A human monoclonal antibody targeting the stem cell factor receptor (c-Kit) blocks tumor cell signaling and inhibits tumor growth

Abstract

Stem cell factor receptor (c-Kit) exerts multiple biological effects on target cells upon binding its ligand stem cell factor (SCF). Aberrant activation of c-Kit results in dysregulated signaling and is implicated in the pathogenesis of numerous cancers. The development of more specific and effective c-Kit therapies is warranted given its essential role in tumorigenesis. In this study, we describe the biological properties of CK6, a fully human IgG1 monoclonal antibody against the extracellular region of human c-Kit. CK6 specifically binds c-Kit receptor with high affinity (EC₅₀ = 0.06 nM) and strongly blocks its interaction with SCF (IC₅₀ = 0.41 nM) in solid phase assays. Flow cytometry shows CK6 binding to c-Kit on the cell surface of human small cell lung carcinoma (SCLC), melanoma, and leukemia tumor cell lines. Furthermore, exposure to CK6 inhibits SCF stimulation of c-Kit tyrosine kinase activity and downstream signaling pathways such as mitogen-activated protein kinase (MAPK) and protein kinase B (AKT), in addition to reducing tumor cell line growth in vitro. CK6 treatment significantly decreases human xenograft tumor growth in NCI-H526 SCLC (T/C% = 57) and Malme-3M melanoma (T/C% = 58) models in vivo. The combination of CK6 with standard of care chemotherapy agents, cisplatin and etoposide for SCLC or dacarbazine for melanoma, more potently reduces tumor growth (SCLC T/C% = 24, melanoma T/C% = 38) compared with CK6 or chemotherapy alone. In summary, our results demonstrate that CK6 is a c-Kit antagonist antibody with tumor growth neutralizing properties and are highly suggestive of potential therapeutic application in treating human malignancies harboring c-Kit receptor.

Introduction

Stem cell factor receptor (c-Kit) is a type III receptor tyrosine kinase family member that mediates cell growth, survival and differentiation signals in response to the ligand stem cell factor (SCF).^1-4 c-Kit is found expressed in normal cells⁵ including hematopoietic progenitor stem cells, mast cells, melanocytes, interstitial cells of Cajal, and germ cells, where it has physiological functions in hematopoiesis, mast cell development, melanogenesis, gut pacemaking, and gametogenesis, respectively.^6-11 The aberrant expression of c-Kit receptor and/or its activation through mutations or SCF/c-Kit autocrine/paracrine signaling loop mechanisms occurs in numerous malignancies and is thought to promote tumor development.^3,4 Examples of cancers with such c-Kit abnormalities are small cell lung carcinoma (SCLC), acute myeloid leukemia (AML), gastrointestinal stromal tumor (GIST), melanoma,x and systemic mastocytosis.^12-16

The structure of c-Kit receptor is characterized by the presence of an extracellular region with five immunoglobulin (Ig)-like motifs of which the first, second, and third are involved in SCF binding and the fourth and fifth in receptor dimerization.¹⁷ This is followed by a single transmembrane spanning domain, an intracellular juxtamembrane domain, a split protein tyrosine kinase domain, and a COOH-terminal region. Similar to the activation mechanism of other growth factor receptors, SCF binding to c-Kit induces receptor dimerization and intrinsic tyrosine kinase activity that leads to receptor autophosphorylation and triggering of downstream signaling cascades that include Ras/mitogen-activated protein kinase (MAPK) and phosphatidylinositol 3-kinase (PI3K)/protein kinase B (AKT) pathways.^18-20 The biological outcome of ligand-dependent c-Kit-mediated stimulation of these intracellular signaling pathways is increased cell proliferation and survival. Furthermore, gain of function c-Kit mutations are frequently located in the juxtamembrane and kinase domains, and these contribute to the constitutive ligand-independent receptor activity which is commonly observed in cancers of GIST,^14,21 AML,^13,22 and subtypes of melanoma (e.g., acral, mucosal).^15,23

