The GABA Paradox

Journal of Neurochemistry Lippincott Williams & Wilkins, Inc., Philadelphia © 1999 International Society for Neurochemistry

Short Review

The GABA Paradox: Multiple Roles as Metabolite, Neurotransmitter, and Neurodifferentiative Agent Helle S. Waagepetersen, *Ursula Sonnewald, and Arne Schousboe PharmaBiotec Research Center, Department of Pharmacology, Royal Danish School of Pharmacy, Copenhagen, Denmark; and *Department of Pharmacology and Toxicology, Medical Faculty, Norwegian University of Science and Technology (NTNU), Trondheim, Norway

Abstract: GABA, which is present in the brain in large amounts, is distributed among distinctly different cellular pools, possibly reflecting its multiple functions as metabolite, neurotransmitter, and neurotrophin. Its metabolic enzymes also exhibit heterogeneity, because glutamate decarboxylase exists in two isoforms with different subcellular distribution and regulatory properties. Moreover, recent evidence points to a more pronounced regulatory role of the tricarboxylic acid cycle than hitherto anticipated in the biosynthetic machinery responsible for formation of GABA from glutamine. Additionally, GABAergic neurons may contain distinct populations of mitochondria having different turnover rates of the tricarboxylic acid cycle with different levels of association with GABA synthesis from 2-oxoglutarate via glutamate. These aspects are discussed in relation to the different functional roles of GABA and its prominent involvement in epileptogenic activity. Key Words: Tricarboxylic acid cycle— Glutamine—Glutamate—Mitochondria—Metabolism— Epilepsy—Glutamate decarboxylase—Compartmentation. J. Neurochem. 73, 1335–1342 (1999).

controlling interconversions of apo- and holo-GAD (Kaufmann et al., 1986, 1991; Kobayashi et al., 1987; Erlander et al., 1991; Martin et al., 1991a,b; Bu et al., 1992). This information has been instrumental in attempts to explain how changes in the activity in GABAergic structures are utilized to regulate the rate of synthesis of neurotransmitter GABA. Thus, GAD65 associated mainly with nerve endings is inhibited by even a minor increase in the neuronal GABA concentration (Porter and Martin, 1984; Liden et al., 1987). Moreover, it has been demonstrated convincingly that GAD, which hitherto has been firmly believed to be expressed only in GABAergic neurons (Roberts, 1991), is also expressed in glutamatergic neurons (Sloviter et al., 1996). This obviously adds to the challenge of understanding GABA functions, particularly in relation to the role of the finetuning of glutamatergic excitatory activity and GABAergic inhibitory activity for initiation of seizures (Roberts, 1991). In this context, it has been debated for a long time what may be the exact role of the different subcellular pools of GABA, and anticonvulsants working via interaction with GABA metabolism are known to affect different subcellular pools of GABA differently (Iadarola and Gale, 1980; Gale et al., 1982; Preece and Cerdan, 1996). In addition, GABA is thought to play an important role as a trophic factor in synaptogenesis (Wolff et al., 1978), again adding to the complexity and diversity of GABA functions. These issues will be addressed in the present review with the main emphasis on a discussion of the role of mitochondria and TCA cycle activity in GABA biosynthesis.

Ever since g-aminobutyrate (GABA) was discovered in the CNS (Roberts and Frankel, 1950), it has been puzzling how a large cellular pool can be compatible with a role as a neurotransmitter. Recent detailed studies of the metabolic reactions in GABAergic neurons have indicated a complex involvement of mitochondrial function, namely, a regulatory role of the tricarboxylic acid (TCA) cycle for synthesis of GABA (Waagepetersen et al., 1998a,b). As discussed thoroughly in a recent review by Martin and Rimvall (1993), there are many issues concerning regulatory mechanisms for biosynthesis of GABA in distinctly different functional pools that need to be clarified. In this context, it has been of major importance that two isoforms of the GABA-synthesizing enzyme, glutamate decarboxylase (GAD65 and GAD67), have been cloned and subsequently characterized with regard to the subcellular distribution and mechanisms

Address correspondence and reprint requests to Prof. A. Schousboe at PharmaBiotec Research Center, Department of Pharmacology, Royal Danish School of Pharmacy, 2 Universitetsparken, DK-2100 Copenhagen, Denmark. Abbreviations used: GABA, g-aminobutyrate; GABA-T, GABAaminotransferase; GAD, glutamate decarboxylase; GVG, g-vinyl GABA; PAG, phosphate-activated glutaminase; TCA, tricarboxylic acid; VPA, valproate.

1335

1336

H. S. WAAGEPETERSEN ET AL. GABA METABOLISM AND COMPARTMENTATION

GABA is a nonprotein amino acid found in virtually all prokaryotic and eukaryotic organisms, including plants. GABA arises via decarboxylation of L-glutamate by GAD and is metabolized subsequently via GABAaminotransferase (GABA-T) to succinic semialdehyde, which is then oxidized to succinate. The enzymes responsible for these conversions have been called the GABA shunt (Bala´zs et al., 1970). This shunt allows four of the five carbons lost from the TCA cycle in the form of 2-oxoglutarate to reenter at the level of succinate. There appears to be some discrepancy as to the estimated activity of this cycle. Cell culture (Westergaard et al., 1995a) and brain tissue (Bala´zs et al., 1970) experiments indicate low (,10%) activity, whereas a recent study in mouse brain reported that the GABA shunt accounts for 17% of the TCA cycle activity (Hassel et al., 1998). Conversion of glutamate to GABA and subsequently to succinate semialdehyde involves exclusively pyridoxal phosphate-requiring enzymes and might thus be a way to obtain energy from glutamate without production of ammonia. It should be kept in mind that this pathway (the GABA shunt) will not provide the same amount of energy as the alternative TCA cycle reactions (8% is lost). This loss might, however, be negligible, because most cells are considered to have a large excess capacity for producing energy (Sokoloff et al., 1977). GABA synthesis in the brain appears to take place in two major compartments, i.e., the cell body and the nerve terminal

FIG. 1. Cultured cerebral cortical neurons were incubated with [U-13C]glucose in the absence or presence of glutamine, and cell extracts were analyzed by 13C NMR spectroscopy. The cycling ratio was calculated as the amount of 13C labeling in glutamate C-3 divided by the amount of glutamate C-4 doublet. With regard to GABA, it corresponds to GABA C-3 divided by the amount of GABA C-2 doublet. The cycling for aspartate was calculated from the amount of C-3 doublet subtracted from the total amount of C-3 labeling, and this is divided by the amount of C-3 doublet (Waagepetersen et al., 1998b). Results are averages 6 SEM. A statistically significant difference between the values of GABA and glutamate in the control situation is indicated by a Latin cross (†p , 0.05). Statistically significant differences between the control situation (white columns) and the presence of glutamine (dark columns) are indicated by asterisks (***p , 0.001; **p , 0.025). Original results are from Waagepetersen et al. (1998b).

