1064

In vivo Detection of Metabolic Turnover of GABA and Glutamate in Human Brain using Dynamically Acquired MEGA-PRESS MRS During 13C-Labeled Glucose Infusion
Masoumeh Dehghani1,2, Pierre Etienne3, and Jamie Near1,2

1Department of Psychiatry, Mcgill university, Montreal, QC, Canada, 2Centre d'Imagerie Cérébrale, Douglas Mental Health University, Montreal, QC, Canada, 3Clinical Research Division, Douglas Institute Psychiatrist, Montreal, QC, Canada

Synopsis

Glucose, the main substrate for cerebral energy metabolism, serves as a metabolic precursor for both glutamate and GABA synthesis. In the current study, we employ a novel approach to investigate 13C-labeling of both glutamate and GABA in the human brain. Specifically, localized homonuclear (1H) J-difference edited (MEGA-PRESS) MRS spectra were acquired dynamically (without heteronuclear decoupling or editing pulses) to detect glutamate and GABA labelling following an infusion of 13C labeled glucose. Despite excellent spectral quality and temporal stability, little or no GABA labeling was observed, raising some questions as to the functional status of the GABA pools detected by MEGA-PRESS.

Introduction

Glutamate (Glu) and -aminobutyric acid (GABA) are the main excitatory and inhibitory neurotransmitters in the brain, respectively1,2. Glucose,the main substrate for cerebral energy metabolism, serves as a metabolic precursor for both Glu and GABA synthesis3,4. NMR spectroscopy can been used to probe the metabolism of Glu and GABA in vivo following the infusion of 13C‐labeled glucose. While many previous studies have examined the turnover of Glu in the human brain, few have focused on the GABA. In the current study, we employ a novel approach to investigate 13C-labeling of both Glu and GABA in the human brain.

Methods

One healthy female volunteer provided informed consent to participate in this study. Experiments were performed on a 3T Siemens Prisma MR scanner with a commercial body transmit-volume coil and 32-channel receive array. Bo field inhomogeneities were minimized within VOI(50x45x35 mm3) positioned over the precuneus/posterior cingulate cortex using the GRE-shim procedure, resulting in a water linewidth of 6 Hz. Localized water suppressed 1H spectra were acquired using MEGA-PRESS J-difference editing sequence5 with with the parameters shown in Fig.1. Prior to infusion, one baseline water-unsuppressed and two water-suppressed scans were acquired. The infusion consisted of a solution of 99% [1-13C] glucose (20% w/w), administrated intravenously(0.30 g/kg body-weight) at a constant rate over 40 min. During and following the infusion, repeated water-suppressed MEGA-PRESS scans were acquired to track dynamic changes in 1H signals due to incorporation of the 13C label. Spectral pre-processing steps, including coil combination, phase and frequency correction and averaging were performed in MATLAB using the FID-A toolkit6. The MEGA-PRESS edit-on and edit-off subspectra at each timepoint were aligned and subtracted to yield the MEGA-PRESS difference spectra. A time-series of labeling-specific signals was then obtained by subtracting the processed pre-infusion MEGA-PRESS difference spectra from each of the subsequent timepoints. Quantification was performed in LCModel using basis sets that were simulated in-house using the FID-A toolkit6. For analysis of the pre-infusion MEGA-PRESS difference spectrum, the basis set included only the spectral shapes of standard 1H metabolites. For analysis of the subsequent labelling timeseries, the basis set took into account the pronounced effect of heteronuclear (13C-1H) scalar coupling on the observed 1H spectra, as described previously7. The fractional enrichment(FE) of Glu and GABA were estimated at each time point.

Results

Fig.1 shows the region of interest for acqusition, along with the MEGA-PRESS edit-on, edit-off and difference spectra. Fig.2 shows the simulated basis spectra used to fit the labelling timeseries in LCModel. Note that large reductions in the signal intensity of the 3.0 ppm GABA resonance are expected to result from labeling at either the C4 position of GABA (due to reduction in endogenous 12C-H4 of GABA) or the C3 positions of GABA (due to inefficacy of the 1.9 ppm editing pulse when the 12C-H3 of GABA resonance is split due to 13C coupling). Fig.3 shows the time-series of labeled-difference spectra obtained by the subtraction of the baseline difference spectrum from post-infusion time-series spectra. As expected, a clear decrease in signal from 12C-bonded protons and an increase in signal from 13C-coupled protons was observed for Glu. However, labeling of GABA was not obvious. In particular, no appreciable reduction in the 3.0 ppm GABA signal was observed, indicating minimal labeling at C3 and C4 positions of GABA (Fig3). The LCModel fit of the final post-infusion labeled-difference spectra, along with the fit residuals and the estimated fit components for Glu labeled at positions C4, C3, C2 and GABA labeled at positions C2 and C3 are shown in Fig.4. The turnover curves of total FE of Glu and GABA are shown in Fig.5. At the end of the acquisition, the total FEs of Glu and GABA were 27 % and 8 %, respectively.

