasmic protein B-tubulin (B-tub, MW 55 kDa) (Figure 2a). Western blot analysis revealed that stress reduced the expression of GLUT-3 (Figure 2b), but not the astrocytic form of GLUT-1 (45 kDa) (Figure 2c) and that treatment with RG and PGJ2 prevented this stress-induced reduction in GLUT-3 expression at both times studied (Figure 2b). The blood brain barrier form of GLUT-1 (55 kDa) was not found in the samples used in this study (not shown).

Figure 2
figure 2

(a) Detection of synaptophysin (Syn) protein presence and absence of β-tubulin (β-tub) in synaptosomal forebrain fractions by Western Blot. (b) Western blot analysis of neuronal GLUT-3 in synaptosomal fractions from forebrain of control and stressed (S7 or S14) rats treated with saline (SS) and with rosiglitazone (3 mg/kg) (RG) or 15d-PGJ2 (120 μg/kg) (PGJ2), respectively at the onset of stress and laser densitometric analysis of the bands (arbitrary units, AU). The data are representative of samples of six different animals. *P<0.05 vs control; #P<0.05 vs S7 or S14 (Newman–Keuls test). (c) Western blot detection of astrocytic GLUT-1 (45 kDa) protein and laser densitometric analysis of the band in synaptosomes of control and stressed rats during 6 h for 7 (S7) or 14 (S14) days with or without rosiglitazone (3 mg/kg) (RG) or 15d-PGJ2 (120 μg/kg) (PGJ2), respectively at the beginning of restraint session. Western data are representative of samples of six different animals.

Full size image

Effects of RG and PGJ2 on Stress-Induced Decrease in Brain ATP Levels

Exposure to repeated stress (S7 and S14) lead to a marked decrease in brain ATP levels (S7:52±6.2%; S14:63±5% vs control, 1.40±0.2 nmol/mg protein, both P<0.05, Figure 3). Treatment with RG and PGJ2 prevented this stress-induced decrease in ATP at both times studied (S7+RG:84±9; S7+PGJ2:88±8, both P<0.05 vs control; S14+RG:95±8; S14+PGJ2:92±7, both P<0.05 vs control) (Figure 3). Control, nonstressed rats were injected with either RG or PGJ2 for 7 or 14 days (3 mg/kg/day or 120 μg/kg/day, respectively), but ATP levels were not modified with the compounds at the tested doses (not shown, n=4).

Figure 3
figure 3

ATP content in brain cortex of control and stressed rats during 6 h for 7 days (S7) or 14 days (S14) receiving vehicle (SS), rosiglitazone (RS; 3 mg/kg) or 15d-PGJ2 (120 μg/kg) (PGJ2) respectively at the onset of stress protocol. The data represent the mean±SEM of eight rats. *P<0.05 vs control (Newman–Keuls test). #P<0.05 vs SS.

Full size image

Effects of RG and PGJ2 on Stress-Induced Decrease in Synaptosomal Glutamate Uptake and in the Expression of Glutamate Transporters

We tested the effect of rosiglitazone and PGJ2 on synaptosomal glutamate uptake. Control, nonstressed rats were injected with either RG or PGJ2 for 7 or 14 days (3 mg/kg/day and 120 μg/kg/day, respectively). Cerebral glutamate uptake was not modified with the compounds at the tested doses (not shown, n=4). Exposure to repeated stress (S7 and S14) leads to a marked decrease in glutamate uptake (S7:45±7; S14:53±4% vs control, 4±0.0015 nmol/min/mg prot. both P<0.05, Figure 4). Treatment with RG and PGJ2 prevented this stress-induced impairment in glutamate uptake function at both times studied (S7+RG:82±15; S7+PGJ2:125±20, both P<0.05 vs control; S14+RG:68±8; S14+PGJ2:78±10, both P<0.05 vs control) (Figure 4).

Figure 4
figure 4

Effect of restraint stress on glutamate uptake in forebrain synaptosomes of control (nonstressed) and stressed rats for 6 h during 7 (S7) or 14 (S14) days receiving vehicle (SS, see Materials and methods for details), rosiglitazone (RG; 3 mg/kg) or 15d-PGJ2 (PGJ2 120 μg/kg), respectively at the onset of restraint. The data represent the mean SEM of eight rats. *P<0.05 vs control; #P<0.05 vs S7 or S14 (Newman–Keuls test).