J. Neurochem., Vol. 73, No. 4, 1999

FIG. 2. An illustration of a two-compartment model of TCA cycle function and association with GABA synthesis. a: Mitochondria having close association with GABA synthesis indicated by the large open arrow. b: Mitochondria having a higher TCA cycle activity, but a much less pronounced GABA synthesis as indicated by the small open arrow.

(Garfinkel, 1966; Patel et al., 1974; Iadarola and Gale, 1981; Martin and Rimvall, 1993). Two forms of GAD have been found in GABAergic neurons: GAD65, which is associated predominantly with the axonal region including nerve endings, and GAD67, which is more widely distributed (Erlander et al., 1991; Gonzales et al., 1991; Henry and Tappaz, 1991; Kaufman et al., 1991). This may be compatible with the recent demonstration of compartmentalized metabolism of glutamate in GABAergic neurons in culture (Westergaard et al., 1995a; Waagepetersen et al., 1998a,b, 1999a,b). Incorporation of 13C from [U-13C]glucose into glutamate in cultured cerebral cortical neurons reached a steady-state level within 4 h, whereas for GABA this was not the case. Moreover, as seen in Fig. 1a and b, the cycling ratio of a TCA cycle equilibrating with glutamate used for GABA synthesis was lower than that of a cycle equilibrating with glutamate, which is less likely to be used for GABA synthesis. The finding that glutamate and GABA are in equilibrium with TCA cycles having different cycling activities underlines the existence of multiple TCA cycles related to synthesis of GABA and glutamate (Waagepetersen et al., 1998b). A schematic representation of different TCA cycle compartments is provided in Fig. 2. Figure 2a shows a compartment with close connection to GABA synthesis and lower TCA cycle activity. Figure 2b illustrates a compartment with a TCA cycle exhibiting higher activity and in which glutamate is less likely to be used for GABA synthesis. Further evidence that neurons may contain subsets of mitochondria with different metabolic characteristics comes from studies of metabolism of [U-13C]lactate and [U-13C]glucose in the presence of glutamine in cultured neocortical neurons. Glutamine was found to affect the percent incorporation of 13C into glutamate from these two acetylCoA-generating substrates differently (Waagepetersen et al., 1998b). Also metabolism of [U-13C]glutamine has exhibited compartmentation of mitochondrial metabolism. Thus, a difference in aspartate isotopomer formation was seen when [U-13C]glutamine or [U-13C]glutamate was used as a precursor in GABAergic neurons

THE GABA PARADOX

1337

(Westergaard et al., 1995a). Mitochondrial heterogeneity has been observed not only in neurons, but also in astrocytes (Schousboe et al., 1993; Sonnewald et al., 1993a), and a further discussion of this phenomenon can be found in a recent review by Sonnewald et al. (1998). GLUTAMINE AS PRECURSOR FOR GABA: ROLE OF THE TCA CYCLE It has been demonstrated convincingly in different brain tissue preparations that both exogenous glutamine and glutamate can function as precursors for the synthesis of GABA, with glutamine often functioning as the preferred substrate (Reubi et al., 1978; Yu et al., 1984; Yudkoff et al., 1989). A comparison of GAD and glutaminase activities in the neostriatum of rats lesioned with varying amounts of kainic acid suggests that some 60% of the activity of PAG is located in GABAergic structures (McGeer et al., 1983), emphasizing the importance of glutamine as a substrate for GABAergic neurons. It should be mentioned, however, that PAG has been reported to have essentially the same activity in glutamatergic and GABAergic neurons in culture (Drejer et al., 1985; Larsson et al., 1985). In cultured cerebral cortical neurons (GABAergic), it has been shown using [U-13C]glutamine or [U-13C]glutamate that the enrichment in GABA (;14%) was independent of the precursor (Westergaard et al., 1995a). Even though the enrichment was similar, the labeling patterns were different. GABA formed from glutamine was derived to a considerable extent from the TCA cycle via 2-oxoglutarate and glutamate, whereas when glutamate was the precursor, direct synthesis involving only GAD was prevailing (Westergaard et al., 1995a). That the synthesis of GABA from glutamine involves the TCA cycle via 2-oxoglutarate was confirmed by Waagepetersen et al. (1998b) using cultured cerebral cortical neurons incubated with unlabeled glutamine in combination with [U-13C]glucose. Figure 1 shows that the presence of glutamine decreases the cycling ratio of TCA cycles equilibrating with GABA, glutamate, and aspartate, strongly supporting the notion that the glutamine carbon skeleton has access to the cycle. This is underlined by the concomitant finding (Fig. 3) that the percent enrichment in the amino acids was essentially unaffected by glutamine, because access to the TCA cycle is a prerequisite for labeling from [U-13C]glucose (Waagepetersen et al., 1998b). Comparable to what was obtained when cerebral cortical neurons were incubated with [U-13C]glutamate or [U-13C]glutamine (Westergaard et al., 1995a), a relatively low enrichment of GABA was observed in synaptosomes using [15N]glutamate or [2-15N]glutamine as precursors (Yudkoff et al., 1989). These observations point to the existence of a large unlabeled GABA pool that may be explained by a slow turnover of GABA, as has been reported by several authors (Bernasconi et al., 1982; Chapman and Evans, 1983; Paulsen and Fonnum, 1987; Martin and Rimvall, 1993). In agreement with

