SEARCH

SEARCH BY CITATION

Keywords:

  • 13C NMR;
  • brain;
  • GABA ;
  • glucose;
  • glutamate;
  • metabolism

Abstract

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgements
  8. References
  9. Supporting Information
Thumbnail image of graphical abstract

Energy metabolism supports both inhibitory and excitatory neurotransmission processes. This study investigated the specific contribution of astrocytic metabolism to γ-aminobutyric acid (GABA) synthesis and inhibitory GABAergic neurotransmission that remained to be ilucidated in vivo. Therefore, we measured 13C incorporation into brain metabolites by dynamic 13C nuclear magnetic resonance spectroscopy at 14.1 T in rats under α-chloralose anaesthesia during infusion of [1,6-13C]glucose. The enhanced sensitivity at 14.1 T allowed to quantify incorporation of 13C into the three aliphatic carbons of GABA non-invasively. Metabolic fluxes were determined with a mathematical model of brain metabolism comprising glial, glutamatergic and GABAergic compartments. GABA synthesis rate was 0.11 ± 0.01 μmol/g/min. GABA-glutamine cycle was 0.053 ± 0.003 μmol/g/min and accounted for 22 ± 1% of total neurotransmitter cycling between neurons and glia. Cerebral glucose oxidation was 0.47 ± 0.02 μmol/g/min, of which 35 ± 1% and 7 ± 1% was diverted to the glutamatergic and GABAergic tricarboxylic acid cycles, respectively. The remaining fraction of glucose oxidation was in glia, where 12 ± 1% of the TCA cycle flux was dedicated to oxidation of GABA. 16 ± 2% of glutamine synthesis was provided to GABAergic neurons. We conclude that substantial metabolic activity occurs in GABAergic neurons and that glial metabolism supports both glutamatergic and GABAergic neurons in the living rat brain.

We performed 13C NMR spectroscopy in vivo at high magnetic field (14.1 T) upon administration of [1,6-13C]glucose. This allowed to measure 13C incorporation into the three aliphatic carbons of GABA in the rat brain, in addition to those of glutamate, glutamine and aspartate. These data were then modelled to determine fluxes of energy metabolism in GABAergic and glutamatergic neurons and glial cells.

Abbreviations used
CMRglc

cerebral metabolic rate for glucose consumption

CRLB

Cramér-Rao lower bound

DEPT

distortionless enhancement by polarization transfer

FE

fractional enrichment

GABA-T

GABA-transaminase

GABA

γ-aminobutyric acid

GAD

glutamate decarboxylase

NMR

nuclear magnetic resonance

PCA

perchloric acid

SSADH

succinic-semialdehyde dehydrogenase

SSA

succinic-semialdehyde

TCA cycle

tricarboxylic acid cycle

The vast majority of synapses in the central nervous system use either glutamate or γ-aminobutyric acid (GABA) to mediate excitatory or inhibitory neurotransmission. Although glutamatergic synapses outnumber GABAergic synapses and the energetic requirements of excitatory neurons are considered higher than those of inhibitory neurons, a substantial amount of glucose, the main substrate for cerebral energy metabolism, has been suggested to be oxidized in inhibitory GABAergic neurons (e.g. Patel et al. 2005; Calvetti and Somersalo 2012; Howarth et al. 2012).

Nuclear magnetic resonance (NMR) spectroscopy has been used to probe metabolic pathways in vivo upon infusion of specific 13C-enriched substrates (Gruetter 2002). However, GABA metabolism in vivo remains difficult to study because of its low concentration. Although several attempts were made to detect 13C turnover curves of GABA, the dynamic measurement of GABA enrichment was restricted to only one or two carbons (Pfeuffer et al. 1999; Yang et al. 2005; van Eijsden et al. 2010). Therefore, studies of brain metabolism mostly focused on glutamatergic function, as glutamate is present in much higher concentrations. Determination of the activity of GABAergic metabolic pathways is important since GABA metabolism has been implicated in neurological disorders (reviewed in Choi and Shen 2012; Waagepetersen et al. 1999): several psychiatric disorders were found to be associated with modification of inhibitory GABAergic function, GABA metabolism and/or GABA levels; in epilepsy, inhibition of GABA-transaminase (GABA-T, EC 2.6.1.19) or GABA transporters increases synaptic GABA availability, which correlates with reduced glucose consumption and improved seizure control.

GABA metabolism occurs in the GABA shunt (e.g. Choi and Shen 2012). Briefly, GABA is synthesized in the cytosol from glutamate by glutamate decarboxylase (GAD, EC 4.1.1.15) and is catabolized in mitochondria, starting with GABA conversion to succinic-semialdehyde (SSA) in mitochondria by GABA-T, with simultaneous amination of 2-oxoglutarate to glutamate, and then SSA catabolism to succinate by SSA dehydrogenase (SSADH, EC 1.2.1.24), for dual use in the tricarboxylic acid (TCA) cycle, and as electron donor of the mitochondrial electron transport chain. After synaptic release, GABA is believed to be mostly cleared by neurons as GABA transporters are less abundant in glial than neuronal membranes (Conti et al. 2004). Nevertheless, uptake of GABA by astrocytes and further metabolism through the GABA shunt, as well as glutamine provision to GABAergic neurons by the GABA-glutamine cycle, have been reported (Schousboe and Waagepetersen 2003, 2007; Bak et al. 2006). In previous 13C NMR spectroscopy studies designed to quantify metabolic fluxes of GABA metabolism in the brain in vivo, astrocytic fluxes were either constrained to those in neurons or determined in parallel experiments using 13C-enriched acetate that is principally oxidized in glia (e.g. Patel et al. 2005; van Eijsden et al. 2010). Therefore, the effective contribution of astrocytic metabolism to GABAergic neurotransmission remains to be addressed in vivo. The aim of this study was to determine compartmentalized metabolic fluxes in GABAergic neurons and connecting astrocytes using the resonances of all three aliphatic carbons of GABA at 14.1 T in the living rat brain.

