A.C.H. Yu, L. Hertz. M.D.Norenberg, E. Sykova and S.G. Waxman (Eds.) Progress in Brain Research. Vol. 94 0 IWZ Elsevier Science Publishers B.V. All rights reserved.
213
CHAPTER 18
Nitrogen metabolism: neuronal-astroglial relationships Marc Yudkoff’, Itzhak Nissiml, Leif Hertz3, David Pleasure1 and Maria Erecinska2
’
Departments of I Pediatrics and Pharmacology, University of Pennsylvania School of Medicine, Philadelphia, PA, U.S.A.; and Department of Pharmacologv, University of Saskatchewan, Saskatoon, Canada
The glutamate-glutamine cycle: general outline The extracellular concentration of glutamic acid in the CNS is much lower than that of peripheral tissues (2 - 3 pM vs 50 pM). This is important for two reasons: (a) glutamate is the premier excitatory neurotransmitter, and a low level in the synaptic cleft amplifies the signal-to-noise ratio upon release of this amino acid from pre-synaptic terminals; and (b) excessive stimulation of certain glutamatergic neurons, particularly those bearing N-methyl-Daspartate (NMDA) receptors, can lead to neuronal injury and/or death (“excitotoxicity”). Intra-synaptic glutamate is kept low by highaffinity, Na+ -dependent re-uptake systems in neurons and astrocytes. The neuronal and glial transporters are not identical (cf. Erecinska and Silver, 1990, for review), and the precise contribution of either to uptake is uncertain, although it is likely that astroglia, which have a greater driving force for uptake (greater Na+ gradient and membrane potential, Erecinska and Silver, 1989), must figure prominently in the overall process. Astrocytes are well equipped to metabolize the glutamate they transport since they are enriched with the glutamine synthetase pathway (Norenberg and Martinez-Hernandez, 1979). Indeed, the importation of glutamate into astrocytes and subsequent formation of glutamine constitutes one limb of the “glutamate-glutamine cycle” which is completed by the export of glutamine to neurons, where it is hydrolyzed to glutamate via phosphate-
-
dependent glutaminase (Hertz, 1979; Shank and Aprison, 1981). The glutamate-glutamine cycle thereby accomplishes the two major requirements of glutamatergic neurotransmission, i.e., the rapid removal of glutamate from the synaptic cleft, and the restoration of glutamate to neurons in the form of glutamine, a non-neuroactive compound which can be shuttled safely between astrocytes and neurons. Problems associated with the glutamate-glutamine cycle Available evidence confirms that astrocytes are a major site of brain glutamine synthesis. However, other evidence indicates that these cells also may play an important role in the utilization of this amino acid. Thus, astrocytes display both the concentrative uptake of glutamine and a fairly high activity of phosphate-dependent glutaminase (Schousboe et al., 1979). The latter metabolic pathway does not appear to be an artifact of culturing in the presence of a high concentration of glutamine (Kvamme et al., 1982). Furthermore, the rate of uptake of glutamine into cultured astrocytes exceeds that noted into neurons, suggesting that there is not necessarily a net flow of glutamine from astrocytes to neurons (Hertz et al., 1980; Ramaharobandro et al., 1982). Nor are these findings restricted to astrocytes in culture: the glutaminase activity of bulkisolated astrocytes exceeds that of synaptosomes (Subbalakshmi and Murthy, 1985).
214
Our own observations of [2-'SN]glutamine metabolism in cultured astrocytes confirm that these cells do not simply synthesize glutamine. As shown in Fig. lB, when astrocytes are cultured in the presence of [2-1SN]glutamine(2 mM) ( - 50.0 atom 070 excess), the enrichment in glutamine steadily declines as this amino acid is synthesized from unlabeled sources and the ISN label is diluted. However, the concentration of glutamine (I4N + lSN) in the medium remains unchanged (Fig. lA), indicating simultaneous synthesis via glutamine synthetase and consumption through glutaminase. This is reflected also in the exponential disappearance (Fig. 1C) of the [2-1SN]glutaminefrom the medium. The rate of disappearance of glutamine (12.7 nmol/min per mg protein) (Table I) is comparable to the rate of synthesis we have measured with similar stable isotope methods ( - 9 nmol/min per mg protein) (Yudkoff et al., 1986, 1988).