FIG. 3. Percent enrichment was calculated from data obtained from NMR spectroscopy and HPLC analysis of cell extracts after incubation of cultured cerebral cortical neurons with [U-13C]glucose in the absence (white columns) or presence of glutamine (dark columns) for 4 h. Results are averages 6 SEM. Original results are from Waagepetersen et al. (1998b).

previous work by Sihra and Nicholls (1987) using isolated synaptosomes and [14C]GABA to label the GABA pools, it might be suggested that a large vesicular pool of GABA separated from a cytoplasmic pool would generate a low enrichment. Use of the GABA-transport blocker tiagabine during depolarization of neocortical neurons (GABAergic) with 55 mM K1 in the presence of Ca21 allowed study of GABA release selectively from the vesicular pool, as demonstrated previously by Belhage et al. (1993). By incubation of cells in [13C]glucose and mass spectrometric determination of labeling in GABA in the cells as well as in the release medium, it could be shown that the vesicular pool of GABA is fully equilibrated with the cytosolic GABA pool labeled from [U-13C]glucose (Waagepetersen et al., 1999a). To study further the pathways for synthesis of GABA from glutamine, this amino acid was present during incubation with [U-13C]glucose and subsequently during depolarization. The findings indicated that synthesis of vesicular GABA from glutamine involved mitochondrial and TCA cycle metabolism and that newly synthesized GABA did not rapidly equilibrate with the vesicular pool (Waagepetersen et al., 1999a). This is illustrated in Fig. 4 showing that percent labeling of GABA from [U-13C]glucose in the presence of glutamine is higher in the releasable, vesicular pool than in the cellular pool. Together with the finding that the labeling of vesicular GABA was independent of the presence of glutamine, this underlines that glutamine could not have been converted directly to vesicular GABA, but must have been metabolized via the TCA cycle (Waagepetersen et al., 1999a). That glutamate formed from glutamine in the PAG-catalyzed reaction is in rapid equilibrium with the TCA cycle pool of a-ketoglutarate is compatible with the very high activity of aspartate aminotransferase (Mason et al., 1995) and the demonstration that PAG is associated exclusively with the mitochondria and most likely in close association with the inner mitochondrial membrane (Kvamme et al., 1988; Laake et al., 1999). J. Neurochem., Vol. 73, No. 4, 1999

1338

H. S. WAAGEPETERSEN ET AL.

FIG. 4. Cultured cerebral cortical neurons were incubated for 4 h in serum-free medium containing [U-13C]glucose. Subsequent to the incubation, the cultures were repetitively depolarized eight times with 55 mM K1 in the presence of Ca21 (1.0 mM) and tiagabine (50 mM), blocking release of GABA from the cytoplasmic pool (a). Glutamine was either present during incubation (b) or throughout the experiment, i.e., during the incubation and the subsequent depolarization (c). Percent 13C labeling in GABA of cell extracts and depolarization buffer was analyzed by gas chromatography–mass spectrometry. Results are averages 6 SEM. A statistically significant difference between cell extracts and the depolarization medium is indicated (**p , 0.025). Original results are from Waagepetersen et al. (1999a).

GABA IN GLUTAMATERGIC CELLS The GABA-synthesizing enzyme GAD, which was originally thought to be localized exclusively in GABAergic neurons (Roberts, 1991), has been detected recently in glutamatergic hippocampal granule cells as well (Schwarzer and Sperk, 1995; Sloviter et al., 1996). The role of GABA in these cells can only be speculated on at this point, but an attractive hypothesis could be that of an inhibitory potential for excitatory neurons. This may be supported by the fact that GAD67 expression in these glutamatergic cells is enhanced after seizures (Schwarzer and Sperk, 1995). GABA-T, the enzyme for GABA degradation, has also been shown to be distributed further than the GABAergic system (McGeer et al., 1983), but this is less surprising, because it is a ubiquitous enzyme present in brain and liver with the same activity (Roberts and Bregoff, 1953). This may be related to the fact that GABA-T also is involved in the transamination of b-alanine, a metabolite from pyrimidine bases (Schousboe et al., 1973). It should be emphasized that high-affinity GABA transporters (e.g., GAT-1) so far have been found mainly in GABAergic neurons or astrocytes (Borden, 1996), and only few reports have appeared indicating the presence of GAT-1 in putative glutamatergic neurons (Minelli et al., 1995; Nishimura et al., 1997; Yasumi et al., 1997). GABA HOMEOSTASIS AND EPILEPSY GABA is released continuously from nerve terminals in relation to neurotransmission. The inactivation of GABA, like that of glutamate, has been the focus of numerous studies. GABA transporters have been identified on neurons and astrocytes, but unlike glutamate, it appears that GABA is taken up predominantly by neuJ. Neurochem., Vol. 73, No. 4, 1999

