Multiple compartments with different metabolic characteristics are involved in biosynthesis of intracellular and released glutamine and citrate in astrocytes

Branched-chain amino acids and neurotransmitter metabolism: expression of cytosolic branched-chain aminotransferase (BCATc) in the cerebellum and hippocampus

In the brain, catabolism of the branched-chain amino acids (BCAAs) provides nitrogen for the synthesis of glutamate and glutamine. Glutamate is formed through transfer of an amino group from BCAA to alpha-ketoglutarate in reaction catalyzed by branched-chain aminotransferases (BCAT). There are two isozymes of BCAT: cytosolic BCATc, which is found in the nervous system, ovary, and placenta, and mitochondrial BCATm, which is found in all organs except rat liver. In cell culture systems, BCATc is found only in neurons and developing oligodendrocytes, whereas BCATm is the isoform in astroglia. In this study, we used immunohistochemistry to examine the distribution of BCATc in the rat brain, focusing on the well-known neural architecture of the cerebellum and hippocampus. We show that BCATc is expressed only in neurons in the adult rat brain. In glutamatergic neurons such as granule cells of the cerebellar cortex and of the dentate gyrus, BCATc is localized to axons and nerve terminals. In contrast, in GABAergic neurons such as cerebellar Purkinje cells and hippocampal pyramidal basket cells, BCATc is concentrated in cell bodies. A common function for BCATc in these neurotransmitter systems may be to modulate amounts of glutamate available either for release as neurotransmitter or for use as precursor for synthesis of GABA. Particularly striking in our findings is the strong expression of BCATc in the mossy fiber pathway of the hippocampal formation. This result is discussed in light of the effectiveness of the anticonvulsant drug gabapentin, which is a specific inhibitor of BCATc.

Cerebral glucose metabolism and the glutamine cycle as detected by in vivo and in vitro 13C NMR spectroscopy

We review briefly 13C NMR studies of cerebral glucose metabolism with an emphasis on the roles of glial energetics and the glutamine cycle. Mathematical modeling analysis of in vivo 13C turnover experiments from the C4 carbons of glutamate and glutamine are consistent with: (i) the glutamine cycle being the major cerebral metabolic route supporting glutamatergic neurotransmission, (ii) glial glutamine synthesis being stoichiometrically coupled to glycolytic ATP production, (iii) glutamine serving as the main precursor of neurotransmitter glutamate and (iv) glutamatergic neurotransmission being supported by lactate oxidation in the neurons in a process accounting for 60-80% of the energy derived from glucose catabolism. However, more recent experimental approaches using inhibitors of the glial tricarboxylic acid (TCA) cycle (trifluoroacetic acid, TFA) or of glutamine synthase (methionine sulfoximine, MSO) reveal that a considerable portion of the energy required to support glutamine synthesis is derived from the oxidative metabolism of glucose in the astroglia and that a significant amount of the neurotransmitter glutamate is produced from neuronal glucose or lactate rather than from glial glutamine. Moreover, a redox switch has been proposed that allows the neurons to use either glucose or lactate as substrates for oxidation, depending on the relative availability of these fuels under resting or activation conditions, respectively. Together, these results suggest that the coupling mechanisms between neuronal and glial metabolism are more complex than initially envisioned.

Intracellular metabolic compartmentation assessed by 13C magnetic resonance spectroscopy

Our understanding of the brain has developed from the theory that it is one continuous cell to the knowledge that there are many brain cells originally termed neurons and, furthermore to the discovery of glial cells and their multiple functions. Thus, an increasing complexity was unraveled and we have not reached a complete understanding of the phenomenon which comprises the compartmentation of metabolic pathways and metabolites. This is an important principle needed to fully understand the metabolic processes of the brain. At the cellular level this concept is well established whereas intracellular compartmentation has yet to be explored. Using magnetic resonance spectroscopy (MRS) for analysis of isotopomer composition combined with quantification of amino acid contents it is possible to construct models that describe intracellular compartmentation. Results of studies of cultures of astrocytes and neurons incubated in media containing [U- 13C]glutamate in the presence or absence of thiopental may be used to propose an intracellular three compartment model of mitochondrial function. Due to the experimental paradigm only certain aspects of metabolism can be described. The present model consists of compartments assigned to CO(2) production, glutamate synthesis from ketoglutarate and finally synthesis of a four-carbon metabolite which is shuttled between compartments. It is likely that metabolism may be far more complex than this and we are only beginning to glimpse some aspects of compartmentation at the cellular level.