32
r
1 0 ' ' 0 4
rn "
I
' '
'
6 12 16 20 24 Time (hr)
1
Time (hr)
0 4
8 12 16 20 24 Time (hr)
Fig. 1. A . The concentration of glutamine in the medium following the addition of 2 mM [2-15N]glutamine.Astrocytes maintained at 37°C in Ham's F-12 medium. B . Isotopic abundance (atom % excess) in [2-1SN]glutamineduring the incubation of the astrocytes in steady-state medium. C . The absolute concentration of [2-1SN]glutamine (nmol ISN/mg protein) in the medium. Value represents the product of isotopic abundance (atom % excess/100) and the glutamine concentration (A). (From Yudkoff et al., 1988.)
Whether net synthesis or consumption occurs is determined in part by the glutamine level of the medium, with synthesis predominating when the level of this amino acid is low and consumption predominating when it is high. As shown in Table I, an important fate of 2-N of glutamine appears to be the formation of alanine, although appreciable ISN is noted in essential amino acids, e.g., phenylalanine, indicating a capacity for the transamination of the cognate alpha-ketoacids.
Glutamine synthesis in neurons via glutaminase The above data underscore the complexity of glial metabolism of glutamine. The fact that these cells can consume glutamine does not exclude the probability that this amino acid is a precursor to neuronal glutamate. As shown in Fig. 2, we have utilized gas chromatography-mass spectrometry as an analytic tool in order to determine flux through the glutaminase pathway in synaptosomes incubated in the presence of [2-1SN]glutamine as a metabolic tracer (Yudkoff et al., 1989). The "flux", representing the sum of lSN resident in [lSN]glutamate, [ISN]aspartate and [lSN]GABA, was measured when the intra-synaptosomal [glutamine] (15 - 30 nmol/mg protein) was high enough to saturate glutaminase ( K , = 1.6-4 mM; McGeer and McGeer, 1979; Benjamin, 1981). As indicated in Table 11, flux through glutaminase is -3.5 nmol/mg protein per minute, or less than 10% of the probable glutaminase activity in synaptosomes (35 - 125 nmol/mg protein per minute; Ward and Bradford, 1979; Kvamme and Lenda, 1982). It also is evident that much of the glutamate so formed becomes transaminated to aspartate and that this process is accentuated if glucose is omitted from the incubation medium, presumably reflecting the importance of glutamine as an alternate metabolic fuel. We have noted a similar phenomenon in studies of [ lSN]glutamate metabolism (Erecinska et al., 1988). Although flux through synaptosomal glutaminase is far removed from the Vmax for this enzyme,
215 TABLE I Utilization and synthesis of glutamine by astrocytes Utilization (nmol/min per mg protein) Appearance of ISN in: Alanine Serine Phenylalanine Tyrosine Valine Leucine
12.74
Synthesis (nmol/min per mg protein)
9.01
2.01 0.26 0.54
0.04 0.07 0.16
Precursor: [2-1SN]glutamine(2 mM). Rate of utilization determined by fitting curve of medium [2-1SN]glutamine disappearance to theexpressiony = Ae-k" + Be-k2'. Values for appearance of I5N in amino acids (sum of medium and cells) determined from theequationy = A - BeCk'. (From Yudkoff et al., 1989.)
