Pharmacokinetics (PK) and toxicology of KOS-1022, an orally administered heat shock protein 90 (HSP90) Inhibitor

3093 To date, Hsp90 inhibitors are administered intermittently by IV dosing to patients. It was hypothesized that more continuous inhibition of this chaperone protein (using oral administration) would result in more sustained degradation of its critical oncogenic client proteins. KOS-1022 (17-dimethylaminoethylamino-17-demethoxygeldanamycin) was therefore identified as an orally bioavailable compound for development. Early testing showed that bioavailability equaled 22-54% (oral gavage to rats) and 36% (clinical capsule formulation to dogs). Oral antitumor activity (Hollingshead et al 2005) had been noted at a tolerable dose of 10 mg/kg/day. Subsequent studies investigated the safety and PK of KOS-1022 in the rat and the dog following repeated oral gavage. Initial dose-range finding studies in the rat demonstrated significant hepatotoxicity and body weight loss upon daily dosing (>4 mg/kg). Given this finding and an elimination half-life in humans of >20 hours upon IV dosing, subsequent studies investigated alternate daily dosing. Oral KOS-1022 was administered at 2, 5, and 7 mg/kg (rats) and 0.4, 0.8, and 1.0 mg/kg (dogs) QOD for 28 days. The mid-dose of both studies equaled a maximally tolerable dose (lethality was observed in 10/42 rats and 1/10 dogs at the highest dose). At the MTD, target organs include kidney (changes in electrolytes, increased urea/creatinine; papillary/cortical necrosis), liver (ALT/AST and bilirubin elevation; congestion, necrosis, and multifocal fibrosis), biliary tract, adrenal (primarily cortex), reproductive organs, gastrointestinal epithelium and lymphoproliferative tissues (thymus, lymph nodes, spleen). Treatment-related changes were resolved by the end of the 4-week recovery period with the exception of selected organ weights (increased adrenals and decreased thymus, pituitary and kidney weights) and microscopic findings in the testes and epididymides. High blood concentrations were seen in these studies (rat: AUC 0-48 3701 ng*h/mL on Day 27 at 7 mg/kg; dog 499 ng/mL*hr on Day 27 at 1.0 mg/kg). Comparing the NOAEL and MTD in these studies, KOS-1022 demonstrated generally dose-proportional pharmacokinetics. Tmax was observed 2-6 hours post-dose (rat) and 1.3-1.8 hours (dog). At non-lethal doses, elimination half-life was 7.8 -10.5 hours (rat) and 11.5 - 22.1 hours (dog). Comparison of the rat and dog MTD and NOAEL by allometric scaling concluded that the dog was the more sensitive species (values of 16 and versus 30 and 12 mg/m2 in the rat). The starting dose for human testing was therefore based on 1/6 MTD in the dog. Based on allometric or mg/kg scaling (assuming a 70 kg person), 5 mg/dose was considered a safe starting dose. Clinical testing of KOS-1022 will begin in late 2005. As the drug did not show significant hepatotoxicity in the dog upon daily oral dosing (x 5 days), both alternate daily and daily dosing may be explored in humans.