Several small molecule multi-targeted tyrosine kinase inhibitors have been developed as cancer therapies in the clinic, some of which (e.g., imatinib, sunitinib, and regorafenib) demonstrated efficacy in GIST patient tumors harboring mutated and constitutively active c-Kit receptors.^24-26 Currently approved agents are non-selective c-Kit inhibitors in that they also block other kinases including BCR-ABL, PDGFR, or VEGFR. These agents also have side effects that limit their potential owing to the emergence of primary and secondary resistance mechanisms,^27,28 encouraging the discovery of more specific c-Kit-targeted therapies including more selective kinase inhibitors or antibodies. Here we report on the development and characterization of CK6, a fully human IgG1 monoclonal antibody with high affinity against the extracellular region of c-Kit. We show strong binding of CK6 to human c-Kit protein and potent blocking of c-Kit interaction with SCF. In c-Kit-expressing tumor cell lines, in vitro exposure to CK6 inhibits SCF stimulated signaling and cell growth responses. Single agent treatment with CK6 reduces growth of SCLC and melanoma tumor xenograft models in vivo. Furthermore, the combination of CK6 with standard of care chemotherapy provides enhanced tumor growth inhibition in these models. Taken together, our findings suggest the potential for c-Kit antibody therapy in SCLC, melanoma, and other c-Kit expressing human tumors.

Results

Characterization of CK6 binding activity to c-Kit

To generate a fully human anti-cKit monoclonal antibody with high affinity receptor binding and neutralizing properties, human immunoglobulin transgenic mice were immunized with recombinant human extracellular c-Kit protein. Antibody clone CK6 was identified through hybridoma screening. CK6 was determined to bind in a dose-dependent manner to immobilized recombinant human c-Kit with an EC₅₀ of 0.06 nM (Fig. 1A) and blocked c-Kit interaction with human SCF in a dose-dependent manner with an IC₅₀ of 0.41 nM (Fig. 1B). Surface plasmon resonance was performed to measure the kinetics of CK6 binding to recombinant human extracellular c-Kit protein. CK6 showed strong binding affinity to human c-Kit with a K_D value of 0.525 nM.

Figure 1. Solid phase CK6 binding and blocking properties. (A) CK6 binding to immobilized human c-Kit protein (EC₅₀ = 0.06 nM) in ELISA. (B) CK6 blocking of human c-Kit protein binding to immobilized human SCF (IC₅₀ = 0.41 nM) in ELISA. One representative experiment is shown with EC₅₀ and IC₅₀ values calculated from two independent runs.

The potential species cross-reactivity of CK6 against mouse, rat, and cynomolgus monkey c-Kit was investigated using cells expressing each species specific receptor. As shown in flow cytometry histograms (Fig. S1A), CK6 bound to 293 cells transiently overexpressing full-length human and monkey c-Kit protein but had no reactivity with mouse or rat c-Kit. A western blot analysis of the corresponding cell lysates was also run as an internal control to show c-Kit protein expression levels in each cell line sample (Fig. S1B). The recognition of monkey c-Kit and not mouse or rat c-Kit protein by CK6 was further confirmed in binding ELISA using species-specific recombinant extracellular c-Kit proteins (data not shown). Thus, CK6 exhibits strong cross-reactivity against human and monkey c-Kit.

Inhibition of SCF signaling in SCLC, melanoma, and leukemia tumor cell lines in vitro

To evaluate the cell surface binding of CK6 to native human c-Kit on tumor cell lines, we conducted flow cytometry analysis. CK6 was found to strongly bind various human tumor cell lines representing SCLC, melanoma, and leukemia as shown in Figure 2. To demonstrate specificity, control human IgG staining was used in all cell lines and showed minimal or no background staining. In addition, DMS-153 and SK-MEL-1 did not bind CK6 and served as c-Kit negative cell controls. The level of c-Kit protein expression for most of the tumor cell line panel was also confirmed by western blots of cell lysates (data not shown).

Figure 2. Analysis of CK6 binding to c-Kit receptor expressing tumor cell lines. Flow cytometry staining was performed with CK6 antibody or control human IgG against a panel of (A) SCLC, (B) melanoma, and (C) leukemia cell lines. Histograms indicate fluorescence signal of unstained cells (solid gray fill) and cells stained with either CK6 (black solid line) or human IgG (black dashed line).

CK6 antibody was tested for in vitro activity on SCF mediated signaling in tumor cell lines derived from patients with SCLC, melanoma and leukemia. Pretreatment with CK6 for 2 h inhibited SCF stimulation of c-Kit signal transduction in several cell lines (Fig. 3) with reduced phospho-cKit, phospho-p44/p42 MAPK and phospho-AKT protein levels in response to CK6 but not control human IgG. CK6 antibody demonstrated dose dependent inhibition of c-Kit expressing cells (e.g., SCLC) that was comparable to that of small molecule tyrosine kinase inhibitors sunitinib and imatinib (Fig. S2; data not shown).

