The Design and Optimization of a First-in-Class Inhibitor of NACK: Illuminating a Druggable Target in Notch

Notch Activation Complex Kinase (NACK) is a key player in Notch-mediated tumorigenesis and an attractive novel target for the treatment of esophageal adenocarcinoma, and other Notch-rich cancers. NACK, annotated in the kinome as SgK223, is an atypical pseudokinase and an established Notch transcriptional co-activator. However, there is no known endogenous or exogenous ligand, no existing co-crystal structure, and no reported biological data. To identify a scaffold for NACK inhibition, a machine learning approach was utilized. Over six million commercially available compounds were screened against established kinase machine learning classifiers, and nearly 8000 were prioritized based on the predicted probability of being active. A homology structure model of the NACK kinase domain was generated and further optimized by all-atom explicit water molecular dynamics (MD) simulations, followed by virtual screening of prioritized compounds. Top-scoring compounds were purchased and screened in in vitro and in vivo assays. Commercially available compound Z271-0326 displayed low micromolar inhibitory activity across several Notch-rich cancer cell-lines and was further validated in rodent models. A robust novel chemical synthesis for Z271-0326 was accomplished in six steps with an overall yield of 26%. Efforts were then aimed towards optimizing Z271-0326 into the first NACK molecular probe. I optimized our virtual NACK kinase domain structure model via MD simulations and performed a structure-guided optimization approach to elucidate putative binding interactions. Analogues were designed and synthesized with the goal of improving inhibitory activity while retaining selectivity. Assay results demonstrate that the hit compound is optimizable as analogue UM-323 demonstrates an EC50 of 20 nM in OE33 cells and displays NACK target engagement via thermal shift assay. In vivo DMPK studies reveal UM-323 half-life was increased six-fold and potency was increased 150-fold. However, I arrived at a molecular probe and a First-in-Class NACK inhibitor.