Motesanib inhibits Kit mutations associated with gastrointestinal stromal tumors

Background Activating mutations in Kit receptor tyrosine kinase or the related platelet-derived growth factor receptor (PDGFR) play an important role in the pathogenesis of gastrointestinal stromal tumors (GIST). Methods This study investigated the activity of motesanib, an inhibitor of vascular endothelial growth factor receptors (VEGFR) 1, 2, and 3; PDGFR; and Kit, against primary activating Kit mutants and mutants associated with secondary resistance to imatinib. Single- and double-mutant isoforms of Kit were evaluated for their sensitivity to motesanib or imatinib in autophosphorylation assays and in Ba/F3 cell proliferation assays. Results Motesanib inhibited Kit autophosphorylation in CHO cell lines expressing primary activating mutations in exon 9 (AYins503-504, IC50 = 18 nM) and exon 11 (V560 D, IC50 = 5 nM; Δ552-559, IC50 = 1 nM). Motesanib also demonstrated activity against kinase domain mutations conferring imatinib resistance (V560D/V654A, IC50 = 77 nM; V560D/T670I, IC50 = 277 nM; Y823 D, IC50 = 64 nM) but failed to inhibit the imatinib-resistant D816V mutant (IC50 > 3000 nM). Motesanib suppressed the proliferation of Ba/F3 cells expressing Kit mutants with IC50 values in good agreement with those observed in the autophosphorylation assays. Conclusions In conclusion, our data suggest that motesanib possesses inhibitory activity against primary Kit mutations and some imatinib-resistant secondary mutations.


Background
Approximately 85% to 90% of all cases of gastrointestinal stromal tumors (GIST) are associated with gain-of-function mutations in the gene KIT [1][2][3][4]. A further 5% to 10% of cases of GIST are associated with activating mutations in the platelet-derived growth factor receptor alpha (PDGFRα) gene [1,4,5]. Activating Kit mutations in GIST occur principally in the extracellular domain, the juxtamembrane domain (which regulates receptor dimerization), kinase domain I, and kinase domain II (or activation loop) [1]. Imatinib, a small-molecule inhibitor of Kit and PDGFRα, represents an effective first-line therapy option for patients with advanced GIST [6]. Imatinib is a potent inhibitor of wild-type Kit and juxtamembrane domain Kit mutants, while Kit activation loop mutants are resistant [1,7]. Secondary imatinib resistance is most commonly associated with the acquisition of a secondary mutation in Kit (either in the kinase domain I or the activation loop) or in PDGFRα [8].
Motesanib is an orally administered small-molecule antagonist of vascular endothelial growth factor receptors (VEGFR) 1, 2, and 3; PDGFR and Kit [9,10]. In clinical studies, motesanib has shown encouraging efficacy in the treatment of patients with advanced solid tumors [10][11][12][13]. In biochemical assays, motesanib potently inhibits the activity of both Kit (50% inhibitory concentration [IC 50 ] = 8 nM) and PDGFR (IC 50 = 84 nM) [9], suggesting that it may have direct antitumor activity in GIST [14,15]. The aim of this study was to characterize the ability of motesanib to inhibit the activity of wild-type Kit in vitro and in vivo, and to investigate differences in the potency of motesanib and imatinib against clinically important primary activating Kit mutants and mutants associated with secondary imatinib resistance. The results suggest that motesanib has inhibitory activity against primary Kit mutations and some imatinib-resistant secondary mutations.

Reagents
Unless specified otherwise all reagents were purchased from Sigma Aldrich; all cell culture reagents were purchased from Invitrogen (Carlsbad, CA).

In Vivo Hair Depigmentation Assay
Female C57B6 mice (6 to 8 weeks old; 20 to 30 g; Charles River Laboratories, Wilmington, MA) were anesthetized, and an area of skin 2 × 2 cm on the right flank was depilated. Oral administration of either 75 mg/kg motesanib (Amgen Inc., Thousand Oaks, CA) or vehicle (water, pH 2.5) was initiated on the same day as depilation and continued for 21 days. On day 21, photographs were taken for assessment of hair depigmentation. The same patch of skin was depilated again on day 28, and photographs for assessment of depigmentation were taken on day 35. All animal experimental procedures were conducted in accordance with the guidelines of the Amgen Animal Care and Use Committee and the Association for Assessment and Accreditation of Laboratory Animal Care standards.

Preparation of Wild-Type and Mutant KIT Constructs
KIT mutants (Table 1) were identified from published reports [8] and generated using PCR-based site-directed mutagenesis. PCR products were cloned into the pcDNA3.1+ hygro vector or the pDSRα22 vector (Amgen Inc), gel purified, and then ligated with a common 5' fragment of human wild-type KIT to yield full-length, mutant constructs in pcDNA3.1+ hygro or pDSRα22 expression vectors.

