Changes in dialysate concentrations of glutamate and GABA in the brain: an index of volume transmission mediated actions?

Authors


Address correspondence and reprint requests to Dr Francisco Mora, Department of Physiology, Faculty of Medicine, Universidad Complutense, Ciudad Universitaria, s/n, 28040 Madrid, Spain. E-mail: francisco-mora@med.ucm.es

Abstract

Brain microdialysis has become a frequently used method to study the extracellular concentrations of neurotransmitters in specific areas of the brain. For years, and this is still the case today, dialysate concentrations and hence extracellular concentrations of neurotransmitters have been interpreted as a direct index of the neuronal release of these specific neurotransmitter systems. Although this seems to be the case for neurotransmitters such as dopamine, serotonin and acetylcholine, the extracellular concentrations of glutamate and GABA do not provide a reliable index of their synaptic exocytotic release. However, many microdialysis studies show changes in extracellular concentrations of glutamate and GABA under specific pharmacological and behavioural stimuli that could be interpreted as a consequence of the activation of specific neurochemical circuits. Despite this, we still do not know the origin and physiological significance of these changes of glutamate and GABA in the extracellular space. Here we propose that the changes in dialysate concentrations of these two neurotransmitters found under specific treatments could be an expression of the activity of the neurone–astrocyte unit in specific circuits of the brain. It is further proposed that dialysate changes of glutamate and GABA could be used as an index of volume transmission mediated actions of these two neurotransmitters in the brain. This hypothesis is based firstly on the assumption that the activity of neurones is functionally linked to the activity of astrocytes, which can release glutamate and GABA to the extracellular space; secondly, on the existence of extrasynaptic glutamate and GABA receptors with functional properties different from those of GABA receptors located at the synapse; and thirdly, on the experimental evidence reporting specific electrophysiological and neurochemical effects of glutamate and GABA when their levels are increased in the extracellular space. According to this concept, glutamate and GABA, once released into the extracellular compartment, could diffuse and have long-lasting effects modulating glutamatergic and/or GABAergic neurone-astrocytic networks and their interactions with other neurotransmitter neurone networks in the same areas of the brain.

Brain microdialysis has become a frequently used method to study the extracellular concentrations of neurotransmitters in specific areas of the brain (Robinson and Justice 1991). In particular, it has been used to investigate glutamatergic, GABAergic and dopaminergic systems regarding their involvement in neurological diseases such as Parkinson's disease and schizophrenia, stress and also in ageing (Carlsson and Carlsson 1990; Goldman-Rakic and Selemon 1997; Segovia et al. 1999, 2001; Del Arco and Mora 2000, 2001, 2002; Del Arco et al. 2001; Mora et al. 2002). For years, and this is still the case today, dialysate concentrations and hence extracellular concentrations of glutamate and GABA have been interpreted by many as a direct index of the synaptic activity of these neurotransmitters. But this simple picture has been challenged as growing evidence suggests that, in contrast to dopamine, dialysate concentrations of glutamate and GABA are not a reliable index of their neuronal release.

The main concern in this respect is the cellular origin of the extracellular concentrations of neurotransmitters measured by means of microdialysis. Dopamine exists only as a neurotransmitter and is mainly synthesized and released from neurones, and dialysate dopamine is therefore considered to originate from neurones and to be an expression of neuronal release. In contrast, glutamate and GABA are found in neurones but also in very important quantities in glial cells (astrocytes). Moreover, glutamate and GABA also participate in metabolic routes. Thus, dialysate glutamate and GABA can arise from neurones as well as from non-neuronal pools. For that reason, and despite several microdialysis studies devoted to establish the origin of their extracellular concentrations (Herrera-Marschitz et al. 1996; Timmerman and Westerink 1997a,b), the meaning of changes in dialysate glutamate and GABA is still a matter of controversy (Timmerman and Westerink 1997a).

Here we propose that, firstly, the changes of extracellular (dialysate) concentrations of glutamate and GABA found under specific treatments could be an expression of the activity of the neurone–astrocyte unit in specific circuits of the brain, and that, secondly, dialysate changes of glutamate and GABA could be an index of volume transmission mediated actions of these neurotransmitters in the brain rather than of their synaptic release. According to the definition of volume transmission by Fuxe and Agnati (1991; see also Agnati et al. 2000), this hypothesis would indicate that the amino acids glutamate and GABA, once released from astrocytes into the extracellular compartment, could diffuse relatively long distances to have specific effects on target cells (this is also referred to as extracellular transmission) in addition to being released from distant astrocytes via intracellular astrocytic calcium waves (Haydon 2001). This proposal is based on the following evidence (see Fig. 1).

