Preclinical Development of an anti-5T4 Antibody-Drug Conjugate: Pharmacokinetics in Mice, Rats, and NHP and Tumor/Tissue Distribution in Mice

The pharmacokinetics of an antibody (huA1)-drug (Auristatin microtubule disrupting MMAF) conjugate, targeting 5T4-expressing cells, were characterized during the discovery and development phases in female nu/nu mice and cynomolgus monkeys after a single dose and in S-D rats and cynomolgus monkeys from multidose toxicity studies. Plasma/serum samples were analyzed using an ELISA-based method for antibody and conjugate (ADC) as well as for the released payload using an LC-MS/MS method. In addition, the distribution of the Ab, ADC, and released payload (cys-mcMMAF) was determined in a number of tissues (tumor, lung, liver, kidney, and heart) in two tumor mouse models (H1975 and MDA-MB-361-DYT2 models) using similar LBA and LC-MS/MS methods. Tissue distribution studies revealed preferential tumor distribution of cys-mcMMAF and its relative specificity to the 5T4 target containing tissue (tumor). Single dose studies suggests lower CL values at the higher doses in mice, although a linear relationship was seen in cynomolgus monkeys at doses from 0.3 to 10 mg/kg with no evidence of TMDD. Evaluation of DAR (drug-antibody ratio) in cynomolgus monkeys (at 3 mg/kg) indicated that at least half of the payload was still on the ADC 1 to 2 weeks after IV dosing. After multiple doses, the huA1 and conjugate data in rats and monkeys indicate that exposure (AUC) increases with increasing dose in a linear fashion. Systemic exposure (as assessed by Cmax and AUC) of the released payload increased with increasing dose, although exposure was very low and its pharmacokinetics appeared to be formation rate limited. The incidence of ADA was generally low in rats and monkeys. We will discuss cross species comparison, relationships between the Ab, ADC, and released payload exposure after multiple dosing, and insights into the distribution of this ADC with a focus on experimental design as a way to address or bypass apparent obstacles and its integration into predictive models.