From Cell to Organism: A Predictive Multiscale Model of Drug Transport.

Abstract

Optimized pharmacokinetic properties of drug candidates are desired to be predicted as early as possible in drug discovery and development. Modeling and simulation have been continuously contributing to facilitating drug discovery and development. A cell-based pharmacokinetic model (1CellPK) was developed to mimic the transport of small molecules through a polarized epithelial cell. Passive transcellular permeability and subcellular distribution of small molecules are predicted by 1CellPK in the presence of an apical-to-basolateral concentration gradient. Input parameters include the physiological parameters of cells and physicochemical properties of small molecules. Basic principles applied in the model are mass conservation, Fick's law of diffusion, Henderson-Hasselbalch equation, and Nernst-Planck equation.

Simulated permeability values showed good correlations with PAMPA, Caco-2, and intestinal permeability measurements for a dataset with thirty-six molecules. Together with a mathematical model that models subcellular localization of small molecules in a non-polarized cell, the cell based pharmacokinetic model could be used to analyze the transcellular permeability and subcellular accumulation in a non-target cell, and optimize distribution to the target site (i.e. cytosol, lysosomes, mitochondria, and extracellular space) in a target cell.

To further validate 1CellPK and demonstrate how it can be used to generate quantitative hypotheses and guide experimental analyses, permeability and total intracellular mass accumulation were measured for a lysosomotropic compound, chloroquine, on MDCK cells. Predicted permeability agrees with observed permeability under various input conditions: adjusted pH values in the donor compartment, adjusted membranes with different size of pores, and various transport directions. However, for mass accumulation, 1CellPK model predicts only for a short time (5 minutes or less), suggesting other mechanisms are involved but not included in the current model for chloroquine uptake.

1CellPK model was further extended to a virtual lung model (the Cyberlung), and the Cyberlung was integrated into whole body physiologically-based pharmacokinetic (PBPK) models to predict lung distribution for three beta-blockers. 1CellPBPK predicted pharmacokinetics in the lung and other organs agrees well with observed data. Successful integration of a single-cell based Cyberlung model with a whole-body PBPK model constitutes an important step towards ab initio single-cell based predictive modeling of drug pharmacokinetics at the whole body level.