rons (Hertz and Schousboe, 1987). However, that neuronally released GABA does enter the astrocytes was demonstrated by Westergaard et al. (1992). In this study, the GABA uptake inhibitor 4,5,6,7-tetrahydroisoxazolo[4,5-c]pyridin-3-ol (Schousboe et al., 1981) was used to block astrocytic GABA uptake in cocultures of astrocytes and neurons, leading to an apparent increased neuronal GABA release due to diminished drain by glial uptake. In hippocampal slices, the GABA transport inhibitor tiagabine (Braestrup et al., 1990) has been shown to enhance GABAergic neuronal activity (Roepstorff and Lambert, 1992). GABA-T activity, which is largely mitochondrial (Salganicoff and De Robertis, 1965), has been found in astrocytes and neurons, although predominantly in neurons (McGeer et al., 1983). From studies of cultured astrocytes and GABAergic neurons, however, it may be concluded that neurons and astrocytes have equally high capacity for GABA metabolism via action of GABA-T (Schousboe et al., 1977; Larsson et al., 1986). It is evident that astrocytic uptake and degradation of GABA require mechanisms for replenishing GABA stores in neurons. Neurons lack the capability of de novo synthesis of GABA from glucose and are thus dependent on metabolites from astrocytes for net synthesis of GABA. This is based on the repeated finding (Yu et al., 1983; Shank et al., 1985; Cesar and Hamprecht, 1995) that the anaplerotic enzyme responsible for de novo synthesis of TCA cycle constituents, pyruvate carboxylase, is only present in astrocytes. In effect, there is extensive cycling of five- and four-carbon units between astrocytes and neurons (for review, see Westergaard et al., 1995b). Candidates such as malate and 2-oxoglutarate have been proposed as GABA precursors, but no direct indications for such an action have been presented (Shank and Campbell, 1984; Kihara and Kubo, 1989; Shank et al., 1989; Hertz et al., 1992). The role of glutamine, which is synthesized in glia only (Norenberg and Martinez-Hernandez, 1979), in this context is well established as discussed above (Reubi et al., 1978; Yu et al., 1984; Battaglioli and Martin, 1990, 1996; Sonnewald et al., 1993b; Westergaard et al., 1995a). An intricate balance of excitation and inhibition is essential for the maintenance of normal function of the CNS. Specific pathways are associated with particular tasks, e.g., the inputs and outputs of the basal ganglia and the cerebellum are responsible for coordinating behavior by integrating sensory and cognitive information about the intention, context, and status of movement. Disturbances in the excitation/inhibition balance in the hypothalamus is thought to be a major factor in the etiology of epilepsy, where GABA is a key player (Schousboe, 1990). Thus, inhibiting the synthesis of GABA causes seizures, as does the administration of GABAA receptor antagonists (Tapia, 1975). It should be noted that in preparations from rat brain, results have been obtained that support the notion that the GABA concentration determines the efficacy of inhibition (Golan et al., 1996). In electrophysiological studies of human temporal-lobe epilepsy, it could be demonstrated that potassium-stim-

THE GABA PARADOX ulated release of GABA was increased, whereas glutamate-induced, calcium-independent release of GABA was markedly decreased. This decrease was shown in amygdala-kindled rats to be due to a decrease in the number of GABA transporters (During et al., 1995). It is generally believed that the cytosolic and, perhaps even more importantly, the nerve ending concentration of a neurotransmitter is an important factor determining its release. That GABA catabolism (GABA-T) and anabolism (GAD) are important for inhibitory efficiency has been shown in several studies (e.g., Gram et al., 1988; Golan and Grossman, 1996). Thus, GAD inhibitors decrease postsynaptic and presynaptic inhibition by reducing both tonic and evoked release (Golan and Grossman, 1996). Antiepileptic treatment, in contrast, is aimed at increasing the GABA concentration (for other strategies, see Shank and Reife, 1999). To claim a GABAergic mechanism of action of a drug, which interferes with the metabolism of GABA, it is not sufficient to demonstrate inhibition of GABA-T and perhaps increased GABA levels in brain (Gale et al., 1982). It must be regarded as an absolute prerequisite to show that the drug causes a selective increase in the GABA transmitter pool, as well as an increased release of GABA following synaptic stimulation (Wood and Kurylo, 1980; Gale et al., 1982). The two antiepileptic drugs g-vinyl GABA (GVG) and valproate (VPA) have been shown to function in this manner in cultured cerebral cortex neurons (Gram et al., 1988), and particularly VPA appears to selectively increase the GABA concentration in the synaptic pool (Iadarola and Gale, 1980; Gale et al., 1982). It should be noted, however, that the concentration at which VPA showed this effect might not be pharmacologically relevant, and it is generally assumed that this drug influences also other inhibitory actions. It is interesting that Sheikh and Martin (1998) found not only an elevation of brain GABA levels with GVG, but also a differential effect on GAD65 and GAD67 expression in various regions of rat brain. Gram et al. (1988) reported that treatment with GVG led to a higher specific radioactivity of the releasable than the cellular pool when cells were incubated with radiolabeled GABA, again suggesting that during depolarization no rapid equilibration is taking place between the cytosolic and vesicular pools (see Waagepetersen et al., 1999a,b). The importance of exogenous GABA for vesicular release is also confirmed in the study by Westergaard et al. (1992), showing that inhibition of astrocytic GABA uptake leads to increased neuronal release. Finally, this strategy has been used successfully in the development of the new antiepileptic drug, tiagabine (Braestrup et al., 1990; Fink-Jensen et al., 1992). GABA AND NEURODIFFERENTIATION Pioneering studies by Wolff et al. (1978) in rat superior cervical ganglia and subsequent work in vivo and in cell cultures by others (for reviews, see Meier et al., 1991; Belhage et al., 1998) have established GABA as an

1339

important neurotrophic and neurodifferentiating signal molecule during early ontogenesis. Although the exact mechanism for this action remains to be elucidated, it clearly involves GABAA receptor activation, as shown by the demonstration that the effect of GABA can be mimicked by GABAA receptor agonists and blocked by GABAA receptor antagonists (Meier et al., 1985; Belhage et al., 1986). Moreover, the action of GABA requires de novo protein synthesis (Belhage et al., 1990). Therefore, it may be argued that this function is related to the neurotransmitter role of GABA, but it should be kept in mind that the final outcome of this type of signaling reaches far beyond that of normal neurotransmission processes. It should also be stressed that this neurodifferentiative action takes place at a developmental stage where vesicular release processes are unlikely to be fully developed (Schousboe and Redburn, 1995). It is therefore likely that nonsynaptic release mechanisms are operating (Gordon-Weeks et al., 1984). In fact, both cerebral cortex and retina contain in early development a population of pioneering cells that contain GABA and that during later developmental periods die or lose their GABAergic phenotype (for references, see Schousboe and Redburn, 1995). In this context, it should also be kept in mind that during early development, i.e., at a period during which GABA neurotransmission is less prominent, GABA might be synthesized through an alternative pathway involving ornithine, as suggested by Seiler (1980). This is underlined by the finding that in avian retina GABA can be found before detection of GAD by monoclonal antibodies (Hokoc et al., 1990), and ornithine decarboxylase, the key enzyme in GABA synthesis from polyamines (Seiler, 1980), is present in retina during early ontogenesis (De Mello et al., 1976). A more detailed discussion of this action of GABA may be found in recent reviews by Belhage et al. (1998) and Carlson et al. (1998). CONCLUDING REMARKS It is apparent that the complexity of GABA biosynthesis has been constantly increasing ever since GAD was originally purified to homogeneity from mouse brain (Wu et al., 1973). Although at that time evidence was presented suggesting different subunits of GAD, the existence of isoforms of GAD was not anticipated (Matsuda et al., 1973). The subsequent cloning studies of GAD, as well as the recent observations suggesting a regulatory role of the TCA cycle (Kaufman et al., 1991; Waagepetersen et al., 1999a,b), may lead to speculations as to the exact regulatory mechanisms governing the rate of synthesis of neurotransmitter and metabolic GABA, respectively. Clearly, further research is needed to elucidate this issue fully. Additionally, as regulation pertinent to neurotransmitter GABA involves GABA transporters both in synaptic membranes and in the vesicles, it is not surprising that the homeostatic mechanisms underlying the maintenance of the neurotransmitter GABA pool are far from being fully understood. J. Neurochem., Vol. 73, No. 4, 1999

