Identification of TKI-Sensitive Point Mutations That Activate c-ABL Kinase Activity and Transformation Potential and Confer in Vitro Resistance to the Allosteric ABL Inhibitor GNF-5

Introduction:

Pathologically activated tyrosine kinases represent attractive therapeutic targets, with deregulated ABL kinase representing one of the best-validated examples. While physiologic activation of c-ABL is tightly controlled by intramolecular interactions, ABL fusion proteins are constitutively activated by oligomerization driven by the fusion partner, as well as by loss of kinase autoinhibition due to removal of an autoinhibitory myristate site at the N-terminus of c-ABL (Nagar et al, Cell, 2003). Recent routine genomic tumor sequencing of various human malignancies has identified point mutations of unknown significance in numerous kinases, including c-ABL, and it is expected that the increasingly common practice of tumor genome sequencing will provide new opportunities for tailored targeted therapy. To maximize therapeutic benefit and avoid unnecessary cost and toxicity, it will be critical to distinguish “driver” from “passenger” mutations.  We therefore sought to: (i) test the transforming potential of select clinically detected c-ABL mutants, (ii) prospectively identify novel additional activating c-ABL point mutations, and (iii) determine if active mutant c-ABL isoforms retain sensitivity to ABL TKIs.

Results:

Two of four clinically identified c-ABL mutations tested caused a dramatic increase in ABL kinase activity and transformed Ba/F3 cells to growth factor independence. A mutagenesis screen of c-ABL identified six additional transforming point mutations in the SH3 and kinase domains, including one that has been reported clinically in breast invasive carcinoma (cBioPortal.org). Relative to BCR-ABL, all c-ABL mutants showed heightened sensitivity to imatinib. Allosteric site inhibitors targeting the myristate-binding pocket in the BCR-ABL kinase domain (KD) have been shown to inhibit BCR-ABL kinase activity, purportedly by mimicking binding of the c-ABL myristate group and establishing c-ABL-like autoinhibitory constraints (Hantschel, Haematologica, 2012). While only two c-ABL mutations from the screen were near the allosteric site, all mutants exhibited decreased sensitivity to the allosteric inhibitor GNF-5. Recent studies have shown that derivative compounds of GNF-5 targeting this site in BCR-ABL are subvertable by on-target resistance mutations. We tested whether three BCR-ABL mutations described to confer in vitro, and in one case clinical, resistance to ABL001 (currently undergoing clinical trial evaluation) could impact c-ABL kinase activity presumably by perturbing this autoinhibitory pocket. In the context of c-ABL, all three ABL001-resistant mutations activated ABL kinase, transformed Ba/F3 cells, and conferred substantial resistance to allosteric inhibition relative to native BCR-ABL. These findings raise the possibility that c-ABL mutations may play a role in resistance to emerging allosteric ABL inhibitors. Importantly, previous work suggested that the genomic breakpoint on chromosome 9 in one third of CML cases is located 5’ to ABL exon 1b, which would be predicted to result in simultaneous KD mutation expression in BCR-ABL and c-ABL (Uphoff et al, Leuk Res, 1999). To definitively demonstrate that co-expression of KD mutations in c-ABL and BCR-ABL can occur in primary patient samples, we analyzed RNA from five CML patients harboring a dominant BCR-ABL T315I mutation, and identified a heterozygous c-ABL 1b T315I mutation in one case.

Conclusions:

We have identified eight novel activating point mutations in c-ABL (three found in clinical isolates) that are potential therapeutic targets of ABL TKIs. Activating c-ABL mutants, including T315I, confer substantial resistance to the allosteric ABL inhibitor GNF-5 regardless of proximity to the inhibitor-binding site, implying resistance mechanisms beyond mere steric hindrance. Thus, mutations may induce distal conformational changes by disrupting a kinase conformation required for allosteric compound inhibition. In a subset of CML cases, c-ABL point mutants are co-expressed from the same allele as BCR-ABL, suggesting that the location of the chromosome 9 breakpoint may negatively impact clinical responsiveness to allosteric ABL inhibitors.  Sequence analysis of c-ABL in patients with resistance to allosteric ABL inhibitors is required to validate c-ABL point mutants as a clinically important mechanism of resistance to this emerging class of targeted therapeutics.