Synopsis

Keywords: Phantoms, Non-Proton

Largely neglected due to its low concentration, yet brain glycogen plays an active role in brain energy metabolism and may be involved in a variety of brain diseases. The application of ¹³C MRS along with ¹³C-glucose intravenous infusion is to date the forefront method for investigating brain glycogen metabolism in vivo. Notably, the C₁-resonances of glycogen and glucose have been detected non-invasively in the conscious human brain by ¹³C MRS at 4T after ¹³C-glucose infusion labelled at the C₁-carbon. In this study, we explore the potential of ¹³C MRS at 7T with broadband ¹H-decoupling towards measuring glycogen and glucose C₁-resonances.

Introduction

Although accepted as the main energy storage in the central nervous system, the role of glycogen in the conscious human brain is largely unknown. Localized carbon-13 magnetic resonance spectroscopy (¹³C MRS) allows the non-invasive detection of human brain glycogen in vivo, which, due to the low concentration of glycogen in the brain, is typically done via ¹³C-glucose intravenous infusion labelled at the C₁-carbon to enhance the ¹³C sensitivity while giving rise to three C₁-resonances: glycogen, glucose-β, and glucose-α^{1, 2}. To further improve both sensitivity and spectral resolution, the use of ultra-high magnetic field (i.e., ≥ 7T) has the clear advantage of separating the resonances to better facilitate their quantification. In this context, the aim of this study was to explore the potential of ¹³C MRS at 7T in conjunction with broadband ¹H-decoupling while adjusting the decoupling scheme parameters towards optimal sensitivity and simultaneous detection of glycogen and glucose C₁-resonances.

Methods

In vitro ¹H-decoupled ¹³C MRS measurements of glycogen and glucose C₁-resonances were performed on a 7T human scanner (Siemens Erlangen/Germany) using a home-built ¹³C-linear/¹H-quadrature RF surface coil. A pulse-acquire sequence for localized ¹³C MRS using broadband ¹H-decoupling during ¹³C signal acquisition was developed using the WALTZ-16 scheme³ (Figure 1). All in vitro measurements were performed using a two-compartment phantom containing 1) 800mM natural abundance of glycogen and 2) 8mM of glucose-C₁ labelled, while all spectra were acquired using uniform adiabatic ¹³C-excitation (2ms)⁴ by placing the carrier frequency at the glucose-β resonance (96.6ppm). The performance of the ¹H-decoupling scheme was investigated by increasing successively the number of WALTZ cycles from 1 to 8 (i.e., 4 WALTZ cycles corresponds to the WALTZ-16 scheme), while adjusting in each case the decoupling duration accordingly to the FID (~96ms). To that purpose, the duration of the main WALTZ-cycle 90°-pulse was successively decreased from 4ms (1 WALTZ cycle) to 0.5ms (8 WALTZ cycles), such that in all cases the decoupling duration matched that of the FID (~96ms). In addition, the spin-lattice (¹³C-T₁) and spin-spin (¹³C-T₂) relaxations times of glycogen and glucose C₁-resonances were measured in vitro using optimized adiabatic an inversion-recovery and Hahn spin-echo⁵ sequences, over a range of inversion times (TI) and echo times (TE), respectively.

Results and Discussion

In vitro ¹H-decoupled ¹³C MRS at 7T revealed three well resolved C₁-resonances including glycogen (100.5ppm), glucose-β (96.6ppm) and glucose-α (92.8ppm), as by comparing spectra without and with broadband ¹H-decoupling using the WALTZ-16 scheme (Figure 2 A left). By successively increasing the number of WALTZ cycles, the ¹³C signal intensities of glucose-β and glucose-α increased or decreased depending on whether an odd or an even number of WALTZ cycles was applied, respectively (Figure 2 A, B), and this could be attributed to the fact that the ¹H-spin ended up either down (Figure 2 A/ green-arrows) or up (Figure 2 A/ red-arrows), respectively. In contrast, no such effect was observed for glycogen in which the ¹³C signal intensity remained fairly constant (Figure 2 A, B). The slight decrease of the glucose-β slope (Figure 2 B) could be attributed either to the presence of sidebands (Figure 2 A bottom/ ocher-arrows) and/or to the proximity of glucose-β ¹H-resonance to that of water (Figure 2 C). In contrast, a slight increase on the glycogen and glucose-α slopes was observed upon increasing the number of WALTZ cycles (Figure 2 B). In fact, despite the observed sidebands in the glucose-β resonance as of 5 WALTZ cycles applied, an improved decoupling efficiency is expected by increasing successively the number of WALTZ cycles beyond 8, as suggested by the slight increase of the slopes of glycogen and glucose-α from 1 to 8 WALTZ cycles. Note that, given the FID duration of the ¹³C signal of glycogen and glucose C₁-resonances (~96ms) and the ¹H chemical shift range comprising the glycogen and glucose H₁-resonances (~0.9ppm, i.e., ~270Hz), a minimum of 3 WALTZ cycles were needed to achieve a decoupling bandwidth broad enough to fully cover glycogen and glucose H₁-resonances, thus, to fully decouple the glycogen and glucose C₁-resonances (Figure 2 A, C). Besides, the ¹³C-T₁ and ¹³C-T₂ of glycogen C₁ in vitro (Figure 3 B, C) were found in close agreement with what is expected at 7T compared to adjacent field strengths⁶, while the small differences of T₂ compared with the literature⁶ may be due to differences in temperature. The ¹³C-T₁ of glucose-α and glucose-β in vitro were found to be rather similar to each other (Figure 3 B), while their ¹³C-T₂ were slightly disparate (Figure 3 C).

Conclusion

In this study, the simultaneous detection and sensitivity of glycogen and glucose C₁-resonances were evaluated by ¹H-decoupled ¹³C MRS in vitro at 7T while adjusting the decoupling parameters of the WALTZ-16 scheme. A clear improved spectral resolution in the detection of glycogen and glucose C₁-resonances at 7T was observed compared to 4T^1,2. The sensitivity of glucose C₁-resonances increased or decreased depending on whether an odd or an even number of WALTZ cycles was applied, respectively, while this effect was not observed for glycogen C₁-resonance. Overall, the use of 3 versus 4 WALTZ cycles may be advantageous in terms of optimal ¹³C sensitivity and low ¹H-power towards measuring human brain glycogen in vivo at 7T.

Acknowledgements

This study was supported by Centre d’Imagerie BioMédicale (CIBM) of the UNIL, UNIGE, HUG, CHUV, EPFL and the Leenaards and Jeantet Foundations.