Amino acid neurotransmitters in the CNS. Relationships between net uptake and exchange in rat brain synaptosomes

Neuronal vs glial glutamate uptake: Resolving the conundrum

Neither normal brain function nor the pathological processes involved in neurological diseases can be adequately understood without knowledge of the release, uptake and metabolism of glutamate. The reason for this is that glutamate (a) is the most abundant amino acid in the brain, (b) is at the cross-roads between several metabolic pathways, and (c) serves as the major excitatory neurotransmitter. In fact most brain cells express glutamate receptors and are thereby influenced by extracellular glutamate. In agreement, brain cells have powerful uptake systems that constantly remove glutamate from the extracellular fluid and thereby limit receptor activation. It has been clear since the 1970s that both astrocytes and neurons express glutamate transporters. However the relative contribution of neuronal and glial transporters to the total glutamate uptake activity, however, as well as their functional importance, has been hotly debated ever since. The present short review provides (a) an overview of what we know about neuronal glutamate uptake as well as an historical description of how we got there, and (b) a hypothesis reconciling apparently contradicting observations thereby possibly resolving the paradox.

Maintaining the presynaptic glutamate supply for excitatory neurotransmission

Glutamate released from synapses during excitatory neurotransmission must be rapidly recycled to maintain neuronal communication. This review evaluates data from physiological experiments at hippocampal CA3 to CA1 synapses and the calyx of Held synapse in the brainstem to analyze quantitatively the rates of release and resupply of glutamate required to sustain neurotransmission. We calculate that, without efficient recycling, the presynaptic glutamate supply will be exhausted within about a minute of normal synaptic activity. We also discuss replenishment of the presynaptic pool by diffusion from the soma, direct uptake of glutamate back into the presynaptic terminal, and uptake of glutamate precursor molecules. Diffusion of glutamate from the soma is calculated to be fast enough to resupply presynaptic glutamate in the hippocampus but not at the calyx of Held. However, because the somatic cytoplasm will also quickly run out of glutamate and synapses can function continually even if the presynaptic axon is severed, mechanisms other than diffusion must be present to resupply glutamate for release. Direct presynaptic uptake of glutamate is not present at the calyx of Held but may play a role in glutamate recycling in the hippocampus. Alternatively, glutamine or tricarboxylic acid cycle intermediates released from glia can serve as a precursor for glutamate in synaptic terminals, and we calculate that the magnitude of presynaptic glutamine uptake is sufficient to supply enough glutamate to sustain neurotransmission. The nature of these mechanisms, their relative abundance, and the co-ordination between them remain areas of intensive investigation.

EAAT2 (GLT-1; slc1a2) glutamate transporters reconstituted in liposomes argues against heteroexchange being substantially faster than net uptake

The EAAT2 glutamate transporter, accounts for >90% of hippocampal glutamate uptake. Although EAAT2 is predominantly expressed in astrocytes, ∼10% of EAAT2 molecules are found in axon terminals. Despite the lower level of EAAT2 expression in glutamatergic terminals, when hippocampal slices are incubated with low concentration of d-aspartate (an EAAT2 substrate), axon terminals accumulate d-aspartate as quickly as astroglia. This implies an unexplained mismatch between the distribution of EAAT2 protein and of EAAT2-mediated transport activity. One hypothesis is that (1) heteroexchange of internal substrate with external substrate is considerably faster than net uptake and (2) terminals favor heteroexchange because of high levels of internal glutamate. However, it is currently unknown whether heteroexchange and uptake have similar or different rates. To address this issue, we used a reconstituted system to compare the relative rates of the two processes in rat and mice. Net uptake was sensitive to changes in the membrane potential and was stimulated by external permeable anions in agreement with the existence of an uncoupled anion conductance. By using the latter, we also demonstrate that the rate of heteroexchange also depends on the membrane potential. Additionally, our data further suggest the presence of a sodium leak in EAAT2. By incorporating the new findings in our previous model of glutamate uptake by EAAT2, we predict that the voltage sensitivity of exchange is caused by the voltage-dependent third Na(+) binding. Further, both our experiments and simulations suggest that the relative rates of net uptake and heteroexchange are comparable in EAAT2.

