Computational Structure-Based Design of Novel Kinase Inhibitors
Protein kinases are essential to nearly every cellular signal transduction pathway. Because of this, and because of their inherent druggability, kinases have become one of the most important drug targets in the past two decades. The druggability of kinases stems from the fact that these enzymes share a strongly-conserved ATP-binding site, which is structurally very amenable to inhibition by the types of flat, aromatic scaffolds that are most accessible to synthetic chemistry. However, the human “kinome” is comprised of ~500 kinases with highly similar active sites, making it difficult to develop selective inhibitors. We have recently developed a comparative modeling approach that allows for rapid and highly-accurate predictions of the three-dimensional structure of arbitrary (Type I) kinases. I will describe how we use these models, together with computational tools to enumerate vast swaths of synthetically-accessible chemical space, as the basis for identifying and optimizing novel and selective inhibitors of key cancer-driving kinases.