Measurement of Glu uptake was initially taken as an index of the presence of putative glutamatergic nerve terminals. In the 1970s, numerous studies were carried out, after selective brain lesions, to identify a decrease in the uptake rate in brain homogenate/synaptosomal preparations, thereby contributing to the identification of the so-called glutamatergic pathways (Fonnum 1984). As well as the potential contribution to basic aspects of EAA signalling, the neuronal transporters possibly located on the plasma membrane of the glutamatergic nerve terminal itself could contribute to a short recycling of the neurotransmitter in the nerve ending by-passing the astrocytes. In that case, however, the exact nature of the transporter involved in the uptake of the neurotransmitter remains unknown. The use of synaptosomes was also considered to provide an opportunity for the study of the dynamic properties of transport in terms of affinity of the uptake systems for neurotransmitters, ionic mechanisms and pharmacological properties. The demonstration of a Na+/Cl– dependence, and definition of a high affinity (Km in the 10−6 m range) together with the existence of compounds able to selectively interfere with the process, contributed to determine precisely the basic properties of the synapse (Danbolt 2001).
The electrogenicity of the transport of Glu into brain synaptosomes was thus rapidly established. However, in such a preparation it was difficult to determine the exact contribution of each ion to the transport. Also, the contribution of glial cells to the transport in brain homogenates brought into question the value of experimental data regarding neuronal transport. In fact, at that time, the situation was a little confused and was only clarified later by the cloning of genes encoding for the transporter proteins (Kanai and Hediger 1992). The stoichiometry was further established in more recent studies using the cell expression of EAAC1/EAAT3 as three sodium ions (Na+) and one hydrogen ion (H+) entering the cell together with one Glu, in exchange for one potassium ion (K+) (Zerangue and Kavanaugh 1996). Transfection experiments in numerous cellular models have further contributed to determine precisely the individual kinetic characteristics of the transporters, with Km values in the micromolar range, and to characterize compounds which could selectively interfere with the different transport systems (Gegelashvili and Schousboe 1998). Unfortunately, no compound was shown to inhibit really selectively one or other of the cloned EAAs transporters.
EAAC1/EAAT3 brain expression
The EAAC1/EAAT3 gene (SLC1A1 gene) shows moderate brain expression compared with that of the glial transporters (Haugeto et al. 1996) and exhibits a rather homogeneous distribution. However, EAAC1/EAAT3 is currently considered to be particularly expressed at cortical level in the hippocampus, basal ganglia structures and the olfactory bulb. Further analysis has shown that the transporter is expressed in glutamatergic neurones and in GABA-containing neurones, but also in cholinergic and aminergic neurones (Rothstein et al. 1994; Coco et al. 1997; Shashidharan et al. 1997; Danbolt et al. 1998; Sepkuty et al. 2002). EAAC1/EAAT3 has also been shown to be significantly expressed in the kidney, placenta, intestine and heart but in these structures, the functional significance remains unclear (Danbolt 2001).
EAAC1/EAAT3 shows a developmental expression profile (Torp et al. 1994; Furuta et al. 1997). Its expression precedes that of GLAST/EAAT1 and, notably, GLT1/EAAT2 during the early stages of brain development in the fetus, and further stabilizes to be constant or slightly reduced in the adult (Sims and Robinson 1999). We recently demonstrated a similar developmental expression profile and uptake activity in rat primary cortical cultures with early EAAC1/EAAT3 expression [from 3 to 14 days in vitro (DIV) with a slight decline after 10 DIV], whereas GLT1/EAAT2 was only detected at 7 DIV (Guillet et al. 2002). These data pointed to the remarkable capability of the neuronal transporter EAAC1/EAAT3 to maintain very efficiently low extracellular Glu levels in the culture medium in the earlier stages of the culture when the glial transporters are not yet efficient.