Materials and methods

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgements
  8. References
  9. Supporting Information

Animals

All experimental procedures involving animals were approved by the local ethics committee (EXPANIM-SCAV, Switzerland). Male Sprague–Dawley rats (267 ± 25 g, = 8, purchased to Charles River Laboratoires, L'Arbresle, France) were prepared as previously detailed in Duarte et al. (2009). Briefly, animals were mechanically ventilated with a mixture of 30% O2 in air using pressure-driven ventilator (MRI-1, CWE incorporated, Ardmore, PA, USA), and catheters were placed into a femoral artery for collection of blood samples and into a femoral vein for infusion of saline solutions containing α-chloralose (Acros Organics, Geel, Belgium) or [1,6-13C]glucose (Sigma-Aldrich, Basel, Switzerland). Light anaesthesia was achieved by administration of α-chloralose (bolus of 80 mg/kg and continuous infusion of 28 mg/kg/h). Arterial blood pressure, heart rate and respiratory rate were continuously monitored with an animal monitoring system (SA Instruments, Stony Brook, NY, USA). Blood samples were collected every 20–30 min to measure gases, pH, lactate and glucose. Arterial pH and pressures of O2 and CO2 were measured using a blood gas analyser (AVL Compact 3; Diamond Diagnostics, Holliston, MA, USA) and were maintained within the normal physiological range by adjusting respiratory rate and volume. Body temperature was maintained between 37.0 and 37.5°C. Plasma glucose and lactate concentrations were quantified with the glucose and lactate oxidase methods, respectively, using two multi-assay analysers (GW7 Micro-Stat; Analox Instruments, London, UK). The [1,6-13C]glucose infusion procedure was previously described in Duarte et al. (2011) and allowed to achieve a sustained plasma fractional enrichment (FE) of 70%, which was confirmed by NMR spectroscopy of collected plasma samples.

NMR spectroscopy

All in vivo NMR experiments were carried out in a DirectDrive spectrometer (Agilent Technologies, Palo Alto, CA, USA) interfaced to a 26-cm horizontal bore 14.1 T magnet (Magnex Scientific, Abingdon, UK), using a homebuilt coil consisting of a 1H quadrature surface coil and a 13C linearly polarized surface coil (Duarte et al. 2011). Animals were immobilized in a home-built holder with a bite bar and two ear inserts to minimize potential motion. The rat brain was positioned in the isocenter of the magnet and fast-spin-echo images with repetition time of 4 s, echo time of 52 ms and echo train length of 8 allowed to identify anatomical landmarks, which were used to place the volume of interest (VOI) of 320 μL in the brain. After FAST(EST)MAP shimming (Gruetter 1993; Gruetter and Tkác 2000), localized 1H NMR spectra were acquired using STEAM with echo time of 2.8 ms, mixing time of 20 ms and repetition time of 4 s (Tkác et al. 2003; Mlynárik et al. 2006). 13C NMR spectra were acquired using semi-adiabatic distortionless enhancement by polarization transfer (DEPT) combined with 3D-ISIS 1H localization (Henry et al. 2003a).

Analysis of 1H and 13C NMR spectra was carried out using LCModel (Stephen Provencher Inc., Oakville, ON, Canada) as described by Mlynárik et al. (2006) and Henry et al. (2003b), respectively. The scaling of dynamically measured 13C concentrations was based on the FE determined in vitro in 13C NMR spectra from perchloric acid extracts of the brain prepared at the end of each experiment (detailed in Duarte et al. 2011). High-resolution in vitro NMR spectroscopy was performed as described in Duarte et al. (2007) using a 14.1 T DRX-600 spectrometer equipped with a 5 mm cryoprobe (Bruker BioSpin SA, Fallenden, Switzerland).

In general, total concentration of brain metabolites was determined by 1H NMR spectroscopy. Since glutamine levels from 1H NMR spectra in vivo were lower than 13C-enriched glutamine in 13C NMR spectra, total glutamine was calculated with the relation between FE and the 13C content at steady state. FE is defined as inline image, which can be written as follows: inline image, where [13C] and [12C] + [13C] represent the concentration of 13C label and the total concentration in a given carbon, respectively. For glutamate and glutamine C3, one can consider that

  • display math

Assuming that the probability of 13C enrichment in C3 through metabolism in the TCA cycle(s) is the same for molecules with or without labelling in C4, FEC3 = C4D34/(C4S + C4D34) for both glutamate (Glu) and glutamine (Gln), where C4D34 and C4S are, respectively, the doublet and singlet in the resonance of C4 from either glutamine or glutamate (e.g. Malloy et al. 1987; Gruetter et al. 1994; Mason et al. 1995). Because C5 does not become labelled upon infusion of [1,6-13C]glucose, the total C4 signal corresponds to C4S + C4D34. Thus, the equation above can be rewritten using 13C measured signals at steady state:

  • display math

Determination of metabolic fluxes and statistical analysis

Kinetic modelling of [1,6-13C]glucose metabolism in glutamatergic, GABAergic and glial compartments was performed by extension of the mathematical model of compartmentalized cerebral metabolism previously described (Duarte et al. 2011). Figure 1 depicts the metabolic pools and fluxes defined in the three-compartment model, which is detailed in Appendix. Variations to this model were tested and are subject of discussion (see below). In particular, the inclusion of a glutamine pool exchangeable through Vexg leads to a major fit improvement. Mathematical models were adjusted to the eleven 13C enrichment curves of glutamate, glutamine and GABA C2, C3 and C4, and of aspartate C2 and C3, using the Levenberg–Marquardt algorithm for non-linear regression, coupled to a Runge-Kutta method for non-stiff systems to obtain numerical solutions of the ordinary differential equations that define the model (detailed in the Appendix). Significance of the fitted parameters (fluxes) was inferred from t-statistics. F-statistics was used for relative assessment of fit quality. Because accuracy in flux estimation is inversely proportional to the noise level of experimental data (Shestov et al. 2007), 13C enrichment curves were averaged from all rats rather than fitting individual time courses. Then, reliability of determined fluxes was evaluated by Monte-Carlo analysis, in which Gaussian noise with the same variance of fit residuals was added to the best fit and initial conditions were randomly generated within confidence interval of the obtained value. Typically, 500 simulated data sets were created for each individual analysis. The estimated metabolic fluxes are shown as mean ± SD, being the SD resulting from fitting a Gamma function to the probability distribution obtained with the Monte-Carlo simulations. All numerical procedures were performed using Matlab (The MathWorks, Natick, MA, USA). Data are presented as mean ± SD of eight experiments, unless otherwise stated.

image

Figure 1. Three-compartment model of brain metabolism. CMRglc is the total cerebral metabolic rate of glucose. Pyruvate originated from glucose consumption is in fast equilibrium with lactate that is exchanged between neurons and glia and is diluted with extra-cerebral lactate through Vout/Vin. VPDHglu accounts for the total TCA cycle flux in glutamatergic neurons, VPDHGABA + VshuntGABA is the total GABAergic TCA cycle, Vg + VPC + Vshuntg is the total glial TCA cycle, VPC is the rate through pyruvate carboxylase. In the glial compartment, the dilution of label at the level of acetyl-CoA (AcCoA) by glial-specific substrates is accounted by Vdil. TCA cycle intermediates oxaloacetate (OAA) and 2-oxoglutarate (OG) exchange with amino acids through the exchange flux VX. The glutamate-glutamine and GABA-glutamine cycles are VNTglu and VNTGABA, respectively, and glutamine synthetase rate is VGS. The GABA synthesis rate catalysed by glutamate decarboxylase (VGAD) was restricted to GABAergic neurons. Vshuntg and VshuntGABA represent the oxidation of GABA in the glial and GABAergic compartments, respectively. Finally, efflux of labelling from the metabolic system occurs through the rate of glial glutamine loss Vefflux. An additional glutamine pool in exchange via Vexg was tested as described in the text. The superscripts g, glu and GABA distinguish metabolic pools or fluxes in the glial, glutamatergic and GABAergic compartments, respectively.