Astrocytic exocytosis vesicles and glutamate: A high-resolution immunofluorescence study

Physiological evidence has demonstrated that cultured astrocytes can release glutamate via Ca2+-dependent mechanisms. Also, glutamate released from astrocytes in the hippocampal slice interferes with synaptic neurotransmission. Since these observations suggest vesicular glutamate release from astrocytes, the presence of glutamate-containing exocytosis vesicles was investigated. We applied immunofluorescence techniques combined with high-performance deconvolution microscopy, which yields a resolution of <200 nm and permits evaluation of double labeling in individual vesicles. Using a well-characterized anti-glutamate antiserum and parameters minimizing fixative-induced autofluorescence, glutamate-immunoreactive (ir) puncta were found all over the astrocyte but were conspicuously dense at the cell boundary and in filopodia. Images were very similar with antibodies against vesicular glutamate transporters (vGluT1 and vGluT2). Labeling for the exocytosis markers rab3, synaptophysin, or synaptobrevin was also punctate, particularly dense at the cell boundary, but disappearing toward the perinuclear region. Sections of the cell boundary were delineated by rab3 immunoreactivity. In double-labeled cells, vesicular colocalization of glutamate and any of the exocytosis markers was frequent in filopodia and at the cell boundary. Within the cell, single-labeled glutamate-ir vesicles prevailed; double-labeled vesicles were infrequently present. By resolving single vesicles, in cultured astrocytes we visualize glutamate-containing vesicles, vesicles displaying vGluT1 or vGluT2, and exocytosis vesicles displaying glutamate-ir. This may provide the morphological correlate of Ca2+-dependent glutamate release from astrocytes, possibly occurring at defined sections of the cell membrane and at filopodia. However, since vGluTs and exocytosis markers are classically restricted to nerve terminals in the CNS, glutamate release from astrocytes in the CNS remains to be studied.

Compartmentation of glutamine, glutamate, and GABA metabolism in neurons and astrocytes: functional implications

The classical concept regarding compartmentation of brain metabolism pertinent to the two neurotransmitter amino acids, glutamate and GABA (gamma-aminobutyrate), operates with different pools of glutamate and glutamine in different cell types, that is, pools that have different sizes and turnover rates. As a result of more sophisticated technology (e.g., nuclear magnetic resonance spectroscopy and mass spectrometry used in relation to cultured neurons and astrocytes), a more complex scenario is emerging. Hence, both neurons and astrocytes exhibit a compartmentalized metabolism that very likely relates to individual cells containing mitochondrial populations having different metabolic roles. Models for this in neurons and astrocytes, respectively, are presented. The functional implications of this for the homeostatic mechanisms regulating the levels of neurotransmitter glutamate and GABA are discussed in relation to development of therapeutic strategies for neurological disorders in which these transmitters are believed to play important roles.

Differential roles of alanine in GABAergic and glutamatergic neurons

Studies in different preparations of neurons and astrocytes of alanine transport and activities of its metabolizing enzyme alanine aminotransferase have led to the proposal that this amino acid is preferentially synthesized in astrocytes and transferred from the astrocytic to the neuronal compartment. From a functional point of view this may well be the case in a GABAergic synapse since theoretically alanine can be utilized as a metabolic fuel in GABAergic neurons where the GABA shunt is operating. Thus, a metabolic scheme is proposed, according to which alanine catabolism is coupled to the TCA cycle where the GABA shunt replaces the alpha-ketoglutarate dehydrogenase/succinyl CoA synthetase reactions. In a glutamatergic synapse in which the large demand for synthesis of neurotransmitter glutamate leads to a large production of ammonia, it is possible that alanine could play a completely different role. Hence, experimental evidence is reviewed suggesting that alanine may serve as a carrier of ammonia nitrogen from the neuronal compartment to the astrocytic compartment using a flux of lactate in the opposite direction to account for transfer of the C-3 carbon skeleton.

