Modeling the subcutaneous pharmacokinetics of antibodies co-administered with rHuPH20.

Abstract

Predicting the subcutaneous (SC) pharmacokinetics (PK) of antibodies in humans is challenging, with clinical data currently being the only reliable data source for modeling SC absorption and bioavailability. Recombinant human hyaluronidase PH20 (rHuPH20) is an enzyme that facilitates SC delivery of high-dose, high-volume therapeutics. Numerous monoclonal antibodies have been co-administered SC with rHuPH20 in a clinical setting, establishing an extensive PK database. The goal of this work is to demonstrate how aggregated clinical data can be leveraged in a universal modeling framework for characterizing SC antibody PK, resulting in parameterization that can be used in predictive simulations of new antibodies. Data for 10 individual antibodies co-administered SC with rHuPH20 were obtained from publicly available sources. PK modeling of each antibody was conducted using the same model structure, but uniquely parameterized. The model structure consisted of a two-compartment model to capture linear kinetics, plus a target-binding mechanism to accommodate nonlinear kinetics driven by antibody-target complex formation and elimination. The clinical PK profiles for all antibodies were accurately described using the universal modeling framework. The SC PK parameters of absorption and bioavailability were consistent across the range of antibody and target properties evaluated. SC administration with rHuPH20 yielded a 30% increase in absorption rate on average and similar or better bioavailability. These parameter values can serve as initial conditions for model-based PK predictions for new antibodies co-administered SC with rHuPH20 to enable evaluation of optimal SC dose and schedule regimens prior to and during clinical development.