Download figure to PowerPoint

Results

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgements
  8. References
  9. Supporting Information

Upon infusion of [1,6-13C]glucose, the FE of plasma glucose was 0.71 ± 0.02 after 5 min and over the whole period of 13C NMR spectra acquisition. When measuring the incorporation of 13C from [1,6-13C]glucose into amino acids in the living rat brain using 13C NMR spectroscopy at 14.1 T (volume of interest is shown in Fig. 2a), 20 carbon resonances from 10 metabolites were detected between 15 and 65 ppm (Fig. 2b), including three GABA peaks that were clearly detectable with 20 min of temporal resolution (Fig. 2c). The analysis of 13C NMR spectra with LCModel allowed to simultaneously quantify not only the three GABA resonances but also their isotopomers (Fig. 3). The multiplets in carbons of GABA arising from the presence of adjacent 13C resonances were observed in the 13C NMR spectra in vivo (Fig. 3a) as well as in vitro in brain PCA extracts (Fig. 3b). Cramér-Rao lower bounds (CRLB) for GABA C2, C4 and C3 were below 30% from 54, 72 and 126 min after the beginning of [1,6-13C]glucose infusion, respectively (Fig. 3c).

image

Figure 2. 13C NMR spectra acquired in vivo at 14.1 T from the rat brain upon infusion of [1,6-13C]glucose. In (a) is represented the volume of interest (VOI) of 5 × 8×8 mm3 in the rat brain. Brain images are 0.8 mm of thickness and spaced by 1.6 mm. The sphere in the top is the [13C]formic acid reference used for calibration of radiofrequency pulses. (b) Shows a representative spectrum acquired for 1 h starting after 5 h of [1,6-13C]glucose infusion, with identification of visible resonances (expansion from 15 to 65 ppm; Gaussian apodization with gf = 0.08, gfs = 0.02). (c) Shows the time course of spectra with 20-min time resolution acquired during 6 h starting at the onset of [1,6-13C]glucose infusion. The top spectrum is the sum of the three last scans, that is, acquired for 1 h (Lorentzian-Gaussian apodization was employed with lb = 5, gf = 0.08, gfs = 0.02). Resonance assignment: Ala, alanine; Lac, lactate; NAA, N-acetylaspartate; GABA, γ-aminobutirate; Gln, glutamine; Glu, glutamate, GSH, glutathione; Asp, aspartate; Ser, serine; Glc, glucose.

Download figure to PowerPoint

image

Figure 3. (a) Depicts a typical fit obtained with LCModel for an in vivo 13C NMR spectrum acquired for 2 h after 4 h of [1,6-13C]glucose infusion. Expansions in the top of (a) show the fitting of individual isotopomers in GABA C2, C3 and C4. These were also observed in the high-resolution spectrum acquired from brain extracts after [1,6-13C]glucose infusion for 6 h (b). High-resolution 13C NMR spectra in vitro were acquired at 600 MHz, pH 7.0 and 12°C to allow separation of NAA C3 and GABA C4 resonances. (c) Shows 13C signal and associated Cramér-Rao lower bound (CRLB) for GABA resonances. Signal intensity is given in arbitrary units (a.u.) and was not corrected for relative DEPT enhancement or distance from the transmit/observe frequency. Data are mean ± SEM of eight rats.

Download figure to PowerPoint

1H NMR spectroscopy was used to determine the total concentration of brain metabolites that become enriched by 13C (Table 1). Brain metabolite levels were found to agree with prior literature reports (Tkác et al. 2003; Duarte et al. 2011). However, total brain levels of glutamine measured by 1H NMR spectroscopy in vivo were found to be below those of 13C-enriched glutamine measured by 13C NMR spectroscopy. Therefore, total glutamine was estimated at 5.3 ± 0.4 μmol/g using the relation between FE and the 13C signals observed at steady-state (Table 1, discussed below).

Table 1. Total cerebral concentration of metabolites observable in 13C NMR spectra in vivo upon infusion of [1,6-13C]glucose, as determined by 1H NMR spectroscopy prior to substrate infusion. Concentrations are shown as mean ± SEM of eight rats
 Concentration (μmol/g)
  1. a

    glutamine concentration determined through FE of glutamate and glutamine C3 in 13C NMR spectra and total glutamate content.

Glutamate8.6 ± 0.4
Glutamine2.7 ± 0.1; a5.3 ± 0.4
Aspartate2.2 ± 0.3
GABA1.3 ± 0.1
GSH0.91 ± 0.09
NAA8.8 ± 0.4
Glucose1.4 ± 0.4
Lactate0.71 ± 0.05
Alanine0.46 ± 0.07

When fitting the mathematical model to the 13C time courses of glutamate, glutamine, aspartate and GABA, generally an excellent fit was observed (Fig. 4), which resulted in the metabolic fluxes in Table 2. The curves generated by the initial model were not a complete representation of the data measured in vivo, which is particularly clear for glutamine C4 (Fig. 4). The goodness-of-fit was improved by the inclusion of an additional exchangeable glutamine pool in the glial compartment (R2 = 0.976 and = 1960, vs. R2 = 0.972 and = 1680), as indicated by the fit residuals, that is, the mean residuals with and without this glutamine pool were −0.0034 ± 0.0244 and 0.0042 ± 0.0268 (dimension of FE) and the residual sum of squares was 0.328 versus 0.395, respectively (Fig. 4b). The exchange flux Vexg was found to be 0.060 ± 0.008 μmol/g/min and significantly different from zero (= 284). The other metabolic fluxes in the model were not substantially affected by the inclusion of this additional glutamine pool (Table 2). Thus, metabolic fluxes are hereafter presented and discussed for the model that includes Vexg.

