Computational methods for drug target profiling and polypharmacology
Thierry Langer, Sharon D. Bryant · 2013
The ‘magic bullet’ concept of hitting a target responsible for a disease with a drug molecule tailored to act as a selective agent, has been a therapeutic goal since the beginning of drug research and one of the driving forces in modern drug discovery for several decades. With the rise of structural biology and molecular pharmacology, and the shift from in vivo to in vitro models in the initial evaluation of biological effects of molecules, the aim of obtaining absolute target specificity had become a goal that seemed within reach. However, there is evidence that drugs interact with many physiological targets, and that polypharmacology bears essential importance on therapeutic efficacy. In this light, discovering compounds exhibiting the ‘right’ selectivity profile (i.e., interaction with several targets or target hubs in a converging biological pathway) has become the holy grail in drug development. Recent examples in the kinase field illustrate this new paradigm. Whereas imatinib (Gleevec®; Novartis, Switzerland) and sunitinib (Sutent®; Pfizer, NY, USA) were designed to be selective, later they were found to be more promiscuous than initially thought [1,2], which could explain why these molecules are successful therapeutically. As recently pointed out, searching for selectively nonselective kinase inhibitors when striking the right balance, can deliver candidates and drugs with superior efficacy compared with inhibitors with high specificity for a single kinase [3].