Characterization Of Liver Metabolism By NMR‐Guided GC‐MS Methodology

Abstract

Non‐Alcoholic Fatty liver disease (NAFLD), Non‐alcoholic Steatohepatitis (NASH), liver cirrhosis and liver cancer affect a combined 80–100 million Americans [Perumpail et al. World J Gastroenterology. 2017, 23(47), 8263]. In order to better diagnose these conditions and monitor treatment progression, several metabolomics approaches have been developed. Amongst these, Mass Spectrometry (MS) and Nuclear Magnetic Resonance (NMR) have been used extensively.In this study, a recently developed method of metabolic profiling coined NMR‐Guided LC‐MS is assessed for its ability to accurately measure metabolites. [Nagana Gowda et al. Anal. Chem. 2018, 90(3), 2001]. The principle of NMR‐Guided MS is to obtain a ratio of a metabolites mass spectral intensity to its assigned concentration from 1H NMR spectra in a reference sample. This ratio is then applied to mass spectral intensities for in other samples to assign a metabolites concentration. For NMR‐Guided MS to be applicable to a metabolite of interest the variance between NMR‐Guided MS and NMR concentrations must be low and constant over a concentration range of biological interest. In this case, we apply a similar normalization methodology to GC‐MS to assess the technical possibility of producing quantitative GC‐MS approaches to metabolomic analysis.To achieve the biological range of interest we extracted metabolites from ex vivo C57/BLKS perfused mouse livers under fed and fasted conditions. Additionally, two common extractions methods, perchloric acid extraction (PCA) and acetonitrile‐Isopropanol‐water (ACN/ISO/H2O) were compared.A significant fraction of high concentration compounds, including lactate and alanine, amino acids such as aspartate and glycine, and some TCA cycle intermediates such as succinate had low variance across fed and fasted samples between NMR‐Guided GC‐MS and NMR. Branched chain amino acids had greater variance by NMR‐Guided GC‐MS than GC‐MS alone when compared to NMR. In the case of β‐Hydroxybutyrate (BHB), the variance between NMRGuided GC‐MS and NMR was acceptable when extracted by PCA but not ACN/ISO/H2O for both tissue extracts and standard mixtures. BHB concentrations assigned by GC‐MS have a high variance relative to NMR regardless of the extraction method. However, for the majority of metabolites, differences in variance between extractions and assignment methods were subtle for livers and apparent only at concentration extremes of standards (Figure I).Altogether the data presented here provides insight about the comparability of GC‐MS and NMR systems, with excellent agreement observed between the methods across most metabolites. We have also demonstrated measurable extraction artifacts between PCA and ACN/ISO/H2O methods for some metabolites. In conclusion, NMR‐Guided GC‐MS may lend to more rapid and accurate metabolite measurements, broadening the swatch of applicability of the relatively inexpensive GC‐MS platform to metabolomics profiling.Support or Funding InformationNIH R01DK105346% coefficient of variance (% CV) for assigned concentrations by GC‐MS or NMR‐Guided MS compared to NMR assigned concentrations are presented as heat maps. The top panel represent standard samples across a range of 1 to 18 μg per metabolite extracted by PCA or ACN/ISO/H2O. While the bottom panel represents fed and fasted liver tissues extracted by PCA or ACN/ISO/H2O. Histograms in the middle provide density information on the %CVs and are color matched to their respective panels.Figure 1