Table 2. Cerebral metabolic fluxes determined with the three-compartment model of [1,6-13C]glucose metabolism. Estimated fluxes are shown in μmol/g/min with associated SD and t-value in parenthesis
  One glial glutamine pool R2 = 0.972, f = 1680With additional glutamine pool R2 = 0.976, f = 1960
DeterminedVPDHGlu (= VTCAGlu)0.36 ± 0.01(995)0.33 ± 0.01(931)
V PDH GABA 0.024 ± 0.005(174)0.017 ± 0.005(141)
V g 0.26 ± 0.02(408)0.29 ± 0.03(409)
VPC (= Vefflux)0.076 ± 0.004(704)0.092 ± 0.005(714)
V dil g 1.1 ± 0.1(189)0.76 ± 0.07(297)
V NT Glu 0.16 ± 0.01(381)0.18 ± 0.01(471)
VNTGABA (= Vshuntg)0.044 ± 0.002(504)0.053 ± 0.003(602)
V GAD GABA 0.098 ± 0.003(886)0.11 ± 0.01(970)
V x Glu 0.41 ± 0.02(568)0.39 ± 0.02(625)
V x GABA 0.0067 ± 0.0038(63.5)0.0068 ± 0.0034(77.5)
V x g 0.010 ± 0.011(28.4)0.018 ± 0.018(35.1)
V ex g   0.060 ± 0.008(284)
CalculatedCMRglc(ox)0.44 ± 0.02 0.47 ± 0.02 
V out 0.16 ± 0.03 0.11 ± 0.03 
V TCA GABA 0.077 ± 0.007 0.070 ± 0.007 
V TCA g 0.38 ± 0.02 0.44 ± 0.03 
V shunt GABA 0.054 ± 0.004 0.053 ± 0.005 
V GS 0.28 ± 0.01 0.33 ± 0.01 
V NT total 0.20 ± 0.01 0.24 ± 0.01 
Vxg + VPC + Vshuntg (OG[RIGHTWARDS ARROW]Glu)0.13 ± 0.01 0.16 ± 0.02 
VxGABA + VshuntGABA (OG[RIGHTWARDS ARROW]Glu)0.060 ± 0.006 0.060 ± 0.006 
image

Figure 4. Fractional enrichment in carbons of glutamate, glutamine, GABA and aspartate, and curves of the best fit of the three-compartment model (a). Data points are the mean of 8 rats, shaded area represents respective SD, blue and red lines depict the best fit to the data of the three-compartment model with one or two glial glutamine pools. Fit residuals of the models (b) with one (blue symbols) or two (red symbols) glial glutamine pools indicate fit improvement by addition of the glutamine pool with low 1H NMR visibility. The lines on the right show the distribution of the residuals for each model.

Download figure to PowerPoint

Total glucose oxidation, that is, cerebral metabolic rate for glucose consumption (CMRglc(ox)), was 0.47 ± 0.02 μmol/g/min. Since glucose consumption was fixed at 0.52 μmol/g/min, as determined previously (see Appendix), net release of lactate (Vout) was 0.11 ± 0.03 μmol/g/min. About half of glucose oxidation was found to occur through pyruvate dehydrogenase in the glial compartment (47 ± 3%), and a considerable fraction because of pyruvate carboxylation: VPC was 10 ± 1% of CMRglc(ox) and 21 ± 1% of VTCAg. In glutamatergic and GABAergic neurons, the fraction of total glucose oxidation was 35 ± 1% and 7 ± 1%, respectively. VNTGABA was 22 ± 1% of total neurotransmitter cycling between neurons and astrocytes. In GABAergic neurons, VGAD was 0.11 ± 0.01 μmol/g/min and VNTGABA was 0.053 ± 0.003 μmol/g/min, which was set to be equal to the GABAergic glutaminase flux. Thus, glial glutamine contributed to 48 ± 5% of GABA synthesis. Glial VTCA and VGS account for GABA recycling through the glial GABA shunt, which was 12 ± 1% of VTCAg and resulted in 16 ± 2% of glutamine synthesis (VGS) to be shuttled to GABAergic neurons for GABA production. In total, the rate of glucose oxidation supporting GABAergic activity was 13 ± 1% (VTCAGABA + Vshuntg). The rate of label exchange between cytosolic amino acids and their mitochondrial counterparts, namely mediated by the malate-aspartate shuttle, were on the order of the TCA cycle rate in each compartment.

Discussion

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgements
  8. References
  9. Supporting Information

This study reports for the first time the in vivo measurement of the 13C enrichment in all three aliphatic carbons of GABA in the rat brain, which contributed to elucidate multiple aspects of metabolic compartmentation of GABA metabolism in vivo. GABA C4 but not C3 and C2 receives labelling from pyruvate through pyruvate carboxylation and thus it allowed to determine the contribution of glial metabolism to GABA synthesis, in addition to metabolic fluxes in GABAergic neurons (depicted in Fig. 1). In particular, we found that a considerable fraction of glucose was oxidized in the GABAergic compartment and connecting glial cells to fulfil the energetic requirements of GABAergic neurotransmission (13 ± 1%), and that glial glutamine accounts for an important fraction of GABA synthesis (48 ± 5%).

Detection of 13C enrichment in GABA in vivo has been mostly accomplished using indirect detection of signals from 13C-coupled protons. The first turnover curve of 13C enriched GABA in vivo was reported by Pfeuffer et al. (1999), who measured GABA C3 in the rat brain during infusion of [1-13C]glucose. Later, other GABA turnover curves were determined in vivo (e.g. Yang et al. 2005; van Eijsden et al. 2010). Although the indirect detection of 13C-coupled protons has a high signal-to-noise ratio of protons, the low chemical shift dispersion, even at 14.1 T (Xin et al. 2010), results in substantial signal overlap. In contrast, the wide chemical shift dispersion of direct 13C NMR spectra at high field allowed to quantify all three aliphatic GABA resonances. This study demonstrates the simultaneous quantification of the three aliphatic carbons of GABA with an average precision (CRLB) of 0.17 μmol/g (Fig. 3c) at a temporal resolution comparable to that of studies using 1H-observed-13C-edited spectroscopy methods (Pfeuffer et al. 1999; Yang et al. 2005; van Eijsden et al. 2010), a possibility that was not hitherto accessible at lower magnetic fields, especially for GABA C4, which partially overlaps with the C3 of N-acetylaspartate (aspartyl moiety) (e.g. Choi and Gruetter 2003). The ability to quantify all GABA signals was ascribed to increased spectral resolution and high signal achieved at 14.1 T with surface coil signal detection, quantification by LCModel and use of [1,6-13C]glucose infusion that leads to a nearly twofold increase in the fractional enrichment of observed metabolites, compared to the use of [1-13C]glucose as substrate.

Cerebral metabolism of GABA and GABA-glutamine cycle

When comparing the fluxes obtained with the present three-compartment model (Fig. 1) to those obtained by fitting the previously published two-compartment model of brain energy metabolism (Duarte et al. 2011), we noted that most metabolic fluxes were not substantially different (Fig. 5a). Glucose oxidation was higher (0.47 ± 0.02 vs. 0.38 ± 0.02 μmol/g/min) and the net release of lactate (Vout) was 62 ± 11% lower than when using the two-compartment model (both with CMRglc = 0.52 μmol/g/min). Thus, a substantial fraction of glucose utilization in the rat brain is involved in maintaining GABAergic function, rather than released as lactate. This indicates that the inclusion of a GABAergic compartment leads to a more complete account for total glucose metabolism.

image

Figure 5. Comparison of metabolic fluxes with those obtained from two-compartment models. (a) Shows fluxes obtained by fitting models of brain metabolism with three (filled bars) and two (open bars) compartments to the present data. The two-compartment model was adapted from Duarte et al. (2011) to include Vexg. (b) Shows the relation of estimated fluxes of total glucose oxidation, neuronal and glial TCA cycles and pyruvate carboxylation to glutamate/GABA-glutamine cycle, that is, total VNT, across several studies using similar methodology in the rat or human brain: (i) rat brain under pentobarbital-induced isoelectricity (Choi et al. 2002), (ii) rat brain under light anaesthesia (Duarte et al. 2011), (iii) human cortex (Gruetter et al. 2001), (iv) current study, (v) awake rat brain (Oz et al. 2004). Metabolic fluxes are shown in μmol/g/min with associated SD. For linear regression, data sets (iv) and (iii) were excluded in VTCAn and VTCAg, respectively.