Stable Transfection of CHO and Ba/F3 Cells With Wild-Type and Mutant KIT
AM-1/D Chinese Hamster Ovary (CHO) cells (Amgen Inc.) were maintained under standard conditions. Cells were transfected with wild-type or mutant KIT using Lipofectamine2000 and Opti-MEM (Invitrogen) follow-ing the manufacturer's instructions. Four days after transfection, cells were transferred into selection medium: Gibco DMEM High Glucose with 10% FBS plus 300 μg/ mL hygromycin (Roche Applied Sciences, Indianapolis, IN) for cells transfected with pcDNA3.1+ hygro; DMEM High Glucose with 10% dialyzed FBS for cells transfected with pDSRα22. Stably transfected CHO cells were selected 2 weeks later and maintained as described above.
Interleukin 3 (IL-3)-dependent Ba/F3 cells were maintained under standard conditions including 3 ng/mL murine IL-3 (Cat # PMC0035; Invitrogen/BioSource). Cells were transfected with wild-type or mutant KIT in the pDSRa22 expression vector along with linearized pcDNA Neo using the Nucleofector Kit V and a Nucleoporator (Lonza; Cologne, Germany) following the manufacturer's instructions. Two to 3 days post transfection, cells were transferred into selection medium (supplemented RPMI medium plus 750 μg/mL G418). Stably transfected Ba/F3 cells were maintained in supplemented RPMI medium plus 3 ng/mL murine IL-3.
Fluorescence activated cell sorting (FACS) was utilized to isolate pools of CHO and Ba/F3 cells stably expressing wild-type and mutant KIT variants. FACS was performed on a FACS Aria cell sorter (BD Biosciences San Jose, CA), under sterile conditions using 488 nm laser excitation. KIT transfected cells were labeled with the anti-Kit monoclonal antibody SR1 (prepared at Amgen Inc.; data on file) followed by incubation with FITC-labeled secondary anti-mouse IgG antibody (SouthernBiotech, Birmingham, AL). Cells were then resuspended in Dulbecco's phosphate-buffered saline with 0.5% bovine serum albumin at a final concentration of 1 × 10 6 cells per mL to ensure a constant and viable sorting rate of 5000 cells/sec. Cells transfected with vector control were used to adjust the baseline instrument settings. Forward and side scatter gating enabled the exclusion of dead cells and debris. The top 10% to 15% of Kit-positive cells within the overall transfected cell population were then isolated to ensure collection of high-expressing cells. Cells were

In Vitro Inhibition of Wild-Type Kit by Motesanib
Motesanib potently inhibited SCF-induced autophosphorylation of Kit in CHO cells stably transfected with the wild-type KIT gene (IC 50 = 36 nM). In comparison, imatinib inhibited wild-type Kit with an IC 50 of 165 nM.

Inhibition of Wild-Type Kit Activity in Mice by Motesanib
Hair depigmentation was used as a surrogate marker to assess the ability of motesanib to inhibit Kit activity in vivo [16]. Following depilation, female C57B6 mice were administered either 75 mg/kg motesanib (n = 8) or vehicle (n = 8) twice daily for 21 days. In mice receiving motesanib, hair regrowth was markedly depigmented compared with mice receiving vehicle ( Figure 1). This effect was reversible. Following the cessation of motesanib treatment on day 21, the mice were depilated again   Figure 2 summarizes the results from the autophosphorylation experiments using CHO cells stably transfected with the wild-type KIT gene or various KIT mutant genes. Tyrosine phosphorylation of wild-type Kit was dose-dependent, with the greatest intensity of autophosphorylation occurring after a 30 minute incubation of the cells with 300 ng/mL of SCF. In contrast, tyrosine phosphorylation of activated Kit mutants occurred in the absence of SCF with no further phosphorylation induced by treatment with SCF.

Activity of Motesanib against Primary Activating Kit Mutants
In CHO cells, motesanib inhibited the autophosphorylation of the primary activating Kit mutants V560 D, Δ552-559, and AYins503-504 (Table 2; Figure 3B). In each instance, motesanib was a more potent inhibitor of Kit autophosphorylation than imatinib. For example, motesanib inhibited the AYins503-504 mutant with an IC 50 of 18 nM, whereas imatinib inhibited this mutant with an IC 50 of 84 nM. Interestingly, the IC 50 values for inhibition of these Kit mutants were lower than the IC 50 for inhibition of wild-type Kit by motesanib. Consistent results were obtained in a functional viability assay utilizing IL-3-independent growth of Ba/F3 cells ( Figure 3C). For example, when testing the AYins503-504 mutant, the IC 50 for motesanib was 11 nM versus 47 nM for imatinib.

