A Brain PBPK Model Prospectively Predicts Fenebrutinib CSF Exposure
Genentech researchers combined in vitro permeability, efflux and binding data in a five-compartment brain model. Predictions were within 1.5-fold of cerebrospinal-fluid concentrations measured later in people with multiple sclerosis.
Building the Prediction from In Vitro Inputs
A Genentech team evaluated a bottom-up approach to predicting central nervous system exposure for fenebrutinib. The work combined in vitro measurements of permeability, efflux-transporter activity and brain binding with physiological inputs in a physiologically based pharmacokinetic (PBPK) model with five brain compartments.
The researchers first assessed the approach using a compound with publicly available clinical cerebrospinal-fluid (CSF) data. This step informed empirical scaling factors intended to account for gaps between in vitro measurements and their in vivo translation. The refined approach was then used to predict fenebrutinib CSF concentrations before the corresponding clinical measurements became available.
Comparing the Forecast with Later Clinical Measurements
The predicted fenebrutinib concentrations were within 1.5-fold of subsequently measured CSF concentrations in patients with multiple sclerosis. This prospective comparison supports the model’s use in early assessment of CNS exposure for this compound.