Exposomic biomonitoring of polyphenols by LC-HRMS via non-targeted analysis and suspect screening

Polyphenols are investigated intensively in nutritional and clinical settings because of their beneficial bioactive properties and prevalence in plant-based foods. Due to their complexity, analysis with untargeted approaches is favorable. These approaches typically use high-resolution mass spectrometry (HRMS) rather than low-resolution mass spectrom-etry (LRMS). The advantages of using HRMS were evaluated in this work by thoroughly testing various untargeted techniques and available online resources. By applying iterative data-dependent acquisition on real-life urine samples, 27 features were identified with spectral libraries, 88 with in silico fragmentation, and 113 features were annotated by their monoisotopic mass with PhytoHub, an online database containing >2000 polyphenols. Moreover, potentially polyphenol-related features were searched for using various non-targeted analysis techniques including MassQL for glucuronide and sulfate neutral losses, and MetaboAnalyst for statistical analysis. As HRMS typically suffers from a sensitivity loss compared to state-of-the-art LRMS used in targeted workflows, the sensitivity gap between the two instrumental approaches was quantified. Both instruments showed feasible sensitivity, with the median limits of detection being 10 - 18 ng/mL for HRMS and 4.8 - 5.8 ng/mL for LRMS in three human matrices (urine, serum, and plas-ma). The applicability of HRMS was further evaluated in real-life urine samples with 24 polyphenols detected. The results demonstrate that despite its intrinsic limitations, HRMS can readily be used for comprehensively investigating human exposure and related health effects of polyphenols and their metabolites as well as toxicological mixture effects with other xenobiotics.