Development of physiologically based liver distribution model that incorporates intracellular lipid partitioning and...

Steady-state volume of distribution can be predicted using tissue-to-plasma partition coefficients (K_p) and tissue volumes. K_p values are important components of physiologically based pharmacokinetic models, allowing for estimation of distribution kinetics and simulation of concentration-time profiles. Many in silico approaches have been developed to predict tissue K_p values based on physicochemical processes that govern drug distribution. However, these methods frequently overpredict or underpredict tissue K_p values, highlighting the need to consider additional mechanisms that can impact drug distribution kinetics. Many drugs have been shown to bind to rat and human fatty acid binding proteins (FABPs) in vitro but the impact of this binding to drug distribution has not been incorporated into K_p predictions. We hypothesized that incorporating intracellular protein binding into tissue K_p predictions will improve K_p prediction accuracy. Using liver as a model organ, 4 physiologically based dynamic liver distribution models (LDMs) were developed to assess the role of distribution processes in K_p predictions. The developed LDMs incorporated known distribution mechanisms and intracellular drug binding to liver FABP (FABP1). The liver K_p values for drugs that bind to FABP1 were accurately predicted using the LDM that incorporates lipid partitioning, albumin distribution, and FABP1 binding but not using LDMs without FABP1 binding. Human FABP1 expression was quantified in 61 human livers and the interindividual variability in tissue FABP1 binding was incorporated into tissue K_p predictions. These simulations showed that intracellular FABP1 binding can cause interindividual variability in K_p values and result in higher cytosolic drug concentrations. SIGNIFICANCE STATEMENT: This study shows that incorporating intracellular protein binding such as binding to fatty acid binding protein 1 into tissue K_p predictions improves accuracy of the predictions. The novel dynamic liver distribution model can be extrapolated to other organs of interest and integrated into full-body physiologically based pharmacokinetic models to predict drug distribution kinetics. With dynamic and saturable distribution mechanisms incorporated into a physiologically based pharmacokinetic model, nonlinear distribution kinetics can be simulated for various drugs.