it can be stimulated markedly by depolarization. As shown in Fig. 3, when synaptosomes are incubated in the presence of veratridine, the uptake of glutamine after the initial 2 min is sharply reduced, but the total amount of [15N]glutamate and [15N]aspartate formed is similar (Erecinska et al., 1990). Thus, depolarization both reduces synaptosomal glutamine uptake, which is Na+dependent (Erecinska et al., 1990), and also stimulates flux through the glutaminase pathway. Enhanced flux results primarily from increased intra-mitochondrial Pi, which is derived from the hydrolysis of high-energy phosphate compounds during and after depolarization (Siesjo, 1978; Dagani and Erecinska, 1987). The phosphate activation curve is very steep in synaptosomes (Bradford and Ward, 1976), emphasizing the sensitivity of the enzyme to activation by phosphate, which reduces theK, (1.4 - 4 mM; Weil-Malherbe, 1969; McGeer and McGeer, 1979; Benjamin, 1981) of the enzyme for glutamine (Kovacevic and McGivan, 1983). As illustrated in Fig. 4, we found that the phosphate effect on glutaminase is modulated by calcium (Erecinska et al., 1990). In medium with high (10 mM) phosphate, calcium (1.27 mM) stimulated the formation of glutamate and aspartate from glutamine, but in the presence of low
(0.1 mM) phosphate, calcium inhibited flux through glutaminase. The likely explanation for these observations is that calcium favors the uptake of Pi into mitochondria (Fiskum and Lehninger, 1982), thereby enhancing stimulation of glutaminase. In contrast, at a low external [Pi],this effect is overshadowed by a calcium-induced reduction in free intra-mitochondria1 phosphate and a resultant inhibition of flux through glutaminase. An important observation is the rapidity with which [15N]glutamate derived from [215N]glutamine appears in the incubation medium (Fig. 5 ) , even when incubations are done in the absence of veratridine (Erecinska et al., 1990). Indeed, the glutamate probably is released into the medium even before it is taken up into the vesicular pool of this amino acid, an equilibration that occurs relatively slowly (Kauppinen et al., 1988). Furthermore, we found that the magnitude of the release of glutamate and aspartate is directly related to the external glutamine concentration, with relatively little glutamate or aspartate appearing in the medium at 0.15 mM but more at 0.5 mM [glutaminelex, (Erecinska et al., 1990; Fig. 5 ) . This finding could 2o
r
213
CHAPTER 18
Nitrogen metabolism: neuronal-astroglial relationships Marc Yudkoff’, Itzhak Nissiml, Leif Hertz3, David Pleasure1 and Maria Erecinska2
’
Departments of I Pediatrics and Pharmacology, University of Pennsylvania School of Medicine, Philadelphia, PA, U.S.A.; and Department of Pharmacologv, University of Saskatchewan, Saskatoon, Canada
The glutamate-glutamine cycle: general outline The extracellular concentration of glutamic acid in the CNS is much lower than that of peripheral tissues (2 - 3 pM vs 50 pM). This is important for two reasons: (a) glutamate is the premier excitatory neurotransmitter, and a low level in the synaptic cleft amplifies the signal-to-noise ratio upon release of this amino acid from pre-synaptic terminals; and (b) excessive stimulation of certain glutamatergic neurons, particularly those bearing N-methyl-Daspartate (NMDA) receptors, can lead to neuronal injury and/or death (“excitotoxicity”). Intra-synaptic glutamate is kept low by highaffinity, Na+ -dependent re-uptake systems in neurons and astrocytes. The neuronal and glial transporters are not identical (cf. Erecinska and Silver, 1990, for review), and the precise contribution of either to uptake is uncertain, although it is likely that astroglia, which have a greater driving force for uptake (greater Na+ gradient and membrane potential, Erecinska and Silver, 1989), must figure prominently in the overall process. Astrocytes are well equipped to metabolize the glutamate they transport since they are enriched with the glutamine synthetase pathway (Norenberg and Martinez-Hernandez, 1979). Indeed, the importation of glutamate into astrocytes and subsequent formation of glutamine constitutes one limb of the “glutamate-glutamine cycle” which is completed by the export of glutamine to neurons, where it is hydrolyzed to glutamate via phosphate-