Activity of Motesanib against Imatinib-Resistant Kit Mutants
Motesanib inhibited the activity of Kit mutants associated with secondary imatinib resistance. In Kit autophosphorylation assays, motesanib inhibited tyrosine phosphorylation of the juxtamembrane domain/kinase domain I double mutants V560D/V654A and V560D/T670I with IC 50 values of 77 nM and 277 nM, respectively. Imatinib had limited activity against the V560D/V654A mutant and no activity against the V560D/T670I mutant at concentrations of up to 3000 nM (Table 3; Figure 4B). Consistent results were obtained in the Ba/F3 cells expressing the V560D/V654A and V560D/T670I mutants with motesanib IC 50 values of 91 nM and 180 nM, respectively. Again, motesanib was a more potent inhibitor of these mutants than imatinib (Table 3; Figure 4C).
Similarly, motesanib inhibited autophosphorylation of the imatinib-resistant activation loop mutant Y823 D (IC50 = 64 nM) more potently than imatinib (IC50 > 3000 nM) (Table 3: Figure 4B). However, neither motesanib nor imatinib inhibited autophosphorylation of the D816V mutant (Table 3). Consistent with these results, motesanib inhibited the growth of Ba/F3 cells transfected with the V560D/V654A, V560D/T670I, or Y823 D mutant more potently than imatinib. Of note, the IC50 of imatinib against the Y823 D mutant when established in the functional viability assay was at least 10-fold lower than the IC50 measured in the autophosphorylation assay. IL-3-independent Ba/F3 cells expressing the D816V Kit mutant could not be established.

Discussion
In this study, motesanib was found to be a potent inhibitor of wild-type Kit, both in vitro and in vivo. In a surrogate marker assay, we observed reversible hair  depigmentation in mice treated with motesanib 75 mg/kg twice daily. This dose is comparable to the doses used in xenograft studies demonstrating antitumor and antiangiogenic properties of motesanib [9,17]. Kit signaling plays an important role in the regulation of hair follicle melanocytes, likely through control of tyrosinase and tyrosinase-related protein 1 (TRP1) expression [16]. Depigmentation has previously been observed in mice treated with anti-Kit antibodies [16,18] or with sunitinib [18]. Importantly, motesanib had inhibitory activity against Kit mutants associated with GIST and inhibited these mutants more potently than imatinib and generally with an IC 50 that was less than or similar to the 24-hour trough concentration of motesanib at therapeutic doses in humans [10]. Motesanib was a more potent inhibitor of the primary activating juxtamembrane domain and extracellular domain Kit mutants V560 D, Δ552-559, and AYins503-504, compared with imatinib. Importantly, motesanib also inhibited the activity of an activation loop mutant (Y823D) associated with imatinib resistance. Imatinib did not inhibit this mutant at concentrations of up to 3000 nM, suggesting that there are marked differences in how the two inhibitors interact with Kit. We previously solved the structure of motesanib bound to the VEGFR2 kinase domain at 2.2 Å resolution (PDB Accession Code 3EFL) [19]. This structure superimposes favorably with that of Kit co-crystallized with imatinib (PDB Accession Code 1T46) [20]. Both inhibitors bind the inactive, auto-inhibited form of the kinases with the backbone of the protein reorganized into the so-called "DFG-out" conformation. Based on the structural similarities and the similar potencies of motesanib against VEGFR2 and Kit, we reasoned that motesanib binds these target kinases in exactly the same fashion.
Modeling studies suggest that motesanib engages Kit via three polar interactions and a multitude of van der Waals contacts ( Figure 5). In the context of this study, the most important of these interactions are those with threonine 670 via a non-classical CH-O pseudo hydrogen bond and interactions with valine 654 through hydropho-  bic contacts. The fifteen-fold loss of motesanib activity (5 nM versus 77 nM) noted with the V560D/V654A double mutant, compared with V560 D alone, is rationalized by the loss of two van der Waals contacts with alanine 654 in a similar fashion to that described for imatinib [21,22]. Motesanib and imatinib have much diminished activity against the activation loop mutant (D816V). The D816V mutant destabilizes the inactivated form of Kit, in a way that the ability of the protein to adopt the "DFG out" (inactive) conformation is much reduced or even eliminated; thus, the mutation prevents both motesanib and imatinib from binding to the ATP pocket [23,24]. The failure to potently inhibit the D816V mutation is a feature of Kit inhibitors in the clinic, with the exception for dasatinib [23,25,26], which binds the "DFG in", or activated form, of the kinase [27]. However, the ability of motesanib to inhibit the Y823 D mutant suggests that its activity may not be entirely restricted to an inactive protein conformation, or alternatively it may reflect that in contrast to the D816V mutation, the conformational equilibrium of the Y823 D mutant is not shifted permanently to the active conformation.
The data from the present study are of translational relevance, supporting evidence indicating that targeted therapy molecules with different binding sites and/or mode of action may be required in the treatment of cancers for which mutations are the primary oncogenic event. A recent study has demonstrated that differences in the conformational structure of Kit mutants influences the ability of sunitinib [28], and imatinib to bind and inhibit receptor autophosphorylation, thus providing a unique mechanism of drug resistance for each mutant that is unlikely to be overcome using a single treatment [23].

Conclusions
In summary, the results of this study demonstrate that different Kit mutations respond differently to motesanib or imatinib. This likely reflects differences in the molecules' mode of action. The data also show that motesanib is active against Kit mutations associated with resistance,  suggesting that it may have clinical utility in the treatment of patients with primary and secondary imatinibresistant GIST.