Abstract
Background
Development of resistance to kinase inhibitors remains a challenge in chronic myeloid leukemia (CML). Kinase domain mutations are a common mechanism of resistance, yet the mechanism of resistance in the absence of mutations remains less clear. Recent evidence suggests that the bone marrow microenvironment provides a sanctuary for leukemia cells, and may be involved in mediating resistance to imatinib – particularly in the absence of BCR-ABL kinase domain mutations. We tested selected cytokines, growth factors, and extracellular matrix proteins expressed by cells in the bone marrow microenvironment for their ability to protect CML cells from imatinib.
Results
We found that fibroblast growth factor 2 (FGF2) was the most protective protein for the K562 CML cell line when exposed to imatinib. FGF2 was not only capable of promoting growth in short-term culture, but uniquely able to promote long-term resistance in vitro (p<0.0001 by 2-way ANOVA analysis). To analyze the mechanism of resistance, we used siRNA to target the FGF receptors 1-4 and found that only siRNA targeting FGFR3 was able to abrogate the protective effect of FGF2. Phospho-chip and Western blot analysis revealed that FGF2 binds FGFR3, which then signals the downstream kinases Ras, c-RAF, MEK1, and ERK1/2 to promote survival in the presence of imatinib. Inhibition of FGFR3 with the specific FGFR inhibitor PD173074 led to dephosphorylation of this signaling cascade, and restored sensitivity to imatinib of FGF2-mediated resistant K562 cells. Resistance could also be overcome with ponatinib, a multi-kinase inhibitor that targets both BCR-ABL and FGFR, whereas imatinib, nilotinib and dasatinib were all ineffective against FGF2-mediated resistant K562 cells.
Although ponatinib was rationally designed to circumvent the BCR-ABL T315I gatekeeper mutation, it was also able to achieve major cytogenetic responses in 62% of patients without detectable kinase domain mutations in the recent PACE trial. We theorized that increased FGF2 may drive resistance in the subset of patients without kinase domain mutations who respond to ponatinib, similar to our in vitro findings. To evaluate this possibility, we identified patients without kinase domain mutations who were responsive to ponatinib and quantified bone marrow FGF2 by immunohistochemistry. In comparison to ponatinib-responsive patients with kinase domain mutations, patients without kinase domain mutations had increased FGF2 in their bone marrow (50.5% versus 36.6%, p=0.033). Moreover, FGF2 in the marrow decreased concurrently with response to ponatinib, further suggesting that FGF2-mediated resistance is interrupted by FGFR inhibition (-15.9% versus 0.8%, when compared to the change in FGF2 of patients with kinase domain mutations, p=0.012). Qualitatively, FGF2 was predominantly localized in supportive stromal cells (consistent with previous reports), supporting a paracrine mechanism of resistance. Furthermore, we also evaluated a single patient without kinase domain mutations who was resistant to ponatinib. In this patient’s marrow, there was no elevation in FGF2 or change in FGF2 with ponatinib treatment. Taken together, inhibition of FGFR appears to be critical for the clinical activity of ponatinib in patients without kinase domain mutations.
Conclusions
In summary, our data supports a model of resistance in which FGF2 production by the marrow stromal cells promotes resistance to multiple ABL kinase inhibitors without the need for mutation of the ABL kinase domain. Resistance occurs via FGF2 ligand-induced activation of the FGFR3/Ras/MAPK pathway, and can be overcome by concomitant inhibition of ABL and FGFR. In combination with recent clinical data with ponatinib, our data suggest that FGF2-mediated resistance is a major mechanism of resistance in CML patients without kinase domain mutations. These results illustrate the clinical importance of ligand-induced resistance to kinase inhibitors and support an approach of developing rational inhibitor combinations to circumvent resistance, particularly in other kinase-driven malignancies that routinely develop resistance to kinase inhibitors.
Disclosures:
Tyner: InCyte Corporation: Research Funding. Druker:Novartis, Bristol-Myers Squibb, & ARIAD: Novartis, BMS & ARIAD clin trial funding. OHSU holds contracts; no salary/lab research funds. OHSU & Druker have financial interest in MolecularMD; technology used in some studies licensed to MolecularMD. This conflict reviewed and managed by OHSU. Other.
Integrating in vitro sensitivity and dose-response slope is predictive of clinical response to ABL kinase inhibitors in chronic myeloid leukemia
