Activation mutations of human c-KIT resistant to imatinib mesylate are sensitive to the tyrosine kinase inhibitor PKC412

c-KIT is a member of the type 3 subclass of transmembrane receptor tyrosine kinases, characterized by 5 immunoglobulin-like domains in the extracellular region, a negative regulatory juxtamembrane (JM) domain and a split adenosine triphosphate-binding and phosphotransferase tyrosine kinase domain (reviewed in Broudy¹ ). c-KIT activation by stem cell factor (SCF) promotes dimerization and transphosphorylation at tyrosine residues, resulting in downstream signaling events leading to cell growth. Mutations of c-KIT have been associated with hematopoietic and nonhematopoietic tumors, including systemic mast cell disease (SMCD),^2-4  acute myeloid leukemia (AML),^5-8  and gastrointestinal stromal tumors (GISTs).^9-12 

The tyrosine kinase inhibitor imatinib mesylate (Gleevec; Novartis Pharma AG, Basel, Switzerland) is efficacious in the majority of patients with GIST harboring c-KIT mutations.^9,13  c-KIT is most commonly activated in GIST tumors by small deletions in the JM that are thought to constitutively activate the tyrosine kinase by loss of autoinhibitory function.¹⁴  All c-KIT JM deletion mutants tested thus far are sensitive to inhibition by imatinib mesylate. These observations have suggested that imatinib mesylate might be of value in the treatment of SMCD or AML that also have activating mutations in c-KIT. However, the most common activating alleles in the context of SMCD and AML are the D816V and D816Y mutations in the c-KIT activation loop, and these mutations have been shown to be highly resistant to imatinib mesylate in vivo¹⁵  and in vitro.^16-19 

PKC412 is a novel staurosporine-derived tyrosine kinase inhibitor that targets protein kinase C (PKC), kinase insert domain-containing receptor (KDR), vascular endothelial growth factor receptor-2 (VEGF-R2), platelet-derived growth factor receptor α (PDGFRα), Fms-like tyrosine kinase 3 (FLT3), and c-KIT.²⁰  PKC412 has previously been shown to be effective against the fusion protein Fip1-like 1 (FIP1L1)-PDGFRα with mutations in the kinase domain that are resistant to imatinib mesylate.²¹  Since inhibitors of D816V/Y c-KIT mutations would potentially have therapeutic activity in SMCD and AML, we have evaluated the effectiveness of PKC412 against a panel of c-KIT mutations identified in patients with SMCD, GIST, and AML.

A 2.9-kB XhoI cDNA fragment encoding human (hu)-c-KIT (lacking amino acids 510-513, GNNK) was cloned into the XhoI site of pCDNA3. c-KIT point mutations (Table 1) were generated (Transformer Site-Directed Mutagenesis Kit; Clontech, Palo Alto, CA) and subcloned into the XhoI site of MSCV-IRES-GFP and MSCV-PGK-NEO. The IL-3-dependent mouse pre-B-cell line Ba/F3 was maintained, and retrovirus was prepared, as previously described.²²  Ba/F3 cells were transduced with viral supernatant, expanded for 2 passages (with 10 ng/mL rmIL-3), then selected for IL-3-independent growth. Ba/F3 cells transduced with Neo-containing vectors were first selected in IL-3 with 100 mg/mL G418. IL-3-independent Ba/F3 cells transduced with wild-type (WT) c-KIT were selected with 20 ng/mL rhSCF.

Patient samples were collected following informed consent in accordance with the institutional review board (IRB)-approved Dana-Farber Cancer Institute protocol 01-206.

Cells (1 × 10⁴) in 100 μL medium with or without IL-3 were incubated with various concentrations of imatinib mesylate or PKC412 in 96-well plates for 48 hours in triplicate. ³H-thymidine (1 μCi [0.037 MBq]) was added, and cells were incubated for 4 hours. Cells were harvested, and ³H-thymidine incorporation was determined. Growth inhibition was plotted as the ratio of the average ³H-thymidine incorporation in drug-treated wells relative to no-drug controls. IC₅₀s for imatinib mesylate or PKC412 were calculated by sigmoidal curve fitting software (OriginLab, Northampton, MA).

Cell lysates, immunoprecipitations, and immunoblotting were performed as previously described.²²  c-KIT was precipitated with 5 μg anti-human c-KIT antibody (Assay Design Systems, Ann Arbor, MI), separated on a 7.5% denaturing polyacrylamide gel and transferred to nitrocellulose. Blots were probed with mouse anti-phosphotyrosine antibody 4G10 (Upstate, Charlottesville, VA), stripped, and reprobed with mouse anti-human c-KIT (E1): SC-17806 (Santa Cruz Biotechnology, Santa Cruz, CA). Whole-cell lysates (1/10th volume immunoprecipitate [IP] input) were similarly separated and probed with rabbit anti-phospho-Stat5 (signal transducer and transcriptional activator 5) or rabbit anti-phospho-Stat3, stripped, and reprobed with rabbit anti-Stat5a or rabbit anti-Stat3 (Cell Signaling Technology, Beverly, MA).