Figure 3. CK6 inhibits SCF mediated signal transduction in vitro. Tumor cell lines representing (A) SCLC, (B) melanoma, and (C) leukemia were serum starved overnight and then pretreated with or without CK6 antibody or control human IgG at indicated doses for 2 h before subsequent stimulation with human SCF (1 μg/mL) for 15 min. Cell lysates were prepared and then analyzed by western blot for levels of phosphorylated cKit, phosphorylated AKT, and phosphorylated p44/p42 MAPK proteins. GAPDH control antibody was used to show relatively equal protein levels in lysate samples.

Neutralization of tumor cell line growth by CK6 in vitro

To determine the ability of CK6 to inhibit SCF induced cell growth in vitro, we ran cell viability assay in the SCF responsive MO7e leukemia cell line.²⁹ MO7e cells were exposed to increasing concentrations of either CK6 or control human IgG for 2 h before incubation with SCF for 72 h. Cell viability was measured and is depicted in Figure 4A. CK6 completely inhibited SCF mediated cell growth response in a dose-dependent manner with an IC₅₀ of 2.799 nM. Control human IgG provided no significant change in viability compared with CK6. To confirm the observed cell growth blocking properties of CK6, we examined its ability to antagonize the growth of tumor cells in soft agar (Fig. 4B). CK6 treatment of NCI-H69 SCLC cells decreased the number of colonies formed resulting in a 47% reduction in colony growth compared with control human IgG. Images from one representative experiment depict lower colony numbers in the CK6 treated wells.

Figure 4. Inhibition of cell growth by CK6 in vitro. (A) MO7e cells were grown in presence or absence of varying concentrations of CK6 or control human IgG for 2 h and then stimulated with or without human SCF (100 ng/mL) for 72 h. Cell viability was assessed using the Promega CellTiter-Glo luminescent assay. One representative experiment is shown with IC₅₀ = 2.799 nM. (B) NCI-H69 cells growing in serum containing media were plated onto soft agar with or without 660 nM CK6 or control human IgG. Growth was measured by counting the number of cell colonies and is displayed as a percentage relative to human IgG control. The mean of three independent experiments is shown. Images of colony growth from one experiment are depicted (duplicate wells).

CK6 inhibits the growth of human tumor xenografts in vivo

To evaluate the antitumor activity of CK6 in vivo, we used female immunodeficient athymic nude mice bearing human tumor xenografts as models. Nude mice with NCI-H526 SCLC tumors were treated with USP saline, CK6, cisplatin/etoposide or CK6+cisplatin/etoposide (Fig. 5A). Antitumor growth response measured as change in mean tumor volume was present in the chemotherapy alone (T/C% = 50%) or CK6 alone (T/C% = 57%) arms compared with the control USP saline group. Enhanced tumor growth inhibition was observed when combining CK6 with chemotherapy (T/C% = 24%). As a second model, nude mice with Malme-3M melanoma tumors were treated in a similar manner with USP saline, CK6, dacarbazine or CK6+dacarbazine (Fig. 5B). Antitumor growth response was detected in study arms treated with single agents dacarbazine (T/C% = 65%) or CK6 (T/C% = 58%) compared with the control USP saline group. Enhanced tumor growth inhibition was apparent when combining CK6 with chemotherapy (T/C% = 38%). For both studies, there were no significant CK6 related changes in body weight. These findings show that CK6 reduces tumor growth as a single agent but has significantly greater efficacy in combination with standard of care chemotherapy agents.

Figure 5. Tumor xenograft growth inhibition by CK6 in vivo. (A) NCI-H526 SCLC and (B) Malme-3M melanoma human tumor xenografts growing in mice were treated with indicated doses of CK6, chemotherapy agents, or CK6 combined with chemotherapy as shown. Single agent CK6 reduced tumor volume and in combination with chemotherapy provided significantly more antitumor growth activity. CK6 (40 mg/kg) was dosed 3 times weekly. Chemotherapy dosing is indicated by arrows; cisplatin (5 mg/kg)/etoposide (40 mg/kg), dacarbazine (80 mg/kg). Tumor volume data are presented as mean ± SEM.

