ORIGINAL ABSTRACT

Metabolomics is a valuable tool to assess glyphosate exposure and its potential impact on human health. However, few studies have used metabolomics to evaluate human exposure to glyphosate or glyphosate-based herbicides (GBHs). In this study, an untargeted and targeted metabolomics approach was applied to human skin fibroblasts exposed to the GBH Roundup (GLYP-R). Cytotoxicity, cell death, and oxidative stress assays were performed to evaluate potential damage caused by GLYP-R in fibroblasts. The herbicide showed a cytotoxic effect at concentrations above 100.0 mg L-1, with IC50 = 164.2 ± 8.7 mg L-1, inducing significant reactive oxygen species (ROS) production and necrosis. A GC×GC/Q-TOFMS method using derivatization with propyl chloroformate/propanol was developed for untargeted analysis, allowing the identification of 400 metabolites of different classes in the samples. The most significant compounds in the discrimination and classification of the samples were fatty acids and amino acids (AA). Based on the relevance of AA in untargeted analysis, a targeted analysis of 21 AA was performed using the same validated GC×GC method. Metabolomic analyses allowed the construction of two biomarker models with performance evaluated by receiver operating characteristic (ROC) curves: an untargeted model formed by four metabolites (methylcysteine, N-acetyl-l-methionine, methyl stearate, and linoleic acid) and a targeted model formed by three AA (l-glutamic acid, l-cysteine, and γ-aminobutyric acid). This study is the first to report the use of metabolomics to evaluate human skin cells exposed to GLYP-R, contributing to the toxicological research on glyphosate.