Figure 1.

Scheme depicting the hypothesis proposed in the text in which the changes of extracellular concentrations of glutamate (GLU), and probably GABA, monitored by microdialysis are an index of the role these neurotransmitters play as volume transmission signals in the brain. Here only glutamate mechanisms are illustrated. Astrocytes could be specifically involved in modulating extracellular concentrations of these amino acids under specific pharmacological treatments or behavioural stimuli. On one hand synaptic glutamate could act on extrasynaptic glutamatergic receptors located on neurones and astrocytes. In the case of astrocytes it is illustrated that the increase of calcium (Ca2+) levels induces astrocytic glutamate release. On the other hand, other neurotransmitters such as dopamine (DA), noradrenaline (NA) or acetylcholine (Ach) could activate non-glutamatergic receptors located on astrocytes and in turn induce (i.e. via intracellular pathways) the release of astrocytic glutamate to the extracellular space where again calcium can play a central role. Glutamate once increased into the extracellular space ([GLU]ex) would diffuse through the extracellular space to reach the microdialysis probe; [GLU]ex would also activate glutamatergic extrasynaptic receptors to modulate the activity of glial-neuronal assemblies around the microdialysis probe. Astrocytic glutamate and GABA, as well as neuronal amines released from non-synaptic sites, would create an extracellular microenvironment that would modulate the activity of neighbouring neurone–astrocyte assemblies via volume transmission. The effect produced by one particular neurotransmitter would depend on other neurotransmitters existing in the extracellular microenvironment through integration of these signals in astrocytes and neurones. Calcium waves in astrocytic networks are not indicated (see text).

  • (i) Functional and structural data suggest that glutamate and GABA do not diffuse from the synaptic cleft to reach the extracellular space in significant quantities and are therefore not monitored by the microdialysis probe.
  • (ii) Astrocytes, functionally linked to neuronal activity, are able to release glutamate and GABA to the extracellular space.
  • (iii) Pharmacological treatments and behavioural stimuli involving the activation of specific circuits of the brain produce changes in dialysate glutamate or GABA.
  • (iv) Recent findings have shown the existence of glutamate and GABA extrasynaptic receptors, which would have functional properties different from those of synaptic receptors.
  • (v) Extracellular glutamate and/or GABA have been shown to produce specific electrophysiological or neurochemical effects in some areas of the brain, which fits well with a volume transmission mediated action of these amino acids.

Based on all these previous findings, glutamate and GABA levels in the extracellular space might change independently of glutamate and GABA levels within the synaptic cleft to take part in modulating neuronal circuits in specific brain areas.

On the origin of extracellular glutamate, GABA and dopamine levels as studied with microdialysis: A –synapses vs. non-junctional varicosities; B – neuronal vs. glial sources

Two classical criteria are often applied to investigate the synaptic vs. the non-synaptic origin of dialysate concentrations of neurotransmitters: involvement of nerve impulse and calcium-dependency (Timmerman and Westerink 1997a). It has been extensively shown that the perfusion of TTX reduces by 70–80% the basal dialysate concentration of dopamine, and basal dopamine shows a clear calcium-dependence (Westerink et al. 1987; Morari et al. 1993; Del Arco and Mora 2001). In contrast to dopamine, most of the studies have reported a very low sensitivity or insensitivity to TTX and low dependence or independence on calcium of the basal dialysate levels of glutamate and GABA (Moghaddam 1993; Morari et al. 1993; Timmerman et al. 1999; Del Arco and Mora 2002). From these results it is interpreted that basal dialysate concentrations of dopamine, but not those of glutamate and GABA, originate from neuronal terminals and reflect neuronal release (mainly from non-junctional varicosities operating via volume transmission; Zoli et al. 1999) and in turn the activity of the dopaminergic system. Other origins, different from neuronal, have been suggested for basal extracellular concentrations of glutamate and GABA (Timmerman and Westerink 1997a; Jabaudon et al. 1999). In particular, very recent evidence suggests that the cystine/glutamate exchanger located in astrocytes may be responsible for most of the basal glutamate measured with microdialysis (Jabaudon et al. 1999; Baker et al. 2001).

