Capillary electrophoresis and hydrophilic interaction chromatography coupled to mass spectrometry for anionic metabolic profiling

Abstract

Metabolomics encompasses the identification and quantification of low-molecular weight endogenous metabolites in biological samples, such as cells, tissues and body fluids. The identities, concentrations and fluxes of metabolites result from gene expression, protein expression and environmental factors (e.g. bacteria, diet, drugs), which together determine the phenotype of an organism. Therefore, profiling of metabolites present in biological samples provides information on biochemical processes and phenotypes of organisms. Various analytical techniques have been developed to analyze different classes of metabolites, including capillary electrophoresis (CE) and hydrophilic interaction chromatography (HILIC) coupled to mass spectrometry (MS), which are eminently suited for profiling highly polar compounds. This thesis describes the optimization, evaluation and comparison of CE–MS and HILIC–MS using negative electrospray ionization for the profiling of anionic metabolites. Particular attention was paid to sensitivity, reproducibility, separation selectivity and metabolite coverage. It is demonstrated that the use of triethylamine in the background electrolyte and sheath liquid in CE–MS is an effective way to improve the limits of detection of anionic metabolites significantly as compared to earlier developed CE–MS methods applying ammonium acetate. The increase in signal intensities allows the detection of low-abundant metabolites, thereby increasing the metabolite coverage. This is of utmost importance in metabolic profiling studies. The optimization of the mobile phase composition, the gradient of the mobile phase and sample solvent in HILIC–MS resulted in a large separation window with good peak shapes and low limits of detection of anionic metabolites. Detection limits obtained with HILIC–MS were up to 80 times lower compared to CE–MS, mainly due to the larger injection volume. Moreover, the migration and retention times of anionic metabolites were considerably different, demonstrating the different separation selectivities of CE–MS and HILIC–MS. The applicability of the optimal CE–MS and HILIC–MS methods has been tested by analyzing urine samples in studies on antibiotic-treated rats and patients with inborn errors of metabolism. Reproducible metabolite profiles were obtained as demonstrated by the repetitive analysis of a quality control sample. In general, more molecular features were obtained with HILIC–MS. However, CE–MS also revealed a large number of metabolite features, which were not detected with HILIC–MS. Using multivariate data analysis, different groups of urine samples could be discerned based on obtained CE–MS and HILIC–MS data, and potential biomarkers were revealed. These discriminatory compounds were putatively identified based on accurate mass and MS/MS experiments. Some of the biomarker candidates were observed with both CE–MS and HILIC–MS. Several anionic metabolites were also uniquely revealed by one of the applied techniques. This demonstrates that these analytical techniques can provide important and complementary information in metabolic profiling studies.