Full size image

We also assessed the possibility that these effects might be due to expressional changes in glutamate transporters. Western blot analysis revealed that stress reduced the expression of EAAT2 (Figure 5a), but not EAAT1 (not shown) or EAAT3 (Figure 5b) and that treatment with RG and PGJ2 prevented this stress-induced reduction in EAAT-2 expression at both times studied (Figure 5a).

Figure 5
figure 5

(a) Western Blot detection of astrocytic EAAT-2 protein and laser densitometric analysis of the band (arbitrary units, AU) in synaptosomes of control and stressed rats during 6 h for 7 (S7) or 14 (S14) days with or without rosiglitazone (3 mg/kg) (RG) or 15d-PGJ2 (120 μg/kg) (PGJ2), respectively at the onset of stress protocol. Western data are representative of samples of six different animals. *P<0.05 vs control; #P<0.05 vs S7 or S14 (Newman-Keuls test) (SS: vehicle, see Materials and methods for details). (b) Western blot analysis of neuronal EAAT-3 protein and laser densitometric analysis of the band (AU) in synaptosomal fraction of control and stressed rats during 6 h for 7 (S7) or 14 (S14) days with or without rosiglitazone (3 mg/kg) (RG) or 15d-PGJ2 (120 μg/kg) (PGJ2), respectively at the beginning of restraint session. Western data are representative of samples of six different animals (SS: vehicle see Materials and methods for details).

Full size image

Absence of Effects of RG and PGJ2 on Stress-Induced Increase in Plasma Corticosterone Levels

To test the possibility that the effects of selective PPARγ activation were related to interference on the general response to stress, we determined plasma corticosterone levels in all of the experimental groups studied. The effects elicited by the PPARγ ligands after stress in the brain are independent of the systemic stress response, as plasma corticosterone levels (at 1500 h) were not modified in rats injected at the onset of stress: control, saline-injected at 0900 h:208.43±15.09 ng/ml; control rats receiving RG 3 mg/kg:221.92±65.03; receiving PGJ2 120 μg/kg:191.03±39.99 ng/ml, both P>0.05 vs control, saline; stress 7d:325.55±16.02 ng/ml, P<0.05 vs control; stress 7d plus RG:297.12±20.02 ng/ml, P>0.05 vs stress 7d; stress 7d plus PGJ2:302.55±25.71 ng/ml, P>0.05 vs stress 7d. Finally, stress 14d:346.95±45.11 ng/ml, P<0.05 vs control; stress 14d plus RG:310.10±32.22 ng/ml, P>0.05 vs stress 14d; stress 14d plus PGJ2:325.62±48.84 ng/ml, P>0.05 vs stress 14d.

DISCUSSION

All of the findings presented here indicate that repeated stress causes an energy-compromised status in the brain which might account for excitotoxicity seen in this condition and that treatment with PPARγ ligands prevents all these changes. In particular, the reported results demonstrate that repeated immobilization stress during 1 or 2 weeks produces a decrease in brain cortical synaptosomal glucose uptake, this effect being prevented by treatment with both natural and synthetic PPARγ ligands by restoring protein expression of the neuronal glucose transporter, GLUT-3, in membrane fraction. Besides, brain ATP production is decreased in rats exposed to stress, and treatment with PPARγ ligands prevents stress-induced ATP loss. Finally, repeated immobilization stress also produces a decrease in brain cortical synaptosomal glutamate uptake, this effect being prevented by treatment with PPARγ ligands by restoring protein expression of the glial glutamate transporter, EAAT2 in synaptosomes.