Molecular structure and topology of the transporter
The molecular weight of the EAAC1/EAAT3 protein is generally accepted to be of 64 kDa but, because of higher molecular mass species in electrophoretic experiments, it has been suggested that the protein may form multimers. Homo-multimers or hetero-multimers could contribute to the maturation of the transporter (Yang and Kilberg 2002) and, further, to the transporter pore. In this respect, it has been suggested from expression of EAAC1/EAAT3 in the Xenopus oocyte that the transport complex is a pentamer forming a pore with Cl– channel activity (Eskandari et al. 2000). The five sodium-dependent high affinity cloned Glu transporters in eukaryotic cells are in fact members of a more general membrane transport protein family also involving prokaryotic transporters. Interestingly, both categories of proteins share a high degree of amino acid homology, similar substrate transport ionic dependency, somewhat similar substrate specificity and close transmembrane topology (Saier 2000). In this respect, the use of prokaryotic cells recently contributed to a study of the structure of the eukaryotic cell Glu transporter, as a crystal structure of the transporter homologue was obtained from Pyrococcus horikosshii (Yernool et al. 2004). Data suggested a trimeric rather than pentameric structure of the Glu transporter, further verifying previous results from the same group obtained from the purification of Bacillus caldotenax and Bacillus stearothermophilus Glu transporters (Yernool et al. 2003). These data are in complete agreement with the conclusions of a previous study emphasizing the fact that trimers and dimers predominate following cross-linking of GLT1 and GLAST (Haugeto et al. 1996).
The Glu transporter topology, however, remains to be clarified. Various models of the transmembrane arrangement have been successively proposed but mainly for the glial transporters. Very little is known about EAAC1/EAAT3. However, because of somewhat high conservation in the molecular structure of the different proteins, it can be proposed that the neuronal transporter exhibits intracellular localizations of N- and C-terminals, as determined for the glial transporters. The protein shows eight to 10 transmembrane (TM) domains thought to be organized in α-helices. Data also show an amino acid identity of the different transporters of about 50%, especially in the carboxy-terminal part of the proteins (Grunewald et al. 2002). Regarding, more specifically, GLT1, topology analysis showed a very special arrangement of the protein, with two oppositely orientated re-entrant loops related to TM7 and TM8. Also, several amino acid residues have been shown to be critical for the function of the transporters either in TM7 or re-entrant loops I and II (Seal et al. 2000). The protein could be glycosylated in the large extracellular hydrophylic loop between TM3 and 4 (two sites) (Seal and Amara 1999) and shows putative consensus phosphorylation sites in its intracellular part, suggesting the possibility of post-traductional regulation of its activity. Furthermore, because of interactions of the transporter with proteins that contribute to its anchorage on the membrane and with proteins located in its immediate vicinity (Lin et al. 2001), there is a strong possibility that these partners contribute to the rapid trafficking of EAAC1/EAAT3 to and from the plasma membrane, as detected in various experimental models (Danbolt 2001).
Due to the somewhat high structural homology it is difficult to act selectively on the different transporter subtypes using specific pharmacological compounds (Shigeri et al. 2004). Some of these compounds, however, show a preference for inhibition of GLT1/EAAT2, but the selective study of the other subtypes, and especially of GLAST/EAAT1 and EAAC1/EAAT3, is difficult. Interestingly, the ability of l-beta-benzyl-aspartate (l-beta-BA) to preferentially inhibit the neuronal EAAC1/EAAT3 transporter was recently demonstrated (Esslinger et al. 2005). In this study, electrophysiological analysis showed that l-beta-BA acts as a non-substrate inhibitor. If verified, the use of such a selective compound, which has been described to have a 10-fold preference for EAAC1/EAAT3 compared with the glial transporters, could contribute substantially to further elucidation of the function of the neuronal transporter.
