The application of mechanistic absorption, distribution, metabolism and excretion studies and physiologically-based pharmacokinetic modeling in the discovery of the next-generation oral selective estrogen receptor degrader camizestrant to achieve an acceptable human pharmacokinetic profile.

Camizestrant (AZD9833) is a potent, next-generation oral selective estrogen receptor (ERα) degrader (SERD) and complete ERα antagonist. It is currently being investigated in multiple phase 3 clinical trials for the treatment of ER+/HER2-negative breast cancer. ER is a clinically validated target, and when camizestrant was progressed into the clinic, the only approved SERD was fulvestrant which is delivered via once-monthly intramuscular injections. Oral delivery of a SERD has proved challenging, often due to poor pharmacokinetic (PK) properties and/or short half-life (t). Camizestrant is designed to overcome these challenges as a low molecular weight, low lipophilicity base with good permeability characteristics. It has extensive tissue distribution (Vdss 4-19 L/kg) and high bioavailability in animals, yet dose-dependent bioavailability (from 2 to >90%) in dogs. The in vitro hepatic metabolic intrinsic clearance (CL) for camizestrant varied across species with CL determined in human hepatocytes (530 μL/min/g liver) lower than rat (5900 μL/min/g liver) and dog (5000 μL/min/g liver). The K determined in dog liver microsomes (0.3 μM) was lower compared to humans (1 μM) and rats (6 μM). Cytochrome P450 phenotyping studies determined metabolism in dogs is primarily mediated by CYP2D15 whereas CYP3A4/5 are the dominant cytochrome P450 isoforms in humans. A physiologically-based pharmacokinetic model was developed and validated based on the integrated mechanistic data which recapitulated the observed dose-dependent exposure in dogs. Human exposures were simulated based on this model which incorporated human-specific parameters and indicated that human clearance was independent of dose at pharmacologically relevant concentrations. The simulated human PK parameters and exposure profile(s) were appropriate for the progression of camizestrant into phase 1 clinical evaluation. Clinical exposure data confirmed the preclinical predictions and have enabled rapid exploration of ERα pharmacology in patients. SIGNIFICANCE STATEMENT: This study demonstrates that understanding interspecies in vitro and in vivo pharmacokinetic properties is imperative to making human-specific PK predictions for candidate drugs. For camizestrant, physiologically-based pharmacokinetic simulations hypothesized concentration-dependent metabolism as the underlying cause for nonlinear PK in the dog. The analysis predicted broadly linear PK in humans at pharmacologically relevant doses. Despite uncertainty raised by the data in dogs, confidence in the predicted human PK profile, using human-relevant data, led to the progression of camizestrant to the clinical.