A three-dimensional pharmacophore modelling of ITK inhibitors and virtual screening for novel inhibitors.

Interleukin-2-inducible T-cell kinase (ITK) is a key member of the Tec family of non-receptor tyrosine kinases, and has been found to be a novel target for a number of inflammatory and autoimmune diseases. A three-dimensional pharmacophore model has been generated for protein ITK from its known inhibitors. The best HypoGen model consisted of four pharmacophore features: one hydrogen bond acceptor, one hydrogen bond donor and two hydrophobic rings. This model showed a correlation coefficient of 0.947, a root mean square deviation of 0.914 and a configuration cost of 16.866. The model was validated using test set prediction and Fischer's test. A test set containing 204 compounds showed an r(2) of 0.745 between estimated activity and activity measured experimentally. Fisher's test gave a confidence level of 95%. The best pharmacophore model (Hypo1) was then employed for virtual screening (3D database searching), including Lipinsiki's filter, to obtain a pool of more drug-like molecules. The molecular pool thus retrieved was subjected to docking analysis with a study protein to remove any molecules showing false positive activity for ITK.