Bioactivation of isothiazoles: minimizing the risk of potential toxicity in drug discovery

Compound 1, (7-methoxy-N-((6-(3-methylisothiazol-5-yl)-[1,2,4]triazolo[4,3-b]pyridazin-3-yl)methyl)-1,5-naphthyridin-4-amine) is a potent, selective inhibitor of c-Met (mesenchymal-epithelial transition factor), a receptor tyrosine kinase that is often deregulated in cancer. Compound 1 displayed desirable pharmacokinetic properties in multiple preclinical species. Glutathione trapping studies in liver microsomes resulted in the NADPH-dependent formation of a glutathione conjugate. Compound 1 also exhibited very high in vitro NADPH-dependent covalent binding to microsomal proteins. Species differences in covalent binding were observed, with the highest binding in rats, mice, and monkeys (1100-1300 pmol/mg/h), followed by dogs (400 pmol/mg/h) and humans (144 pmol/mg/h). This covalent binding to protein was abolished by coincubation with glutathione. Together, these in vitro data suggest that covalent binding and glutathione conjugation proceed via bioactivation to a chemically reactive intermediate. The cytochrome (CYP) P450 enzymes responsible for this bioactivation were identified as cytochrome P450 3A4, 1A2, and 2D6 in human and cytochrome P450 2A2, 3A1, and 3A2 in rats. The glutathione metabolite was detected in the bile of rats and mice, thus demonstrating bioactivation occurring in vivo. Efforts to elucidate the structure of the glutathione adduct led to the isolation and characterization of the metabolite by NMR and mass spectrometry. The analytical data confirmed conclusively that the glutathione conjugation was on the 4-C position of the isothiazole ring. Such P450-mediated bioactivation of an isothiazole or thiazole group has not been previously reported. We propose a mechanism of bioactivation via sulfur oxidation followed by glutathione attack at the 4-position with subsequent loss of water resulting in the formation of the glutathione conjugate. Efforts to reduce bioactivation without compromising potency and pharmacokinetics were undertaken in order to minimize the potential risk of toxicity. Because of the exemplary pharmacokinetic/pharmacodynamic (PK/PD) properties of the isothiazole group, initial attempts were focused on introducing alternative metabolic soft spots into the molecule. These efforts resulted in the discovery of 7-(2-methoxyethoxy)-N-((6-(3-methyl-5-isothiazolyl)[1,2,4]triazolo[4,3-b]pyridazin-3-yl)methyl)-1,5-naphthyridin-4-amine (compound 2), with the major metabolic transformation occurring on the naphthyridine ring alkoxy substituent. However, a glutathione conjugate of compound 2 was produced in vitro and in vivo in a manner similar to that observed for compound 1. Furthermore, the covalent binding was high across species (360, 300, 529, 208, and 98 pmol/mg/h in rats, mice, dogs, monkeys, and humans, respectively), but coincubation with glutathione reduced the extent of covalent binding. The second viable alternative in reducing bioactivation involved replacing the isothiazole ring with bioisosteric heterocycles. Replacement of the isothiazole ring with an isoxazole or a pyrazole reduced the bioactivation while retaining the desirable PK/PD characteristics of compounds 1 and 2.