The data discussed above (TTX- and calcium-dependence) fit well with the neuroanatomical evidence showing that the neuronal compartment of dopamine, but not that of glutamate and GABA, is linked to the extracellular space. In fact, the synaptic arrangement (open vs. closed synapses) of these neurotransmitters is clearly different (Fuxe and Agnati 1991; Zoli and Agnati 1996). In particular, the existence of non-junctional dopamine varicosities, which release dopamine far from postsynaptic sites, and the low density of dopamine transporters located close to the synaptic cleft (Zoli and Agnati 1996; Jansson et al. 2002) allow the diffusion of synaptic dopamine through the extracellular space to be sampled by the dialysis probe. In contrast, a significant diffusion of glutamate or GABA from the synaptic cleft to the extracellular space seems unlikely because the high density of glutamate transporters located on astrocytic processes forming a sheath around synapses makes that possibility unlikely (Pfrieger and Barres 1996; Diamond and Jahr 2000). Thus, glutamatergic and GABAergic neuronal systems seem not to be designed to allow glutamate or GABA diffusion from synapses, at least not far enough to reach the extracellular space and to be detected by the microdialysis probe. This last assertion is strongly supported by studies showing that specific stimulation of presynapses, though inducing postsynaptic excitatory potentials, did not increase dialysate concentrations of glutamate (Segovia et al. 1997b; Obrenovitch et al. 2000). Moreover, in vivo induction of long-term potentiation in the dentate gyrus of the hippocampus by tetanic stimulation of the perforant pathway was not accompanied by changes in extracellular glutamate measured with microdialysis (Jay et al. 1999).

Interestingly, there are some microdialysis studies showing increases of extracellular concentrations of glutamate and GABA produced by specific drugs, which can be prevented by TTX (neuronal activity) and are dependent on calcium in the perfusion medium (Osborne et al. 1990; Moghaddam 1993; Grobin and Deutch 1998; Antonelli et al. 2000), although basal concentrations are not. Moreover, some microdialysis studies have shown that the electrical or chemical stimulation of specific glutamatergic or GABAergic pathways produces increases of dialysate glutamate or GABA in the terminal areas that are blocked by TTX (Bianchi et al. 1996; Timmerman and Westerink 1997b; Rossetti et al. 1998). These data have been interpreted as a neuronal release of glutamate and/or GABA under conditions of high stimulation, in which glutamate and/or GABA could diffuse from the synaptic cleft and reach the microdialysis probe. These data could fit well with studies showing that glutamate and GABA can spill-over from the synaptic cleft to perisynaptic or extrasynaptic regions (Isaacson et al. 1993; Scanziani et al. 1997; Rusakov and Kullmann 1998; Bergles et al. 1999; Scanziani 2000). In this respect recent evidence indicates that glutamate and/or GABA spill-over can be specifically regulated either by changing the activity of high-affinity glutamate and GABA transporters around the synaptic cleft (Isaacson et al. 1993; Scanziani et al. 1997; Rusakov and Kullmann 1998; Bergles et al. 1999) or by modifying the astrocytic coverage of neurones (Oliet et al. 2001), suggesting that in particular physiological circumstances more glutamate and/or GABA could reach the extracellular space and activate extrasynaptic receptors (see Scanziani 2002). However it is still unclear whether the amount of glutamate or GABA diffusing from the synaptic cleft to perisynaptic regions can reach the microdialysis probe. Thus, and despite the spill-over of neuronal glutamate and GABA that may contribute to dialysate concentrations of these amino acids, it seems unlikely that it accounts for the main source of the glutamate and GABA measured by microdialysis according to the evidence discussed in the previous paragraph (see also below).

