Development and evaluation of a comprehensive workflow for suspect screening of exposome-related xenobiotics and phase II metabolites in diverse human biofluids.

Suspect and non-target screening (SNTS) methods are being promoted in order to decode the human exposome since a wide chemical space can be analysed in a diversity of human biofluids. However, SNTS approaches in the exposomics field are infra-studied in comparison to environmental or food monitoring studies. In this work, a comprehensive suspect screening workflow was developed to annotate exposome-related xenobiotics and phase II metabolites in diverse human biofluids. Precisely, human urine, breast milk, saliva and ovarian follicular fluid were employed as samples and analysed by means of ultra-high performance liquid chromatography coupled with high resolution tandem mass spectrometry (UHPLC-HRMS/MS). To automate the workflow, the "peak rating" parameter implemented in Compound Discoverer 3.3.2 was optimized to avoid time-consuming manual revision of chromatographic peaks. In addition, the presence of endogenous molecules that might interfere with the annotation of xenobiotics was carefully studied as the employment of inclusion and exclusion suspect lists. To evaluate the workflow, limits of identification (LOIs) and type I and II errors (i.e., false positives and negatives, respectively) were calculated in both standard solutions and spiked biofluids using 161 xenobiotics and 22 metabolites. For 80.3 % of the suspects, LOIs below 15 ng/mL were achieved. In terms of type I errors, only two cases were identified in standards and spiked samples. Regarding type II errors, the 7.7 % errors accounted in standards increased to 17.4 % in real samples. Lastly, the use of an inclusion list for endogens was favoured since it avoided 18.7 % of potential type I errors, while the exclusion list caused 7.2 % of type II errors despite making the annotation workflow less time-consuming.

Sample preparation for suspect screening of persistent, mobile and toxic substances and their phase II metabolites in human urine by mixed-mode liquid chromatography.

Persistent, mobile and toxic substances have drawn attention nowadays due to their particular properties, but they are overlooked in human monitorization works, limiting the knowledge of the human exposome. In that sense, human urine is an interesting matrix since not only parent compounds are eliminated, but also their phase II metabolites that could act as biomarkers. In this work, 11 sample preparation procedures involving preconcentration were tested to ensure maximum analytical coverage in human urine using mixed-mode liquid chromatography coupled with high-resolution tandem mass spectrometry. The optimized procedure consisted of a combination of solid-phase extraction and salt-assisted liquid-liquid extraction and it was employed for suspect screening. Additionally, a non-discriminatory dilute-and-shoot approach was also evaluated. After evaluating the workflow in terms of limits of identification and type II errors (i.e., false negatives), a pooled urine sample was analysed. From a list of 1450 suspects and in-silico simulated 1568 phase II metabolites (i.e. sulphates, glucuronides, and glycines), 44 and 14 substances were annotated, respectively. Most of the screened suspects were diverse industrial chemicals, but biocides, natural products and pharmaceuticals were also detected. Lastly, the complementarity of the sample preparation procedures, columns, and analysis conditions was assessed. As a result, dilute-and-shoot and the Acclaim Trinity P1 column at pH = 3 (positive ionization) and pH = 7 (negative ionization) allowed the maximum coverage since almost 70 % of the total suspects could be screened using those conditions.