In vivo 13C NMR measurements of cerebral glutamine synthesis as evidence for glutamate-glutamine cycling

Abstract

The cerebral tricarboxylic acid (TCA) cycle rate and the rate of glutamine synthesis were measured in rats in vivo under normal physiological and hyperammonemic conditions using 13C NMR spectroscopy. In the hyperammonemic animals, blood ammonia levels were raised from control values of approximately 0.05 mM to approximately 0.35 mM by an intravenous ammonium acetate infusion. Once a steady-state of cerebral metabolites was established, a [1-13C]glucose infusion was initiated, and 13C NMR spectra acquired continuously on a 7-tesla spectrometer to monitor 13C labeling of cerebral metabolites. The time courses of glutamate and glutamine C-4 labeling were fitted to a mathematical model to yield TCA cycle rate (V(TCA)) and the flux from glutamate to glutamine through the glutamine synthetase pathway (V(gln)). Under hyperammonemia the value of V(TCA) was 0.57 +/- 0.16 micromol/min per g (mean +/- SD, n = 6) and was not significantly different (unpaired t test; P > 0.10) from that measured in the control animals (0.46 +/- 0.12 micromol/min per g, n = 5). Therefore, the TCA cycle rate was not significantly altered by hyperammonemia. The measured rate of glutamine synthesis under hyperammonemia was 0.43 +/- 0.14 micromol/min per g (mean +/- SD, n = 6), which was significantly higher (unpaired t test; P < 0.01) than that measured in the control group (0.21 +/- 0.04 micromol/ min per g, n = 5). We propose that the majority of the glutamine synthetase flux under normal physiological conditions results from neurotransmitter substrate cycling between neurons and glia. Under hyperammonemia the observed increase in glutamine synthesis is comparable to the expected increase in ammonia transport into the brain and reported measurements of glutamine efflux under such conditions. Thus, under conditions of elevated plasma ammonia an increase in the rate of glutamine synthesis occurs as a means of ammonia detoxification, and this is superimposed on the constant rate of neurotransmitter cycling through glutamine synthetase.

Figure 1

Selected time points from sets of ¹H-decoupled ¹³C NMR spectra of rat brain, in vivo, at 7.0 T, acquired during a [1-¹³C]glucose infusion (pre-infusion baseline scans subtracted). The regions of the ¹³C spectrum encompassing the metabolite resonances of interest are shown for a control (a) and hyperammonemic (b) rat, respectively. Labeled resonances are glutamate (Glu), and glutamine (Gln). Time interval between spectra shown is ≈20 min (first and last spectra are ≈10 min and ≈210 min into the glucose infusion, respectively).

Figure 2

Fits to the in vivo ¹³C data for glutamate (○) and glutamine (×) C-4, for a control (a) and hyperammonemic (b) rat. Zero time represents the start of the [1-¹³C]glucose infusion. Peak heights were converted to fractional enrichments as part of the fitting procedure, and the lines represent the best fit to the data generated by the mathematical model (see Methods). The fit to the glutamate C-4 data gives V_TCA, and the fit to the glutamine C-4 data gives the mzximum estimate of V_gln.

Figure 3

Schematic representation of the proposed model. Glc, glucose; Glu, glutamate; Gln, glutamine; NH₃, ammonia; V_trans, rate of ammonia transport into the brain; V_efflux, rate of glutamine efflux from the brain; V_pc, anaplerotic flux; V_cycle, rate of neurotransmitter cycling; V_gln, rate of glutamine synthesis.