Novel Bioactivation of Isoxazole‐containing Bromodomain and Extra Terminal Domain (BET) Inhibitors

Abstract

Bromodomain and extra-terminal (BET) inhibitors comprise a growing class of anti-cancer agents. New generations of BET inhibitors seek to capitalize on the 3,5-dimethylisoxazole motif and incorporate more molecular diversity for higher selectivity. Isoxazole metabolism and bioactivation resulting in toxicity is low but importantly, not absent and woefully understudied for BET inhibitors. We hypothesize that BET inhibitor substituents modulate risks for isoxazoles to undergo bioactivation into reactive metabolites as potential initiators of toxicity. As a test, we carried out a novel iterative modeling strategy to predict bioactivation pathways involving isoxazoles and then validated them experimentally for selected BET inhibitors early and late in development. Specifically, we predicted possible bioactivation pathways for 32 isoxazole-containing BET inhibitors varying in molecular structure and selectivity toward bromodomain 4. We then selected two drug leads, OXFBD02 and OXFBD04, and one late in development, I-BET151, for more complete predictions by coupling our models for quinone formation, metabolite structure and reactivity. Those efforts demonstrated all three leads were highly susceptible to bioactivation into traditional quinones and novel extended quinone-methide involving the isoxazole that readily reacted with glutathione to form adducts. For validating model predictions, we carried out steady-state kinetic studies to identify and assess bioactivation pathways by trapping quinones with dansyl glutathione. Analysis of adducts by MS characterized structures, while dansyl fluorescence provided quantitative yields for their formation over time. The resulting kinetics corroborated reported extensive metabolism of OXFBD02 and revealed three bioactivation pathways preferentially yielding six traditional quinone isomers and as a minor pathway, the proposed, novel extended quinone-methide. For OXFBD04, substitution of the OXFBD02 phenyl group with pyridine significantly reduced metabolism and favored a single bioactivation pathway generating two traditional quinone isomers and the novel extended quinone-methide. Overall suppressed OXFBD04 metabolism relative to OXFBD02 was consistent with expected deactivation of aromatic rings to metabolism; however, the pyridine ring was not the target for metabolism making the observed decreased metabolism and bioactivation a surprising. I-BET151 underwent minimal metabolism and no bioactivation. For both OXFBD leads, modeling correctly predicted the extended quinone-methide, but incorrectly scaled its relative importance in the other more important, bioactivation pathways. To date, there are no reported studies investigating whether bioactivation of isoxazole containing molecules contributes to toxicity. Nevertheless, knowledge of the relative significance of bioactivation pathways for 3,5-dimethyl isoxazole-containing BET inhibitors could aid in targeting future, more detailed studies to assess those issues and their relevance to toxicity.