Biochemical characteristics of the gamma-aminobutyric acid system in the insulinoma cell lines HIT-T15, RIN-m5F, betaTC3, and comparison with rat brain

BACKGROUND

gamma-aminobutyric acid (GABA) is the most abundant inhibitory neurotransmitter in the mammalian brain. Both GABA and its synthesizing enzyme, L-glutamate decarboxylase (GAD), are also present in the insulin-secreting pancreatic beta cells, in which its physiologic role is unclear. We have studied several aspects of the GABA system in the insulinoma cell lines HIT-T15, RIN-m5F, and betaTC3 in comparison with rat brain tissue.

METHODS

Insulinoma cell lines and embryonic rat brain cortex neurons were cultured. GAD activity was determined by a radioenzymatic method and the presence of GAD(67) protein was assessed by immunocytochemistry. Amino acid content and the effect of different conditions on the release of endogenous GABA were measured by HPLC and fluorometric detection after o-phthaldialdehyde derivatization. [3H]GABA was used for measuring the uptake of the amino acid in the insulinoma cultures and in rat forebrain synaptosomes.

RESULTS

The three insulinoma lines possess GABA and GAD activity at levels of approximately 20% compared with adult rat brain cortex. Dissimilar from the latter, in insulinoma cultures enzyme activity was not enhanced by addition of an excess of the coenzyme pyridoxal-5'-phosphate. Immunocytochemical visualization of GAD showed that the cells in both neuronal cultures and insulinoma lines were GAD(67)-positive, similar to Purkinje cell somata of adult rat cerebellar cortex. [3H]GABA uptake in the cell lines was approximately 10% of that in rat forebrain synaptosomes and showed less ionic and temperature dependence. In both cultured cerebral neurons and RINm5F cells, the addition of arginine induced the release of GABA, whereas neither high K(+) concentration nor glucose had any effect.

CONCLUSIONS

The insulinoma cell lines studied possess the same GAD(67) form of the enzyme present in brain. RIN line cells are capable of transporting glutamate. In these cells as well as in cultured cortical neurons, arginine stimulates the release of GABA and glutamate probably as the result of its electrogenic transport. Insulinoma cell lines may therefore be useful to study GABA metabolism and function in pancreatic beta cells.

Glutamate uptake

Brain tissue has a remarkable ability to accumulate glutamate. This ability is due to glutamate transporter proteins present in the plasma membranes of both glial cells and neurons. The transporter proteins represent the only (significant) mechanism for removal of glutamate from the extracellular fluid and their importance for the long-term maintenance of low and non-toxic concentrations of glutamate is now well documented. In addition to this simple, but essential glutamate removal role, the glutamate transporters appear to have more sophisticated functions in the modulation of neurotransmission. They may modify the time course of synaptic events, the extent and pattern of activation and desensitization of receptors outside the synaptic cleft and at neighboring synapses (intersynaptic cross-talk). Further, the glutamate transporters provide glutamate for synthesis of e.g. GABA, glutathione and protein, and for energy production. They also play roles in peripheral organs and tissues (e.g. bone, heart, intestine, kidneys, pancreas and placenta). Glutamate uptake appears to be modulated on virtually all possible levels, i.e. DNA transcription, mRNA splicing and degradation, protein synthesis and targeting, and actual amino acid transport activity and associated ion channel activities. A variety of soluble compounds (e.g. glutamate, cytokines and growth factors) influence glutamate transporter expression and activities. Neither the normal functioning of glutamatergic synapses nor the pathogenesis of major neurological diseases (e.g. cerebral ischemia, hypoglycemia, amyotrophic lateral sclerosis, Alzheimer's disease, traumatic brain injury, epilepsy and schizophrenia) as well as non-neurological diseases (e.g. osteoporosis) can be properly understood unless more is learned about these transporter proteins. Like glutamate itself, glutamate transporters are somehow involved in almost all aspects of normal and abnormal brain activity.