The role of EAAC1/EAAT3 has to be considered with regard to its early expression during brain development compared with that of the glial transporters, but also taking into account the fact that at the excitatory synapses, the contribution of astrocytes to Glu uptake is actually related to the presence of such glial cells. In this respect, for example at the cerebellar cortex level, the excitatory synapses corresponding to parallel and climbing fibres are surrounded by numerous glial cells of the Bergmann subtype, which supports a major role for the glial transporters in the removal of the neurotransmitter from the synapse. Conversely, as mentioned above, at the hippocampal and cerebral cortex levels the number of astrocytes surrounding the nerve terminals is limited, in this case further suggesting a more important contribution of the neuronal transporter to excitatory synapse physiology. In such a situation, the contribution of EAAC1/EAAT3 is questionable, but it is clear that it does not play a major role in clearing Glu from the extracellular space, as suggested from experiments using a down-expression of the transporter (Rothstein et al. 1996). Consequently, it is worth mentioning that all the Glu could not be taken up by the transporters and could possibly diffuse to reach, for example, the extra-synaptic NMDA receptors (Bergles et al. 1999) and NMDA receptors present at neighbouring synapses (Rusakov and Kullmann 1998). Such a mechanism could be highly relevant with regards to excitotoxicity as it has been further proposed that if NMDA receptor-mediated synaptic activity is neuroprotective, excessive extra-synaptic NMDA activity would presumably contribute to neuronal cell death (Hardingham et al. 2002). In such a case, lack of the correct EAAC1/EAAT3 activity could be deleterious for the neurones.
When expressed by post-synaptic GABA-containing neurones, EAAC1/EAAT3 may also help provide neurones with Glu used as an immediate precursor for GABA synthesis. In the case of expression by EAA-containing neurones, the transporters could provide the neurones with releasable Glu but, when located on the dendritic part of the neurones, their contribution could be to regulate putative release of the neurotransmitter from the dendrites and soma of the cells. Interestingly, the presence of the transporter on the membrane of cholinergic or monoaminergic neurones further emphasizes the possible contribution to neuroprotective mechanisms. In fact, because of high affinity for cysteine, EAAC1/EAAT3 could represent a means of providing cysteine to the neurones (Shanker et al. 2001), which contributes to the synthesis of glutathione, a major anti-oxidant agent (Dringen 2000). In agreement with such a proposal, it has been demonstrated from cortical neurones in culture that pharmacological inhibition of Glu uptake limits neuronal glutathione synthesis in the presence of cysteine in the culture medium (Chen and Swanson 2003; Himi et al. 2003). Because pharmacologically-induced blockade of glutathione biosynthesis can cause neurodegeneration (Jain et al. 1991; Schulz et al. 2000), alterations in neuronal cysteine uptake through EAAC1/EAAT3 could also contribute to neuronal death. Interestingly, another major source of glutathione is the cystine/glutamate transporter, also designated as system Xc–. Such a transporter in the adult was expressed not in neurones but mainly by glial cells (Sato et al. 2002). Because mice lacking one of the proteins contributing to the transport system (XCT–/– mice) show no apparent behavioural phenotype and no change in brain glutathione content (Sato et al. 2005), it is suggested that in such a situation the production of glutathione to assure basic anti-oxidant protection is, at least in part, under the control of EAAC1/EAAT3.
However, experimental blockade of EAAC1/EAAT3 expression does not induce neuronal degeneration, although dendritic swelling is associated with the loss of the transporter (Rothstein et al. 1996). Interestingly, although previous studies using EAAC1/EAAT3 gene knock-out (KO) mice have shown no neuronal degeneration (Peghini et al. 1997), a recent study reports neuronal glutathione deficiency and age-dependent neurodegeneration in a similar EAAC1/EAAT3-deficient mouse (Aoyama et al. 2006). Brain atrophy in the aged mice was correlated with an increased volume of cerebral ventricles, decreased thickness of the cerebral cortex and decreased number of axones in the corpus callosum, although the loss of neurones in the hippocampal CA1 field was limited. In such KO mice, these structural changes were shown to be prevented by administration of membrane-permeable N-acetylcysteine, further suggesting that EAAC1/EAAT3 is a major source of cysteine and contributes to the protection of neurones against oxidative stress.
