Understanding and reducing the experimental variability of in vitro plasma protein binding measurements.

The experimental measurement of plasma protein binding is a useful in vitro Absorption Distribution Metabolism and Excretion(ADME) assay currently conducted in both screening and definitive early development candidate modes. The fraction unbound is utilized to calculate important pharmacokinetic (PK) parameters such as unbound clearance and unbound volume of distribution in animals that can be used to make human PK and dose predictions and estimate clinically relevant drug–drug interaction potential. Although these types of assays have been executed for decades, a rigorous statistical analysis of sources of variability has not been conducted because of the tedious nature of the manual experiment. Automated conduct of the incubations using a 96-well equilibrium dialysis device as well as high-throughput liquid chromatography–mass spectrometry quantitation has now made this level of rigor accessible and useful. Sources of variability were assessed including well position, day-to-day, and site-to-site reproducibility. Optimal pH conditions were determined using a design of experiments method interrogating buffer strength, CO2% and device preparation conditions. Variability was minimized by implementing an in-well control that is concurrently analyzed with new chemical entity analytes. Data acceptance criteria have been set for both the in-well control and the range of analyte variability, with a sliding scale tied to analyte-binding characteristics.