1340

H. S. WAAGEPETERSEN ET AL.

Acknowledgment: The authors would like to thank Dr. Uwe Sonnewald for fruitful discussion. This research was supported by the Danish State Biotechnology Programs (1991–95 and 1995–99), the Danish Medical Research Council (9700761), the NORMOX NorFa Grant, the Research Council of Norway (SIP 782008.01), the Lundbeck, NOVO Nordisk, Blix, and SINTEF Foundations, and the Foundation for Special Medical Applications (RiT).

REFERENCES Bala´zs R., Machiyama Y., Hammond B. J., Julian T., and Richter D. (1970) The operation of the gamma-aminobutyrate bypath of the tricarboxylic acid cycle in brain tissue in vitro. Biochem. J. 116, 445– 461. Battaglioli G. and Martin D. L. (1990) Stimulation of synaptosomal g-aminobutyric acid synthesis by glutamate and glutamine. J. Neurochem. 54, 1179 –1187. Battaglioli G. and Martin D. L. (1996) Glutamine stimulates g-aminobutyric acid synthesis in synaptosomes but other putative astrocyte-to-neuron shuttle substrates do not. Neurosci. Lett. 209, 129 –133. Belhage B., Meier E., and Schousboe A. (1986) GABA-agonists induce the formation of low-affinity GABA-receptors on cultured cerebellar granule cells via preexisting high affinity GABA-receptors. Neurochem. Res. 11, 599 – 606. Belhage B., Hansen G. H., Meier E., and Schousboe A. (1990) Effects of inhibitors of protein synthesis and intracellular transport on the g-aminobutyric acid agonist-induced functional differentiation of cultured cerebellar granule cells. J. Neurochem. 55, 1107–1113. Belhage B., Hansen G. H., and Schousboe A. (1993) Depolarization by K1 and glutamate activates different neurotransmitter release mechanisms in GABAergic neurons: vesicular versus non-vesicular release of GABA. Neuroscience 54, 1019 –1034. Belhage B., Hansen G. H., Elster L., and Schousboe A. (1998) Effects of gamma-aminobutyric acid (GABA) on synaptogenesis and synaptic function. Perspect. Dev. Neurobiol. 5, 235–246. Bernasconi R., Maitre L., Martin P., and Raschdorf F. (1982) The use of inhibitors of GABA-transaminase for the determination of GABA turnover in mouse brain regions: an evaluation of aminooxyacetic acid and gabaculine. J. Neurochem. 38, 57– 66. Borden L. A. (1996) GABA transporter heterogeneity: pharmacology and cellular localization. Neurochem. Int. 29, 335–356. Braestrup C., Nielsen E. B., Sonnewald U., Knutsen L. J., Andersen K. E., Jansen J. A., Frederiksen K., Andersen P. H., Mortensen A., and Suzdak P. D. (1990) (R)-N-[4,4-Bis(3-methyl-2-thienyl)but3-en-1-yl]nipecotic acid binds with high affinity to the brain g-aminobutyric acid uptake carrier. J. Neurochem. 54, 639 – 647. Bu D.-F., Erlander M. G., Hitz B. C., Tillakaratne N. J. K., Kaufman D. L., Wagner-McPherson C. B., Evans B. A., and Tobin A. J. (1992) Two human glutamate decarboxylases, 65-kDa GAD and 67-kDa GAD, are each encoded by a single gene. Proc. Natl. Acad. Sci. USA 89, 2115–2119. Carlson B. X., Elster L., and Schousboe A. (1998) Pharmacological and functional implications of developmentally-regulated changes in GABA(A) receptor subunit expression in the cerebellum. Eur. J. Pharmacol. 352, 1–14. Cesar M. and Hamprecht B. (1995) Immunohistochemical examination of neural rat and mouse primary cultures using monoclonal antibodies raised against pyruvate carboxylase. J. Neurochem. 64, 2312–2318. Chapman A. G. and Evans M. C. (1983) Cortical GABA turnover during bicuculline seizures in rats. J. Neurochem. 41, 886 – 889. De Mello F. G., Bachrach U., and Nirenberg M. (1976) Ornithine and glutamate decarboxylase activities in developing chick retina. J. Neurochem. 27, 847– 851. Drejer J., Larsson O. M., Kvamme E., Svenneby G., Hertz L., and Schousboe A. (1985) Ontogenetic development of glutamate metabolizing enzymes in cultured cerebellar granule cells and in cerebellum in vivo. Neurochem. Res. 10, 49 – 62.

