Optimisation of Biofluid and Tissue Metabolite and Lipid Extraction for Clinical Metabolic Phenotyping

Introduction Clinical metabolic phenotyping aims to detect and measure 1000s of metabolites and lipids within clinical samples (e.g. biofluids and tissues) to identify changes between metabolic phenotypes (e.g. disease status) and to understand biochemical mechanisms driving the phenotype. Sample extraction is a critical step in clinical metabolic phenotyping: it must be reproducible and give a high extraction yield of metabolites and lipids. Technological and methodological innovation We tested multiple monophasic and biphasic metabolite/lipid extraction methods for biofluids (urine/plasma) and tissue (heart/kidney/liver). We also tested solvent-biofluid incubation time/temperature. Extracts were analysed by UHPLC-MS assays: HILIC (urine, plasma, tissue polar extracts); C18 aqueous reversed phase [RP] (urine polar extracts); C18 reversed phase (plasma & tissue lipid extracts). Each method was assessed for yield, reproducibility and class of extracted metabolites. Results and impact Based on yield and reproducibility the best methods were: plasma/urine HILIC– monophasic 50:50 methanol (MeOH):acetonitrile (ACN); urine RP – any tested monophasic method except 100% ACN; plasma lipids – monophasic 100% isopropanol (IPA). Altering solvent-biofluid incubation time/temperature had little effect on yield. For tissue, MeOH/CHCl3/H2O was the best all-round method; however, for some specific compounds other methods performed better, e.g. cardiolipins were better extracted by 100% IPA.