Experimental and Computational Predictions of Drug Solubility in Human Gastrointestinal Fluids

Abstract: The aqueous solubility of a drug is viewed as a pivotal property for its oral absorption since only dissolved molecules can permeate the gut wall and reach the systemic circulation. The fluids in the intestine, however, do not only consist of water and therefore poor water solubility may not necessarily imply a poor solubility in the intestinal fluids and resulting low bioavailability. This thesis addresses the determination of drug solubility and dissolution rates in biorelevant dissolution media (BDM) with the aim of applying these methods to the early stages of drug discovery, where there is a need to reduce the volume of the medium and the amount of solid drug used in testing. The thesis also addresses the need for computational methods for predicting solubility in intestinal fluids and, hence, allowing in silico screening of drugs yet to be synthesized. The apparent solubility and dissolution behavior of large series of lipophilic and other diverse compounds in BDM were studied using a miniaturized method developed herein. The media used in the experimental design provided an opportunity to assess the effects of charge, solubilization in mixed lipid aggregates, and ethanol in BDM. Highly lipophilic and uncharged drugs were efficiently solubilized by aggregates in the BDM while solubilization was decreased with charge. The decrease was more pronounced for negatively charged drugs. The solubility of anionic and neutral drugs was significantly increased by the addition of ethanol to the medium and absorption simulations showed that intake of alcohol could lead to increased plasma concentrations of neutral compounds. Statistical models based on calculated molecular descriptors that accurately predicted the apparent solubility in fasted-state simulated intestinal fluid and in aspirated human intestinal fluid were also developed. In summary, the work undertaken in this thesis has resulted in new experimental and computational models for assessment of the dissolution and solubility of poorly water-soluble compounds in BDM. The models are applicable in the early discovery and development phases for predicting physiologically relevant solubility and the effects thereof on drug absorption.