Because of the electrogenicity of the transport, which can increase the excitability of the cell, and of the possible dependence of GABA synthesis on Glu neuronal uptake, alterations in EAAC1/EAAT3 expression, activity or regulatory processes may influence brain excitability locally and thus contribute to paroxistic activity, as detected in epilepsy. Indeed, the use of antisense oligonucleotides to block EAAC1/EAAT3 expression was shown in vivo to increase paroxistic activity in the adult rat (Rothstein et al. 1996), which could be related to deficient GABA inhibition. Although, in the EAAC1/EAAT3-deficient mice, relatively limited impairments in spontaneous motor activity probably associated with amino aciduria were initially found (Peghini et al. 1997), further suggesting a possible primary role for the transporter in metabolic regulation of GABA and cysteine biosynthesis, and possible compensatory processes related to the blockade of the transporter gene expression, conversely the recent study in the aged EAAC1/EAAT3-deficient mice (Aoyama et al. 2006) showed marked behavioural alterations. Increased aggressiveness and impaired self-grooming was reported in association with decreased performance in the Morris water maze test and reduced spontaneous locomotor activity. In this situation, these behavioural changes could be related to the structural changes described at the cortical and hippocampal level, reinforcing the idea of an essential role for the neuronal transporter in a protective action against oxidative stress-increased neurodegeneration associated with ageing.
However, regarding the very early expression of the neuronal transporter during development, the situation could be different from that seen in the adult. One interesting feature is the excitatory role of GABA played during the earlier developmental stages in both hippocampus and cerebral cortex, for example. This could be related to the early expression of the Cl– transporter protein, NKCC1, which contributes to the accumulation of Cl– in the neurones. Consequently, in numerous immature brain structures the activation of the GABAA receptors contributes to a net efflux of Cl– from the cell (Ben Ari 2002). Interestingly, at hippocampal level, for example, the maturation of GABA-containing interneurones precedes that of excitatory glutamatergic pyramidal neurones and astrocytes. Consequently, the activity-dependant plasticity contributing to the maturation of hippocampal networks is primarily related to the depolarizing action of GABA (Represa and Ben Ari 2005). In this respect, the nearly parallel expression of EAAC1/EAAT3 may contribute to the provision of the neurones with Glu involved in the essential excitatory action of GABA during the early developmental stages.
Regulation of Glu uptake involving EAAC1/EAAT3
In the early 1980s, we contributed to the development of the concept of possible short-term regulation of the rate of Glu uptake with regards to changes in neuronal activity. Using synaptosomal preparations, we demonstrated that in vivo electrical stimulation of the frontal cortex in the rat increased striatal Glu uptake (Nieoullon et al. 1983). Such a concept of activity-dependent regulation of the uptake of the neurotransmitter at the synapse was further extended to demonstrations that stress also modulates the rate of Glu uptake (Moghaddam 1993), and that modulation of the rate of uptake occurs during the oestrous cycle (Mitrovic et al. 1998). Numerous studies have later confirmed such a proposal (Danbolt 2001). In this respect, it is interesting to note the increase in hippocampal Glu uptake during the early stages of LTP that specifically concerns EAAC1/EAAT3, further emphasizing the role of the adaptive rate of the uptake process with regard to synaptic plasticity (Levenson et al. 2002). Regulation at the transcriptional level is suggested by experiments in rodents which showed that increasing locomotor activity was associated to a concomitant increase in EAAC1/EAAT3 gene expression at hippocampal level (Molteni et al. 2002).