Perspectives on the metabolic management of epilepsy through dietary reduction of glucose and elevation of ketone bodies

Brain cells are metabolically flexible because they can derive energy from both glucose and ketone bodies (acetoacetate and beta-hydroxybutyrate). Metabolic control theory applies principles of bioenergetics and genome flexibility to the management of complex phenotypic traits. Epilepsy is a complex brain disorder involving excessive, synchronous, abnormal electrical firing patterns of neurons. We propose that many epilepsies with varied etiologies may ultimately involve disruptions of brain energy homeostasis and are potentially manageable through principles of metabolic control theory. This control involves moderate shifts in the availability of brain energy metabolites (glucose and ketone bodies) that alter energy metabolism through glycolysis and the tricarboxylic acid cycle, respectively. These shifts produce adjustments in gene-linked metabolic networks that manage or control the seizure disorder despite the continued presence of the inherited or acquired factors responsible for the epilepsy. This hypothesis is supported by information on the management of seizures with diets including fasting, the ketogenic diet and caloric restriction. A better understanding of the compensatory genetic and neurochemical networks of brain energy metabolism may produce novel antiepileptic therapies that are more effective and biologically friendly than those currently available.

Protective role of glucose-6-phosphate dehydrogenase activity in the metabolic response of C6 rat glioma cells to polyunsaturated fatty acid exposure

Polyunsaturated fatty acids (PUFAs) can influence tumor growth and migration, both in vitro and in vivo. The PUFA gamma-linolenic acid (GLA) has been reported to improve the poor prognosis associated with human gliomas, although its effects at sublethal concentrations on residual cells postsurgery are poorly understood. The study investigated the effects sublethal PUFA doses (90 or 150 microM) may have on rat C6 glioma cell energy metabolism, since an adequate energy supply is essential for cell proliferation, migration, and apoptosis. Of note was the identification of mitochondrial heterogeneity in relation to the mitochondrial membrane potential (MMP), which has been suggested but unproven in previous studies. GLA and eicosapentaenoic acid (EPA) caused significant changes in cellular fatty acid composition and increased the percentage of cells with a low MMP after a 96-h exposure period. The presence of PUFAs inhibited C6 cell proliferation and migration, although apoptosis was not induced. The protein expression and activity of glucose-6-phosphate dehydrogenase was increased after 96-h incubation with 90 microM GLA and EPA and would allow redox regulation through increased NADPH production, permitting the maintenance of adequate intracellular reduced glutathione concentrations and limiting rates of lipid peroxidation and reactive oxygen species generation. Neither NADP(+)-isocitrate dehydrogenase nor NADP(+)-malate dehydrogenase activity responded to PUFAs, suggesting it is glucose-6-phosphate dehydrogenase that is the principal source of NADPH in C6 cells. These data compliment studies showing that higher concentrations of GLA induced glioma cell death and tumor regression and suggest that GLA treatment could be useful for the inhibition of residual cell proliferation and migration after surgical removal of the tumor mass.

Inhibition of glutamine transport depletes glutamate and GABA neurotransmitter pools: further evidence for metabolic compartmentation

The role of glutamine and alanine transport in the recycling of neurotransmitter glutamate was investigated in Guinea pig brain cortical tissue slices and prisms, and in cultured neuroblastoma and astrocyte cell lines. The ability of exogenous (2 mm) glutamine to displace 13C label supplied as [3-13C]pyruvate, [2-13C]acetate, l-[3-13C]lactate, or d-[1-13C]glucose was investigated using NMR spectroscopy. Glutamine transport was inhibited in slices under quiescent or depolarising conditions using histidine, which shares most transport routes with glutamine, or 2-(methylamino)isobutyric acid (MeAIB), a specific inhibitor of the neuronal system A. Glutamine mainly entered a large, slow turnover pool, probably located in neurons, which did not interact with the glutamate/glutamine neurotransmitter cycle. This uptake was inhibited by MeAIB. When [1-13C]glucose was used as substrate, glutamate/glutamine cycle turnover was inhibited by histidine but not MeAIB, suggesting that neuronal system A may not play a prominent role in neurotransmitter cycling. When transport was blocked by histidine under depolarising conditions, neurotransmitter pools were depleted, showing that glutamine transport is essential for maintenance of glutamate, GABA and alanine pools. Alanine labelling and release were decreased by histidine, showing that alanine was released from neurons and returned to astrocytes. The resultant implications for metabolic compartmentation and regulation of metabolism by transport processes are discussed.

