INTRODUCTION

The hippocampus plays a critical role in linking brain energetics and behavior typically observed in the response to stress. With increased interests in the hippocampus as a potential regulator of the hypothalamic-pituitary-adrenocortical (HPA) axis, the final common pathway in the stress response,¹ differentiating glucose metabolism between dorsal hippocampus and hypothalamus may lay foundation for investigating potential mechanisms of HPA sensitivity regulation by the hippocampus.
¹³C Magnetic Resonance Spectroscopy (MRS) has shown to be a very promising tool to study brain metabolism in vivo. Indirect carbon spectroscopy (¹H-[¹³C] MRS) is a technique that allows measuring ¹³C-labelled metabolites with higher sensitivity.² Recently, metabolic fluxes including TCA cycle in mouse hypothalamus were assessable using ¹H-[¹³C] MRS upon infusion of [1,6-¹³C₂] glucose at 14.1T.³
Thus we aimed to compare glucose metabolism of the mouse dorsal hippocampus and hypothalamus upon [1,6-¹³C₂] glucose.

METHODS

Fourteen adult male C57BL/6 mice (25±2g) were scanned under isoflurane anesthesia (1-2%) in a horizontal 14.1T/26cm Varian magnet (Agilent Inc., USA). A homemade ¹H surface coil in quadrature combined with a linear ¹³C coil was designed specificially for ¹H-[¹³C] MRS. The bilateral dorsal hippocampus (2×6×1.5 mm³) and hypothalamus (2×2.7×2.2 mm³) were localized using a set of T₂-weighted FSE images. Field inhomogeneity was adjusted using FASTMAP to reach a water linewidth <25Hz¹. On the target VOI, localized indirect ¹H-[¹³C] detection was applied using the full signal intensity BISEP-SPECIAL sequence (TE/TR=2.8/4000ms) together with OVS and water suppression.^2,3 Glucose metabolism of both bilateral dorsal hippocampus (n=7) and hypothalamus (n=7) was evaluated upon a bolus of 99% enriched 20% (w/v) [1,6-¹³C₂] glucose and followed by continuous infusion of 70% enriched 20% (w/v) [1,6-¹³C₂] glucose up to 4 hours. Spectra were frequency corrected and summed (~5.5min of hippocampus and ~11min for hypothalamus) for LCModel quantification. Non-edited ¹H MR spectra contain ¹H resonances coupled to both ¹²C and ¹³C and thus can be quantified with a standard basis set for neurochemical profiles of mouse hippocampus.^2,3 The ¹³C-editing spectra were quantified using another simulated basis set as previously.^2,3 The fraction of isotope enrichment (FE) in lactate (Lac, LacC3), glutamate (Glu, GluC4), glutamine (Gln, GlnC4), the sum of Glu and Gln (Glx, GlxC3) and γ-aminobutyric acid (GABA, GABAC2 and GABAC3) was obtained. The FEs of plasma glucose (Glc), Lac and acetate (Ace) were measured on three mice with the identical infusion protocol without any MR measurements. All results were then fitted to a one compartment model of glucose metabolism (Figure 1) using MATLAB nonlinear regression methods.^2,3 The cerebral metabolic rate of glucose (CMR_g), tricarboxylic acid cycle (V_TCA), a dilution flux from blood lactate (V_dilⁱⁿ and V_dil^out) and from blood acetate (V_dil^g), a transmitochondrial flux (V_x), apparent neurotransmission rate (V_NT), pyruvate carboxylase flux (V_PC), a Gln efflux (V_eff), GABA flux (V_GABA), and exchange between two GABA pools and two Gln pools (V_ex^g and V_exⁱ) were estimated. Monte–Carlo simulation was used to evaluate the errors of all adjusted metabolic fluxes._^2,3

RESULTS AND DISCUSSION

Based on anatomical images with satisfactory quality, bilateral dorsal hippocampus and hypothalamus were clearly identified (Figure 2). The typical ¹H-[¹³C] MR non-edited spectra of hippocampus (Figure 2a, SNRs=20±2, metabolic linewidths=18±2Hz) and hypothalamus (Figure 2b, SNRs=16±3, metabolic linewidths =17±2Hz) were obtained in jointly with edited spectra of both hippocampus (Figure 2c, SNRs=8.3±0.6) and hypothalamus (Figure 2d, SNRs=5.3±0.3) after 4 hours of [1,6-¹³C₂] glucose infusion, showing clear metabolites labeling. For instance, ¹³C-coupled ¹H resonances of LacC3, GABAC3, GluC3, GlnC3, GABAC2, GluC4, GlnC4, GlxC2, GlcC6 and AspC3 were clearly visible (Figure 2c and 2d).
GlcC6 rose rapidly, nearly in a step-wise fashion, shortly reaching a plateau enrichment at 0.66±0.05 (0.69±0.09 in the first half an hour) for hypothalamus and at 0.78±0.06 (0.69±0.01 in the first half an hour) for hippocampus, respectively. In Figure 3, LacC3 also rose quickly and reached a plateau around 15min after the bolus, with FEs at 0.51±0.02 for hypothalamus and at 0.52±0.02 for hippocampus, After about 100-130 minutes of glucose infusion, Glu and Gln reached their perspective FE steady-states, i.e. 0.51±0.02 for GluC4 and 0.51±0.01 for GlnC4 in hypothalamus and different from 0.50±0.01 for GluC4 (p=0.004) and 0.46±0.02 for GlnC4 (p<0.0001) in hippocampus. FEs of hippocampal GlxC3 and GlxC2 were 0.42±0.02 and 0.35±0.02, while FEs in hypothalamus were 0.46±0.03 (p=0.001) and 0.34±0.03, respectively. FEs of AspC3 were 0.34±0.06 for hippocampus and 0.38±0.09 for hypothalamus. GABAC2 reached a plateau FE approximately at 150 min, 0.43±0.03 in hypothalamus and 0.42±0.03 in hippocampus. A stable FE of GABAC3 was reached only towards the end of the measurements, i.e. 0.36±0.03 in hypothalamus and 0.38±0.03 in hippocampus, respectively.
The averaged FE time courses of GlcC6, LacC3, GluC4, GlnC4, GlxC3 and GABAC2 and GABAC3 were reliably measured and therefore used for modeling (Figure 1). The resulting fits of time courses of ¹³C accumulation (Figure 3) are shown in Figure 4 and the fluxes are summarized in Table 1.

This study shows for the first time an in vivo comparison of metabolic fluxes in the mouse dorsal hippocampus and hypothalamus using indirect ¹H-[¹³C] MRS upon [1,6-¹³C₂]-glucose infusion. This study provides foundation for investigating relevant evidence to the underlying mechanism of hippocampal regulation of HPA axis.

Acknowledgements

This work was supported by the CIBM of the UNIL, UNIGE, HUG; CHUV, EPFL and Leenaards and Louis-Jeantet Foundations.