J. Neurochem., Vol. 73, No. 4, 1999

During M. J., Ryder K. M., and Spencer D. D. (1995) Hippocampal GABA transporter function in temporal-lobe epilepsy. Nature 376, 174 –177. Erlander M. G., Tillakaratne N. J. K., Feldblum S., Patel N., and Tobin A. J. (1991) Two genes encode distinct glutamate decarboxylases. Neuron 7, 91–100. Fink-Jensen A., Suzdak P. D., Swedberg M. D., Judge M. E., Hansen L., and Nielsen P. G. (1992) The gamma-aminobutyric acid (GABA) uptake inhibitor, tiagabine, increases extracellular brain levels of GABA in awake rats. Eur. J. Pharmacol. 220, 197–201. Gale K., Iadarola M. J., Casu M., and Keating R. F. (1982) Relationship between GABA levels in vivo and anticonvulsant activity: importance of cellular compartments and regional localization in brain, in Problems in GABA Research from Brain to Bacteria (Okada Y. and Roberts E., eds), pp. 159 –181. Excerpta Medica, Amsterdam. Garfinkel D. (1966) A simulation study of the metabolism and compartmentation in brain of glutamate, aspartate, the Krebs cycle, and related metabolites. J. Biol. Chem. 241, 3918 –3929. Golan H. and Grossman Y. (1996) Block of glutamate decarboxylase decreases GABAergic inhibition at the crayfish synapses: possible role of presynaptic metabotropic mechanisms. J. Neurophysiol. 75, 2089 –2098. Golan H., Talpalar A. E., Schleifstein-Attias D., and Grossman Y. (1996) GABA metabolism controls inhibition efficacy in the mammalian CNS. Neurosci. Lett. 217, 25–28. Gonzales C., Kaufman D. L., Tobin A. J., and Chesselet M.-F. (1991) Distribution of glutamate decarboxylase (Mr 67,000) in the basal ganglia of the rat: an immunohistochemical study with a selective cDNA-generated polyclonal antibody. J. Neurocytol. 20, 953– 961. Gordon-Weeks P. R., Lockerbie R. O., and Pearce B. R. (1984) Uptake and release of [3H]GABA by growth cones isolated from neonatal rat brain. Neurosci. Lett. 52, 205–210. Gram L., Larsson O. M., Johnsen A. H., and Schousboe A. (1988) Effects of valproate, vigabatrin and aminooxyacetic acid on release of endogenous and exogenous GABA from cultured neurons. Epilepsy Res. 2, 87–95. Hassel B., Johannessen C. U., Sonnewald U., and Fonnum F. (1998) Quantification of the GABA shunt and the importance of the GABA shunt versus the 2-oxoglutarate dehydrogenase pathway in GABAergic neurons. J. Neurochem. 71, 1511–1518. Henry S. and Tappaz M. (1991) Relative enrichment of the lighter 59 kDa form of glutamic acid decarboxylase in nerve endings: an immunoblotting study in pituitary neurointermediate lobe. Neurosci. Lett. 131, 253–256. Hertz L. and Schousboe A. (1987) Primary cultures of GABAergic and glutamatergic neurons as model systems to study neurotransmitter functions. I. Differentiated cells, in Model Systems of Development and Aging of the Nervous System (Vernadakis A., Privat A., Lauder J. M., Timiras P. S., and Giacobini E., eds), pp. 19 –31. Martinus Nijhoff Publishing, Boston. Hertz L., Yu A. C., and Schousboe A. (1992) Uptake and metabolism of malate in neurons and astrocytes in primary cultures. J. Neurosci. Res. 33, 289 –296. Hokoc J. N., Ventura A. L., Gardino P. F., and De Mello F. G. (1990) Developmental immunoreactivity for GABA and GAD in the avian retina: possible alternative pathway for GABA synthesis. Brain Res. 532, 197–202. Iadarola M. J. and Gale K. (1980) Evaluation of increases in nerve terminal-dependent vs. nerve terminal-independent compartments of GABA in vivo. Brain Res. Bull. 5, S213–S219. Iadarola M. J. and Gale K. (1981) Cellular compartments of GABA in brain and their relationship to anticonvulsant activity. Mol. Cell. Biochem. 39, 305–330. Kaufman D. L., McGinnes J. F., Krieger N. R., and Tobin A. J. (1986) Brain glutamate decarboxylase cloned in lgt-11: fusion protein produces g-aminobutyric acid. Science 232, 1138 –1140. Kaufman D. L., Houser C. R., and Tobin A. J. (1991) Two forms of the g-aminobutyric acid synthetic enzyme glutamate decarboxylase have distinct intraneuronal distributions and cofactor interactions. J. Neurochem. 56, 720 –723.

