The overall aim of the project is to validate, explore the basis of, update, and disseminate the approach developed under CA46732 for assessing the nature and intensity of empirical drug interactions (Loewe synergism, Loewe antagonism, Loewe additivity, synergism, antagonism and coalism), called the universal response surface approach (URSA). This proposal for the continuation of CA46732 emphasizes the critical need for laboratory experiments to explore a set of questions which came about as a result of the mathematical/statistical modeling work on drug interactions completed during the last three years. To accomplish the singular, focused aim of facilitating URSA to make a large positive impact on the conduct of research worldwide into the efficacy of anticancer drug combinations, several subprojects will be completed: l. The validity and utility of a set of theoretical/empirical mathematical/statistical models, which were derived to describe concentration-effect (dose-response) phenomena and agent interaction, will be investigated in laboratory studies, specifically designed for this goal, of anticancer agents and combinations of agents with three assays, a total growth assay (TGA), a colony count assay (CCA), and an individual colony formation assay (iCFA). 2. Based upon the results of subproject #1, older models will be modified when appropriate, and newer models will be created and tested when needed. Improvements will also be made to design and model-fitting techniques. 3. In order to investigate the mechanistic basis of anticancer drug interactions, the relationship between the empirical models developed under CA46732, with a set of theoretical models of intracellular metabolism will be explored. 4. The first version of the software package, SYNFIT, developed under CA46732, will be distributed, comments solicited from users, and appropriate improvements and enhancements made. Specific enhancements already planned include: a) the incorporation of generalized nonlinear modeling procedures, after rival numerical algorithms for generalized nonlinear modeling have been compared; b) the incorporation of D-optimal and other statistical experimental design procedures, after rival numerical algorithms for these procedures have been compared; c) the incorporation of 3-D graphical procedures.
Showing the most recent 10 out of 18 publications
