High-Resolution Mass Spectrometry Coupled with XCMS Online for High-throughput Detection and Identification of Drug Metabolites

Abstract

ID 25135 Poster Board 182 Metabolite identification (MetID) studies provide critical information to elucidate the biotransformation pathways of a new chemical entity during the early stages of drug discovery and development. Liquid chromatography-high resolution mass spectrometry (LC-HRMS) is the gold-standard technique for prediction of metabolite structures, however, the current MetID studies rely on manual data processing. The goal of our study was to analyze LC-HRMS data using a high-throughput metabolomics-based MetID workflow by utilizing an in-house theoretical metabolite predictor and open-access XCMS Online (xcmsonline.scripps.edu).In particular, we applied this approach to detect and identify biotransformation products of cimetidine, probenecid, and imatinib in blood and urine samples collected from three rat pharmacokinetic studies. Briefly, rats (n= 4 to 8) were dosed with placebo (control) or drugs [cimetidine (100 mg/kg, IP), probenecid (1000 mg/kg, IV), and imatinib (30 mg/kg, PO)], and the blood and urine samples were collected at various time points/intervals. The control and drug-treated samples were individually processed by protein precipitation followed by drying and reconstitution in LC-MS compatible solvents. The pooled blood (0.5, 1, and 2 h) and urine (0-4 h) samples of cimetidine and probenecid, and pooled blood (1, 2, and 4 h) samples of imatinib were individually analyzed using nanoLC coupled Thermo Q-Exactive HF mass spectrometer. Data-independent acquisition method was applied to allow detection of the low level metabolites. Cimetidine, probenecid, and imatinib treated samples and the corresponding controls were analyzed by XCMS Online software. Lists of all potential theoretical metabolites and their high-resolution masses were created for these three drugs using a novel in-house theoretical metabolite predictor. The output excel file was first processed to shortlist the potential metabolites using the following optimized criteria: a) metabolite elevated in the treatment group by fold difference of at least 2.5, b) accurate mass within ± 5 ppm of predicted theoretical mass, c) no detectable peak in any control samples and distinguishable peak in all treatment samples, and d) mass-defect shift and the relative retention time within a reasonable range of theoretically possible value. Then, the shortlisted data were searched against the theoretical metabolite list in a semi-automated fashion to identify drug metabolites (Figure 1). Cimetidine-treated rats showed 5 and 6 metabolite hits (3 common) in blood and urine, respectively, which included novel dihydroxyl, carboxyl, and N-acetylcysteine derivatives of cimetidine. Similarly, 6 and 14 potential metabolites (4 common) were detected in the probenecid-treated rat blood and urine, respectively. Of these, dihydroxyl, glycine and glucosylated derivatives of probenecid were detected for the first time. Five reported metabolites were detected and identified in imatinib-treated rat blood samples. Thus, a sensitive and semi-automated workflo...