Metastatic colorectal cancer (mCRC) is the second leading cause of cancer-related mortality in the United States, and annually accounts for nearly 500,000 deaths worldwide. Currently, the small molecule kinase inhibitor (KI) regorafenib is the primary second line therapy for metastatic CRC that is not treatable with immunotherapy or anti-EGFR therapies. However, regorafenib generally provides only modest improvements in survival? typically months?and often at the cost of significant side effects. Proposed targets for regorafenib include kinases that act within tumor cells as well as non-autonomously; however, with over 500 possible targets in the human kinome, the exact mechanism by which this compound operates remains controversial and not fully known. This presents a daunting challenge; without a verifiable target or mechanism, no clear path exists to guide the development of improved therapies for mCRC. Here, we propose an alternative approach to drug development that focuses on kinase networks in the context of the whole animal. Specifically, we will take a multidisciplinary approach to define kinases that are beneficial to inhibit (?pro-targets?) or avoid (?anti-targets?) in the context of KRAS-variant CRC. Using Drosophila and mammalian models, we will identify kinases that?when reduced?alter the efficacy of regorafenib and similar compounds. We will also conduct extensive structure-activity relationship analyses, evaluating how modifications in already identified lead compounds impact changes in efficacy and therapeutic index. Finally, we will use computational structural biology to convert our chemical genetic insights into highly optimized and precise polypharmacological leads. In this final step, we generate new analogs to selectively eliminate putative anti-target activity while maintaining or increasing inhibitory activity against other beneficial targets. We have used our chemical genetic platform to identify a promising lead compound, APS5-86-2, that demonstrates significant activity relative to regorafenib in several mCRC models, including human patient derived xenografts (PDX). Comparative analysis suggests that the improved activity of APS5-86-2 relative to regorafenib derives from distinct polypharmacology on several RTKs and critical cancer drivers, including CDK9, AURKA, EGFR, BRAF, and RAF1. In this proposal, we examine the mechanism and importance of these and other putative pro- and anti-target kinases using genetic analysis and in vivo target engagement. The objective is to identify the kinase networks that mediate KRAS-variant mCRC by combining chemical biology with genetics, and to then derive inhibitors that best attack these networks through structure-based drug design. We have been successful previously with a similar approach, but in less complex tumor models (Dar et al., Nature, 2012; Sonoshita et al., Nature Chem. Bio., 2018); here we seek to extend our platform to a more prevalent disease with the goal of directly impacting mCRC by creating new, highly differentiated, and improved drugs.
Synthetic tailoring of polypharmacological drugs is often difficult, as the most appropriate targets may not be readily apparent and therefore few roadmaps exist to guide chemistry. In this application we propose a multidisciplinary approach for accessing novel target and chemical space so as to derive highly precise polypharmacological compounds. By combining chemical and genetic modifier screens, with computational modelling, we will identify distinct kinases that can strongly enhance (?pro-targets?) or limit (?anti-targets?) the activity of kinase inhibitors within models of KRAS variant metastatic colorectal cancer, and then harness this information to predict and synthesize novel compounds optimized towards whole animal networks.
