Pharmacologic inhibition of BCR-ABL, using clinically active ATP-competitive inhibitors imatinib, nilotinib, and dasatinib, has been used to investigate BCR-ABL kinase activated signaling pathways. However, these agents show cross reactivity with other kinases (e.g. Kit, PDGFR, SRC family members), and their multi-targeted nature complicates assigning biological effects to the inhibition of a specific kinase target. Allosteric kinase inhibitors modulate the catalytic activity of protein kinases by binding to a site distant from the active site and inducing a protein conformation that inhibits kinase activity. These agents show promise as clinical agents and may offer advantages over ATP-competitive inhibitors in studying the function of specific kinases because they exploit binding sites and regulatory mechanisms that are unique to a particular kinase. GNF-2, a mono-selective BCR-ABL inhibitor that targets wild-type BCR-ABL and many clinically relevant imatinib resistant mutants, was recently discovered and provided the first demonstration that c-ABL kinase activity could be modulated by an inhibitor that binds outside the ATP or substrate binding sites. GNF-2 binds to a myristoyl-binding pocket in the C-lobe of the c-ABL kinase domain but its mechanism of inhibiting specific BCR-ABL kinase targets remains unclear. We previously reported that BCR-ABL activates an autocrine IGF-1 pro-survival signaling pathway in CML blast crisis cells through HCK-mediated activation of STAT5b. As GNF-2 is known to inhibit STAT5b phosphorylation, and HCK myristolyation is known to regulate its cellular localization, we hypothesized that GNF-2 inhibits BCR-ABL activation of HCK by binding to the ABL myristoyl-binding pocket and blocking access to the HCK myristoyl moiety. In support of this hypothesis, we now show that GNF-2 inhibits HCK phosphorylation and IGF-1 activation, but not HCK binding to BCR-ABL. To confirm the importance of the HCK myristoyl moiety in HCK activation, we mutated the myristoyl attachment site at position 2 in HCK from glycine to alanine. The mutant G2A HCK still interacted with BCR-ABL in co-immunoprecipitation assays but showed significantly lower levels of phosphorylation compared to wild-type HCK. To confirm that the decrease in phosphorylation was not due to mislocalization of G2A HCK, we mutated the myristoylation binding pocket of BCR-ABL by changing glutamic acid at position 505 to lysine. Similar to G2A HCK, E505K BCR-ABL still interacted with HCK, but the phosphorylation levels of HCK were dramatically reduced. To confirm that the HCK myristoyl moiety directly interacted with the ABL myristoyl-binding pocket, we used fluorescent spectroscopy to measure the ability of a myristoylated peptide corresponding to the six N-terminal amino acids of HCK to displace GNF-2. The fluorescence of GNF-2 is enhanced when it associates with the myristoyl-binding pocket of ABL. Using this assay, we calculated the Kd of GNF-2 to be 180 nM. We then assayed the ability of myristolyated HCK peptide to displace GNF-2 from ABL. We calculated the IC50 of the myristolyated HCK peptide to be 25 μM when ABL was saturated with 300 nM GNF-2. Myristate showed an IC50 of 213 μM, which is ∼ 10-fold higher than the myristoylated peptide. No binding was detected between the non-myristoylated peptide and ABL. Together, our study highlights a novel acquired function resulting from the fusion of BCR to the N-terminus of ABL, which converts the myristoyl-binding pocket in ABL from a negative regulator of kinase activity to an HCK activation motif that activates downstream IGF-1 signaling. These results also reveal the mechanism of action of the mono-selective BCR-ABL inhibitor GNF-2 and highlight the ABL myristoyl-binding pocket as a therapeutic target for inhibiting BCR-ABL activity.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.