Purpose

Brain activity-associated energy requirements are satisfied by efficient supply of glucose and oxygen from the blood stream. Despite the effort to understand the compartmentalization of cerebral energy metabolism between neurons and astrocytes, the actual glial contribution to oxidative metabolism is still matter of debate [1]. In this study we took advantage of the columnar characteristics of the Tupaia belangeri primary visual cortex (V1) to measure metabolic changes induced by continuous stimulation of V1.

Methods

9 tree shrews (7 females, 200±56 g, 0.5-6 years old, and 2 males, 211±25 g, 6 months old) under light isoflurane anesthesia (0.5-0.7%) were randomly allocated to experimental groups (n=5 in the stimulation and n=4 in the resting group, respectively). All acquisitions were performed on a 14.1 T/26 cm horizontal bore magnet with homebuilt ¹H quadrature transmit/receive and ¹H quadrature transmit/receive-combined ¹³C linearly polarized surface coils. After anatomical magnetic resonance imaging (MRI), each animal underwent the three MR modalities, namely blood oxygenation level-dependent functional MRI (BOLD fMRI), ¹H and ¹³C magnetic resonance spectroscopy (MRS) localized in V1, either at rest or during visual stimulation. The visual stimulation device was composed of 2 matrices of 64 light-emitting diodes (LED) each that allows delivering visual stimuli at high magnetic field. The paradigm consisted of delivering lines in 4 orientations and 2 directions (Figure 1C) at 5 and 7 Hz randomly switched. Each pattern with spatial frequency 0.04-0.05 cycle/degree was presented for at least 5 s. The luminosity contrast was set to maximum, corresponding to 48 ± 4 LUX at 1 cm distance from the eyes of the animal. fMRI was performed using gradient echo echo planar imaging (GE-EPI) (TR/TE=2.0 s/18 ms; FOV=23×23 mm²; matrix=64×64; slice thickness=0.8 mm ; bandwidth=200 kHz) and allowed mapping cortical activation using the paradigm 30 s ON – 30 s OFF at 5 Hz or 7 Hz. After FAST(EST)MAP shimming, V1-localized ¹H functional MRS (fMRS) and ¹³C MRS (during infusion of [1,6-¹³C]glucose) were performed with SPECIAL [2] (TR=4 s and TE=2.8 ms, VOI = 35 µL) and semi-adiabatic distortionless enhancement by polarization transfer (DEPT) combined with 3D-ISIS for ¹H localization [3] (VOI = 105 µL), respectively. ¹H fMRS was performed during 20 min rest, 20 min of continuous stimulation (at 5 and 7 Hz randomly switched) and 20 min recovery period, and¹³C MRS was performed either at rest or during 4 h of continuous stimulation (at 5 and 7 Hz randomly switched). LCModel was used for analysis of both ¹H and ¹³C spectra [4]. The scaling of ¹³C fractional enrichment (FE) curves was based on MRS of brain extracts [5]. Data was fitted to a two-compartment model and variance of parameters was determined by Monte-Carlo analyses [5][6].

Results

Visual stimulation resulted in a relatively large activated area in V1 that allowed localized MRS (Figure 1A-B). Cortical brain activity resulted in a decrease in both brain glucose concentration (-17% ; -0.34 µmol/g ; P<0.001 one-way ANOVA with Tuckey post-test for multiple comparisons) and phosphocreatine/creatine ratio (-8% ; -0.07; P<0.05) after 15 minutes of stimulation onset (Figure 2). Total creatine concentration did not vary during stimulation as compared to rest. The combinations of high detection sensitivity along with robust cortical activation state allowed observing marked changes in the measured metabolic rates by ¹³C MRS at both individual (Figure 3) and group levels. At the individual level, close relationships between the neurotransmission rate (V_NT) and total cerebral metabolic rate of glucose oxidation (CMR_glc(ox), R²=0.679, P=0.006), glial (V_TCA^g, R²=0.66, P=0.008) and neuronal (V_TCAⁿ, R²=0.40, P=0.066) oxidative metabolism was measured (Figure 3C). At the group level, 20% increase in V_NT (∆V_NT +0.038±0.042 µmol/g/min, P=0.077, permutation test and Holm-Bonferroni correction for multiple comparisons) resulted in a 24% (∆V_TCA^g=0.063±0.057 µmol/g/min; P=0.007) and 12% (∆V_TCAⁿ=0.061±0.032 µmol/g/min; P<0.001) increase in glial and neuronal TCA cycle activity, respectively, resulting in a net 14% increase in CMR_glc(ox) (∆CMR_glc(ox) = 0.058±0.032 µmol/g/min; P<0.001).

Discussion

Taking advantage of the columnar characteristics of tree shrew visual cortex rendered possible continuous stimulation of V1. The fact that we performed continuous stimulation kept metabolism at steady-state during stimulation. Cortical brain activity resulted in a significant increase in CMR_glc as assessed with both ¹H fMRS and ¹³C MRS, which was associated to an increase of similar amplitude of both glial and neuronal oxidative metabolism that scaled with the glutamate-glutamine cycle rate.

Conclusion

These results support the active partnership between neurons and astrocytes as both respond to and support glutamatergic neurotransmission by increasing in the same proportions their oxidative metabolism.

Acknowledgements

This work was supported by National Competence Center in Biomedical Imaging, Swiss National Science Foundation, and Centre d’Imagerie BioMédicale of the UNIL, UNIGE, HUG, CHUV, EPFL, the Leenaards and Jeantet Foundations.