Modelling and prediction of drug transport processes with experimental and calculated molecular properties A multivariate approach

Sammanfattning: Less than 2% of the lead compounds generated by the pharmaceutical industry enter the exploratory drug-development phase, from which point they stand only a 10% chance of becoming a commercial medicine. A large proportion of the compounds fail due to poor biopharmaceutical properties, such as permeability and solubility. The main theme of this thesis is, therefore, the development of better experimental and computational methods for modelling and predicting the biopharmaceutical properties of drug candidates. Immobilised liposome chromatography (ILC) was used for studying drug-membrane interactions and for the prediction of passive drug transport. For the drugs studied in this thesis, ILC and octanol/water partitioning showed a similar performance with regard to the prediction of passive drug transport.The theoretical work was directed at the modelling and prediction of drag transport processes using calculated molecular properties and PLS analysis. In the initial studies, the molecular properties were calculated with an advanced computational tool (MolSurf) that takes the three-dimensional structural information and electronic properties of the compound into consideration. Statistical models well suited to the prediction of drug transport processes such as Caco-2 cell permeability, intestinal absorption and CNS penetration were derived.This approach was also successfully applied to the modelling of the interaction of drugs with P-glycoprotein. Subsequently, rapidly calculated descriptors based on two-dimensional structural information and PLS analysis were also found to give good predictive models of drug transport properties. The preferred use of the latter models is for screening compound collections and virtual libraries. It can be concluded that calculated molecular properties in conjunction with PLS analysis are well suited to the modelling and prediction of drug transport processes and to identifying compounds with potential biopharmaceutical problems at an early stage of the drug-development process.