Fast removal of synaptic glutamate by postsynaptic transporters

Glutamate transporters are believed to remove glutamate from the synaptic cleft only slowly because they cycle slowly. However, we show that when glutamate binds to postsynaptic transporters at the cerebellar climbing fiber synapse, it evokes a conformation change and inward current that reflect glutamate removal from the synaptic cleft within a few milliseconds, a time scale much faster than the overall cycle time. Contrary to present models, glutamate removal does not require binding of an extracellular proton, and the time course of transporter anion conductance activation differs from that of glutamate removal. The charge movement associated with glutamate removal is consistent with the majority of synaptically released glutamate being removed from the synaptic cleft by postsynaptic transporters.

Multiple types of binding sites for atrial natriuretic peptide in rat olfactory bulb membranes and synaptosomes

The binding of atrial natriuretic peptide (ANP) to rat olfactory bulb membranes and synaptosomes was examined. [125I]ANP (rat, 99-126) bound specifically to a single class of binding site on olfactory bulb membrane preparation with dissociation constant (Kd) of 106 pM and maximum binding capacity (Bmax) of 13.6 fmol/mg protein. Comparable results were obtained when the binding was characterized using displacement and kinetic experiments. The ring deleted analog of ANP, C-ANP (rat, 4-23) displaced [125I]ANP only minimally from its binding site in the membrane preparation. Saturation, displacement and blocking experiments on [125I]ANP binding to rat olfactory bulb synaptosomes revealed the presence of two distinct binding sites. Simultaneous analysis of homogeneous and heterogeneous displacement curves and blocking experiments revealed the quantitative characteristics of these receptors to be: Kd1 = 44 pM, Bmax1 = 42 fmol/mg protein and Kd2 = 1050 pM, Bmax2 = 173 fmol/mg protein, for the high and low affinity binding sites, respectively. Kinetic experiments further confirmed the differences between the receptors present in the membranes and the synaptosomes preparations. The demonstration of multiple ANP binding sites in olfactory bulb synaptosomes but not membrane preparations raises the possibility of a particular function of ANP in nerve terminals.

Cerebral aspartate utilization: near-equilibrium relationships in aspartate aminotransferase reaction

The pathways of nitrogen transfer from 50 microM [15N]aspartate were studied in rat brain synaptosomes and cultured primary rat astrocytes by using gas chromatography-mass spectrometry technique. Aspartate was taken up rapidly by both preparations, but the rates of transport were faster in astrocytes than in synaptosomes. In synaptosomes, 15N was incorporated predominantly into glutamate, whereas in glial cells, glutamine and other 15N-amino acids were also produced. In both preparations, the initial rate of N transfer from aspartate to glutamate was within a factor of 2-3 of that in the opposite direction. The rates of transamination were greater in synaptosomes than in astrocytes. Omission of glucose increased the formation of [15N]-glutamate in synaptosomes, but not in astrocytes. Rotenone substantially decreased the rate of transamination. There was no detectable incorporation of 15N from labeled aspartate to 6-amino-15N-labeled adenine nucleotides during 60-min incubation of synaptosomes under a variety of conditions; however, such activity could be demonstrated in glial cells. The formation of 15N-labeled adenine nucleotides was marginally increased by the presence of 1 mM aminooxyacetate, but was unaffected by pretreatment with 1 mM 5-amino-4-imidazolecarboxamide ribose. It is concluded that (1) aspartate aminotransferase is near equilibrium in both synaptosomes and astrocytes under cellular conditions, but the rates of transamination are faster in the nerve endings; (2) in the absence of glucose, use of amino acids for the purpose of energy production increases in synaptosomes, but may not do so in glial cells because the latter possess larger glycogen stores; and (3) nerve endings have a very limited capacity for salvage of the adenine nucleotides via the purine nucleotide cycle.

