Integrating In-vitro Permeability Assays within PBPK Modeling to Predict CNS Distribution of Standard and Engineered...

The typically low, molecule-dependent permeability of antibodies across the blood-brain barrier (BBB) has driven the development of engineered constructs with optimized BBB transcytosis to facilitate efficient brain delivery. Physiologically-based pharmacokinetic (PBPK) modeling provides a framework to predict brain disposition and inform drug development; however, current models lack true a priori predictive capability and remain dependent on in-vivo data. In-vitro brain endothelial cell permeability (P_app) assays provide antibody-specific estimates of brain transport, though these assays have yet to be integrated into PBPK models. This study developed and cross-validated an integrated in-vitro-in-vivo-extrapolation (IVIVE) PBPK framework that uses P_app values to predict antibody-specific cerebrospinal fluid (CSF) concentrations a priori. Seven antibodies were evaluated, including four standard (non-FC5-fused) and three TMEM30A-binding constructs (FC5-fused). PBPK modeling was conducted using PK-Sim®/MoBi®. P_app values were used to estimate antibody-specific brain transport parameters. For comparison, a conventional modeling approach was implemented, where brain transport was inferred to be similar to a reference antibody (trastuzumab). Model performance was assessed by comparing predicted versus observed CSF exposures (area-under-the-concentration-time-curve) in rats following intravenous administration. Integration of P_app data substantially improved brain exposure predictions, reducing the absolute average percentage prediction error for CSF exposure from 296.1% (conventional approach) to 53.4% (IVIVE-PBPK). The framework accurately captured the brain disposition of both standard and FC5-fused antibodies without requiring pre-existing in-vivo CSF data. Overall, this P_app-informed PBPK modeling approach enables a priori, mechanistic prediction of antibody brain exposure, supporting candidate selection and reducing reliance on animal studies in brain drug development.