THE GABA PARADOX Kihara M. and Kubo T. (1989) Aspartate aminotransferase for synthesis of transmitter glutamate in the medulla oblongata: effect of aminooxyacetic acid and 2-oxoglutarate. J. Neurochem. 52, 1127– 1134. Kobayashi Y., Kaufman D. L., and Tobin A. J. (1987) Glutamic acid decarboxylase cDNA: nucleotide sequence encoding an enzymatically active fusion protein. J. Neurosci. 7, 2768 –2772. Kvamme E., Svenneby G., and Torgner I. A. (1988) Glutaminases, in Glutamine and Glutamate in Mammals, Vol. 2 (Kvamme E., ed), pp. 54 – 63. CRC Press, Boca Raton, Florida. Laake J. H., Takumi Y., Eidet J., Torgner I. A., Roberg B., Kvamme E., and Ottersen O. P. (1999) Postembedding immunogold labelling reveals subcellular localization and pathway-specific enrichment of phosphate activated glutaminase in rat cerebellum. Neuroscience 88, 1137–1151. Larsson O. M., Drejer J., Kvamme E., Svenneby G., Hertz L., and Schousboe A. (1985) Ontogenetic development of glutamate and GABA metabolizing enzymes in cultured cerebral cortex interneurons and in cerebral cortex in vivo. Int. J. Dev. Neurosci. 3, 177–185. Larsson O. M., Gram L., Schousboe I., and Schousboe A. (1986) Differential effect of gamma-vinyl GABA and valproate on GABA-transaminase from cultured neurones and astrocytes. Neuropharmacology 25, 617– 625. Liden E., Karlson I., and Sellstrom A. (1987) Is the concentration of g-aminobutyric acid in the nerve terminal regulated via product inhibition of glutamic acid decarboxylase? Neurochem. Res. 12, 489 – 493. Martin D. L. and Rimvall K. (1993) Regulation of g-aminobutyric acid synthesis in the brain. J. Neurochem. 60, 395– 407. Martin D. L., Martin S. B., Wu S. J., and Espina N. (1991a) Cofactor interactions and the regulation of glutamate decarboxylase activity. Neurochem. Res. 16, 243–249. Martin D. L., Martin S. B., Wu S. J., and Espina N. (1991b) Regulatory properties of brain glutamate decarboxylase (GAD): the apoenzyme of GAD is present principally as the smaller of two molecular forms of GAD in brain. J. Neurosci. 11, 2725–2731. Mason G. F., Gruetter R., Rothman D. L., Behar K. L., Shulman R. G., and Novotny E. J. (1995) Simultaneous determination of the rates of the TCA cycle, glucose utilization, alpha-ketoglutarate/glutamate exchange, and glutamine synthesis in human brain by NMR. J. Cereb. Blood Flow Metab. 15, 12–25. Matsuda T., Wu J.-Y., and Roberts E. (1973) Electrophoresis of glutamic acid decarboxylase (EC 4.1.1.15) from mouse brain in sodium dodecyl sulphate polyacrylamide gels. J. Neurochem. 21, 167–172. McGeer E. G., McGeer P. L., and Thompson S. (1983) GABA and glutamate enzymes, in Glutamine, Glutamate and GABA in the Central Nervous System (Hertz L., Kvamme E., McGeer E. G., and Schousboe A., eds), pp. 3–17. Alan R. Liss, Inc., New York. Meier E., Hansen G. H., and Schousboe A. (1985) The trophic effect of GABA on cerebellar granule cells is mediated by GABA-receptors. Int. J. Dev. Neurosci. 3, 401– 407. Meier E., Hertz L., and Schousboe A. (1991) Neurotransmitters as developmental signals. Neurochem. Int. 19, 1–15. Minelli A., Brecha N. C., Karschin C., DeBiasi S., and Conti F. (1995) GAT-1, a high-affinity GABA plasma membrane transporter, is localized to neurons and astroglia in the cerebral cortex. J. Neurosci. 15, 7734 –7746. Nishimura M., Sato K., Mizuno M., Yoshiya I., Shimada S., Saito N., and Tohyama M. (1997) Differential expression patterns of GABA transporters (GAT1–3) in the rat olfactory bulb. Brain Res. Mol. Brain Res. 45, 268 –274. Norenberg M. D. and Martinez-Hernandez A. (1979) Fine structural localization of glutamine synthetase in astrocytes of rat brain. Brain Res. 161, 303–310. Patel A. J., Johnson A. L., and Bala´zs R. (1974) Metabolic compartmentation of glutamate associated with the formation of g-aminobutyrate. J. Neurochem. 23, 1271–1279. Paulsen R. and Fonnum F. (1987) Comparison of results obtained with different methods for estimating GABA turnover in rat neostriatum. Biochem. Pharmacol. 36, 1539 –1544.

1341

Porter T. G. and Martin D. L. (1984) Evidence for feedback regulation of glutamate decarboxylase by g-aminobutyric acid. J. Neurochem. 43, 1464 –1467. Preece N. E. and Cerdan S. (1996) Metabolic precursors and compartmentation of cerebral GABA in vigabatrin-treated rats. J. Neurochem. 67, 1718 –1725. Reubi J. C., Van den Berg C. J., and Cue´nod M. (1978) Glutamine as precursor for the GABA and glutamate transmitter pools. Neurosci. Lett. 10, 171–174. Roberts E. (1991) Living systems are tonically inhibited, autonomous optimizers, and disinhibition coupled to variability generation is their major organizing principle: inhibitory command-control at levels of membrane, genome, metabolism, brain, and society. Neurochem. Res. 16, 409 – 421. Roberts E. and Bregoff H. M. (1953) Transamination of g-aminobutyric acid and b-alanine in brain and liver. J. Biol. Chem. 201, 393–398. Roberts E. and Frankel S. (1950) g-Aminobutyric acid in brain: its formation from glutamic acid. J. Biol. Chem. 187, 55– 63. Roepstorff A. and Lambert J. D. (1992) Comparison of the effect of the GABA uptake blockers, tiagabine and nipecotic acid, on inhibitory synaptic efficacy in hippocampal CA1 neurones. Neurosci. Lett. 146, 131–134. Salganicoff L. and De Robertis E. (1965) Subcellular distribution of the enzymes of the glutamic acid, glutamine and g-aminobutyric acid cycles in rat brain. J. Neurochem. 12, 287–309. Schousboe A. (1990) Neurochemical alterations associated with epilepsy or seizure activity, in Comprehensive Epileptology (Dam M. and Gram L., eds), pp. 1–16. Raven Press, New York. Schousboe A. and Redburn D. A. (1995) Modulatory actions of gamma aminobutyric acid (GABA) on GABA type A receptor subunit expression and function. J. Neurosci. Res. 41, 1–7. Schousboe A., Wu J. Y., and Roberts E. (1973) Purification and characterization of the 4-aminobutyrate–2,ketoglutarate transaminase from mouse brain. Biochemistry 12, 2868 –2873. Schousboe A., Svenneby G., and Hertz L. (1977) Uptake and metabolism of glutamate in astrocytes cultured from dissociated mouse brain hemispheres. J. Neurochem. 29, 999 –1005. Schousboe A., Larsson O. M., Hertz L., and Krogsgaard-Larsen P. (1981) Heterocyclic GABA analogues as selective inhibitors of astroglial GABA uptake. Adv. Biochem. Psychopharmacol. 29, 135–141. Schousboe A., Westergaard N., Sonnewald U., Petersen S. B., Huang R., Peng L., and Hertz L. (1993) Glutamate and glutamine metabolism and compartmentation in astrocytes. Dev. Neurosci. 15, 359 –366. Schwarzer C. and Sperk G. (1995) Hippocampal granule cells express glutamic acid decarboxylase-67 after limbic seizures in the rat. Neuroscience 69, 705–709. Seiler N. (1980) On the role of GABA in vertebrate polyamine metabolism. Physiol. Chem. Phys. 12, 411– 429. Shank R. P. and Campbell G. L. (1984) a-Ketoglutarate and malate uptake and metabolism by synaptosomes: further evidence for an astrocytes-to-neuron metabolic shuttle. J. Neurochem. 42, 1153– 1161. Shank R. P. and Reife R. A. (1999) Agents for the treatment of epilepsy, in Quick Look: Pharmacology (Raffa R., ed), in press. Fence Creek Publishing, Lyndell, Pennsylvania. Shank R. P., Bennett G. S., Freytag S. O., and Campbell G. L. (1985) Pyruvate carboxylase: an astrocyte-specific enzyme implicated in the replenishment of amino acid neurotransmitter pools. Brain Res. 329, 364 –367. Shank R. P., Baldy W. J., and Ash C. W. (1989) Glutamine and 2-oxoglutarate as metabolic precursors of the transmitter pools of glutamate and GABA: correlation of regional uptake by rat brain synaptosomes. Neurochem. Res. 4, 371–376. Sheikh S. N. and Martin D. L. (1998) Elevation of brain GABA levels with vigabatrin (gamma-vinyl GABA) differentially affects GAD65 and GAD67 expression in various regions of rat brain. J. Neurosci. Res. 52, 736 –741.

