Designing Compounds that Probe the Phosphatidylinositol 3-Kinase Signalosome

Abstract

Phosphoinositide 3-kinases (PI3K) are major regulators of many cellular functions, and its hyperactivation is a major target for anticancer drug discovery. PI3Kα is the isoform most implicated in cancer, and this thesis is dedicated to investigating how specific protein-ligand interactions in the active site affect selectivity and affinity of inhibitors; how ATP-competitive inhibitors affect membrane binding of PI3Kα; how known membrane binding inhibitors of coagulation factor V and VIII C2 domains affect PI3Kα- membrane interactions; and how compounds predicted to occupy a newly discovered drug binding site affect activity and signaling of PI3K. To identify the contributions of different regions of the active site to PI3Kα-selectivity and affinity, structure-based design by molecular modeling were used to merge chemical groups from highly selective and high-affinity compounds. Water was included in the docking to explore different chemical units that mediate ligand-protein interactions in the affinity pocket, and lipid kinase assay results showed that inhibitor potency increased significantly with the predicted increase in affinity pocket interactions. Cell signaling analysis showed that PI3Kα-selective inhibitors were more active in the H1047R mutant bearing cell lines SK-OV-3 and T47D compared with the E545K mutant MCF-7 cell line. Using biophysical methods that reported on protein-membrane interactions, ATPcompetitive inhibitors were also investigated for their effect on the enzyme’s ability to bind membranes, the location of its lipid substrate, PIP2. The results showed that GSK2126458, a potent pan-PI3K inhibitor was able to affect the PI3Kα-membrane interaction for PI3Kα WT, but less so for the H1047R oncogenic mutant, which has higher membrane-binding affinity, suggesting that these two proteins occupy distinct conformational ensembles. The results also suggested that there may be specific protein-ligand interactions that are responsible for the observed effect. Known inhibitors of coagulation factor V and VIII were also investigated for their effects on PI3Kα activity and membrane binding, however, their PAINS-like properties prevented definite conclusions from being drawn. Lastly, compounds proposed to bind to a newly discovered pocket were able to significantly decrease AKT phosphorylation through an undetermined mechanism. Together, these data may lead to new mechanisms of blocking PI3K by interfering with membrane-binding and formation of the PI3K signalosome.