Download figure to PowerPoint

A small but significant fraction of glucose was directly utilized in GABAergic neurons (7 ± 1% of total oxidation), as found by Patel et al. (2005) with quantification of 13C GABA in brain extracts, suggesting that excitation plays an energetically dominant role. Anatomical and physiological data support the concept that inhibitory neurotransmission overall requires less energy compared to glutamatergic activity. Indeed, glutamatergic synapses outnumber GABAergic synapses by fivefold, with less than 30% of cortical neurons being inhibitory GABAergic interneurons (DeFelipe and Fariñas 1992; Sahara et al. 2012), and the energetic requirements of excitatory neurons have been suggested to be nearly three times those of inhibitory neurons (Howarth et al. 2012). Although it has been suggested that the recycling of GABA may not be coupled to glucose utilization in astrocytes, even though astrocytes respond to GABA by increasing intracellular sodium content (Chatton et al. 2003) according to the stechiometry of co-transport with two Na+ and one Cl (Gadea and López-Colomé 2001), this does not preclude that activity of GABAergic nerve terminals per se does not involve energy consumption in neighbouring astrocytes. When providing simultaneously [U-13C]glucose and [2-13C]acetate to superfused hippocampal slices we previously found lower but significant contribution of acetate metabolized in glia to the multiplets of GABA, compared to those of glutamate, which further increases upon stimulation (Duarte et al. 2007).

It should be noted that the inclusion of GABAergic activity in the model lead to a substantial increment in the glial VTCA to 47 ± 3% of total glucose oxidation (Fig. 5a), as part of it is required for GABA oxidation through the GABA shunt (12 ± 1% of VTCAg), which resulted in synthesis of glutamine to be shuttled to GABAergic neurons for GABA production (16 ± 2% of VGS). In this study, when accounting for this fraction of glial TCA cycle that is involved in supporting GABA recycling, the total glucose oxidation attributed to GABAergic function was 13 ± 1% of total CMRglc(ox).

Astrocytes lack the enzymatic machinery for GABA synthesis and, therefore, virtually all glial GABA originates from neuronal release (Schousboe and Waagepetersen 2003). In fact, we tested the presence of VGAD in the glial compartment and this flux was consistently not significantly different from zero (not shown). Although not being able to synthesize GABA, astrocytes play an important role in GABA metabolism (reviewed in Schousboe et al. 1992; Choi and Shen 2012). In the present model, the flux through GAD was set to be equal to the sum of the glial and neuronal GABA shunts (thus respecting steady-state mass balance) that are associated with the respective GABA-T activities initiating GABA catabolism in mitochondria. Our results for VshuntGABA and Vshuntg (Table 2) suggest that nearly half of GABA produced in GABAergic neurons is further oxidized in astrocytes through the glial GABA shunt and that half of GABA synthesis relies on glutamine provided by astrocytes, which is consistent with studies in cultured cortical neurons (Westergaard et al. 1995). This does not mean, however, that equal amounts of synaptically released GABA are taken up by astrocytes and neurons, while glutamate transporters in astrocytes largely exceed those in neurons, GABA transporters are mostly concentrated in membranes of GABAergic neurons (Conti et al. 2004), suggesting preferential uptake of glutamate by astrocytes and of GABA by neurons (e.g. Schousboe and Waagepetersen 2003, 2007; Bak et al. 2006). Thus, it may be suggested that a substantial amount of GABA released into the synapse must be taken up by same nerve terminal and directly re-incorporated into synaptic vesicles for subsequent release.

Glial metabolism supports neurotransmission

Since early studies of brain energy metabolism, it has been postulated that metabolic pathways leading to the production of energy are under direct control of cerebral activity (van den Berg et al. 1969). Aiming at elucidating how glucose metabolism is related to both glutamatergic and GABAergic synaptic transmission, we further compared the present results with previous reports in rodents and humans that considered all metabolic fluxes as independent parameters (Gruetter et al. 2001; Choi et al. 2002; Oz et al. 2004; Duarte et al. 2011). In general, the present results are in line with previous studies in either rodents or humans (Fig. 5b). Interestingly, in addition to glucose oxidation (Sibson et al. 1998), glial pyruvate carboxylation was related to the total neuron-glia neurotransmitter cycling (Fig. 5b). In fact, the activity-induced increase in glial anaplerosis through VPC is consistent with increased influx of bicarbonate into astrocytes upon neuronal release of K+ (Chesler 1990; Brookes and Turner 1994) that may stimulate pyruvate carboxylation (Gamberino et al. 1997) along with modulation of glycolysis, glycogenolysis and lactate production (Choi et al. 2012). An intrinsic coupling between VPC, VGS and Vefflux was defined in the mathematical model, which finds support in the fact that increased brain activity leads to ammonia accumulation that likely requires detoxification through synthesis of glutamine that diffuses to the blood stream (suggested by Merle and Franconi 2012).

The total flux through pyruvate dehydrogenase in glia, depicted by VTCAg, also increased with VNT in the rat brain. Because oxaloacetate produced from both pyruvate carboxylation and GABA oxidation require equivalent amount of acetyl-CoA for condensation, VTCAg was defined as Vg + VPC + Vshuntg (see Appendix), with Vg (that was 66 ± 8% of VTCAg) denoting the flux through glial pyruvate dehydrogenase corresponding to the complete oxidation of pyruvate (Gruetter et al. 2001). Interestingly, we did not observe a significant modification of Vg with VNT (not shown), suggesting that pyruvate carboxylation and GABA oxidation are the main drive for the relation between glial VTCA and VNT, and that a large fraction of glial energy production is probably dedicated to non-signalling processes. In line with this, energy consumption for tasks unrelated to neurotransmission could account for up to ~ 50% of total energy expenditure in the whole rat brain (discussed in Howarth et al. 2012).

The relationship between the rates of glutamatergic neurotransmitter cycling and glucose oxidation in neurons is well established in the literature (e.g. Hyder et al. 2013). In agreement, oxidative metabolism in neurons was related with VNT only for glutamatergic neurons, probably because GABA re-uptake is less energy demanding and GABAergic neurons receive substantial amount of glutamine synthesized de novo by astrocytes (discussed above). Increase of glial VTCA with VNT seemed to be species specific, being apparent that the data set from the human brain is not in line with those from rats.