Internalization of CK6

Many neutralizing antibodies exert some of their potency via internalization and degradation of the antigen. Given the observed neutralizing properties of CK6 in vitro and in vivo, we decided to investigate the internalization properties of CK6 upon binding c-Kit by flow cytometry analysis of fluorescently labeled CK6. SCLC cell lines NCI-H526, NCI-H889 and NCI-H69 were selected and either untreated or exposed to Alexa Fluor 488 conjugated CK6 or Alexa Fluor 488 human IgG1 control at 37 °C for 1, 3, 6, and 24 h respectively (or at 4 °C as positive control) followed by acid wash step to remove cell surface fluorescent signal before analysis by flow cytometry. Levels of intracellular CK6 antibody, measured by mean fluorescence intensity (MFI) values, increased over time in each of the three tumor cell lines compared with the human IgG1 control and reached highest levels at 24 h (Fig. S3). However, internalization and degradation of the c-Kit receptor was not observed by flow cytometry or western blot in several tumor cell lines and CK6 exposure did not reduce amounts of c-Kit protein in NCI-H526 and MO7e cells (Fig. S4A and B). These findings indicate that CK6 is internalized but treatment does not result in c-Kit receptor degradation.

c-Kit expression in primary human lung tumor tissue

Given our findings of c-Kit expression in tumor cell lines and sensitivity to CK6 treatment in vitro and in vivo, we were interested in the prevalence of c-Kit in human primary tumor tissues. A commercial anti-cKit antibody (A4502, Dako) was utilized to assess c-Kit protein expression by immunohistochemistry in lung formalin-fixed paraffin-embedded (FFPE) tumor tissue microarrays (TMAs) containing 78 SCLC samples and 38 non-small cell lung carcinoma (NSCLC) samples. Forty-six percent (36 of 78 samples) of SCLC samples were c-Kit positive (Fig. 6A). All non-neoplastic lung tissues examined were negative and only a single NSCLC sample representing an atypical carcinoid tumor was positive (1 of 38 samples). Representative immunohistochemistry images of c-Kit positive SCLC, c-Kit negative NSCLC (adenocarcinoma sample) and non-neoplastic lung tissue are shown in Figure 6B.

Figure 6. c-Kit expression in primary lung tumor tissue samples. (A) Human FFPE lung tumor tissue microarrays (TMAs) were immunostained with c-Kit antibody A4502 (Dako). The immunolabeling of the lung tumor cores was scored by the pathologist. The number of c-Kit positive samples is shown as a percentage of the total SCLC, NSCLC or non-neoplastic lung samples. (B) Representative images of c-Kit immunostaining in lung tumor and non-neoplastic lung tissue captured at 20× magnification using Aperio ScanScope XT. Positive c-Kit staining (red signal) is depicted in SCLC (two separate examples) and not in non-neoplastic lung or NSCLC (adenocarcinoma example).

Discussion

Aberrant activation of the SCF/c-Kit signaling axis has been implicated in tumor development and progression.^3,4 In c-Kit driven GIST cancer patients, small molecule tyrosine kinase inhibitors imatinib and sunitinib have shown good clinical activity but effects may be limited due to intrinsic and acquired resistance (e.g., up to 15% primary and 50% secondary resistance to imatinib) resulting in disease progression in most patients.^24,25,27,28 Furthermore, the efficacy of these agents in other c-Kit positive tumors is not as evident, especially in SCLC and leukemia (AML) where clinical trials have failed in patients with relapsed or refractory disease,^30-32 and in melanoma where modest drug activity was reported predominantly in patients with mutant c-Kit tumors.^33,34 A function blocking antibody is an alternative approach to targeting c-Kit expressing cancers. Previous reports of c-Kit antibodies (e.g., YB5.B8, SR-1),^35,36 lacked extensive in vivo characterization and demonstrated potential for agonistic activity.³⁷ In the present study, we describe the in vitro and in vivo antitumor growth neutralizing properties of CK6, a fully human monoclonal antibody to c-Kit, and provide rationale for its potential therapeutic use in multiple c-Kit positive cancers that include SCLC, melanoma, and leukemia.

To characterize the binding properties of CK6 antibody, we used solid phase protein assays and flow cytometry. CK6 showed specific high affinity binding to human c-Kit protein and species cross-reactivity with monkey c-Kit. In addition, CK6 strongly blocked the interaction between c-Kit and SCF. The putative extracellular epitope of c-Kit that contacts CK6 is located within the third Ig-like domain (data not shown), a region known to be involved in SCF binding,¹⁷ which may explain the potent CK6 mediated blocking effect. Furthermore, CK6 antibody recognized native c-Kit protein present on the cell surface of SCLC, melanoma, leukemia, and other tumor cell lines (Fig. 2; data not shown). The extent of binding varied across the cell lines, most of which are c-Kit wild type. Interestingly, CK6 bound to the mutant c-Kit harboring Kasumi-1 leukemia line (N882K mutant). Our findings indicate strong CK6 binding to both wild type and mutant forms of c-Kit receptor that would enable specific targeting of c-Kit positive cancers with CK6 antibody.

