Biochemical Context Drives Inhibitor Selectivity in Aurora A Kinase

Abstract

The Aurora family of mitotic kinases control events necessary for proper separation of chromosomes throughout mitosis. The family is highly conserved and structurally similar but differs in their activation pathways and interacting partners. For example, Aurora A kinase can be activated by phosphorylation on its activation loop, a dynamic regulatory structure in all protein kinases, or through allosteric binding of one of its partner proteins, Tpx2, and each of these mechanisms regulates separate functions of the kinase in cells. Furthermore, overexpression patterns of Aurora kinases have been observed in a number of human cancers that have driven a wealth of inhibitor development targeting Aurora activity, but no compounds have been FDA approved. Recently, Aurora A was shown to provide a non-catalytic stabilizing scaffold for N-Myc, the protein product of the MYCN gene, which is heavily amplified in many childhood neuroblastomas and late-stage neuroendocrine prostate cancers. A strategy to block growth of cells harboring MYCN amplification is to disrupt complex formation between N-Myc and Aurora A, as this complex protects N-Myc from ubiquitin-mediated proteolysis. However, no current compounds can effectively dissociate this complex in cellular models, and further inhibitor development is hindered by a lack of complete structural detail of the complex. An understanding of the dynamics that exist between Aurora inhibitors and Aurora binding partners, both catalytic and non-catalytic, would provide important details for improved development of inhibitors with increased selectivity in Aurora-driven cancers. My thesis work has focused on the conformational effects of inhibitors and understanding how they can lead to selectivity patterns in Aurora A. Using a new fluorescence technique that can track and quantify conformational rearrangements of the kinase domain in solution, I have profiled a panel of clinically relevant Aurora kinase inhibitors against Aurora A in several of its biochemical contexts. The results show that inhibitors display a wide range of conformational preferences, with all inhibitors promoting either the active DFG-in state or inactive DFG-out states, but to widely differing extents. For example, DFG-out inhibitors preferentially bind the more-dynamic form of Aurora A that is activated by phosphorylation, but DFG-in inhibitors preferentially bind to Aurora A constrained in the DFG-in state by its allosteric activator Tpx2. In contrast to what has been seen with Tpx2, there is not an obvious relationship between the conformational effect of N-Myc on the kinase domain and the interaction with inhibitors, with some inhibitors that were predicted to weaken the complex actually strengthening it. This demonstrates that despite the similar crystal structures of the Aurora A bound to Tpx2 and to N-Myc, these binding partners actually have quite different allosteric effects on Aurora A. Quantification of the conformational interplay between ATP-competitive inhibitors and N-Myc demonstrates that many Aurora inhibitors cannot fully disrupt the AurA:N-Myc complex, indicating that existing ATP-competitive inhibitors alone are likely ineffective at treating MYCN-amplified cancers. However, cooperativity patterns revealed clues to potential allosteric interactions between N-Myc and Aurora A that mimic the interaction of INCENP with Aurora B and helped to generate a model of the complete interaction of N-Myc with Aurora A that will help guide future efforts that aim to dissociate the Aurora A:N-Myc complex. Taken together, the cooperativity patterns observed in both the AurA:Tpx2 and AurA:N-Myc complexes suggest that many inhibitors currently in clinical development may be capable of differentiating between Aurora A signaling pathways implicated in normal mitotic control and in melanoma, neuroblastoma and prostate cancer. Based on these results, it is clear that the context of biochemical and allosteric interactions with the kinase domain is the key driver of inhibitor selectivity in Aurora A. Application of the technology developed during these studies can also be applied to a wide range of clinically important kinases. A similar understanding of inhibitor cooperativity and structure-activity information in other kinases will lead to the development of inhibitors that exploit differences in conformational dynamics to enhance the selectivity of kinase inhibitors.