PURPOSEAGM-130 is a cyclin-dependent kinase inhibitor that exhibits dose-dependent efficacy in xenograft mouse models. During preclinical pharmacokinetic (PK) studies, mice and rats showed comparable PK parameters while dogs showed unusually high clearance (CL), which has made human PK prediction challenging. To address this discrepancy, we performed a human microdosing PK and developed a mouse PK/PD model in order to guide the first-in-human studies.METHODSA microdose of AGM-130 was given via intravenous injection to healthy subjects. Efficacy data obtained using MCF-7 breast cancer cells implanted in mice was analyzed using pre-existing tumor growth inhibition models. We simulated a human PK/PD profile with the PK parameters obtained from the microdose study and the PD parameters estimated from the xenograft PK/PD model.RESULTSThe human CL of AGM-130 was 3.08 L/h/kg, which was comparable to CL in mice and rats. The time-courses of tumor growth in xenograft model was well described by a preexisting model. Our simulation indicated that the human doses needed for 50 and 90% inhibition of tumor growth were about 100 and 400 mg, respectively.CONCLUSIONSThis is the first report of using microdose PK and xenograft PK/PD model to predict efficacious doses before the first-in-human trial in cancer patients. In addition, this work highlights the importance of integration of all of information in PK/PD analysis and illustrates how modeling and simulation can be used to add value in the early stages of drug development.