Glucose transport from blood into the brain is mediated by a facilitated diffusion-type transport system, that is members of the GLUT supergene family of integral membrane proteins (Maher et al, 1994; Vannucci et al, 1997; Shepherd and Kahn, 1999). Tissue-specific expression of one or more members determines the rate of glucose entry into the cell. In brain, GLUT-1 and GLUT-3 seem to be the most important. GLUT-1, which is expressed and localized at the endothelial cells of the blood-brain barrier (in a glycosylated form with 55 kDa), is responsible for the majority of glucose uptake and utilization in the brain (Duelli and Kuschinsky, 2001) and takes part in the first step in the transport of glucose from the blood into the brain (Pardridge et al, 1990; Boado and Pardridge, 1990). On the membrane of end feet around blood vessels, astrocytes express a specific form of GLUT-1, with 45 kDa (reviewed in Pellerin (2005)). The next step of glucose transport from extracellular space into neuronal cells is taken by GLUT-3, localized at the neuronal cell membrane (Maher, 1995). GLUT-3 possesses higher affinity (lower kilometers) for glucose than GLUT-1 (reviewed in Bolaños et al (2004)). Taking into account the high affinity and dependence of neurons for glucose, it has been postulated that the activity of this transporter would play a beneficial role during glucose deprivation, hypoxic episodes or other metabolic compromising situations (Burant and Bell, 1992; Gerhart et al, 1992; Fattoretti et al, 2001; Burkhalter et al, 2003). Thus, GLUT-3 activity affords neuroprotection (Hara et al, 1989; Lee and Bondy, 1993; Urabe et al, 1996; Uehara et al, 1997), and its malfunctioning is related with neuronal damage (reviewed in McEwen and Reagan (2004)). We have shown that after repeated stress, at the stress intensity and duration evaluated in this article, the neuronal transporter GLUT-3 membrane expression results clearly affected supporting the preventive effects demonstrated of the PPARγ agonists used in this particular experimental setting (García-Bueno et al, 2005a, 2005b).

The results of the present study indicate that astrocytic GLUT-1 expression is not modified after 1 or 2 weeks of repeated stress. Membrane preparations such as synaptosomes are free from BBB cells. Thus, GLUT-1 should come from astrocytes, as verified by identification of the astrocytic form (45 kDa), but not the 55 kDa form, in the samples used in this study. The presence of this specific isoform of GLUT-1 has been previously demonstrated in rat brain synaptosomal preparations by Bhattacharyya and Brodsky (1988).

In our study, plasma corticosterone levels are beyond physiological levels, which is in agreement with the fact that glucocorticoids dose-dependently inhibit brain glucose uptake when they are present in the brain at concentrations above the high physiological range in in vitro (Virgin et al, 1991) and in vivo (Doyle et al, 1994). Interestingly, similar GC treatment caused a decrease in the affinity of glutamate uptake by astrocytes. Studies have shown that in hippocampal cell cultures glucocorticoids significantly inhibit glucose transport and glutamate uptake by both neurons and astrocytes (Horner et al, 1990; Virgin et al, 1991). This latter observation suggests that GCs might impair the ability of astrocytes to aid neurons (ie, by impairing their ability to remove damaging glutamate from the synapse), which may be the case of our model.

Two possible mechanisms have been proposed to explain the stress-induced decrease in GLUT-3 expression: modification in binding to plasma membrane (Horner et al, 1987) or oxidative attack by mediators released during stress (Reagan et al, 2000). The first possible mechanism of GLUTs regulation is mediated by translocation from the plasma membrane to intracellular sites, as demonstrated previously after GC treatment (Horner et al, 1987). Many studies have shown that fusion of GLUT-3 vesicles with the plasma membrane increases glucose uptake (Uemura and West Greenlee, 2001). The fact that the effects observed here were seen on an enriched membrane preparation (synaptosomes) suggests that this may be the case in stress-induced decrease in glucose uptake. The second possible mechanism for the inhibitory stress effects on glucose uptake is the oxidative damage of GLUT-3 evidenced by conjugation of 4-hydroxynonenal (HNE) (Mark et al, 1997) or other lipid peroxidation products released in brain in the same stress model used in this paper (daily immobilization for 6 h during 7 days) (Reagan et al, 2000). This has also been seen in other settings such as cultured hippocampal neurons exposed to Abeta (increased HNE production and conjugation to GLUT-3) (Mark et al, 1997). In the stress model used here, the production of oxidative and nitrosative mediators and the accumulation of lipid peroxidation markers in the brain have been widely documented (Liu et al, 1996; Madrigal et al, 2001).

The reported decrease in glucose uptake mediated by stress or by high levels of GCs could place neurons in an energy-compromised environment, which could detrimentally affect neuronal responsiveness to pathophysiological events. In fact, glucose transport impairment precedes ATP depletion in brain, increasing neuronal vulnerability to excitotoxicity by compromising function of ion-motive ATPases (Mark et al, 1995), as it has been seen to occur in some neurodegenerative processes such Ab-induced toxicity, schizophrenia and others (Novelli et al, 1988; Kalaria and Harik, 1989; Sims, 1990; Mark et al, 1997; McDermott and De Silva, 2005) in which the reduction in glucose uptake occurs as an early step in the disease process prior to neuronal degeneration (Pettegrew et al, 1994; Reiman et al, 1996). On the other hand, GCs promote reduction in ATP levels (Tombaugh and Sapolsky, 1992; Lawrence and Sapolsky, 1994). One of the possible ATP-dependent mechanisms compromised by stress are glutamate transporters.

