Inverse in silico screening for identification of kinase inhibitor targets

Protein kinases are clinically relevant, attractive drug targets for cancer. One major problem with kinase inhibitors is broad promiscuity, causing off-target actions and side effects. In silico prediction of targets of a compound would immensely facilitate and accelerate drug development. Using a virtual "inverse" screening approach, where single compounds are docked into protein structures from a database, we identify among known targets of indirubin derivatives phosphoinositide-dependent kinase 1 (PDK1) as a target of one derivative (6BIO) in particular. This prediction is functionally supported by an in vitro kinase assay, inhibition of intracellular phosphorylation of PDK1-substrates, and inhibition of endothelial cell migration, which highly depends on PDK1. Virtual inverse screening combined with biological tests, thus, is proposed as a valuable tool for the drug discovery process and re-examination of already established kinase inhibitors.