Subcellular compartmentation of glial glutamine

In this study, we found that, when approaching isotopic steady-state, glutamine enrichment in C4 was similar to that in C2. This implies that glial dilution occurs and, therefore, Vdil was introduced to represent the utilization of glial-specific substrates that enter brain metabolism at the level of acetyl-CoA (see Duarte et al. 2011 and references therein). In addition, even when steady state of 13C enrichment was reached for the mainly neuronal glutamate, GABA and aspartate, a significant and continuous increase in glutamine enrichment occurs (Fig. 4a) in the presence of a nearly constant total glutamine concentration, as measured in vivo using 1H NMR spectroscopy under identical experimental conditions (Duarte et al. 2009; Duarte and Gruetter 2012). While pyruvate carboxylation could explain this effect in glutamine C3 and C2, the C4 only receives label from acetyl-CoA and thus should reach the steady-state enrichment together with the remaining amino acids. This continuous increase in brain glutamine enrichment upon 13C-enriched glucose administration was also observed in previous studies using dynamic 13C NMR spectroscopy in vivo in both rodents (Duarte et al. 2011) and humans (Shen et al. 1999).

Brain 13C-enriched glutamine was found higher than total concentration of glutamine measured by 1H NMR spectroscopy in vivo. This suggests that a fraction of brain glutamine is not visible in the 1H NMR spectrum in vivo. In brain extracts, levels of glutamine relative to glutamate were similar to those measured in vivo (data not shown), suggesting that this putative glutamine pool with low NMR visibility (as suggested by Hancu and Port 2011) may be associated with macromolecules that precipitate upon perchloric acid extraction. The eventual existence of glutamine pools with different physical properties impelled us to test an alternative model with a fourth glutamine pool of the size of the non-visible glutamine, that is, difference between 5.3 ± 0.4 μmol/g and glutamine concentration measured by 1H NMR spectroscopy. This pool was assumed to be in direct exchange with glial glutamine indicated by Vexg (see Fig. 1).

The inclusion of a fourth glutamine pool in exchange with one of the other three (the glial pool) resulted in an approximation of the model curves to the measured data, namely glutamine carbons. This pool was found to be associated with a slow turnover and thus characterized by a nearly continuous increase in 13C enrichment over the whole analysed time course. Astrocyte morphology is compatible with the existence of different functional domains and evidence from in vitro studies in cultured astrocytes using 13C-labelled compounds may help unveil the existence of multiple glial glutamine pools: in addition to mitochondrial heterogeneity and glutamine synthesis from multiple TCA cycles (Sonnewald et al. 1998; Waagepetersen et al. 2006), intracellular and released glutamine displayed distinct labelling patters from metabolism of 13C-enriched glucose and lactate (Waagepetersen et al. 2001). Together with our results, these observations support the existence of a glial glutamine pool with slow turnover that is not released to neurons but could be used for protein synthesis or eventually metabolized for energy production (discussed in McKenna 2007). In alternative, it is possible that metabolic heterogeneity occurs within the astrocytic syncytium or that multiple glial compartments are present in the volume of interest encompassing distinct brain areas. This scenario is however difficult to model because of the large number of parameters to be estimated from the observed 13C enrichment curves.

Mitochondrial membrane transport

Labelling of brain glutamate from 13C-enriched oxidative substrates requires transfer of the label from the mitochondrial TCA cycle intermediate 2-oxoglutarate to cytosolic glutamate, mainly via the 2-oxoglutarate-malate and glutamate-aspartate carriers in the mitochondrial membrane, which also sustains the transfer of reducing equivalents from cytosol to mitochondria. In the compartments where the GABA shunt occurs, that is, glia and GABAergic neurons, a substantial part of 2-oxoglutarate transamination to glutamate occurred with GABA (Vshunt), yielding SSA that is further metabolized by the TCA cycle. Mitochondrial transport of 13C label in glia, that is, VXg + VPC + Vshuntg that includes GABA transamination and net glutamate synthesis, was much lower than the neuronal counterpart, which is in agreement with the finding that carriers exchanging aspartate and glutamate in the malate-aspartate shuttle are predominantly expressed in neurons rather than astrocytes (Ramos et al. 2003; Berkich et al. 2007; Pardo et al. 2011). In contrast, Li et al. recently found comparable amounts of Aralar protein, the main mitochondrial carrier for aspartate/glutamate in the brain, in freshly isolated neurons and astrocytes (Li et al. 2012), suggesting that the transference of reducing equivalents across the mitochondrial membrane is not a limiting step for adaptation to altered metabolic or energetic demand in astrocytes.

Although the rate of label transfer from 2-oxoglutarate to glutamate is frequently thought to occur at a more rapid rate than that of TCA cycle intermediate oxidation (e.g. Mason et al. 1995; Sibson et al. 1998; Yang et al. 2009), metabolic modelling of 13C NMR data from humans (Gruetter et al. 2001) and rats (Choi et al. 2002; Oz et al. 2004; Duarte et al. 2011) with independent flux estimation found mitochondrial exchange fluxes on the order of the TCA cycle rate, as observed in this study. The present data were only consistent with mitochondrial exchange fluxes in the same order of the respective TCA cycle rates and the constraint of VX to be 10 times greater than VTCA in each compartment lead to large fit residuals (not shown). This conclusion is further strengthened by the finding that under near physiological conditions the rate of 2-oxoglutarate synthesis in brain mitochondria was similar to that of α-ketoglutarate/glutamate efflux (Berkich et al. 2005). Furthermore, 2-oxoglutarate but not glutamate was detected in the rat brain by 13C NMR spectroscopy upon infusion of hyperpolarized acetate (Mishkovsky et al. 2012).

Conclusion

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgements
  8. References
  9. Supporting Information

We conclude that GABAergic neurons consume a substantial fraction of glucose and that glutamine of glial origin is an important precursor for GABA synthesis in the rat brain in vivo. This is in line with the metabolic cooperation between astrocytes and neurons for maintenance of not only excitatory but also inhibitory neurotransmission. Furthermore, subcellular metabolic compartmentation, namely in astrocytes, can improve description of the experimental data with mathematical models of brain metabolism.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgements
  8. References
  9. Supporting Information

This study was supported by Swiss National Science Foundation (grant 131087) and by Centre d'Imagerie BioMédicale (CIBM) of the UNIL, UNIGE, HUG, CHUV, EPFL and the Leenaards and Jeantet Foundations. The authors are grateful to Patricia M. Nunes (Radboud University Nijmegen Medical Centre, Nijmegen) and Dr. Bernard Lanz (LIFMET) for fruitful discussions and to Jessica A.M. Bastiaansen (LIFMET) for help with preparation of figures. The authors declare no conflicts of interest.

