Skip to main content
eScholarship
Open Access Publications from the University of California

UCSF

UC San Francisco Electronic Theses and Dissertations bannerUCSF

Predicting Drug Disposition by Integrating In Vitro and In Silico Methodology

Abstract

The safety and efficacy of drugs depend upon appropriate dosing of drugs made possible by understanding the dispositional profile a drug will follow. A drug’s disposition includes its absorption from an administration site, its distribution throughout the body, and its elimination from the body, characterized by metabolism and excretion. Disposition is often mediated by drug metabolizing enzymes and drug transporters. Alterations in the expression or activity of metabolizing enzymes and transporters can therefore affect the safety or efficacy of a drug and it is necessary to characterize their impact on every drug. The Biopharmaceutics Drug Disposition Classification System (BDDCS) uses the extent of metabolism and solubility of drugs to predict drug disposition, including when transporters and metabolizing enzymes are clinically relevant. Here, we utilized observations from this system to predict the three major routes of drug elimination (metabolism, renal excretion of unchanged drug, and biliary excretion of unchanged drug). These predictions were made by integrating in vitro measurements of permeability rate to predict the extent of metabolism with an in silico logistic regression model we developed that uses calculated polarizability and predicted metabolic stability to predict when poorly metabolized compounds will be eliminated in the urine or the bile. This approach correctly identified 72 ± 9%, 85 ± 2%, and 73 ± 2% of extensively metabolized, biliarily eliminated, and renally eliminated drugs, respectively. We discuss the physiological context through which permeability, polarizability, and metabolic stability may inform the major elimination route. We further developed a model predicting BDDCS class using commercially available in silico models of permeability rate to predict the extent of metabolism and dose number to predict the solubility class. This approach correctly identified 54.1%, 57.8%, 69.3%, and 45.2% of class 1, 2, 3, and 4 drugs, respectively, while in vitro approaches predict with greater accuracy. We correct previously misclassified drugs, discuss reasons for misclassification, incorporate more than 175 additional drugs into the system, and discuss how BDDCS can self-correct when observed and predicted dispositional effects are not aligned. We conclude by reflecting on the demonstrated and potential applications of BDDCS and the importance of predicting drug disposition.

Main Content
For improved accessibility of PDF content, download the file to your device.
Current View