-
dependent glutaminase (Hertz, 1979; Shank and Aprison, 1981). The glutamate-glutamine cycle thereby accomplishes the two major requirements of glutamatergic neurotransmission, i.e., the rapid removal of glutamate from the synaptic cleft, and the restoration of glutamate to neurons in the form of glutamine, a non-neuroactive compound which can be shuttled safely between astrocytes and neurons. Problems associated with the glutamate-glutamine cycle Available evidence confirms that astrocytes are a major site of brain glutamine synthesis. However, other evidence indicates that these cells also may play an important role in the utilization of this amino acid. Thus, astrocytes display both the concentrative uptake of glutamine and a fairly high activity of phosphate-dependent glutaminase (Schousboe et al., 1979). The latter metabolic pathway does not appear to be an artifact of culturing in the presence of a high concentration of glutamine (Kvamme et al., 1982). Furthermore, the rate of uptake of glutamine into cultured astrocytes exceeds that noted into neurons, suggesting that there is not necessarily a net flow of glutamine from astrocytes to neurons (Hertz et al., 1980; Ramaharobandro et al., 1982). Nor are these findings restricted to astrocytes in culture: the glutaminase activity of bulkisolated astrocytes exceeds that of synaptosomes (Subbalakshmi and Murthy, 1985).
214
Our own observations of [2-'SN]glutamine metabolism in cultured astrocytes confirm that these cells do not simply synthesize glutamine. As shown in Fig. lB, when astrocytes are cultured in the presence of [2-1SN]glutamine(2 mM) ( - 50.0 atom 070 excess), the enrichment in glutamine steadily declines as this amino acid is synthesized from unlabeled sources and the ISN label is diluted. However, the concentration of glutamine (I4N + lSN) in the medium remains unchanged (Fig. lA), indicating simultaneous synthesis via glutamine synthetase and consumption through glutaminase. This is reflected also in the exponential disappearance (Fig. 1C) of the [2-1SN]glutaminefrom the medium. The rate of disappearance of glutamine (12.7 nmol/min per mg protein) (Table I) is comparable to the rate of synthesis we have measured with similar stable isotope methods ( - 9 nmol/min per mg protein) (Yudkoff et al., 1986, 1988).
32
r
1 0 ' ' 0 4
rn "
I
' '
'
6 12 16 20 24 Time (hr)
1
Time (hr)
0 4
8 12 16 20 24 Time (hr)
Fig. 1. A . The concentration of glutamine in the medium following the addition of 2 mM [2-15N]glutamine.Astrocytes maintained at 37°C in Ham's F-12 medium. B . Isotopic abundance (atom % excess) in [2-1SN]glutamineduring the incubation of the astrocytes in steady-state medium. C . The absolute concentration of [2-1SN]glutamine (nmol ISN/mg protein) in the medium. Value represents the product of isotopic abundance (atom % excess/100) and the glutamine concentration (A). (From Yudkoff et al., 1988.)
Whether net synthesis or consumption occurs is determined in part by the glutamine level of the medium, with synthesis predominating when the level of this amino acid is low and consumption predominating when it is high. As shown in Table I, an important fate of 2-N of glutamine appears to be the formation of alanine, although appreciable ISN is noted in essential amino acids, e.g., phenylalanine, indicating a capacity for the transamination of the cognate alpha-ketoacids.