Alternatively to spill-over, recent findings about the functional interaction between neurones and astrocytes give rise to an exciting possibility to explain the specific increases of extracellular concentrations of glutamate and GABA obtained in the above-mentioned microdialysis studies. Thus, it has been shown that astrocytes are able to release glutamate, and also GABA, into the extracellular space (Attwell et al. 1993; Araque et al. 1999; Carmignoto 2000; Liu et al. 2000). Furthermore, it has been proposed that the astrocytic glutamate release can be produced through glutamate receptors by previous activation of adjacent neurones (Araque et al. 1999), and that once released this astrocytic glutamate can act on neuronal extrasynaptic receptors to modulate neurotransmission (Araque et al. 1999; Carmignoto 2000; Haydon 2001). These data suggest that neuronal exocytotic glutamate release after electrical stimulation may induce an astrocytic glutamate release into the extracellular space (see Fig. 1), which therefore would be TTX-sensitive and calcium dependent. In addition, it has been shown that the release of glutamate from astrocytes is preceded by a rise of intracellular calcium that can be propagated throughout the astrocytic network via gap-junctions making up calcium waves (Carmignoto 2000; Haydon 2001). This suggests that glutamate can also be released by astrocytes located far away from the initial point of stimulation (Araque et al. 2000; Bezzi and Volterra 2001; Haydon 2001). As a consequence, the astrocytic release of glutamate can be amplified, and more glutamate accumulated in the extracellular compartment, through the neurone–astrocytic network in a reaction-diffusion process (Carmignoto 2000). We suggest that most of the glutamate increase in the extracellular space that is monitored with microdialysis could arise from astrocytes through this mechanism. The same origin is also suggested for dialysate GABA, but unlike astrocytic glutamate, the functional neurone–astrocytic interaction regulating the release of astrocytic GABA into the extracellular space has not been investigated in detail and this suggestion should therefore be interpreted cautiously. Interestingly, there are some microdialysis studies suggesting the involvement of astrocytes in the changes in extracellular glutamate and GABA (Honma et al. 1996; Miele et al. 1996; Pierce et al. 1996; Márquez de Prado et al. 2000). For instance, Pierce et al. (1996) have suggested that the ability of cocaine treatment to increase dialysate concentrations of glutamate in the nucleus accumbens are mediated by astrocytes. Likewise, astrocytes have been involved in the changes in extracellular levels of glutamate and/or GABA induced by circadian rhythms in the suprachiasmatic nucleus (Honma et al. 1996) and striatum (Márquez de Prado et al. 2000).

Astrocytes can also be directly activated by a multitude of other neurotransmitters and neuromodulators in addition to glutamate (Carmignoto 2000; Parpura and Haydon 2000; Bezzi and Volterra 2001). In fact, neurotransmitters such as dopamine, noradrenaline or GABA can induce increases of intracellular calcium in astrocytes (Parpura and Haydon 2000; Araque et al. 2000). These data open up the possibility that other neurotransmitters can induce glutamate or GABA release from astrocytes into the extracellular space. The mechanism by which astrocytes release glutamate or GABA into the extracellular space remains to be elucidated. Among other mechanisms, exocytosis, the reversion of the high-affinity glutamate transporter and the involvement of glutamate exchangers (i.e. cystine/glutamate exchanger) have been suggested (Attwell et al. 1993; Araque et al. 2000; Kalivas and Baker 2001).

Changes in dialysate glutamate and GABA: an index of volume transmission mediated actions

As suggested in the last section, changes in the activity of neurone–astrocytic units might parallel specific and physiologically relevant changes in the dialysate concentrations of glutamate and GABA. Data reported in Tables 1 and 2 show some of the most relevant studies related to the monitoring of glutamate or GABA by microdialysis under specific pharmacological treatments and behavioural stimuli. Indeed, the studies shown in Tables 1 and 2 were selected according to the experimental protocols in which extracellular changes of glutamate and GABA could presumably reflect the activation of specific circuits of the brain.

Table 1.  Summary of some studies showing the effects of different drug treatments (intraperitoneal injections) on dialysate concentrations of glutamate (GLU) and GABA in different areas of the rat brain
Drug treatmentDialysate GLU/GABABrain areaReference
  1. VTA, ventrotegmental area.

Amphetamines↑ GLUStriatumNash and Yamamoto 1993
N. AccumbensXue et al. 1996Reid et al. 1997
VTAXue et al. 1996
 ↓ GABAN. AccumbensHurd et al. 1992
Ventral PallidumBourdelais and Kalivas 1990
Cocaine↑ GLUN. AccumbensReid et al. 1997
Prefrontal cortexReid et al. 1997
Clozapine↑ GLUPrefrontal cortexDaly and Moghaddam 1993
↓ GABAPrefrontal cortexBourdelais and Deutch 1994
Haloperidol↓ GABAPrefrontal cortexBourdelais and Deutch 1994
↑ GABAGlobus PallidusOsborne et al. 1990
Apomorphine↑ GABAPrefrontal cortexGrobin and Deutch 1998
Ketamine and phencyclidine↑ GLUPrefrontal cortexMoghaddam et al. 1997
N. AccumbensMoghaddam and Adams 1998
Nicotine↑ GLUHippocampusFedele et al. 1998
Ethanol↓ GLUN. AccumbensYang et al. 1998
Lithium↓↑ GLUPrefrontal cortexAntonelli et al. 2000
↑ GABAPrefrontal cortexAntonelli et al. 2000
Glucocorticoids↑ GLUHippocampusVenero and Borrell 1999
Modafinil↑ GLU–↓ GABAStriatumFerraro et al. 1998
↓ GABAVentral PallidumFerraro et al. 1998
Table 2.  Summary of some studies showing effects of different behavioural stimuli on dialysate concentrations of glutamate (GLU) and GABA in different areas of the rat brain
Behavioural stimuliDialysate GLU/GABABrain areaReference
  1. VTA, ventrotegmental area.

