Beyond the Gatekeeper: Imatinib- and Dasatinib-Resistant BCR-ABL/F317 Mutations Confer Cross-Resistance to VX-680 but Are Sensitive to a Structural Derivative of VX-680

The management of chronic phase CML has been revolutionized by selective ABL tyrosine kinase inhibitor (TKI) therapy. Despite the effectiveness of these targeted agents, long-term control of blast phase CML and Ph+ ALL has been elusive, where the majority of patients relapse within 6–12 months. For blast phase CML and Ph+ ALL, two TKIs are currently approved: imatinib and dasatinib. While head-to-head comparisons of these agents have not been performed, it is generally believed that dasatinib is the more active agent for these phases of disease. In most cases, loss of response to these agents is driven by BCR-ABL kinase domain mutations. While more than 70 mutations have been associated with clinical resistance to imatinib, dasatinib appears vulnerable primarily to five mutations: V299L, T315A, T315I, F317I, and F317L. Of these, T315I and F317L are cross-resistant to imatinib. For the achievement of long-term remissions in blast phase CML and Ph+ ALL, a combination of TKIs that can collectively suppress all resistant BCR-ABL kinase domain mutations holds therapeutic promise. The BCR-ABL/T315I mutation, which confers a high degree of resistance to all approved BCR-ABL TKIs, has been referred to as a “molecular gatekeeper”, as it restricts access to a deeper hydrophobic pocket within the ABL kinase domain and makes an important stabilizing H-bond with imatinib, dasatinib and nilotinib. The Aurora kinase inhibitor VX-680 was the first compound to have activity against BCR-ABL/T315I in vitro, as well as clinically. To determine the promise of a kinase inhibitor combination of dasatinib and VX-680, we assessed the activity of VX-680 against the five dasatinib-resistant mutations using a cell-based flow cytometric assay of BCR-ABL kinase activity. While three mutants are sensitive, mutations at F317 demonstrated a high degree of resistance. We tested a number of other Aurora kinase inhibitors of different chemotypes and found that each of these had similar difficulty at inhibiting the kinase activity of BCR-ABL/F317 mutants. Based upon the co-crystal structure of VX-680 complexed with ABL, we have performed structure-activity relationship studies of 12 VX-680 scaffold derivatives, and have successfully identified structural modifications that increase kinase inhibitory activity against F317 mutants. Moreover, one of these derivatives increases the selectivity for ABL relative to Aurora kinases, which may help reduce the likelihood of suppressing normal hematopoiesis, a dose-limiting toxicity of Aurora kinase inhibitors that may substantially limit their effectiveness for the management of hematologic malignancies such as blast phase CML and Ph+ ALL. Lastly, we have performed structural studies of ABL/F317 mutants complexed with select VX-680 derivatives in an effort to understand how F317 mutations confer resistance to a broad range of ABL and Aurora kinase inhibitors. Interestingly, a recent study reported the successful selection of Aurora kinase inhibitor-resistant clones derived from a human colon cancer cell line (Girdler et al, 2008). While no resistance-conferring mutations were isolated at L154, the Aurora kinase gatekeeper residue, mutations were detected at Y156 in Aurora B, which corresponds to F317 in ABL. Aurora B Y156 mutations were found to confer resistance to a number of Aurora kinase inhibitors, including VX-680. As Aurora kinase inhibitors are being studied in a variety of non-hematologic malignancies, there is an increasing need to understand and overcome the mechanisms whereby mutations at this residue confer resistance to these agents. It is hoped that our studies will lead not only to the development of an effective adjunctive kinase inhibitor for the treatment of blast phase CML and Ph+ ALL, but will also shed light on the growing problem of resistance conferred by mutations at residues that correspond to BCR-ABL/F317 in other kinases.