Reactive metabolite trapping screens and potential pitfalls: bioactivation of a homomorpholine and formation of an unstable thiazolidine adduct.

Successful early attrition of potential problematic compounds is of great importance in the pharmaceutical industry. The lead compound in a recent project targeting neuropathic pain was susceptible to metabolic bioactivation, which produced reactive metabolites and showed covalent binding to protein. Therefore, as a part of the backup series for this compound several structural modifications were explored to mediate the reactive metabolite and covalent binding risk. A homomorpholine containing series of compounds was identified without compromising potency. However, when these compounds were incubated with human liver microsomes in the presence of GSH, Cys-Gly adducts were identified, instead of intact GSH conjugates. This article examines the formation of the Cys-Gly adduct with AZX ([M+H](+) 486) as a representative compound for this series. The AZX-Cys-Gly-adduct ([M+H](+) 662) showed evidence of ring contraction by formation of a thiazolidine-glycine and was additionally shown to be unstable. During its isolation for structural characterization by H-1 NMR spectroscopy, it was found to have decomposed to a product with [M+H](+) 446. The characterization and identification of this labile GSH-derived adduct using LC-MS/MS and H-1 NMR are described, along with observations around stability. In addition, various structurally related trapping reagents were employed in an attempt to further investigate the reaction mechanism along with a methoxylamine trapping experiment to confirm the structure of the postulated reactive intermediate.