Ischemia-induced accumulation of extracellular amino acids in cerebral cortex, white matter, and cerebrospinal fluid

In a global model of brain ischemia, accumulation of amino acids was studied in the extracellular space of the auditory cortex and the internal capsule using microdialysis, and in CSF of halothane anesthetized cats. In both brain regions, blood flow determined by hydrogen clearance decreased below 10 ml/100 g/min after extracranial multiple-vessel occlusion, and extracellular potassium activity (Ke) measured in the dialysate increased significantly. A delayed rise in Ke was observed in CSF. In contrast, ischemic amino acid accumulation differed markedly between the two brain regions investigated. In cortex, transmitter amino acids glutamate, aspartate, and gamma-aminobutyric acid (GABA) rose almost immediately after onset of ischemia, and increased 30-, 25-, and 250-fold, respectively, after 2 h of ischemia. The nontransmitter amino acids taurine, alanine, and serine increased 10-, seven-, and fourfold, respectively, whereas glutamine and essential amino acids (valine, phenylalanine, isoleucine, and leucine) increased only 1.5-fold. In the internal capsule, increases in amino acids, if any, were delayed and much smaller than in cortex. The largest alteration was a fivefold elevation of GABA. In CSF, changes in amino acids were small and comparable to those in the internal capsule. Our results demonstrate that ischemia-induced extracellular amino acid accumulation is a well localized phenomenon restricted to gray matter structures that possess release and reuptake systems for these substances. We assume that amino acids diffuse slowly into adjacent while matter structures, and into CSF.

Effects of ethanol on Na+-dependent amino acid uptake: dependence on rat age and Na+, K+-ATPase activity

Acute effects of ethanol on Na(+)-dependent transport of gamma-aminobutyric acid (GABA) and glutamic acid (GLU) were investigated in crude synaptosomal preparations from rat cerebral cortex. In experiments with 30-40-day-old (peripubertal) rats, the overall dose responses of the GABA and GLU transport systems to ethanol were biphasic. Stimulation was observed at ethanol concentrations (40-160 mM) relevant to intoxication. Inhibition was observed at higher concentrations of ethanol. The stimulatory phase of the dose response was not observed in 60-100-day-old (adult) rats. In preparations from peripubertal rats, other alcohols also had biphasic dose response curves with stimulation at low alcohol concentrations. The relative efficacy of the different alcohols appeared to correlate with the relative membrane-buffer partition coefficient. In synaptosomal membrane vesicles, where artificial ion concentration gradients rather than Na+,K(+)-ATPase activity provide the driving force for uptake, ethanol did not stimulate GABA uptake. In direct measures of Na+,K(+)-ATPase activity, both Rb+ uptake and ATP hydrolysis were enhanced by 80 mM ethanol. We conclude that stimulation of Na(+)-dependent uptake of amino acids by ethanol was secondary to enhanced Na+,K(+)-ATPase activity and may be associated with a specific developmental stage in the rat.

Displacement of endogenous glutamate with D-aspartate: an effective strategy for reducing the calcium-independent component of glutamate release from synaptosomes

D-aspartate was used in the present study to partially deplete the cytosolic pool of glutamate, which is released independent of extracellular Ca2+, prior to measuring the K(+)-evoked release of this endogenous acidic amino acid from rat hippocampal mossy fiber synaptosomes. This pretreatment is known to be an effective method for substantially reducing the Ca(2+)-independent component of glutamate release. The rate of glutamate efflux is dependent on the concentration of sodium ions in the external medium and can be stimulated by exposure of hippocampal mossy fiber synaptosomes to external D-aspartate (50 microM). Following the partial displacement of this cytosolic pool of glutamate with D-aspartate, the K(+)-evoked release of the residual, presumably vesicular, pool of endogenous glutamate has a strict requirement for external calcium and is highly dependent on the extent to which depolarization elevates the level of free cytosolic calcium. It is concluded that the protocol described in this study for the displacement of cytosolic glutamate with D-aspartate provides a useful alternative method of controlling for the Ca(2+)-independent component of glutamate release in synaptosomal preparations.

