A validated multi-matrix platform for metabolomic fingerprinting of human urine, feces and plasma using ultra-high performance liquid-chromatography coupled to hybrid orbitrap high-resolution mass spectrometry

In recent years, metabolomics has surfaced as an innovative research strategy in human metabolism, whereby selection of the biological matrix and its inherent metabolome is of crucial importance. However, focusing on a single matrix may imply that relevant molecules of complementary physiological pathways, covered by other matrices, are missed. To address this problem, this study presents a unique multi-matrix platform for polar metabolic fingerprinting of feces, plasma and urine, applying ultra-high performance liquid-chromatography coupled to hybrid quadrupole-Orbitrap high-resolution mass spectrometry, that is able to achieve a significantly higher coverage of the system's metabolome and reveal more significant results and interesting correlations in comparison with single-matrix analyses. All three fingerprinting approaches were proven ‘fit-for-purpose’ through extensive validation in which a number of endogenous metabolites were measured in representative quality control samples. For targeted and untargeted validation of all three matrices, excellent linearity (coefficients of determination R2 ≥ 0.99 or 0.90 respectively), recovery and precision (coefficients of variance ≤ 15% or 30% respectively) were observed. The potential of the platform was demonstrated by subjecting fecal, urine and plasma samples (collected within one day) from ten healthy volunteers to metabolic fingerprinting, yielding respectively 9 672, 9 647, and 6122 components. Orthogonal partial least-squares discriminant analysis provided similar results for feces and plasma to discriminate according to gender (p-value, R2(X), R2(Y) and Q2(Y)), suggesting feces as an excellent alternative biofluid to plasma. Moreover, combining the different matrices improved the model's predictivity, indicating the superiority of multi-matrix platforms for research purposes in biomarker detection or pathway elucidation and in the selection of the most optimal matrix for future clinical purposes.