Our group also focused on another concept related to post-traductional short-term regulation of the Glu uptake process. We suggested that Glu uptake might be modulated in response to activation of neurotransmitter receptors, further emphasizing the possibility of a post-traductional regulation of the transport as a mechanism functionally associated with neuronal activity and release of the neurotransmitter. For example, we demonstrated that activation of dopaminergic receptors in the rat striatum, which was previously shown to reduce Glu release, was also able to selectively inhibit Glu uptake (Kerkerian et al. 1987). Such a concept was further highlighted by the demonstration that EAAT have putative consensus phosphorylation sites in their molecular structures. Consequently, a series of pharmacological experiments has been conducted, both in vivo and in vitro, aimed at pharmacologically promoting or blocking the activity of various protein kinases. Activation of protein kinase C (PKC) was initially shown to stimulate Glu uptake from brain homogenates (Casado et al. 1993; Dowd and Robinson 1996) and was recently reported to selectively increase EAAC1/EAAT3 activity, whereas it contributed to reduce GLT1/EAAT2 function (Gonzales and Robinson 2004). We further showed the possible contribution of PKA (Pisano et al. 1996; Lortet et al. 1999) and more recently, of the PI3 kinase (PI3K). In our laboratory, in primary cultures from rat cerebral cortex when GLT1/EAAT2 is not yet expressed, pharmacological inhibition of PKC and PKA reduced apparent Glu uptake. In such a developmental model, phosphatidylinositol-3-kinase (PI3K) inhibition was only active to reduce the uptake process at 14 DIV when GLT1/EAAT2 is fully active (Guillet et al. 2005). The situation, however, is not clear enough and the mechanisms by which the kinases influence transport have yet to be clarified. In particular, it remains to be established whether the direct phosphorylation of the transporter protein alters the transport or, alternatively, whether the change in transport rate is due to changes in protein–protein interactions because, as above mentioned, transport implicates a complex association of protein partners at the plasma membrane that also display phosphorylable elements.
These studies emphasized the possibility of an influence of protein kinase regulation on the trafficking of the transporters, possibly through the activation of membrane receptors such as for neurotensin (Najimi et al. 2002). Interestingly, the activation of PKC, which was shown to increase Glu transport (see above), was demonstrated in hippocampal neurones to promote the association of EAAC1/EAAT3 with the αPKC isoform (Gonzales et al. 2002, 2003; Fournier et al. 2004). Such an interaction was proposed to contribute to the stabilization of the transporter at the plasma membrane from intracellular stores (Davis et al. 1998), thereby increasing the transport rate of the neurotransmitter. In canine kidney cell lines where the distribution of the transporter is polarized, as well as on the dendritic tree of hippocampal cells, EAAC1/EAAT3 was indeed associated with a sorting motif on the C-terminal of the transporter protein (Cheng et al. 2002), which could also contribute to its stabilization at the plasma membrane. The C-terminal part of EAAC1/EAAT3 protein could also be involved in the regulation of transport rate as it represents a key site for interaction with regulatory cytosolic proteins, such as GTRAP3-18, which was shown to reduce Glu transport when associated with the transporter. Such a mechanism would help decrease affinity of the transporter for Glu without any change in the trafficking of the transporter protein (Lin et al. 2001; Butchbach et al. 2002). GTRAP3-18 would help modulate EAAC1/EAAT3 activity by promoting a partial deglycosylation of the transporter (Ruggiero et al. 2003). Moreover, possible interactions with pre-synaptic proteins that might stabilize the transporter at the membrane are not excluded as in transfected oocytes, syntaxin 1A was also shown to regulate EAAC1/EAAT3 expression and activity (Zhu et al. 2005).
