ABSTRACT
Drug discovery and development are expensive undertakings and any technology that can decrease the length of time prior to New Drug Application (NDA) submission, or restrict the number of experimental procedures required for compound selection and development, will save money. The philosophy of ‘kill early, kill often’ during preclinical drug development is a mantra frequently chanted by pharmaceutical industry managers today. The observed oral bioavailability of a particular therapeutic agent can be broken down into components that reflect delivery to the intestine (gastric emptying, intestinal transit, pH, food), absorption from the lumen (dissolution, lipophilicity, particle size, active transport), intestinal metabolism, active efflux, and subsequent first-pass hepatic extraction (Hall etal., 1999). Experimental in vitro and estimated in silico biopharmaceutical properties can be used to predict drug absorption, distribution, metabolism, excretion, and toxicity (ADMET). This chapter will focus on in silico approaches that have the potential to save valuable resources in the drug discovery and development process. The discussion of property estimation will begin by defining some of the statistical terms and pitfalls commonly observed in theoretical model building. A discussion of the literature associated with the futile attempts to build statistical models of drug absorption directly from structure will follow. Next, will be the presentation of some of the more successful recent attempts to build models of individual biopharmaceutical properties, followed by a review of the theoretical methods employed to simulate gastrointestinal (GI) absorption. Finally, we will discuss our results in estimating biopharmaceutical properties and simulating GI absorption using Biopharmaceutical property Estimation and the Advanced Compartmental Absorption and Transit (BEstACAT) model.
