Uniform sensitivity of FLT3 activation loop mutants to the tyrosine kinase inhibitor midostaurin

Fms-like tyrosine kinase 3 (FLT3), a cell surface receptor tyrosine kinase, is among the most commonly mutated genes in acute myeloid leukemia (AML).¹  Stimulation of FLT3 activates signal transduction pathways such as signal transducer and activator of transcription 5 (STAT5), RAS/mitogen-activated protein kinase (RAS/MAPK), phosphoinositide 3-kinase (PI3K), src homologous and collagen gene (SHC), SH2-containing inositol-5-phosphatase (SHIP), and cytoplasmic tyrosine phosphatase with 2 Src-homology 2 (SH2) domains (SHP2), which play important roles in cellular proliferation, differentiation, and survival.^2,3 

There are 2 types of activating mutations in FLT3 described in patients with leukemia. These include a spectrum of internal tandem duplications (ITD) occurring within the auto-inhibitory juxtamembrane domain,^4,–6  and activation loop mutations that include Asp835Tyr (D835Y), Asp835Val (D835V), Asp835His (D835H), Asp835Glu (D835E), Asp835Ala (D835A), Asp835Asn (D835N), Asp835 deletion and Ile836 deletion.^7,,–10  These activating mutations result in constitutive phosphorylation and activation of FLT3, and subsequent activation of downstream targets.^10,11 

The importance of FLT3 mutations in the pathogenesis of leukemias has been well established and, in most studies, these have been shown to confer a poor prognosis, with decreased survival.^12,–14  Therefore, attention has been focused on developing small molecule inhibitors that target FLT3. Midostaurin (formerly known as PKC412) is a selective inhibitor of FLT3, as well as vascular endothelial growth factor receptor (VEGFR), platelet-derived growth factor receptor (PDGFR), c-kit receptor tyrosine kinase (KIT), and fibroblast growth receptor 1 (FGFR-1).^15,–17  In vitro, midostaurin induces apoptosis in Ba/F3 cells that have been transformed to IL-3-independent growth by FLT3-ITD.¹⁸  Furthermore, midostaurin is effective in FLT3-ITD-induced disease in a murine bone marrow transplantation model.¹⁸ 

Midostaurin has been investigated in phase 1 and phase 2 studies, and is generally well tolerated. In a phase 2 study in patients with relapsed or refractory AML, single-agent midostaurin resulted in significant reductions in peripheral blood and bone marrow blasts, with a 35% response rate.¹⁹  Midostaurin also showed activity in a phase 1 trial in combination with daunorubicin and cytarabine in newly diagnosed AML, in particular those patients with FLT3-ITD mutations.²⁰  Based on these promising findings, a randomized phase 3 Cancer and Leukemia Group B (CALGB)/Southwest Oncology Group (SWOG)/Eastern Cooperative Oncology Group (ECOG) Intergroup trial is planned to assess efficacy of induction chemotherapy with midostaurin in patients with mutant FLT3.

Studies have analyzed the sensitivity of FLT3-activating mutations to small molecule inhibitors, and have demonstrated that certain activation loop mutations simultaneously activate the kinase and confer resistance to small molecule inhibitors.²¹  Therefore, defining the precise spectrum of sensitivity of the range of FLT3 activation loop mutations to specific kinase inhibitors may be critical to determine patient eligibility for treatment, and evaluation of response to therapy. Here we report that midostaurin is a potent inhibitor of 8 activation loop mutations tested.

DNA constructs of FLT3-ITD, and FLT3 with D835A, D835E, D835H, D835N, D835V, D835 deletion, I836 deletion, and D835Y point mutations were created and cloned into the murine stem cell virus (MSCV)-neo vector as previously described.^14,21  IL-3-dependent murine hematopoietic Ba/F3 cells were transduced as previously described.^21,22  In addition, a DNA construct consisting of a FLT3 G697R kinase domain point mutation was cloned into the MSCV-FLT3-ITD vector and transduced into Ba/F3 cells.²³  Transduced cells were grown in absence of IL-3 to confirm factor independence.

Midostaurin (Novartis Pharma, Basel, Switzerland) was prepared in a 10mM stock solution in DMSO and stored at −20°C. Serial dilutions were made in DMSO to obtain the final concentrations used for immunoblotting and cell growth assays.

Each cell line (1.5 × 10⁵ cells/mL) was grown with varying concentrations of midostaurin in DMSO for 48 hours. The number of viable cells was then determined by a colorimetric assay.²¹  Results are expressed as a percentage of viable cells after 48 hours in the presence of DMSO only.

