Introduction

Oxidative stress plays a central role in the development of both acute and chronic liver injury [1]. Glutathione is an endogenous anti-oxidant in the liver and the rate it is replenished determines the degree of injury, its potential to progress and final outcomes [2]. Magnetic Resonance provides the only non-invasive techniques available for measuring hepatic mitochondrial function in vivo [3]. In previous work, a novel 13C labelling strategy using glycine was proposed as a method of measuring glutathione production in the liver [4]. In this study, we extended previous work [5] by developing a method to quantify in vivo concentrations and a robust model of glycine to glutathione flux rates. Results are compared in a test-retest protocol.

Methods

Test day protocol: Eight healthy male participants were scanned on two visits (>7 day wash out period between). During test days, participants were given a [2-13C] glycine solution (99% labelled, 4g in 50ml water) at t=0 mins and every 30 minutes following for 6 doses. At baseline, and following each dose, 13C MRS was acquired and blood samples were taken to measure 13C glycine enrichment.
MR Protocol: All spectra were acquired on a Philips Achieva 3T using a Pulseteq 12cm single loop 13C surface coil with quadrature decoupling [6]. A 13C Urea reference sample was attached at the coil centre. Initial scout images were acquired to confirm correct coil placement and measure relative liver-to-coil positioning, followed by 13C MRS acquisition (adiabatic pulse-acquire,104 averages,512 datapoints,TR=5500ms,total scan time ~10 minutes,narrow band decoupling [5]).

Quantification Development: To quantify 13C concentrations, an average human sized liver phantom was produced (1.1 litres, plastic mould) and filled with a [2-13C] glycine solution (13 mmol/l). Spectra were acquired using in vivo protocol whilst the phantom was systematically repositioned.

MRS Analysis and Quantification: All spectra were phase corrected and fitted to Gaussians to estimate peak areas (Matlab). Scout images were used to measure the distance from the reference sample (coil centre) to (1) the liver edge in AP direction; and (2) the top of the phantom/liver (at the liver-lung boundary) in the FH direction. Peak areas were scaled to the reference peak, plotted against distance in the AP and FH directions and fitted to produce curves of signal change with distance (normalized to maximum signal). The final concentration calculation was modified to include distance correction factors:

L_glycine=k_AP k_FH(R_liver/R_phantom)Ph_glycine

were L_glycine and Ph_glycine are the concentrations in liver and phantom; R_liver and R_phantom are the MRS glycine-to-reference peak ratios in the liver and phantom; and k_AP and k_FH are the coil-to-liver correction factors in the AP and FL directions. The final concentration values were plotted against time and the linear gradient of glutathione concentration was calculated from t=0 to t=135 mins. The participant CV% were also calculated and compared.

Results

Figure 1 shows the change in phantom signal as the coil-to-phantom distance was changed.
The participants change in [2-13C] glycine and [2-13C] glutathione concentration is shown in figures 2-3. Blood glycine levels rose from 0.33±0.07mmol/l at baseline to 1.37±0.27mmol/l at t=180 mins and reduced back to 0.76±0.33mmol/l by t=360 mins. Blood 13C glycine enrichment levels reached a maximum of 79±5%.

The peak hepatic [2-13C] glycine concentration across participants was 15±3mmol/l and 16±5mmol/l for visit 1 and 2 respectively, and the peak [2-13C] glutathione concentration was 3.5±0.5mmol/l and 3.6±0.8mmol/l for visits 1 and 2 respectively. Participant variabilities are shown in figure 4. Without correcting for coil-to-liver distance, the between participant coefficient of variation increased (peak Glutathione_UNCORRECTED CV%=21 and 30%; peak Glycine_UNCORRECTED CV%=24 and 31% for visit 1 and 2 respectively)

Discussion

This study describes in vivo quantification of hepatic glycine-to-glutathione synthesis. The coil-to-liver distance had a significant effect on final values. Future studies should account for this, especially with expected variability in patient geometries (e.g. subcutaneous fat).
Baseline blood glycine concentrations were similar to expected values (Human Metabolite Data Base, https://hmdb.ca/metabolites/HMDB0000123) and rose by a factor >4, leading to an enrichment of ~80%. This is consistent with the high hepatic [2-13C] glycine concentrations observed (up to 15 ± 3 mmol/l), which were ~10 times larger than physiological concentrations reported elsewhere [7] implying that this test-day protocol does indeed flood the hepatic glycine pool as intended. This also suggests a very high fractional enrichment of 13C glutathione in the liver was achieved due to the higher dosing compared to a previous pilot study (3g x 4) [4].

In contrast, the maximum enriched [2-13C] glutathione concentrations were similar to average physiological hepatic glutathione levels [8]. Glutathione synthesis is catalysed by two enzymes, with Glutamate Cysteine Ligase (GCL) thought to be rate limiting so that the flux rate is not dependent on glycine concentration (figure 5). The observation in this study of high glycine concentrations but physiological glutathione levels confirm the assumptions of this metabolic model.

The gradient of glutathione synthesis with time had a much lower variability within participants compared to between participants, indicating that glutathione flux rates measured using this technique are repeatable and provide a reliable means to measure oxidative stress. This can be used in future studies to model metabolic flux in different patient groups and to investigate the effects of drug intervention.

Acknowledgements

No acknowledgement found.