Novelty↑ GLUStriatumShinohara et al. 2000
Induced grooming↑ GLUStriatumMiele et al. 1996
Feeding↑ GLUGlobus PallidusPerez et al. 2000
Handling-stress↑ GLUVTATimmerman et al. 1999
Locus Coeruleus 
Restraint-stress↑ GLUPrefrontal cortexMoghaddam 1993
Striatum 
Hippocampus 
N. Accumbens 
Swimming-stress↑ GLUStriatumMoghaddam 1993
N. Accumbens 
Tail-shock stress↑ GLU–↑GABAStriatumKeefe et al. 1993
Immobilization-stress↑ GLUHippocampusLowy et al. 1993
Conditioned emotional response↑ GLU–↑ GABAN. AccumbensSaulskaya and Marsden 1995
Circadian periodicity↑ GLU–↑ GABA (dark) ↓ GLU–↓ GABA (light)StriatumMárquez de Prado et al. 2000
Conditioned stimuli↑ GLUN. AccumbensHotsenpiller et al. 2001
Conditioned taste aversion↑ GLUAmygdalaTucci et al. 1998

In addition to glutamate and GABA, dialysate concentrations of other amino acids such as arginine and serine, or metabolites such as lactate or glucose have been shown to change under physiological circumstances (Fellows et al. 1992; Demestre et al. 1997; Timmerman and Westerink 1997a). Although these findings have argued against the physiological specificity of dialysate changes of glutamate and GABA (Timmerman and Westerink 1997a; Obrenovitch 1999), they are actually in agreement with the possibility of astrocyte-released signals having a role in modulating neurotransmission. For instance, the amino acid d-serine, specifically released from astrocytes, has been suggested to play a role in neurotransmission as an endogenous modulator of NMDA receptors via the glycine binding site (Snyder and Kim 2000). Similarly, lactate originating from astrocytes and reaching the extracellular space can be taken up by neurones and modulate neuronal activity and metabolism (Magistretti and Pellerin 2000). However, a major difference arises from the comparison between these metabolites or amino acids and the neurotransmitters glutamate or GABA: the latter, once released from astrocytes into the extracellular space may act on glutamatergic or GABAergic receptors as full agonists having then specific receptor-mediated functions. In this respect in vivo electrophysiological studies showing specific effects on neuronal excitability with exogenous glutamate and GABA applied by iontophoresis would be in agreement with this assumption (Hu and White 1997; Kiyatkin and Rebec 1999). In fact reports show that glutamate, GABA and also dopamine can interact reciprocally to modulate the activity of neurones when these neurotransmitters are simultaneously increased in the extracellular space (Hu and White 1997; Kiyatkin and Rebec 1999).

We suggest that glutamate and GABA monitored in microdialysis studies could be acting as volume transmission signals. This would mean that glutamate and GABA, once released from astrocytes into the extracellular space, would diffuse from their sites of release in order to modulate, through glutamate and GABA receptors, the activity of neuronal–glial assemblies around the microdialysis probe. The above iontophoretic-electrophysiological studies would actually fit well with this hypothesis because exogenously applied glutamate and GABA have to diffuse from the pipette through the extracellular space to reach their specific receptors and modify the neuronal activity in the area of the brain studied. Moreover, astrocytes have been previously suggested to be a major source of volume transmission signals (Fuxe and Agnati 1991; Zoli et al. 1999; Syková and Chvátal 2000). It is also suggested that dialysate glutamate and GABA would preferentially activate extrasynaptic receptors in order to play a role in neurotransmission. Thus, when glutamate concentrations in the synaptic cleft reach 1–3 mm (Clements 1996), extracellular glutamate monitored in microdialysis studies (in the µm range) would not be able to enter the synaptic cleft against the gradient of concentration and reach receptors located inside the glutamatergic synapse. Moreover the transporters located around the synaptic arrangement are very active taking up glutamate (Pfrieger and Barres 1996; Jabaudon et al. 1999). Therefore, glutamate and GABA released from astrocytes into the extracellular space would diffuse and reach extrasynaptic receptors, thereby modulating the activity of neurone–astrocyte units in specific circuits of the brain.