Role of specific aminotransferases in de novo glutamate synthesis and redox shuttling in the retina

In this study aminotransferase inhibitors were used to determine the relative importance of different aminotransferases in providing nitrogen for de novo glutamate synthesis in the retina. Aminooxyacetate, which inhibits all aminotransferases, blocked de novo glutamate synthesis from H(14)CO(3)(-) by more than 60%. Inhibition of neuronal cytosolic branched chain amino acid transamination by gabapentin or branched chain amino acid transport by the L-system substrate analog, 2-amino-bicyclo-(2,2,1)-heptane-2-carboxylic acid, lowered total de novo synthesis of glutamate by 30%, suggesting that branched chain amino acids may account for half of the glutamate nitrogen contributed by transamination reactions. L-cycloserine, an inhibitor of alanine aminotransferase, inhibited glutamate synthesis less than 15% when added in the presence of 5 mM pyruvate but 47% in the presence of 0.2 mM pyruvate. Although high levels of pyruvate blunted the inhibitory effectiveness of L-cycloserine, the results indicate that, under physiological conditions, alanine as well as branched chain amino acids are probably the predominant sources of glutamate nitrogen in ex vivo retinas. The L-cycloserine results were also used to evaluate activity of the malate/aspartate shuttle. In this shuttle, cytosolic aspartate (synthesized in mitochondria) generates cytosolic oxaloacetate that oxidizes cytosolic NADH via malate dehydrogenase. Because L-cycloserine inhibits cytosolic but not mitochondrial aspartate aminotransferase, L-cycloserine should prevent the utilization of aspartate but not its generation, thereby increasing levels of (14)C-aspartate. Instead, L-cycloserine caused a significant decline in (14)C-aspartate. The results suggest the possibility that shuttle activity is low in retinal Müller cells. Low malate/aspartate shuttle activity may be the molecular basis for the high rate of aerobic glycolysis in retinal Müller cells.

Elucidation of the quantitative significance of pyruvate carboxylation in cultured cerebellar neurons and astrocytes

Pyruvate carboxylation was studied in cerebellar astrocytes and granule neurons. The cells were incubated in medium containing [U-(13)C]glucose (2.5 mM) and [U-(13)C]lactate (1 mM) and varying amounts of 3-nitropropionic acid (3-NPA) plus/minus aspartate. 3-NPA alone clearly stopped tricarboxylic acid (TCA) cycle activity at the succinate dehydrogenase step in both culture types as evidenced by a buildup of succinate. Labeling of aspartate and glutamate was abolished in neurons in the presence of 3-NPA. In astrocytes, however, labeled glutamate and glutamine derived from pyruvate carboxylation was detected. Unchanged glucose and lactate metabolism in the absence of a functioning malate aspartate shuttle indicates the importance of the glycerol-3-phosphate shuttle in brain cells. To compensate for the loss of oxaloacetate in the presence of 3-NPA, unlabeled aspartate (0.25 mM) was added. In this case [1,2-(13)C] and [3,4-(13)C]aspartate were observed in neurons but not in astrocytes. This labeling pattern in aspartate occurs after a full turn of the TCA cycle and thus indicates only partial inhibition by 3-NPA in the neurons when aspartate is present. In astrocytes, however, aspartate derived from uniformly labeled pyruvate was observed clearly indicating pyruvate carboxylation. The present study has unequivocally demonstrated a quantitatively important pyruvate carboxylation in astrocytes but it was not possible to demonstrate the presence of such carboxylation in neurons. Based on the present results it may be safely concluded that neuronal pyruvate carboxylation is unlikely to be of quantitative significance.