Immunoblotting was performed as previously described.^21,24  Briefly, after incubating cells in the presence of DMSO alone or with various concentrations of midostaurin, cell lysates were prepared. Lysates were separated using SDS-PAGE and transferred onto a nitrocellulose membrane. For immunoblotting, the following primary antibodies were used: anti-phospho-FLT3 (Tyr 591; Cell Signaling, Beverly, MA), anti-FLT3/Flk-2 (S-18; Santa Cruz Biotech, Santa Cruz, CA), anti-phospho-STAT5 (Tyr 694; Cell Signaling), anti-STAT5a (L-20; Santa Cruz Biotech), anti-phospho-p44/42 MAPK (Thr202/Tyr204; Cell Signaling), anti-p44/42 MAPK (Cell Signaling). Antirabbit-immunoglobulin horseradish peroxidase (Amersham Biosciences, Piscataway, NJ) was used as a secondary antibody. Detection was performed by enhanced chemiluminescence.

We confirmed that each of the activation loop mutations resulted in constitutive activation of FLT3 and its downstream effectors STAT5 and ERK. Figure 1A demonstrates that FLT3 was constitutively phosphorylated in all mutants. In addition, blotting with phospho-specific antibodies for STAT5 and ERK also demonstrated constitutive activation of these targets of activated FLT3.

Each of the transduced Ba/F3 cell lines was then tested for sensitivity to midostaurin. Ba/F3 cells expressing each of the activation loop mutations as well as FLT3-ITD were incubated with varying concentrations of midostaurin. At 48 hours, viable cells were quantitated using a colorimetric assay. There was inhibition of growth in all cell lines with increasing concentrations of midostaurin (Figure 1B). Inhibition of growth was abrogated by the addition of IL-3, supporting that inhibition by midostaurin is specific to the FLT3 pathway (Figure 1C). The cellular concentration that inhibited response by 50% (IC₅₀) was less than 10 nM for all of the mutants tested, with a range of 1 nM to 10 nM (Table S1, available on the Blood website; see the Supplemental Materials link at the top of the online article), demonstrating that midostaurin resulted in a decrease in cellular proliferation of all of the Ba/F3 constructs at relatively low concentrations. Pharmacokinetic studies in phase 1 and 2 trials of midostaurin have shown that these are values that should be readily achieved in patients.²⁵ 

To assess the effect of midostaurin on FLT3 phosphorylation, lysates were prepared from cells incubated with various concentrations of midostaurin, followed by immunoblotting with anti-phospho-FLT3 antibody. There was inhibition of FLT3 phosphorylation in all of the activation loop mutants at relatively low concentrations of midostaurin (Figure 1D), with an IC₅₀ for inhibition of phosphorylation comparable to that observed for inhibition of cell growth. We also observed inhibition of phosphorylation of STAT5 and ERK, at comparably low concentrations of midostaurin (Figures 1E,F).

Thus, all of the activation loop mutants tested were sensitive to inhibition with midostaurin. However, midostaurin is known to inhibit kinases other than FLT3, and thus it was possible that the observed results, although consistent with FLT3 as a target of midostaurin, could be explained by simultaneous inhibition of other kinases essential for cell growth. To demonstrate that the effect of midostaurin was wholly attributable to inhibition of FLT3, and not due to off-target effects, we repeated the experiments using Ba/F3 cells stably transduced with a FLT3-ITD G697R allele (Ba/F3 FLT3-ITD-G697R cells), that we have previously reported to confer resistance to midostaurin.²³  In the cell growth assay (Figure 2A), there was no inhibition of cellular proliferation at doses of midostaurin up to 300 nM. In addition, immunoblot assays (Figure 2B) demonstrated that midostaurin did not inhibit phosphorylation of FLT3, STAT5, or ERK in Ba/F3 FLT3-ITD-G697R cells. These findings indicate that inhibition of the respective mutant FLT3 activating alleles in this cellular context is due to specific inhibition of FLT3 by midostaurin.

These findings have a direct bearing on phase 3 trials that are currently being planned in the Intergroup setting for treatment of newly diagnosed AML patients with FLT3-ITD mutations with midostaurin and induction chemotherapy. In contrast to other small molecule FLT3 inhibitors, midostaurin inhibits the full spectrum of FLT3 activation loop mutations. Thus, these data indicate that AML patients harboring such activation loop mutations in FLT3 should be considered as candidates for enrollment, in addition to those with FLT3-ITD alleles.

This work was supported in part by National Institutes of Health grants DK50654 and CA66996 (D.G.G.), Belgian Federation Against Cancer grant SCIE2006-34 (J.C.), and by the Leukemia and Lymphoma Society. D.G.G. is a Doris Duke Foundation Distinguished Clinical Scientist and an Investigator of the Howard Hughes Medical Institute.

National Institutes of Health

Contribution: E.V.B. wrote the paper, J.J.C. designed the research and analyzed data, J.C. and J.R. contributed reagents, and D.G.G. designed the research, analyzed data, and wrote the paper.

Conflict-of-interest disclosure: J.R. is employed by Novartis Pharma. All other authors declare no competing financial interests.

Correspondence: Elly V. Barry, MD, Dana-Farber Cancer Institute, Department of Pediatric Oncology, 44 Binney St, SW320, Boston, MA 02115; e-mail:elly_barry@dfci.harvard.edu.