J. Neurochem., Vol. 73, No. 4, 1999

1342

H. S. WAAGEPETERSEN ET AL.

Sihra T. S. and Nicholls D. G. (1987) g-Aminobutyrate can be released exocytotically from guinea-pig cerebral cortical synaptosomes. J. Neurochem. 49, 261–267. Sloviter R. S., Dichter M. A., Rachinsky T. L., Dean E., Goodman J. H., Sollas A. L., and Martin D. L. (1996) Basal expression and induction of glutamate decarboxylase and GABA in excitatory granule cells of the rat and monkey hippocampal dentate gyrus. J. Comp. Neurol. 373, 593– 618. Sokoloff L., Reivich M., Kennedy C., DesRosiers M. H., Patlak C. S., Pettigrew K. D., Sakurada O., and Shinobara M. (1977) The [14C]deoxyglucose method for the measurement of local cerebral glucose utilization: theory, procedure, and normal values in the conscious and anesthetized albino rat. J. Neurochem. 28, 897–916. Sonnewald U., Westergaard N., Hassel B., Muller T. B., Unsgard G., Fonnum F., Hertz L., Schousboe A., and Petersen S. B. (1993a) NMR spectroscopic studies of 13C acetate and 13C glucose metabolism in neocortical astrocytes: evidence for mitochondrial heterogeneity. Dev. Neurosci. 15, 351–358. Sonnewald U., Westergaard N., Schousboe A., Svendsen J. S., Unsgård G., and Petersen S. P. (1993b) Direct demonstration by [13C]NMR spectroscopy that glutamine from astrocytes is a precursor for GABA synthesis in neurons. Neurochem. Int. 1, 19 –29. Sonnewald U., Hertz L., and Schousboe A. (1998) Mitochondrial heterogeneity in the brain at the cellular level. J. Cereb. Blood Flow Metab. 18, 231–237. Tapia R. (1975) Biochemical pharmacology in CNS, in Handbook of Psychopharmacology, Vol. 4 (Iversen L. L., Iversen S. D., and Snyder S. H., eds), pp. 1–58. Plenum Press, New York. Waagepetersen H. S., Bakken I. J., Larsson O. M., Sonnewald U., and Schousboe A. (1998a) Metabolism of lactate in cultured GABAergic neurons studied by 13C-NMR spectroscopy. J. Cereb. Blood Flow Metab. 18, 109 –117. Waagepetersen H. S., Bakken I. J., Larsson O. M., Sonnewald U., and Schousboe A. (1998b) Comparison of lactate and glucose metabolism in cultured neocortical neurons and astrocytes using 13C NMR spectroscopy. Dev. Neurosci. 20, 310 –320. Waagepetersen H. S., Sonnewald U., Larsson O. M., and Schousboe A. (1999a) Synthesis of vesicular GABA from glutamine involves TCA cycle metabolism in neocortical neurons. J. Neurosci. Res. 57, 342–349.

J. Neurochem., Vol. 73, No. 4, 1999

Waagepetersen H. S., Sonnewald U., Larsson O. M., and Schousboe A. (1999b) Compartmentation of TCA cycle metabolism in cultured neocortical neurons revealed by 13C MR spectroscopy. Neurochem. Int. (in press). Westergaard N., Larsson O. M., Jensen B., and Schousboe A. (1992) Synthesis and release of GABA in cerebral cortical neurons cocultured with astrocytes from cerebral cortex or cerebellum. Neurochem. Int. 20, 567–575. Westergaard N., Sonnewald U., Petersen S. B., and Schousboe A. (1995a) Glutamate and glutamine metabolism in cultured GABAergic neurons studied by 13C NMR spectroscopy may indicate compartmentation and mitochondrial heterogeneity. Neurosci. Lett. 185, 24 –28. Westergaard N., Sonnewald U., and Schousboe A. (1995b) Metabolic trafficking between neurons and astrocytes: the glutamate/glutamine cycle revisited. Dev. Neurosci. 17, 203–211. Wolff J. R., Joo F., and Dames W. (1978) Plasticity in dendrites shown by continuous GABA administration in superior cervical ganglion of adult rat. Nature 274, 72–74. Wood J. D. and Kurylo E. (1980) Drug induced changes in the GABA content of nerve terminals determined using a synaptosomal model. Brain Res. Bull. 5, S925–S927. Wu J.-Y., Matsuda T., and Roberts E. (1973) Purification and characterization of glutamate decarboxylase from mouse brain. J. Biol. Chem. 248, 3029 –3034. Yasumi M., Sato K., Shimada S., Nishimura M., and Tohyama M. (1997) Regional distribution of GABA transporter 1 (GAT1) mRNA in the rat brain: comparison with glutamic acid decarboxylase67 (GAD67) mRNA localization. Brain Res. Mol. Brain Res. 44, 205–218. Yu A. C. H., Drejer J., Hertz L., and Schousboe A. (1983) Pyruvate carboxylase activity in primary cultures of astrocytes and neurons. J. Neurochem. 41, 1484 –1487. Yu A. C. H., Fisher T. E., Hertz E., Tildon J. T., Schousboe A., and Hertz L. (1984) Metabolic fate of [14C]-glutamine in mouse cerebral neurons in primary cultures. J. Neurosci. Res. 11, 351–357. Yudkoff M., Zaleska M. M., Nissim I., Nelson D., and Erecin´ska M. (1989) Neuronal glutamine utilization: pathways of nitrogen transfer studied with [15N]glutamine. J. Neurochem. 53, 632– 640.