Abstract 296: Structure-based and property-based drug design of AZD5305, a highly selective PARP1 inhibitor and trapper

Abstract Since the approval of olaparib in 2014 for BRCA mutated (BRCAm) ovarian cancer, many PARP inhibitors have been developed and have seen widespread success. However, as a class, these drugs are not without adverse events which have limited their ability to be combined with chemotherapy. Most first generation PARP inhibitors were developed and optimized before the concept of PARP1-DNA trapping was discovered as the mechanism by which PARP inhibitors exert their synthetic lethal effects on BRCAm cells. Moreover, the first generation PARP inhibitors were not optimized for selectivity across the PARP family potentially driving undesirable side effects, including intestinal toxicity from tankyrase inhibition or hematological toxicity from PARP2 inhibition. With this in mind, we set out to discover a best-in-class, second generation PARP inhibitor that was highly selective for PARP1 over the other 16 members of the PARP family, as well as a highly potent PARP1-DNA trapper. PARP1 and PARP2 have a highly similar amino acid sequence, and most of the residues around the nicotinamide binding site are identical. However, there are some key residue differences in the helical domain which serves a regulator of the nicotinamide binding pocket. The publication of NMS-P118 in 2015 by Nerviano Medical Sciences showed that a highly selective PARP1 inhibitor could be found. This work inspired us to screen an extensive list of previously reported PARP inhibitors for selectivity against PARP2 and we found that FR257516 met the selectivity criteria as previously reported, but lacked the ability to trap PARP1 to DNA and hence lacked any activity in a cell colony formation assay in DLD-1 BRCA2-/- cells. Using parallel chemistry to generate diverse analogs, X-ray crystallography to enable structure-based design, and exploration of multiple nicotinamide mimetic cores, we were able to generate lead compound AZ4554, which was a PARP1 selective PARP1-DNA trapper with potent activity in BRCAm cells. Using concepts of property-based drug design, we were able to optimize lead compound AZ4554 into candidate drug AZD5305, making key improvements in secondary pharmacology, including reducing hERG activity, and intrinsic clearance in human microsomes through the introduction of polar atoms to lower logD without compromising permeability or oral bioavailability. AZD5305 is a highly selective binder of PARP1 over PARP2 and other PARP enzymes by fluorescence polarization, surface plasmon resonance, and single molecule spectroscopy. It is highly potent against DLD-1 BRCA2-/- cells, while sparing isogenic BRCA WT cells. The secondary pharmacology of AZD5305 is remarkably clean, with hERG activity >40 µM. AZD5305 has a very favorable pre-clinical PK profile, low predicted human dose, and has shown efficacy in an MDA-MB-436 mouse xenograft model. Citation Format: Sudhir Hande, Amber Balazs, Sébastien L. Degorce, Kevin Embrey, Avipsa Ghosh, Sonja J. Gill, Anders Gunnarsson, Giuditta Illuzzi, Jordan Lane, Carrie Larner, Elisabetta Leo, Andrew Madin, Lisa McWilliams, Mark J. O'Connor, Jonathan Orme, Fiona Pachl, Martin Packer, Andy Pike, Philip Rawlins, Marianne Schimpl, Anna D. Staniszewska, Andrew Zhang, Xiaolan Zheng, Jeffrey W. Johannes. Structure-based and property-based drug design of AZD5305, a highly selective PARP1 inhibitor and trapper [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 296.