Introduction

Previously, a ¹H magnetic resonance spectroscopy (MRS) method has been described to measure kinetics of brain glutamate metabolism following a ¹³C-glucose infusion^1-3. This method involves quantification of the reduction in ¹²C-coupled glutamate ¹H resonance amplitudes and hence does not require specialized ¹³C-dedicated MRS hardware. This highly translational approach has been implemented with nonhuman^1,2 and human³ primate subjects at 3T. Here we demonstrate implementation of this procedure in rodents at 7T.

Methods

Three male Wistar rats (273, 279 and 281 g) fitted with MRI compatible indwelling jugular catheters (Instech, Plymouth Meeting, PA) were maintained on standard chow. After overnight fasting, animals were anesthetized with 1-3% isoflurane and maintained in a heated environment in the scanner with pulse and respiration monitoring. 1H MRS spectra were collected from an 8x3x3 mm³ voxel subtending the thalamus (Figure 1) with a Bruker Biospin 7T system using a standard Bruker PRESS sequence with the transmitter frequency set to 2.35 ppm (256 spectra with outer volume water suppression, TE/TR=16/2500ms, 1024 points for each FID, spectral bandwidth = 4000Hz). ¹³C1-glucose (0.25g/kg) was intravenously administered with an infusion pump over 60 minutes. A total of 14 ~10-minute acquisitions were collected: 3 prior to infusion (baseline), followed by 7 during 13C1-glucose administration, plus 4 post-infusion acquisitions. The total glucose infusion volume was ~3.4cc.

Spectra were analyzed to quantify fractional enrichment of ¹³C4 glutamate by monitoring the reduction in amplitude of ¹²C4-bonded ¹H signal amplitudes. For each animal the average of the baseline spectra was used as a template for frequency alignment for each 14 sets of 256 spectra prior to quantification. Figure 2a shows a time series for one animal over the course of the infusion, and Figure 2b shows corresponding difference spectra. The ¹H4 difference amplitudes were estimated through fitting time-domain data to exponentially-damped sinusoidal functions as in Miller et al.². The amplitude of the glutamate ¹H4 signal in the baseline spectra was quantified through modeling as a Lorentzian line following LCModel4 pre-processing to account for eddy current correction and baseline removal. Specifically, LCModel basis functions representing 24 spin systems, including glutamate, were fitted, and “isolated glutamate spectra” were derived from the LCModel analysis by removal of the fitted spectra of the other 23 spin systems (Figure 3). The percent enrichment at each time point is the ratio of the glutamate ¹H4 difference spectrum amplitude and the baseline isolated ¹H4 amplitude.

Results

The solid traces in Figures 2b and 3 show Lorentzian models of the difference spectra and baseline spectra, respectively. Figure 4 shows the percent glutamate ¹³C4 enrichment for the 3 rats over the 2-hour post-infusion period.

Discussion

For all three rats, the glutamate ¹³C4 enrichment approached an asymptotic value of 25-35% within approximately one hour following ¹³C-labeled glucose infusion. Both the rate of enrichment and the asymptotic value of approximately 30% are similar to prior findings in primate studies conducted with a 3T MR system, utilizing the ¹H MRS method adopted here^1-3, as well as rodent studies utilizing ¹³C MRS methods⁵. Therefore, this report demonstrates the feasibility of extending this technique to small animal research subjects. One challenge for future work will be to quantitatively interpret enrichment kinetics in terms of metabolic flux through biochemical pathways. Specifically, the appearance of ¹³C-coupled glutamate ¹H3 nuclei, following multiple turns of the tricarboxylic acid cycle, is expected to interfere with C4 enrichment quantification due to the near equivalence of the chemical shift difference between glutamate H3 and H4 nuclei, and half the 1-bond ¹³C-H coupling constant at a field strength of 7T.

Conclusion

This work demonstrates feasibility of using ¹H MRS for quantifying the kinetics of glutamate ¹³C4 enrichment following intravenous administration of ¹³C-labeled glucose in a rodent species, using a high-field small-animal MRI system. Given the ease of implementation of this approach in humans and nonhuman primates, on clinical MRI systems, it is anticipated that this approach will provide a valuable translational tool in comparative studies utilizing rodent models of human disease.

Acknowledgements

Supported by National Institutes of Health grants AA-013521, AA-005965, and AA-025473.