Increasing evidence supports the existence of glutamate and GABA receptors located extrasynaptically (Isaacson et al. 1993; Scanziani et al. 1997; Nusser et al. 1998; Rusakov and Kullmann 1998; Banks and Pearce 2000; Sattler et al. 2000). These extrasynaptic receptors can be located perisynaptically on their own glutamate/GABA synapses (autoreceptors), on astrocytes and on other neurotransmitter systems terminals such as dopaminergic terminals (heteroreceptors). Extrasynaptic high-affinity glutamate and GABA metabotropic receptors seem to be very relevant in sensing the actions mediated by the diffusion of glutamate and GABA. In fact, there are several studies showing that glutamate and GABA do modulate neurotransmission of each other through metabotropic extrasynaptic receptors (Isaacson et al. 1993; Scanziani 2000; Semyanov and Kullmann 2000), although ionotropic glutamate and GABA receptors are also involved in this kind of extrasynaptic interaction (Nusser et al. 1998; Rusakov and Kullmann 1998; Sattler et al. 2000).

Like extrasynaptic receptors, glutamate and GABA heterotransporters can also account for the modulatory effects mediated by glutamate and GABA extracellular diffusion. In fact, it has been shown that glutamate and GABA, taken up by their respective high-affinity specific transporters, can stimulate the release of neurotransmitters such as dopamine, noradrenaline or acetylcholine as well as glutamate and GABA (Bonanno and Raiteri 1994; Vizi 2000). This effect seems to be mediated by the increase of intracellular Na+ that is cotransported together with glutamate and GABA (Vizi and Kiss 2000). Therefore, these high-affinity transporters may play an important role in regulating volume transmission actions mediated by extracellular glutamate and GABA. Firstly, they are indeed involved in the regulation of glutamate and GABA extracellular levels and secondly, this regulation may implicate a signalling pathway through which extracellular glutamate and GABA can modulate specifically the releasing activity of neurotransmitter systems.

Recent findings from our laboratory support the possibility of extracellular glutamate modulating neurotransmission through extrasynaptic receptors. In a series of microdialysis studies we have investigated the effects of increasing extracellular concentrations of endogenous glutamate on the concentration of dopamine in the striatum and nucleus accumbens of the conscious rat (Segovia et al. 1997a, 1999, 2001). The uptake of glutamate was blocked by the perfusion of the specific and potent inhibitor l-trans-2,4-pyrrolidine dicarboxilic acid (PDC) through the microdialysis probe. Perfusion of PDC increased the extracellular concentration of dopamine, and these increases were significantly correlated with the increases of extracellular glutamate (see Fig. 2). Moreover, the increases in dopamine could be blocked by specific glutamatergic ionotropic receptor antagonists. The striatum and nucleus accumbens are regions of convergence of forebrain glutamatergic and midbrain dopaminergic pathways (Smith and Bolam 1990). However, synaptic contacts have not been described between glutamatergic and dopaminergic terminals in these areas (Bouyer et al. 1984; Sesack and Pickel 1992). Thus, the stimulating effects of extracellular glutamate on dopamine release seem to be mediated by extrasynaptic receptors located on dopaminergic terminals. Recently, it has been shown that the injection of PDC into the nucleus accumbens increases locomotor activity (Kim and Vezina 1999). Moreover, local administration of glutamatergic agonists into the nucleus accumbens also produce behavioural activation, which seems to be mediated by dopamine as this effect can be blocked by dopamine receptor antagonists (Kim and Vezina 1997). These results suggest that the extrasynaptic interactions between glutamate and dopamine shown in our studies may be physiologically relevant and could be dominant over synaptic-mediated actions after PDC perfusion.

Figure 2.