Interestingly, the interaction of the transporter with αPKC points to lipid raft implication as the kinase was shown to be primarily localized in such a membrane compartment (Mineo et al. 1998). Thus, αPKC appears to be in a position to trigger the redistribution of EAAC1/EAAT3 to the plasma membrane. Moreover, the association of the transporter with αPKC suggests possible translocation of EAAC1/EAAT3 in intracellular endosomes or even in caveola, which could represent a membrane site essential for modulation of signal transduction at the cell surface, although the presence of caveola in neurones is still under discussion (Mineo et al. 1998). In a recent study, we further analysed the influence of changing protein kinase activity on the cell surface expression of the Glu transporters in neurone-enriched cultures in vitro. Biotinylation and immunoblotting experiments showed that EAAC1/EAAT3 membrane expression was significantly reduced by H89 and wortmannin, which are considered, respectively, as inhibitors of PKA and PI3K. Conversely, phorbol 12-myristate 13 acetate (PMA), a powerful activator of PKC, was shown to increase the presence of the transporter at the plasma membrane (Guillet et al. 2005) as it activates Glu transport. Confocal microscopy analysis further revealed a wortmannin-induced clustering of EAAC1/EAAT3 in the intracellular compartment. Together with changes affecting the transporters GLT1/EAAT2 and GLAST/EAAT1 under similar experimental conditions, these data confirm that trafficking of Glu transporters can be differentially regulated by intracellular signalling pathways and could therefore regulate the uptake process. In line with these data, the rate of trafficking of the protein transport between the membrane compartments was recently estimated and the half-life of EAAC1/EAAT3 at the plasma membrane proposed to be around 5–7 min (Fournier et al. 2004). Furthermore, the implication of the PI3K signalling pathway in cell trafficking of EAAC1/EAAT3 was recently confirmed by the fact that activation of Akt, which represents a downstream target of PI3K, increased the surface expression of the endogenous transporter in transfected cells and, further, Glu transport (Krizman-Genda et al. 2005). These regulatory processes, which may involve some other post-traductionnal processing such as palmitoylation or glycosylation/deglycosylation processes (Huang and El-Husseini 2005), may represent rapid adaptive responses to changes in the cellular environment, for example to prevent possible excitotoxic events (Robinson 2002).
Transporter expression and activity were also shown to be modulated by soluble factors acting in the environment of the synapses, suggesting a close collaboration between cells in synapse functioning. A lot of information has been obtained from cultured astrocytes showing, for example, that the uptake of Glu and GABA is increased when the astrocytes are cultured in the presence of conditioned media from neuronal cultures (Danbolt 2001). Some data have shown that soluble factors secreted by neurones, such as pituitary adenylate cyclase activating polypeptide (PACAP), may contribute to this effect through PK signalling pathways (Harmar et al. 1998), but the nature of the factors involved in such an effect is not known, constituting a new field of active investigation to further our understanding of the complex relationships existing between astrocytes and neurones (Gegelashvili et al. 2000). Conversely, astrocytes have been shown to increase the number of synapses in culture and are required for their maintenance (Pfriger and Barres 1997; Ullian et al. 2001). Platelet-derived growth factor (PDGF) was also shown to increase cell surface expression of EAAC1/EAAT3 in C6 glioma cells through activation of PI3K (Sims et al. 2000). We investigated this further in cortical primary cultures almost exclusively containing neurones without astrocytes (percentage of astrocytes less then 0.5%). These neuronal cultures were exposed to a glia-conditioned medium (GCM) corresponding to the medium without any cells used for culturing the astrocytes. Results showed an increase in Glu uptake in neurones exposed to GCM compared with control (Canolle et al. 2004). Such an increase in the transport of Glu was concomitant with a massive decrease in extracellular Glu concentration and with an increase in EAAC1/EAAT3 cell expression.
In order to identify possible factors secreted by astrocytes that might influence EAAC1/EAAT3 expression/activity in neurones, we tested the hypothesis of a putative role of cholesterol, as it has been previously shown that such a factor influences synapse development (Mauch et al. 2001). Using our experimental model based on the quasi exclusive presence of neurones in the cortical cultures without any astrocytes, we were able to show, in the presence of cholesterol, a dose-dependent increase in the transport of Glu. Interestingly, the exposure of neurones to a GCM obtained from astrocytes treated with mevastatin, an inhibitor of cholesterol synthesis (so-called ‘low cholesterol-GCM’), did not induce a change in Glu transport (Canolle et al. 2004). Thus, these results suggested that cholesterol secreted by the astrocytes might regulate the activity of neuronal Glu transport. Because cholesterol is also involved in the regulation of lipid rafts (Hering et al. 2003), we used methyl-β-cyclodextrine (mBCD) to disorganize the lipid rafts and limit the action of cholesterol. Data showed that in the presence of mBCD, the effects of GCM on Glu transport were considerably reduced, further suggesting that EAAC1/EAAT3 is associated with lipid rafts. We concluded that soluble factors secreted from astrocytes, such as cholesterol, could fine tune Glu transport into the neurones in the same way as neuronal factors could contribute to the promotion of glial Glu uptake. This effect on the neurones could involve lipid rafts.