References

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgements
  8. References
  9. Supporting Information
  • Bak L. K., Schousboe A. and Waagepetersen H. S. (2006) The glutamate/GABA-glutamine cycle: aspects of transport, neurotransmitter homeostasis and ammonia transfer. J. Neurochem. 98, 641653.
  • van den Berg C. J., Krzalić L., Mela P. and Waelsch H. (1969) Compartmentation of glutamate metabolism in brain. Evidence for the existence of two different tricarboxylic acid cycles in brain. Biochem. J. 113, 281290.
  • Berkich D. A., Xu Y., LaNoue K. F., Gruetter R. and Hutson S. M. (2005) Evaluation of brain mitochondrial glutamate and alpha-ketoglutarate transport under physiologic conditions. J. Neurosci. Res. 79, 106113.
  • Berkich D. A., Ola M. S., Cole J., Sweatt A. J., Hutson S. M. and LaNoue K. F. (2007) Mitochondrial transport proteins of the brain. J. Neurosci. Res. 85, 33673377.
  • Brookes N. and Turner R. J. (1994) K+-induced alkalinization in mouse cerebral astrocytes mediated by reversal of electrogenic Na+-HCO3- cotransport. Am. J. Physiol. 267, C1633C1640.
  • Calvetti D. and Somersalo E. (2012) Ménage à trois: the role of neurotransmitters in the energy metabolism of astrocytes, glutamatergic, and GABAergic neurons. J. Cereb. Blood Flow Metab. 32, 14721483.
  • Chatton J. Y., Pellerin L. and Magistretti P. J. (2003) GABA uptake into astrocytes is not associated with significant metabolic cost: implications for brain imaging of inhibitory transmission. Proc. Natl Acad. Sci. USA 100, 1245612461.
  • Chesler M. (1990) The regulation and modulation of pH in the nervous system. Prog. Neurobiol. 34, 401427.
  • Choi I. Y. and Gruetter R. (2003) In vivo 13C NMR assessment of brain glycogen concentration and turnover in the awake rat. Neurochem. Int. 43, 317322.
  • Choi I.-Y. and Shen J. (2012) In vivo GABA metabolism, in Neural Metabolism In Vivo (Choi I.-Y. and Gruetter R., eds), pp. 10951116. Springer, New York.
  • Choi I. Y., Lei H. and Gruetter R. (2002) Effect of deep pentobarbital anesthesia on neurotransmitter metabolism in vivo: on the correlation of total glucose consumption with glutamatergic action. J. Cereb. Blood Flow Metab. 22, 13431351.
  • Choi H. B., Gordon G. R., Zhou N. et al. (2012) Metabolic communication between astrocytes and neurons via bicarbonate-responsive soluble adenylyl cyclase. Neuron 75, 10941104.
  • Conti F., Minelli A. and Melone M. (2004) GABA transporters in the mammalian cerebral cortex: localization, development and pathological implications. Brain Res. Brain Res. Rev. 45, 196212.
  • DeFelipe J. and Fariñas I. (1992) The pyramidal neuron of the cerebral cortex: morphological and chemical characteristics of the synaptic inputs. Prog. Neurobiol. 39, 563607.
  • Duarte J. M. N. and Gruetter R. (2012) Characterization of cerebral glucose dynamics in vivo with a four-state conformational model of transport at the blood-brain-barrier. J. Neurochem. 121, 396406.
  • Duarte J. M. N., Cunha R. A. and Carvalho R. A. (2007) Different metabolism of glutamatergic and GABAergic compartments in superfused hippocampal slices characterized by nuclear magnetic resonance spectroscopy. Neuroscience 144, 13051313.
  • Duarte J. M. N., Carvalho R. A., Cunha R. A. and Gruetter R. (2009) Caffeine consumption attenuates neurochemical modifications in the hippocampus of streptozotocin-induced diabetic rats. J. Neurochem. 111, 368379.
  • Duarte J. M. N., Lanz B. and Gruetter R. (2011) Compartmentalized cerebral metabolism of [1,6-13C]glucose Determined by in vivo 13C NMR spectroscopy at 14.1 T. Front. Neuroenergetics 3, 3.
  • van Eijsden P., Behar K. L., Mason G. F., Braun K. P. and de Graaf R. A. (2010) In vivo neurochemical profiling of rat brain by 1H-[13C] NMR spectroscopy: cerebral energetics and glutamatergic/GABAergic neurotransmission. J. Neurochem. 112, 2433.
  • Gadea A. and López-Colomé A. M. (2001) Glial transporters for glutamate, glycine, and GABA: II. GABA transporters. J. Neurosci. Res. 63, 461468.
  • Gamberino W. C., Berkich D. A., Lynch C. J., Xu B. and LaNoue K. F. (1997) Role of pyruvate carboxylase in facilitation of synthesis of glutamate and glutamine in cultured astrocytes. J. Neurochem. 69, 23122325.
  • Gruetter R. (1993) Automatic, localized in vivo adjustment of all first- and second-order shim coils. Magn. Reson. Med. 29, 804811.
  • Gruetter R. (2002) In vivo 13C NMR studies of compartmentalized cerebral carbohydrate metabolism. Neurochem. Int. 41, 143154.
  • Gruetter R. and Tkác I. (2000) Field mapping without reference scan using asymmetric echo-planar techniques. Magn. Reson. Med. 43, 319323.
  • Gruetter R., Novotny E. J., Boulware S. D., Mason G. F., Rothman D. L., Shulman G. I., Prichard J. W. and Shulman R. G. (1994) Localized 13C NMR spectroscopy in the human brain of amino acid labeling from D-[1-13C]glucose. J. Neurochem. 63, 13771385.
  • Gruetter R., Seaquist E. R. and Ugurbil K. (2001) A mathematical model of compartmentalized neurotransmitter metabolism in the human brain. Am. J. Physiol. Endocrinol. Metab. 281, E100E112.
  • Hancu I. and Port J. (2011) The case of the missing glutamine. NMR Biomed. 24, 529535.
  • Henry P. G., Tkác I. and Gruetter R. (2003a) 1H-localized broadband 13C NMR spectroscopy of the rat brain in vivo at 9.4 T. Magn. Reson. Med. 50, 684692.
  • Henry P. G., Oz G., Provencher S. and Gruetter R. (2003b) Toward dynamic isotopomer analysis in the rat brain in vivo: automatic quantitation of 13C NMR spectra using LCModel. NMR Biomed. 16, 400412.
  • Howarth C., Gleeson P. and Attwell D. (2012) Updated energy budgets for neural computation in the neocortex and cerebellum. J. Cereb. Blood Flow Metab. 32, 12221232.
  • Hyder F., Fulbright R. K., Shulman R. G. and Rothman D. L. (2013) Glutamatergic function in the resting awake human brain is supported by uniformly high oxidative energy. J. Cereb. Blood Flow Metab. 33, 339347.
