Obtaining comprehensive, untargeted metabolic profiles for complex solid samples, e.g., animal tissues, requires sample preparation and access to information-rich analytical methodologies such as mass spectrometry (MS). Here we describe a practical two-step process for tissue samples that is based on extraction into 'aqueous' and 'organic' phases for polar and nonpolar metabolites. Separation methods such as ultraperformance liquid chromatography (UPLC) in combination with MS are needed to obtain sufficient resolution to create diagnostic metabolic profiles and identify candidate biomarkers. We provide detailed protocols for sample preparation, chromatographic procedures, multivariate analysis and metabolite identification via tandem MS (MS/MS) techniques and high-resolution MS. By using these optimized approaches, analysis of a set of samples using a 96-well plate format would take ~48 h: 1 h for system setup, 8-10 h for sample preparation, 34 h for UPLC-MS analysis and 2-3 h for preliminary/exploratory data processing, representing a robust method for untargeted metabolic screening of tissue samples.

Alcoholism is a complex disorder that, in man, appears to be genetically influenced, although the underlying genes and molecular pathways are not completely known. Here the intragastric alcohol feeding model in rodents was used together with high mass accuracy LC/MSn analysis to assess the metabonomic changes in nonpolar metabolite profiles for livers from control and alcohol treated rats and mice. Ion signals with a peak area variance of less than 30% (based on repeat analysis of a pooled quality control sample analysed throughout the batch) were submitted to multivariate statistical analysis (using principal components analysis, PCA). PCA revealed robust differences between profiles from control and alcohol-treated animals from both species. The major metabolites seen to differ between control and alcohol-treated animals were identified using high accuracy MSn data and verified using external search engines (http://www.lipidmaps.org; http://www.hmdb.ca; http://www.genome.jp/kegg/) and authentic standards. The main metabolite classes to show major changes in the alcoholic liver-derived samples were fatty acyls, fatty acid ethyl esters, glycerolipids and phosphatidylethanol homologues. Significant metabolites that were up-regulated by alcohol treatment in both rat and mouse livers included fatty acyls, metabolites such as octadecatrienoic acid and eicosapentaenoic acid, a number of fatty acid ethyl esters such as ethyl arachidonate, ethyl docosahexaenoic acid, ethyl linoleate and ethyl oleate and phosphatidylethanol (PEth) homologues (predominantly PEth 18:0/18:2 and PEth 16:0/18:2; PEth homologues are currently considered as potential biomarkers for harmful and prolonged alcohol consumption in man). A number of glycerophospholipids resulted in both up-regulation (m/z 903.7436 [M+H]+ corresponding to a triglyceride) and down-regulation (m/z 667.5296 [M+H]+ corresponding to a diglyceride). Metabolite profiles were broadly similar in both mouse and rat models. However, there were a number of significant differences in the alcohol-treated group particularly in the marked down-regulation of retinol and free cholesterol in the mouse compared to the rat. Unique markers for alcohol treatment included ethyl docosahexaenoic acid. Metabolites were identified with high confidence using predominantly negative ion MSn data for the fatty acyl components to match to www.lipidmaps.org MS and MS/MS databases; interpreting positive ion data needed to take into account possible adduct ions which may confound the identification of other lipid classes. The observed changes in lipid profiles were consistent with alcohol induced liver injury in humans.