Interrelationship between glutamate and membrane-bound ATPases in nerve cells

Plasma membrane potential generated by Na+, K(+)-ATPase provides the driving force for high-affinity, Na(+)-dependent uptake of glutamate into the cytoplasm of glutamatergic nerve endings and glial cells. Ca2(+)-calmodulin-dependent ATPase in the plasma membrane and Ca2(+)-ATPase in the endoplasmic reticulum influence the intracellular [Ca2+] and, therefore, the exocytotic release of neurotransmitter glutamate. The membrane potential across the membrane of the synaptic vesicles, generated by a H(+)-ATPase, provides the driving force for synaptic vesicular uptake of glutamate as well as that of GABA and glycine. Hypoxia and ischemia lead to release of glutamate, perhaps in consequence of an increased endogenous pool of glutamate and/or lack of substrate (ATP) for the ATPases. This release, rather than being exocytotic, is believed to result mainly from a reversal of the Na(+)-dependent high-affinity glutamate transporter in the plasma membrane.

Neuronal glutamine utilization: pathways of nitrogen transfer studied with [15N]glutamine

Gas chromatography-mass spectrometry was used to evaluate the metabolism of [15N]glutamine in isolated rat brain synaptosomes. In the presence of 0.5 mM glutamine, synaptosomes accumulated this amino acid to a level of 25-35 nmol/mg protein at an initial rate greater than 9 nmol/min/mg of protein. The metabolism of [15N]glutamine generated 15N-labelled glutamate, aspartate, and gamma-aminobutyric acid (GABA). An efflux of both [15N]glutamate and [15N]aspartate from synaptosomes to the medium was observed. Enrichment of 15N in alanine could not be detected because of a limited pool size. Elimination of glucose from the incubation medium substantially increased the rate and amount of [15N]aspartate formed. It is concluded that: (1) With 0.5 mM external glutamine, the glutaminase reaction, and not glutamine transport, determines the rate of metabolism of this amino acid. (2) The primary route of glutamine catabolism involves aspartate aminotransferase which generates 2-oxoglutarate, a substrate for the tricarboxylic acid cycle. This reaction is greatly accelerated by the omission of glucose. (3) Glutamine has preferred access to a population of synaptosomes or to a synaptosomal compartment that generates GABA. (4) Synaptosomes maintain a constant internal level of glutamate plus aspartate of about 70-80 nmol/mg protein. As these amino acids are produced from glutamine in excess of this value, they are released into the medium. Hence synaptosomal glutamine and glutamate metabolism are tightly regulated in an interrelated manner.

Endogenous amino acid release from cultured cerebellar neuronal cells: effect of tetanus toxin on glutamate release

Endogenous amino acid release was measured in developing cerebellar neuronal cells in primary culture. In the presence of 25 mM K+ added to the culture medium, cerebellar cells survived more than 3 weeks and showed a high level of differentiation. These cultures are highly enriched in neurons, and electron-microscopic observation of these cells after 12 days in vitro (DIV) confirmed the presence of a very large proportion of cells with the morphological characteristics of granule cells, making synapses containing many synaptic vesicles. Synaptogenesis was also confirmed by immunostaining the cells with antisera against synapsin I and synaptophysin, two proteins associated with synaptic vesicles. From these cultures, endogenous glutamate release stimulated by 56 mM K+ was already detected after only a few days in culture, the maximal release value (1,579% increase over basal release) being reached after 10 DIV. In addition to that of glutamate, the release of aspartate, asparagine, alanine, and, particularly, gamma-aminobutyric acid (GABA) was stimulated by 56 mM K+ after 14 DIV, but to a lesser extent. No increase in serine, glutamine, taurine, or tyrosine release was observed during K+ depolarization. The effect of K+ on amino acid release was strictly Ca2+-dependent. Stimulation of the cells with veratridine resulted in a qualitatively similar effect on endogenous amino acid release. In the absence of Ca2+, 30% of the veratridine effect persisted. The Ca2+-dependent release was quantitatively similar after stimulation by veratridine and K+. Treatment of cerebellar cells with tetanus toxin (5 micrograms/ml) for 24 h resulted in a total inhibition of the Ca2+-dependent component of the glutamate release evoked by K+ or veratridine. It is concluded that glutamate is the main amino acid neurotransmitter of cerebellar cells developed in primary culture under the present conditions and that glutamate is probably mainly released through the exocytosis of synaptic vesicles.