Discussion

In the present study, we demonstrate the use of MEGA-PRESS editing sequence to follow the fate of 13C label from infused [1-13C] glucose in the human brain, in the absence of complicated hardware and heteronuclear decoupling RF pulses. The presence of the 13C label uptake was clearly detectable in Glu, owing to the pronounced effect of heteronuclear (13C-1H) scalar coupling on the observed 1H spectra. However, the FE of GABA was lower than expected. This finding leads us to question the degree to which the 3.0 ppm MEGA-PRESS GABA resonance truly reflects GABA that is synthesized from glucose via Glu.

Conclusion

These preliminary results suggest that MEGA-PRESS editing sequence has the potential to clearly detect the conversion of 13C labeled glucose into Glu in the brain. However, detection of GABA labelling using this approach remains a work in progress.

Acknowledgements

We would like to thank Holly Newbold-Fox for her assistance with the infusion and MR scanning. This work is supported by the National Engineering and Sciences Research Council (NSERC, RGPIN-2014-06072, J.N.), and the Fonds de recherche du Quebec – Santé (FRQS, J.N.).

References

  1. Puts NAJ, Edden RAE. In vivo magnetic resonance spectroscopy of GABA: a methodological review. Prog Nucl Magn Reson Spectrosc. 2012;60:29–41.
  2. Meldrum BS. Glutamate as a neurotransmitter in the brain: review of physiology and pathology. J Nutr. 2000;130:1007S-15S.
  3. Bak LK, Schousboe A, Waagepetersen HS. The glutamate/GABA-glutamine cycle: aspects of transport, neurotransmitter homeostasis and ammonia transfer. J Neurochem. 2006;98(3):641–53.
  4. Calvetti D, Somersalo E. Ménage à trois: the role of neurotransmitters in the energy metabolism of astrocytes, glutamatergic, and GABAergic neurons. J Cereb Blood Flow Metab. 2012;32(8):1472–83.
  5. Tkác I, Starcuk Z, Choi IY, Gruetter R. In vivo 1H NMR spectroscopy of rat brain at 1 ms echo time. Magn Reson Med. 1999;41(4):649–56.
  6. Near J, Evans CJ, Puts NA, Barker PB, Edden RA. J-difference editing of GABA: simulated and experimental multiplet patterns. J Soc Magn Reson Med. 2013;70(5).
  7. Simpson R, Devenyi GA, Jezzard P, Hennessy TJ, Near J. Advanced processing and simulation of MRS data using the FID appliance (FID-A)-An open source, MATLAB-based toolkit. Magn Reson Med. 2017;77(1):23–33.
  8. Boumezbeur F, Besret L, Valette J, Vaufrey F, Henry P-G, Slavov V, et al. NMR measurement of brain oxidative metabolism in monkeys using 13C-labeled glucose without a 13C radiofrequency channel. Magn Reson Med. 2004;52(1):33–40.

Figures

Figure 1. (a) Three-dimensional T1-weighted anatomical image was acquired using MPRAGE. Localized VOI (5×4.5×3.5 cm3) was placed on the precuneus/posterior cingulate cortex. (b) Pre-infusion 1H MR spectra acquired in vivo from the human brain at 3T using edit-off and edit-on MEGA-PRESS sequence. Acquisition parameters included TR/TE = 3000/68 ms; 80 averages; refocusing pulse bandwidth 1.4 kHz; 14 ms editing pulses placed alternately at 1.88 ppm and 7.5 ppm. (c) Difference spectrum was obtained by subtracting two subspectra edit-off and edit-on (scaled by 10). LCModel fit of the difference spectra was shown for peaks corresponding to GABA and Glu.

Figure 2. The simulated lineshapes accounting for labeling of GABA and Glu at positions C4, C3, and C2. Reductions in the MEGA-PRESS edited GABA signal at 3 ppm are expected to result from 13C labelling at either C4 or C3 positions of GABA.

Figure 3. Stacked plot of labeled-difference spectra obtained from subtraction of the pre-infusion difference baseline from post-infusion time-series difference spectra (2 Hz filtering was applied). 13C labeling of Glu at positions C4, C3 and C2 is evidenced by a reduction in signal from 12C-bonded protons at 2.1, 2.35, 3.7 ppm and an increase in the satellite peaks from 13C-coupled protons spread around 12C-bonded peaks.

Figure 4. LCModel fitting of the labeled-difference spectra obtained from subtraction of baseline pre-infusion and post-infusion difference spectra obtained at the end of infusion. LCModel fit, fit residual, and fitted contributions from Glu-C4, Glu-C3, Glu-C2, GABA-C2, GABA-C3 and the baseline are shown.

Figure 5. Estimated time courses of 13C labeling of Glu and GABA in the precuneus/posterior cingulate cortex of human brain. Each time point represents the total fractional enrichment of Glu and GABA obtained from summing the labeling of Glu at positions C4, C3, and C2 and summing the labeling of GABA at positions C2 and C3, respectively. The fractional enrichments of Glu and GABA were calculated by comparing the post-infusion labelling time-series to the Glu and GABA concentrations estimtated from the pre-infusion baseline scan.

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)
1064