Numerous observations, primarily from animal models, have implicated alteration in Glu transport in various brain diseases (Maragakis and Rothstein 2004). For example, changes in EAAC1/EAAT3 have been reported to correlate with experimental models of epilepsy (Crino et al. 2002; Sepkuty et al. 2002; Furuta et al. 2003). Interestingly, transiently increasing synaptic activity during infancy has been shown to induce permanent changes that could, in response, increase the risk of epilepsy later in life. Data from experiments in rats showed that pharmacological induction of a status epilepticus at post-natal day 10 increased seizure occurrence in the adult concomitant with an increase in glutamate receptor 2 mRNA expression and EAAC1/EAAT3 protein levels in the dentate gyrus (Zhang et al. 2004). This further illustrates either the contribution of the transporter to adaptive changes to limit seizure susceptibility, or its direct implication in the process. In the adult, however, down-expression of the EAAC1/EAAT3 transporter induced in the rat using intracerebral administration of antisense nucleotides resulted in the occurrence of seizures (Rothstein et al. 1996).
Experimental anoxia also induced in vivo alterations in EAAC1/EAAT3 expression (Martin et al. 1997; Rao et al. 2001). Transient ischaemia experiments provided further evidence for a contribution of the neuronal transporter to an adaptive response to increased Glu extracellular concentrations; the expression of EAAC1/EAAT3 was increased 8 h and 1 day after ischaemia in CA1 pyramidal neurones and pyramidal neurones of the cerebral cortex, respectively (Gottlieb et al. 2000). In this case, it is worth noting that immunohistochemical labelling of the EAAC1/EAAT3 transporter protein was increased even 28 days after ischaemia in a subset of oligodendroglial progenitor cells in subcortical white matter. An active role of the neuronal Glu transporter in the production of anoxic depolarization and related neuronal damage following increases in extracellular Glu concentrations was also demonstrated. Interestingly, in hippocampal slices from EAAC1/EAAT3 KO mice submitted in vitro to experimental hypoxia, the latency of anoxic depolarization was enhanced compared with the control. Because, in such a preparation, the pharmacological blockade of glial Glu transport reduced the onset of anoxic responses, it has been suggested that reversed transport of Glu from neurones is an essential step in anoxic depolarization production (Gebhardt et al. 2002). Consequently, the EAAC1/EAAT3 transporter could be actively involved in brain damage associated with hypoxia–anoxia. Laboratory experiments in progress are investigating such a hypothesis further using primary cortical cultures subjected to controlled oxygen–glucose deprivations.
Finally, in an animal model of Alzheimer's disease with mutant amyloid precursor protein (APP) overexpression, a decrease in the protein EAAC1/EAAT3 was observed (Masliah et al. 2000). Thus, the neuronal transporter may play a role in the disease but its contribution in humans remains to be confirmed. Preseniline 1, which contributes to physiological processing of APP, was recently shown to modulate plasma membrane expression and activity of EAAC1/EAAT3 (Yang et al. 2004), whereas β-amyloid peptide decreased Glu transport from ovocytes transfected with the neuronal transporter (Gu et al. 2004). In this respect, a better knowledge of the mechanisms and functions of EAA transport in neurones will contribute to the development of new therapeutic approaches of the diseases.