Glucose and synaptosomal glutamate metabolism: studies with [15N]glutamate

The metabolism of [15N]glutamate was studied with gas chromatography-mass spectrometry in rat brain synaptosomes incubated with and without glucose. [15N]Glutamate was taken up rapidly by the preparation, reaching a steady-state level in less than 5 min. 15N was incorporated predominantly into aspartate and, to a much lesser extent, into gamma-aminobutyrate. The amount of [15N]ammonia formed was very small, and the enrichment of 15N in alanine and glutamine was below the level of detection. Omission of glucose substantially increased the rate and amount of [15N]aspartate generated. It is proposed that in synaptosomes (a) the predominant route of glutamate nitrogen disposal is through the aspartate aminotransferase reaction; (b) the aspartate aminotransferase pathway generates 2-oxoglutarate, which then serves as the metabolic fuel needed to produce ATP; (c) utilization of glutamate via transamination to aspartate is greatly accelerated when flux through the tricarboxylic acid cycle is diminished by the omission of glucose; (d) the metabolism of glutamate via glutamate dehydrogenase in intact synaptosomes is slow, most likely reflecting restriction of enzyme activity by some unknown factor(s), which suggests that the glutamate dehydrogenase reaction may not be near equilibrium in neurons; and (e) the activities of alanine aminotransferase and glutamine synthetase in synaptosomes are very low.

Glucose deprivation depolarizes plasma membrane of cultured astrocytes and collapses transmembrane potassium and glutamate gradients

Primary cultures of astrocytes were used to investigate the effects of glucose deprivation on plasma membrane potential, on the respiration and on the energy status of these cells. Plasma membrane potential, as monitored with a cyanine dye, 3,3'-diethylthiadicarbocyanine, hyperpolarized by about 100% when glucose was added to substrate-deprived cells. The effect of glucose was prevented by iodoacetate or ouabain. In the absence of glucose, cellular adenosine triphosphate/adenosine diphosphate ratio was extensively reduced and pyruvate was unable either to restore energy status or to hyperpolarize the plasma membrane of astrocytes, although it was the preferential substrate for mitochondria within the cells. Glucose deprivation and inhibition of glycolysis or respiration in the presence of glucose caused dramatic decrease in transmembrane potassium ion and L-glutamate gradients. The gradients were not restored in the presence of pyruvate. Thus, aerobic glycolysis, rather than oxidation of pyruvate, is required to maintain maximal plasma membrane potential, adenosine triphosphate/adenosine diphosphate ratios as well as K+ and L-glutamate gradients. This evidence, together with the unresponsiveness of astrocyte respiration to ouabain, indicates a functional dissociation between energy dissipation at the plasma membrane and mitochondrial synthesis of adenosine triphosphate. The results are discussed with regard to the vulnerability of glia at low levels of blood glucose and the contribution of glial dysfunction to development of hypoglycaemic encephalopathy.

Occurrence of a large Ca2+-independent release of glutamate during anoxia in isolated nerve terminals (synaptosomes)

Isolated rat cerebral cortical synaptosomes made anoxic by addition of cyanide developed an inhibition of the Ca2+-dependent release of glutamate 2 min after the addition of the metabolic inhibitor when the intrasynaptosomal ATP/ADP ratio decreased below 1.7. In contrast, cyanide induced a continuous efflux of glutamate through a Ca2+-independent pathway that accounted for the release of 25% of total intrasynaptosomal glutamate in 5 min. The results suggest that a Ca2+-independent release of glutamate could be implicated in the neurotoxic action of this amino acid during anoxia.