  • Li B., Hertz L. and Peng L. (2012) Aralar mRNA and protein levels in neurons and astrocytes freshly isolated from young and adult mouse brain and in maturing cultured astrocytes. Neurochem. Int. 61, 13251332.
  • Malloy C. R., Sherry A. D. and Jeffrey F. M. (1987) Carbon flux through citric acid cycle pathways in perfused heart by 13C NMR spectroscopy. FEBS Lett. 212, 5862.
  • 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, 1225.
  • McKenna M. C. (2007) The glutamate-glutamine cycle is not stoichiometric: fates of glutamate in brain. J. Neurosci. Res. 85, 33473358.
  • Merle M. and Franconi J.-M. (2012) Brain metabolic compartmentalization, metabolism modeling, and cerebral activity-metabolism relationship, in Neural Metabolism In Vivo (Choi I.-Y. and Gruetter R., eds), pp. 947992. Springer, New York.
  • Mishkovsky M., Comment A. and Gruetter R. (2012) In vivo detection of brain Krebs cycle intermediate by hyperpolarized magnetic resonance. J. Cereb. Blood Flow Metab. 32, 21082113.
  • Mlynárik V., Gambarota G., Frenkel H. and Gruetter R. (2006) Localized short-echo-time proton MR spectroscopy with full signal-intensity acquisition. Magn. Reson. Med. 56, 965970.
  • Oz G., Berkich D. A., Henry P. G., Xu Y., LaNoue K., Hutson S. M. and Gruetter R. (2004) Neuroglial metabolism in the awake rat brain: CO2 fixation increases with brain activity. J. Neurosci. 24, 1127311279.
  • Pardo B., Rodrigues T. B., Contreras L., Garzón M., Llorente-Folch I., Kobayashi K., Saheki T., Cerdan S. and Satrústegui J. (2011) Brain glutamine synthesis requires neuronal-born aspartate as amino donor for glial glutamate formation. J. Cereb. Blood Flow Metab. 31, 90101.
  • Patel A. B., de Graaf R. A., Mason G. F., Rothman D. L., Shulman R. G. and Behar K. L. (2005) The contribution of GABA to glutamate/glutamine cycling and energy metabolism in the rat cortex in vivo. Proc. Natl Acad. Sci. USA 102, 55885593.
  • Pfeuffer J., Tkác I., Choi I. Y., Merkle H., Ugurbil K., Garwood M. and Gruetter R. (1999) Localized in vivo 1H NMR detection of neurotransmitter labeling in rat brain during infusion of [1-13C] D-glucose. Magn. Reson. Med. 41, 10771083.
  • Ramos M., del Arco A., Pardo B. et al. (2003) Developmental changes in the Ca2+-regulated mitochondrial aspartate-glutamate carrier aralar1 in brain and prominent expression in the spinal cord. Dev. Brain Res. 143, 3346.
  • Sahara S., Yanagawa Y., O'Leary D. D. and Stevens C. F. (2012) The fraction of cortical GABAergic neurons is constant from near the start of cortical neurogenesis to adulthood. J. Neurosci. 32, 47554761.
  • Schousboe A. and Waagepetersen H. S. (2003) Role of astrocytes in homeostasis of glutamate and GABA during physiological and pathophysiological conditions. Adv. Mol. Cell Biol. 31, 461474.
  • Schousboe A. and Waagepetersen H. S. (2007) GABA: homeostatic and pharmacological aspects. Prog. Brain Res. 160, 919.
  • Schousboe A., Westergaard N., Sonnewald U., Petersen S. B., Yu A. C. and Hertz L. (1992) Regulatory role of astrocytes for neuronal biosynthesis and homeostasis of glutamate and GABA. Prog. Brain Res. 94, 199211.
  • Shen J., Petersen K. F., Behar K. L., Brown P., Nixon T. W., Mason G. F., Petroff O. A., Shulman G. I., Shulman R. G. and Rothman D. L. (1999) Determination of the rate of the glutamate/glutamine cycle in the human brain by in vivo 13C NMR. Proc. Natl Acad. Sci. USA 96, 82358240.
  • Shestov A. A., Valette J., Uğurbil K. and Henry P. G. (2007) On the reliability of 13C metabolic modeling with two-compartment neuronal-glial models. J. Neurosci. Res. 85, 32943303.
  • Sibson N. R., Dhankhar A., Mason G. F., Rothman D. L., Behar K. L. and Shulman R. G. (1998) Stoichiometric coupling of brain glucose metabolism and glutamatergic neuronal activity. Proc. Natl Acad. Sci. USA 95, 316321.
  • Sonnewald U., Hertz L. and Schousboe A. (1998) Mitochondrial heterogeneity in the brain at the cellular level. J. Cereb. Blood Flow Metab. 18, 231237.
  • Tkác I., Rao R., Georgieff M. K. and Gruetter R. (2003) Developmental and regional changes in the neurochemical profile of the rat brain determined by in vivo 1H NMR spectroscopy. Magn. Reson. Med. 50, 2432.
  • Waagepetersen H. S., Sonnewald U. and Schousboe A. (1999) The GABA paradox: multiple roles as metabolite, neurotransmitter, and neurodifferentiative agent. J. Neurochem. 73, 13351342.
  • Waagepetersen H. S., Sonnewald U., Larsson O. M. and Schousboe A. (2001) Multiple compartments with different metabolic characteristics are involved in biosynthesis of intracellular and released glutamine and citrate in astrocytes. Glia 35, 246252.
  • Waagepetersen H. S., Hansen G. H., Fenger K., Lindsay J. G., Gibson G. and Schousboe A. (2006) Cellular mitochondrial heterogeneity in cultured astrocytes as demonstrated by immunogold labeling of alpha-ketoglutarate dehydrogenase. Glia 53, 225231.
  • Westergaard N., Sonnewald U., Petersen S. B. and Schousboe A. (1995) Glutamate and glutamine metabolism in cultured GABAergic neurons studied by 13C NMR spectroscopy may indicate compartmentation and mitochondrial heterogeneity. Neurosci. Lett. 185, 2428.
  • Xin L., Mlynárik V., Lanz B., Frenkel H. and Gruetter R. (2010) 1H-[13C] NMR spectroscopy of the rat brain during infusion of [2-13C] acetate at 14.1 T. Magn. Reson. Med. 64, 334340.
  • Yang J., Li C. Q. and Shen J. (2005) In vivo detection of cortical GABA turnover from intravenously infused [1-13C]D-glucose. Magn. Reson. Med. 53, 12581267.
  • Yang J., Xu S. and Shen J. (2009) Fast isotopic exchange between mitochondria and cytosol in brain revealed by relayed 13C magnetization transfer spectroscopy. J. Cereb. Blood Flow Metab. 29, 661669.

Supporting Information

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgements
  8. References
  9. Supporting Information
FilenameFormatSizeDescription
jnc12333-sup-0001-AppendixS1.pdfapplication/PDF65KAppendix S1. Three-compartment model of [1,6-13C]glucose metabolism in the brain.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.