Feature-Based Molecular Networking Combined with Multivariate Analysis for the Characterization of Glutathione Adducts as a Smoking Gun of Bioactivation.

Abstract

Feature-based molecular networking (FBMN) is a powerful analytical tool for mass spectrometry (MS)-based untargeted metabolomics data analysis. FBMN plays an important role in drug metabolism studies, enabling the visualization of complex metabolomics data to achieve metabolite characterization. In this study, we propose a strategy for the characterization of glutathione (GSH) adducts formed via in vitro metabolic activation using FBMN assisted by multivariate analysis (MVA). Acetaminophen was used as a model substrate for method development, and the practical potential of the method was investigated by its application to 2-aminophenol (2-AP) and 2,4-dinitrochlorobenzene (DNCB). Two 2-AP GSH adducts and one DNCB GSH adduct were successfully characterized by forming networks with GSH even though the mass spectral information obtained for the parent compound was deficient. False positives were effectively filtered out by the variable influence on projection cutoff criteria obtained from orthogonal partial least-squares-discriminant analysis. The GSH adducts formed by enzymatic or nonenzymatic reactions were intuitively distinguished by the pie chart of FBMN results. In summary, our approach effectively characterizes GSH adducts, which serve as compelling evidence of bioactivation. It can be widely utilized to enhance risk assessment in the context of drug metabolism.