Novel labeling approaches for the assessment of human hepatic metabolism by in vivo magnetic resonance spectroscopy


  • Martin Krššák

    Corresponding author
    1. From the Division of Endocrinology and Metabolism, Department of Internal Medicine III, and High Field MR Centre, Department of Biomedical Imaging and Image Guided Therapy, Medical University of Vienna, Vienna, Austria
    • Address reprint requests to: Martin Krššák, Ph.D., Division of Endocrinology and Metabolism, Department of Internal Medicine III, Medical University of Vienna, Währinger Gürtel 18-20, A-1090 Wien, Austria. E-mail:; fax: +43-1-40400 7631.

    Search for more papers by this author

  • Potential conflict of interest: Nothing to report.

  • See Article on Page 2321


adenosine triphosphate


hepatocellular lipid


magnetic resonance spectroscopy


tricarboxylic acid.

The metabolism of carbohydrates, lipids, and protein and the detoxification of an organism are the primary roles of the liver. These processes are crucial for the well-being of the human body and a precise knowledge of these mechanisms is essential for understanding the reactions of the liver tissue to chronic or acute pathologic conditions. Biochemical or histopathological analysis of a liver tissue biopsy can supply ultimate answers to many of the diagnostic questions, but its invasive nature inhibits its application in cross-sectional investigations of a broad population, the frequent follow-ups required to monitor therapy. Further on, fast degradation of many metabolites in invasively sampled tissue prevents human studies directly assessing the metabolic fluxes in healthy and diseased states. However, for this purpose in vivo magnetic resonance spectroscopy (MRS), as shown by Skamarauskas et al.[1] in this issue of Hepatology, can serve as a unique tool for direct assessment of metabolic fluxes or measuring metabolite concentration in the hepatic tissue. No other analytic method provides a noninvasive access to such broad biochemical information.

One of the most straightforward applications, currently available on the majority of clinical MR systems, is the measurement of hepatocellular lipid (HCL) content by single voxel localized proton (1H) MRS. The method is sensitive enough to measure HCL levels below 1% (volume fat fraction) and it has already been validated against gold-standard histological and biochemical analysis of tissue biopsy.[2, 3] The application of this noninvasive method enabled studies on larger populations, showing that almost a third of the Dallas County, Texas, population presented with hepatic steatosis.[4] Improvements in the hardware and optimization of the measurement protocol also allow for absolute quantification of the total amount of choline-containing compounds,[5] which can further be studied by phosphorus (31P) MRS. Excellent 31P MRS signal separation can be achieved at magnetic fields ≥3T[6] and changes in the pattern of 31P MR spectra have already been shown to accompany nonalcoholic steatosis and steatohepatitis.[7] An advanced dynamic method of 31P MRS, called saturation transfer, can even assess the chemical exchange between the γ-phosphorus of adenosine triphosphate (ATP) and the inorganic phosphate enabling insight into the kinetics of ATP synthesis adjacent to oxidative and glycolytic hepatic metabolism.[8]

The spatial resolution of 31P and 13C MR experiments is hampered by low intrinsic sensitivity and, in the case of carbon, by a low natural abundance of the isotope (1.1%) as well. To overcome these problems, a volume of interest of up to 100 cm3 is chosen, frequently by applying localization based on the sensitive volumes of radiofrequency (RF) surface coil probes[1, 9] and volume-selective schemes.[6] Using this approach, the detection limit is about 0.1 mmol/L of tissue for 31P and about 20-50 mmol/L of tissue for 13C MRS.[9, 10] Healthy liver presents with a glycogen concentration in a range between 200 and 400 mmol/L of tissue. A large homogenous liver volume close to the body surface can be well covered by the appropriate MR surface coil. Thus, net glycogen accumulation and glycogenolysis can be monitored under various conditions.[9, 11, 12]

Advanced approaches for increasing the signal-to-noise ratio of nonproton (X-nuclei) spectra include: 1) the use of so-called proton decoupling, special hard- and software experimental set-ups eliminating signal splitting of X nuclei due to magnetic interaction with surrounding protons; 2) selective suppression of unwanted signals; and 3) increasing the abundance of isotope of interest by systemic infusion of substrate metabolites in the case of 13C MRS. The infusion of 13C labeled glucose or the ingestion of [1-13C] enriched glucose (10-99%) increases the sensitivity and accuracy of the measurement of hepatic glycogen synthesis.[13] Sequential infusions of enriched and unlabeled glucose, so-called 13C-pulse-12C-chase experiments,[13] also enables the assessment of rates of glycogen synthesis and simultaneous glycogenolysis in humans.[12]

Currently, two different approaches have been introduced to observe metabolic fluxes and the response to antioxidative stress in liver in vivo. Labeling hepatic metabolism by the infusion of [1-13C]-acetate enables the assessment of tricarboxylic acid (TCA) and anaplerotic flux. TCA labeling from [2-13C]-acetate to the C4 position of glutamate has been performed and observed in muscle previously,[14] but a large amount of lipids in the hepatic tissue obscure the spectral appearance of C4-glutamate, and necessitated a different approach to flux labeling and modeling. The 1-13C label from acetate incorporates into the C5 and C1 positions of glutamate, thereby assessing and quantifying the TCA and anaplerotic fluxes in hepatocytes.[15]

Increased synthesis of glutathione is a marker for hepatic response to oxidative stress. As glutathione is a tri-peptide consisting of glycine, cysteine, and glutamate, marking any of the precursors should be observed as an increase in the label in the glutathione pool. This was recently achieved by protocols including infusion and/or ingestion of [2-13C]-glycine in experimental animals and also in humans. Translating the model of glycine to glutathione labeling from studies on neoplastic lesions,[16] Skamarauskas et al.[1] quantified the glutathione synthesis in rats in conditions of acute and chronic oxidative stress, specifically showing increased rates of synthesis in the status after carbon tetrachloride (CCl4) insult.

Applications in humans used oral ingestion of multiple doses of labeled [2-13C]-glycine providing the label delivery to the liver by way of the portal vein. Even this discontinuous delivery approach led to stable isotope enrichment in the plasma sufficient for labeling the glutathione pool in the liver within 8 hours of the experimental protocol. Although the impact of the results obtained on humans is limited due to the small number of volunteers included, quantitative analysis showed that the variability of the glutathione synthesis rate constant is comparable to that of previously mentioned noninvasively assessed rate constants of hepatic glycogen metabolism.[12] Thus, further studies in a larger population that includes patients with acute liver response to potentially toxic treatment or chronic conditions of diffuse liver disease, such as nonalcoholic fatty liver, metabolic syndrome, or Type 2 diabetes, will be of great interest. The combination of different MRS-based techniques in one experiment could be useful to link steady-state metabolite concentrations to dynamic measures of hepatic metabolism.

  • Martin Krššák, Ph.D.

  • Division of Endocrinology and Metabolism Department of Internal Medicine III, and High Field MR Centre Department of Biomedical Imaging and Image Guided Therapy Medical University of Vienna Vienna, Austria