Glutamine synthesis in neurons via glutaminase The above data underscore the complexity of glial metabolism of glutamine. The fact that these cells can consume glutamine does not exclude the probability that this amino acid is a precursor to neuronal glutamate. As shown in Fig. 2, we have utilized gas chromatography-mass spectrometry as an analytic tool in order to determine flux through the glutaminase pathway in synaptosomes incubated in the presence of [2-1SN]glutamine as a metabolic tracer (Yudkoff et al., 1989). The "flux", representing the sum of lSN resident in [lSN]glutamate, [ISN]aspartate and [lSN]GABA, was measured when the intra-synaptosomal [glutamine] (15 - 30 nmol/mg protein) was high enough to saturate glutaminase ( K , = 1.6-4 mM; McGeer and McGeer, 1979; Benjamin, 1981). As indicated in Table 11, flux through glutaminase is -3.5 nmol/mg protein per minute, or less than 10% of the probable glutaminase activity in synaptosomes (35 - 125 nmol/mg protein per minute; Ward and Bradford, 1979; Kvamme and Lenda, 1982). It also is evident that much of the glutamate so formed becomes transaminated to aspartate and that this process is accentuated if glucose is omitted from the incubation medium, presumably reflecting the importance of glutamine as an alternate metabolic fuel. We have noted a similar phenomenon in studies of [ lSN]glutamate metabolism (Erecinska et al., 1988). Although flux through synaptosomal glutaminase is far removed from the Vmax for this enzyme,
215 TABLE I Utilization and synthesis of glutamine by astrocytes Utilization (nmol/min per mg protein) Appearance of ISN in: Alanine Serine Phenylalanine Tyrosine Valine Leucine
12.74
Synthesis (nmol/min per mg protein)
9.01
2.01 0.26 0.54
0.04 0.07 0.16
Precursor: [2-1SN]glutamine(2 mM). Rate of utilization determined by fitting curve of medium [2-1SN]glutamine disappearance to theexpressiony = Ae-k" + Be-k2'. Values for appearance of I5N in amino acids (sum of medium and cells) determined from theequationy = A - BeCk'. (From Yudkoff et al., 1989.)
it can be stimulated markedly by depolarization. As shown in Fig. 3, when synaptosomes are incubated in the presence of veratridine, the uptake of glutamine after the initial 2 min is sharply reduced, but the total amount of [15N]glutamate and [15N]aspartate formed is similar (Erecinska et al., 1990). Thus, depolarization both reduces synaptosomal glutamine uptake, which is Na+dependent (Erecinska et al., 1990), and also stimulates flux through the glutaminase pathway. Enhanced flux results primarily from increased intra-mitochondrial Pi, which is derived from the hydrolysis of high-energy phosphate compounds during and after depolarization (Siesjo, 1978; Dagani and Erecinska, 1987). The phosphate activation curve is very steep in synaptosomes (Bradford and Ward, 1976), emphasizing the sensitivity of the enzyme to activation by phosphate, which reduces theK, (1.4 - 4 mM; Weil-Malherbe, 1969; McGeer and McGeer, 1979; Benjamin, 1981) of the enzyme for glutamine (Kovacevic and McGivan, 1983). As illustrated in Fig. 4, we found that the phosphate effect on glutaminase is modulated by calcium (Erecinska et al., 1990). In medium with high (10 mM) phosphate, calcium (1.27 mM) stimulated the formation of glutamate and aspartate from glutamine, but in the presence of low
(0.1 mM) phosphate, calcium inhibited flux through glutaminase. The likely explanation for these observations is that calcium favors the uptake of Pi into mitochondria (Fiskum and Lehninger, 1982), thereby enhancing stimulation of glutaminase. In contrast, at a low external [Pi],this effect is overshadowed by a calcium-induced reduction in free intra-mitochondria1 phosphate and a resultant inhibition of flux through glutaminase. An important observation is the rapidity with which [15N]glutamate derived from [215N]glutamine appears in the incubation medium (Fig. 5 ) , even when incubations are done in the absence of veratridine (Erecinska et al., 1990). Indeed, the glutamate probably is released into the medium even before it is taken up into the vesicular pool of this amino acid, an equilibration that occurs relatively slowly (Kauppinen et al., 1988). Furthermore, we found that the magnitude of the release of glutamate and aspartate is directly related to the external glutamine concentration, with relatively little glutamate or aspartate appearing in the medium at 0.15 mM but more at 0.5 mM [glutaminelex, (Erecinska et al., 1990; Fig. 5 ) . This finding could 2o
r
