An Optimization of Liquid-Liquid Extraction of Urinary Volatile and Semi-Volatile Compounds and Its Application for Gas Chromatography-Mass Spectrometry and Proton Nuclear Magnetic Resonance Spectroscopy.

Urinary volatile compounds (VCs) have been recently assessed for disease diagnoses. They belong to very diverse chemical classes, and they are characterized by different volatilities, polarities and concentrations, complicating their analysis via a single analytical procedure. There remains a need for better, lower-cost methods for VC biomarker discovery. Thus, there is a strong need for alternative methods, enabling the detection of a broader range of VCs. Therefore, the main aim of this study was to optimize a simple and reliable liquid-liquid extraction (LLE) procedure for the analysis of VCs in urine using gas chromatography-mass spectrometry (GC-MS), in order to obtain the maximum number of responses. Extraction parameters such as pH, type of solvent and ionic strength were optimized. Moreover, the same extracts were analyzed using Proton Nuclear Magnetic Resonance Spectroscopy (1H-NMR), to evaluate the applicability of a single urine extraction for multiplatform purposes. After the evaluation of experimental conditions, an LLE protocol using 2 mL of urine in the presence of 2 mL of 1 M sulfuric acid and sodium sulphate extracted with dichloromethane was found to be optimal. The optimized method was validated with the external standards and was found to be precise and linear, and allowed for detection of >400 peaks in a single run present in at least 50% of six samples-considerably more than the number of peaks detected by solid-phase microextracton fiber pre-concentration-GC-MS (328 ± 6 vs. 234 ± 4). 1H-NMR spectroscopy of the polar and non-polar extracts extended the range to >40 more (mainly low volatility compounds) metabolites (non-destructively), the majority of which were different from GC-MS. The more peaks detectable, the greater the opportunity of assessing a fingerprint of several compounds to aid biomarker discovery. In summary, we have successfully demonstrated the potential of LLE as a cheap and simple alternative for the analysis of VCs in urine, and for the first time the applicability of a single urine solvent extraction procedure for detecting a wide range of analytes using both GC-MS and 1H-NMR analysis to enhance putative biomarker detection. The proposed method will simplify the transport between laboratories and storage of samples, as compared to intact urine samples.

UHPLC-ESI-MS/MS assay for quantification of endocannabinoids in cerebrospinal fluid using surrogate calibrant and surrogate matrix approaches.

Endocannabinoids are endogenous lipids with the main function recognized to act as neuromodulators through their cannabinoid receptors. Dysregulation of the endocannabinoid system is implicated in various pathologies, such as inflammatory and neurodegenerative diseases. In this study we describe a sensitive UHPLC-MS/MS method for the analysis of trace levels of 7 endocannabinoids in cerebrospinal fluid samples. The analytes covered comprised 1- and 2-arachidonoylglycerol 1- and 2-AG (which were analysed as sum due to their interconversion), 2-arachidonylglycerol ether 2-AGE, anandamide AEA, N-linoleoyl ethanolamide LEA, N-palmitoyl ethanolamide PEA and N-oleoyl ethanolamide OEA. Analytes were extracted from the biofluid by a simple monophasic procedure involving protein precipitation with acetonitrile (MeCN). The analytical method is based on chromatographic separation of the analytes with solid-core (core-shell, superficially porous) particle column Cortecs C18+ . Gradient elution with changing proportion of water and acetonitrile and constant concentration of formic acid provided reasonable separation of analytes, close elution of analytes and their internal standards and minimized matrix effects in biological samples. For specific detection of the endocannabinoids a triple-quadrupole tandem mass spectrometer with electrospray ionisation (ESI) and selected reaction monitoring (SRM) mode was used, and it provided good assay selectivity. The developed method required a minute volume of the biological samples (50 µL) and achieved excellent sensitivity (the lower limit of detection was between 4.15 and 30.18 pM of the biological sample). Linear calibration was achieved in the range from 25 to 10,545 pM for AEA, 90-3802 pM for 1-AG, 90-724 pM for 2-AG, 12-5226 pM for LEA, 33-13,942 for OEA, 34-23,850 pM for 2-AGE, 72-30,190 for PEA and 10-4218 for AEA-d4 in CSF. The method was validated and revealed relative errors in the range of - 14.7 to + 12.3% at LLOQ and - 14.1 to + 14.2% for the remaining validation range. Precisions were in the acceptable range (< 20% RSD at LLOQ, and <15% for the remaining levels) as well. It was finally used to quantify endocannabinoids in human cerebrospinal fluid obtained from 118 donors. Accurate quantification of endogenous compounds in biological samples was achieved by using two different principal approaches (surrogate matrix for AEA, 2-AG, OEA, 2-AGE, LEA and PEA, and surrogate calibrant for AEA only) and they were evaluated by use of the Passing-Bablok regression. Concentrations (median) of CSF samples of patients suffering from CNS infection and controls were found to be around 160 pM for 1- and 2-AG, 86 pM for AEA, 62 for 2-AGE, 58 for LEA, 93 pM for PEA, and 83 pM for OEA.