A likely functional implication of the stress effects on cerebral glucose transport is an impairment of glutamate uptake (Virgin et al, 1991). Indeed, stress has been shown to increase extracellular glutamate concentrations in many brain areas (Lowy et al, 1993; Moghaddam et al, 1994; Stein-Behrens et al, 1994), an effect that is proposed to result from compromised activity of the energy-dependent excitatory amino acid transporters. On the other hand, it has recently been shown that glutamate release is mainly due to reversed operation of neuronal glutamate transporters in processes such as brain ischemia and others (Warner et al, 1996; Jabaudon et al, 2000; Rossi et al, 2000). Thus, in stress, this is one among the various mechanisms, which, alone or combined, may be responsible for glutamate release (Lawrence and Sapolsky, 1994) although we have not seen modifications in glutamate uptake in neurons (EAAT-3). The effects elicited by the type of stress used in this study seem to be restricted to astrocytic mechanisms: a large percentage of the neurotransmitter is removed from the synapse via this route. Astrocyte glutamate uptake is a high affinity process regulated by EAAT-2 (Hertz et al, 1983), and the fact that stress decreases EAAT-2 expression observed here indicates a specific damaging effect at this level. Thus, the excess of extracellular glutamate could be a feedback mechanism, as has been proposed between stress, glucose and glutamate in the brain, as glutamate stimulates the HPA axis promoting a continuous circle (Gabr et al, 1995). Interestingly, this change seems to be time-dependent and region-specific: after repeated stress (21–40 days), an increase in EAAT-2 has been found in hippocampus, which has been proposed as a counter-regulatory mechanism (Reagan et al, 2004; Fontella et al, 2004).

The demonstration of the PPARγ agonists effects used in this study, PGJ2 and rosiglitazone (RG), after stress has important implications for possible pharmacological management of stress-related neuropsychopathologies, in which energy status may be affected. Rosiglitazone is a member of the peroxisome proliferator activated receptor gamma (PPARγ) agonists of the thiazolidinedione family, which are widely used as antidiabetic agents. Rosiglitazone therapy improves insulin sensitivity and glucose uptake in patients with uncomplicated type 2 diabetes. In addition to the effects on glucose metabolism, both TZDs and natural PPARγ ligands have effects on lipid metabolism, inflammatory responses, and cellular proliferation (reviewed in Kostadinova et al (2005)). Finally, several studies have demonstrated an enhancement of glucose uptake and glucose transporter expression in adipocytes and myocyte membranes treated with TZDs (Ciaraldi and Henry, 1997).

Treatment with these drugs is effective in overcoming stress inhibition of neuronal glucose uptake, increasing glucose utilization and GLUT-3 expression. One of the possible mechanisms is the prevention of cerebral GLUT-3 oxidation, which occurs in stress (Reagan et al, 2000) as it has been demonstrated to occur in stressed brain (García-Bueno et al, 2005a, 2005b).

In conclusion, our results demonstrate that TZDs can increase the brain glucose metabolism, suggesting that this class of drugs may be therapeutically useful in conditions in which brain glucose levels or availability are limited.

Besides, rosiglitazone has been found to prevent the loss of ATP in in vitro hearts (Sidell et al, 2002). Our findings open up a line of research directed toward the elucidation of other roles of thiazolidindiones and 15d-PGJ2 by affecting the ATP levels in brain cells, which may have important implications after stress at the level of glutamate uptake and release.

To our knowledge, this is the first work that demonstrates the antiexcitotoxic effect of PPARγ agonists at the level of the expression and activity of the EAATs family. Recently, other authors have showed neuroprotective antiexcitotoxic properties of these compounds blocking the NMDA-receptor-mediated Ca2+ entry (Zhao et al, 2006), preventing oxidative stress (Aoun et al, 2003) or regulating the PI3-kinase; (Nishijima et al, 2001), effects that took place in our model of stress.

In summary, our present findings show that these drugs exert direct actions on cerebral glucose and glutamate metabolism added to its known antiinflamatory/antioxidant effects, adding new therapeutic implications in the management of patients at risk of stressful events.