To assess constitutive tyrosine kinase autophosphorylation, cells were grown in RPMI plus 10% fetal calf serum (FCS), and 4 mg cell lysate was immunoprecipitated. To assess PKC412 inhibition of c-KIT autophosphorylation, cells were washed 3 times in RPMI without FCS or IL-3, then incubated with or without IL-3 and PKC412 (0-500 nM) for 4 hours. WT Ba/F3 cells were starved of IL-3 and serum for 4 hours prior to addition of IL-3 and PKC412. Cell equivalents (1 × 10⁷) of cell lysate were immunoprecipitated.

The IL-3-dependent murine pre-B-cell line Ba/F3 was transduced with retroviruses expressing WT c-KIT or 1 of 14 c-KIT mutants and selected for IL-3-independent growth (Table 1; Figure 1A). Ten of 14 mutant forms of c-KIT transformed Ba/F3 cells to IL-3-independent growth in the absence of rhSCF. Transforming mutations mapped to the extracellular domain (delTYD (417-419) + I,⁸  delTYD (417-419) + RG,⁸  and A502Y503dup (AYdup)¹⁰ ), the juxtamembrane domain (V559D,¹¹  V560G,²  and delV559V560 (delVV)¹¹ ), and the kinase domain (R634W, D816Y,³  D816V,²  and N822K⁷ ). This is the first report of the kinase domain mutation R634W, identified from a patient with chronic myelomonocytic leukemia. Consistent with a report that E839K inactivates the kinase domain,¹²  E839K did not transform Ba/F3 cells to IL-3-independent growth. K642E has been shown to be constitutively phosphorylated in a GIST-derived cell line^10,23  but did not transform Ba/F3 cells in this study. D820G⁴  and V530I,⁸  each identified in a single patient, also did not transform Ba/F3 cells.

Immunoprecipitated c-KIT was constitutively phosphorylated in all transformed cell lines and Ba/F3 cells expressing WT c-KIT grown with 20 ng/mL rhSCF (Figure 1B). We observed differential phosphorylation of 2 c-KIT bands of approximately 160 and 145 kDa, representing the fully glycosylated cell surface receptor, and incompletely processed internalized forms of c-KIT, respectively. Our findings confirm reports that Stat3 is aberrantly phosphorylated by D816V and other c-KIT-activating mutations.^24,25  Stat5 was constitutively phosphorylated by all c-KIT-activating mutations grown in the absence of IL-3.

The effect of imatinib mesylate or PKC412 on cell growth of c-KIT-expressing cell lines was evaluated in a 3H-thymidine incorporation assay (Table 1; Figure 1C; data not shown). Consistent with previous reports, D816V was resistant to imatinib mesylate (IC₅₀, > 10 000 nM).^16-19  D816Y was also observed to be resistant to imatinib mesylate (IC₅₀, > 800 nM). Four mutations, delTYD + I, AYdup, R634W, and N822K, showed IC₅₀s for imatinib mesylate similar to Ba/F3-WT-c-KIT + 20 ng/mL rhSCF (100-300 nM). As previously reported, V560G¹⁹  was hypersensitive to imatinib mesylate, as were the mutations delTYD + RG, V559D, and delVV (IC₅₀s < 100 nM).

PKC412 inhibited growth of all c-KIT-transformed Ba/F3 lines, with IC₅₀s segregating into 2 groups. The imatinib mesylate-resistant D816V and D816Y mutants, as well as delTYD + RG, V560G, R634W, and N822K, were more sensitive to PKC412 (IC₅₀s, 33-95 nM) than was the wild-type receptor stimulated with rhSCF (IC₅₀, 138 nM). The remaining mutations had higher IC₅₀s for PKC412 (146-265 nM).

The level of tyrosine autophosphorylation of the imatinib mesylate-resistant D816V and D816Y c-KIT mutants, as well as phosphorylation of Stat3 and Stat5, was dose dependent on PKC412 over a wide concentration range (5-500nM) and correlated with inhibition of cell growth (Table 1; Figure 1D). Similar findings were observed for representative imatinib mesylate-sensitive extracellular (del-TYD + RG) and JM (V560G) mutants. Although Stat5 phosphorylation by D816V and D816Y was inhibited by PKC412 (Figure 1D-E), an increased level of Stat5 phosphorylation induced by stimulation with IL-3 induction was not inhibited in these cells (Figure 1E). IL-3 did not rescue PKC412 inhibition of phosphorylation of Stat3 (Figure 1E) and did not inhibit IL-3-induced Stat5 phosphorylation in WT Ba/F3 Cells (Figure 1E-F).

These data indicate that among the reported c-KIT mutations in human hematologic malignancies, 10 of 14 confer factor independent growth to Ba/F3 cells. There is variable sensitivity of these mutations to inhibition by imatinib mesylate or PKC412. Notably, however, the most common mutations D816V and D816Y are highly resistant to imatinib mesylate but sensitive to inhibition by PKC412. These findings suggest that PKC412 may be an effective therapeutic agent for treatment of patients with certain imatinib mesylate-resistant c-Kit mutations.

D.G.G is an investigator of the Howard Hughes Medical Institute.