Endogenous glutamate monitored with the microdialysis technique increase extracellular concentrations of dopamine in the nucleus accumbens of the conscious rat (adapted from Segovia et al. 1999). See text for further comments. The figure represents the correlation between increasing concentrations of dialysate glutamate (GLU) (obtained by local perfusion of the glutamate uptake blocker l-trans-pyrrolidine-2,4-dicarboxylic acid, PDC) and dialysate concentrations of dopamine (DA) in young rats. Insert: Effects of the NMDA antagonist 3-[(R)-2-carboxypiperazin-4-yl]-propyl-1-phosphonic acid (CPP) on the actions of PDC on dialysate dopamine.

On the functional role of synaptic vs. volume transmission in glutamate and GABA actions in the brain

The existence of two different types of transmission has been suggested (Zoli et al. 1999): synaptic transmission (major part of wiring transmission), very fast and typically phasic, and volume transmission (extracellular transmission), slow and typically tonic. It has also been suggested that changes of extracellular concentrations of glutamate and GABA, as measured by microdialysis, are probably not involved directly in fast synaptic transmission, but could have widespread effects mediated by extrasynaptic receptors (Miele et al. 1996).

A dual transmission (synaptic, extracellular) has already been shown for some neurotransmitters such as dopamine, acetylcholine and serotonin, which exert their major actions through the extracellular space. For instance, and apart from their synaptic actions (from terminals making synaptic contacts; Descarries et al. 1997), an extracellular sphere of influence for dopamine (Zoli et al. 1998) and/or an extracellular ambient level for acetylcholine (Descarries et al. 1997) have been suggested. This diffuse extracellular action of both dopamine or acetylcholine (influence sphere or ambient level) seems to be related to long-lasting effects on vast neuronal and glial assemblies, probably modifying membrane potential and intracellular cascades via extrasynaptic receptors and in this way enhancing or reducing neuronal excitability and gene expression (Descarries et al. 1997; Zoli et al. 1998; Jansson et al. 2002). The volume transmission is also involved in brain metabolism and in the regulation of the brain–blood flow and the brain–blood barrier (see Fuxe and Agnati 1991; Agnati et al. 2000).

The changes in extracellular glutamate and GABA observed in microdialysis studies might indicate the participation of glutamate and GABA in these extracellular actions of neurotransmitters. These extracellular, volume transmission mediated actions would be different, though parallel and integrated with the synaptic actions. In agreement with this possibility, recent studies have suggested that glutamate and GABA extrasynaptic receptors may have different roles and functional characteristics vs. synaptic glutamate and GABA receptors (Nusser et al. 1998; Sattler et al. 2000; Mody 2001). In particular, the GABAA receptor type located outside synapses has been reported to have functional properties different from those of the GABAA receptor type located at synapses (Nusser et al. 1998; Banks and Pearce 2000). Moreover, a role for extracellular GABA acting as a volume transmission signal modulating tonically the signalling at these extrasynaptic GABAA receptors has recently been suggested (Mody 2001). Thus, like extracellular dopamine and other monoamine neurotransmitters, the tonic actions mediated by extracellular levels of glutamate and GABA could be involved in long-lasting effects on the activity of neuronal–astrocytic networks. In fact, brain communication may allow a match of neuronal and glial activity also enabling appropriate metabolic and blood flow responses in the neurone–astrocytic unit, where volume transmission plays a major role. As a speculation and among other possibilities, extracellular glutamate and GABA could accomplish this role by recruiting existing silent synapses (Malgaroli 1999), regulating the expression of cell surface receptors (Bernard et al. 1999) and/or controlling the oscillatory activity of particular populations of neuronal/astrocytic cells (Scanziani 2000; Rose and Konnerth 2001).

The existence of an extracellular microenvironment of different neurotransmitters implies that the effect of one neurotransmitter would depend on the other types of neurotransmitters by becoming integrated in a specific circuit of the brain and under particular circumstances (Barbour and Häusser 1997). Thus, it might be considered that dysregulations and/or pathological disorders in the brain results, directly or indirectly, from alterations in the volume transmission of glutamate and GABA as well as of other neurotransmitters, rather than in the cessation of their actions at synaptic release sites (Meshul et al. 1999; Marti et al. 2000; Syková and Chvátal 2000). The possibility should be also considered that extrasynaptic glutamate and GABA receptors, and presumably high-affinity transporters, could be the target of new therapeutic drugs aimed to restore the physiological glutamate and GABA extracellular signalling, and thus the volume transmission actions mediated by these two neurotransmitters in the brain (Zoli et al. 1999; Vizi 2000; Soltesz and Nusser 2001).

