Dissection of Bcr-abl Structural Domains Relating to Kinase Auto-Inhibition Using a Forward Mutational Screen with the Non-ATP Competitive Inhibitor GNF-2.

The Bcr-abl kinase is the causative agent for chronic myeloid leukemia (CML) and has been established as the primary clinical target for treatment of the disease through extensive use of Imatinib. Imatinib is the defining member of a class of ATP-binding site competitive inhibitors that lock Bcr-abl in an inactive conformation. Mutational screens of Bcr-abl using Imatinib and its derivatives as probes have been highly informative in prediction of clinically relevant mutations of Bcr-abl as well as in revealing the structure/function relationship of the kinase in general. Using compounds with a distinct mechanism of action from the Imatinib class to interrogate Bcr-abl will contribute to both more complete understanding of kinase function as well as to potential combination therapies for more effective treatment of CML. GNF-2, a recently identified inhibitor of Bcr-abl, establishes a new class of non-ATP competitive Bcr-abl family kinase inhibitors that may be developed as therapeutic agents for CML. GNF-2 effectively impairs the in vivo kinase activity of Bcr-abl and the growth of Bcr-abl transformed cells. GNF-2 functions at least in part through association with the myristate binding pocket of Bcr-abl. In order to further elucidate the mechanism of GNF-2 action as well as clinically relevant GNF-2 resistant mutants of Bcr-abl, a mutational screen coupling Bcr-abl mutagenesis to selection of drug resistance was performed using GNF-2 as probe. A number of functionally distinct resistant Bcr-abl mutations were recovered. Over half of all GNF-2 resistant clones harbored Bcr-abl mutations affecting the myristate binding pocket or the abl-SH3 domain, suggesting two potential methods of mutational resistance.

1. The myristate binding domain mutants support a direct resistance model whereby GNF-2 association with Bcr-abl is impaired by disruption of the myristate binding pocket.

2. Given a previous report that GNF-2 cannot inhibit Bcr-abl kinase activity in vitro, a novel model emerges for indirect resistance to GNF-2 by SH3 mutants that lose affinity for an inhibitory associated protein.

The indirect resistance model specifically suggests that GNF-2 association confers a structural state of wildtype Bcr-abl which facilitates association to a putative inhibitory binding partner, thereby affecting inhibition. Indeed, the strongest of several candidate inhibitory binding partners, the Abl-SH3 domain binding inhibitor Abi-2 was observed to co-immunoprecipitate with Bcr-abl in the presence of GNF-2. This association correlated with reduced Bcr-abl auto-phosphorylation levels. These observations provide preliminary support for an indirect mechanism of Bcr-abl inhibition by GNF-2. Additional experiments involving shRNA knockdown of Abi-2 are being completed to determine the requirement of this Bcr-abl binding partner for GNF-2 activity. Further characterization of the SH3 and myristate binding domain mutants in the context of Abi-2 and GNF-2 binding affinities may establish a previously undescribed indirect mechanism of Bcr-abl inhibition by an allosteric non-ATP inhibitor.