Future perspectives

From the considerable scientific literature published every year using microdialysis, it becomes evident that this technique is one of the most useful and suitable approaches to study some aspects of functional neurochemistry. However, microdialysis may not be reliable enough to better understand the synaptic neurotransmission of the amino acids glutamate and GABA, the two most abundant neurotransmitters in the brain. This limitation is leading to two different outcomes in the field. On one hand, high-sensitivity methods of analysis of dialysate glutamate have been used to increase temporal resolution (seconds) of the microdialysis technique in order to ‘unmask’ glutamate and GABA synaptic release from the extracellular pool (Jay et al. 1999; Perez et al. 2000; Watson and Kennedy 2001). However, as far as we understand, these approaches still may not differentiate between neuronal or astrocytic glutamate and GABA.

On the other hand, the limitation of this technique in detecting directly synaptic glutamate and GABA demands a change in the paradigm with regard to the interpretation of the physiological significance of extracellular levels of glutamate and GABA as measured by brain microdialysis. Thus, as outlined in the present review, dialysate levels of glutamate and GABA might well be interpreted as extracellular volume transmission signals correlated to the functioning of neurone–astrocytic networks. Thus, as the neurone–astrocytic unit underlies the metabolic changes found after the activation of particular areas of the brain, as observed by means of positron emission tomography (PET) or functional magnetic resonance imaging (Magistretti and Pellerin 2000), the specific changes of extracellular glutamate and GABA detected by microdialysis observed under particular behavioural or pharmacological manipulations could indicate the specific activation of neurone–astrocytic networks. Because astrocytes as well as glutamate and GABA seem to be involved in neuropathological diseases such as schizophrenia (Goldman-Rakic and Selemon 1997; Moghaddam and Adams 1998), Parkinson's disease (Meshul et al. 1999; Marti et al. 2000) or epilepsy (During et al. 1995), and also ageing (Segovia et al. 1999, 2001) dialysate glutamate and GABA levels could be useful to better understand the operation of neurone–astrocytic networks in pathological states.

Future studies must be performed to provide more direct support to the hypothesis proposed in the present review that dialysate levels of glutamate and GABA are extracellular signals correlated to the functioning of neurone–astrocytic networks with relevance to physiology and pathology. In particular, one a priori suitable possibility to test this hypothesis would be to combine in vivo microdialysis with drugs that interfere with the astrocyte–astrocyte communication. Thus, for instance, the electrical stimulation of glutamatergic or GABAergic pathways should give rise to lower dialysate increases of either glutamate or GABA in the terminal areas when astrocytic gap-junctions are disrupted. A different approach would be to perfuse directly by reverse dialysis drugs that stimulate specifically the release of astrocytic glutamate and study the neurochemical and/or electrophysiological effects of these increases in the surrounding tissue. In general, a better understanding of this issue will develop with the availability of new drugs able to target specific points of comunication either in the neurone–astrocytic network or in the extracellular signalling pathways (i.e. extrasynaptic receptors). Some of these drugs are currently available (Araque et al. 1999; Carmignoto 2000; Bezzi and Volterra 2001; Haydon 2001).

Concluding remarks

Growing evidence suggests that extracellular concentrations of glutamate and GABA, measured by microdialysis, do not provide a good index of the direct synaptic exocytotic release of these neurotransmitters. In fact, functional and structural evidence suggests that neuronal glutamate and GABA are only involved in very short distance signalling and therefore may not reach the microdialysis probe. However, as the extracellular concentrations of these neurotransmitters change under specific pharmacological treatments and behavioural stimuli, which involve the activity of specific brain circuits, an astrocytic origin of dialysate glutamate and GABA could be suggested in the brain. Taking into account the close functional relationship between neurones and astrocytes as well as the multitude of neurotransmitter receptors localized specifically on astrocytes, extracellular (dialysate) changes of glutamate and GABA could indicate the activation of neurone–astrocytic networks. Thus, there may exist glutamate and GABA volume transmission signalling that differs from synaptic glutamate and GABA signalling, modulating neuronal circuits in specific areas of the brain.

Acknowledgements

The authors thank Dr Tiho Obrenovitch and Dr Alfonso Araque for their suggestions and criticism of the manuscript. The research reported in this review has been supported by CM99-08.5/0031/99, CM2000-08.5/0070.1/2000 and DGINV-SAF-2000/0112. The work has also been supported by Vetenskapsrådet (OMX-715).

Ancillary