Ligand-dependent activation of c-Kit receptor by SCF stimulates cell growth and survival.^3,4,18-20 Through the use of c-Kit expressing tumor cell lines (e.g., SCLC, melanoma, leukemia), we have shown that CK6 antibody prevents SCF mediated induction of c-Kit tyrosine phosphorylation in vitro in addition to reducing downstream activation of phospho-MAPK and phospho-AKT, two key effectors of cell growth and survival. Agonist activity in response to CK6 exposure was not detected (data not shown). Furthermore, we observed that the block in c-Kit signaling results in reduced cell growth of the MO7e leukemia cell line and decreased soft agar colony formation of the NCI-H69 SCLC cells both of which express wild-type c-Kit. These findings demonstrate sensitivity of c-Kit expressing and SCF responsive tumor cell lines to the antagonistic effects of CK6 which may potentially occur through neutralization of SCF/c-Kit autocrine/paracrine loop pathways.³⁸ We also considered the possibility of additional CK6 antibody mediated mechanisms of action, such as receptor degradation. CK6 internalizes within several hours of binding to c-Kit positive cells but does not induce significant receptor degradation (Fig. S4).

The CK6 antibody demonstrated in vivo antitumor growth efficacy in c-Kit expressing human tumor xenograft models of SCLC (NCI-H526) and melanoma (Malme-3M). CK6 monotherapy showed a response similar to that of cisplatin/etoposide in SCLC (CK6 T/C% = 57, chemotherapy T/C% = 50) and dacarbazine in melanoma (CK6 T/C% = 58, chemotherapy T/C% = 65). CK6 showed additive antitumor efficacy when combined with these chemotherapies (SCLC T/C% = 24%, P < 0.0001; melanoma T/C% = 38%, P < 0.0001). Previous studies have failed to demonstrate in vivo efficacy of imatinib in SCLC model NCI-H526.³⁹ There are limited reports describing c-Kit antibody behavior in tumor models in vivo, in particular the SR-1 antibody was shown to inhibit tumor growth in GIST models.⁴⁰ In the future, it will be important to delineate whether c-Kit expression levels and/or mutation status predict in vivo efficacy of CK6 since both the NCI-H526 and Malme-3M selected models expressed wild-type c-Kit and a model with lower receptor levels was not tested. Furthermore, additional in vivo studies are necessary to gain deeper insight into mechanism of CK6 single agent and combination therapy antitumor effects.

Tumor expression profiling or prevalence testing of cancer targets is important for the development of targeted therapies and potential biomarker discovery. In SCLC, c-Kit RNA and protein is found overexpressed in tumor cell lines and primary tumor tissues, and it is hypothesized that there is an SCF/c-Kit autocrine loop involved in this cancer.^12,38,41,42 Our data highly support the function of c-Kit in this cancer type given that CK6 antagonized signaling in SCLC tumor cell lines and blocked their growth in vitro and in vivo. We analyzed c-Kit expression in primary lung tumor samples by immunohistochemistry and found 46% of SCLC tumors were c-Kit positive, consistent with previous reports of 37 to 86% of SCLC tumors being positive for c-Kit.⁴³ Interestingly, NSCLC subtypes representing adenocarcinoma, squamous, large cell and others were predominantly c-Kit negative which agrees with reported lower prevalence in NSCLC than SCLC. Taken together, our data suggest rationale for selecting c-Kit positive SCLC tumors to evaluate CK6 activity in this cancer type which has limited therapy options and poor patient prognosis (e.g., 5-y survival of 1% in extensive disease).⁴⁴ Similarly, in melanoma there are reports of aberrant c-Kit expression and presence of activating mutations in the acral and mucosal tumor subtypes.^15,23,33,34 The observed in vitro and in vivo sensitivity of c-Kit expressing melanomas to CK6 suggests this is an additional indication for antibody mediated therapy. A more extensive evaluation of c-Kit expression across a broader panel of primary tumor samples is warranted (along with tumor molecular profiling) to identify additional c-Kit positive cancer types eligible for CK6 preclinical efficacy testing which ultimately will provide a better understanding of patient population most suitable to receive and likely benefit from CK6 therapy in the clinic.

In summary, we describe a fully human monoclonal antibody with high affinity to the human c-Kit receptor and potent antitumor activity in vitro and in vivo. These data indicate favorable c-Kit targeting with an antibody approach and demonstrate promising anticancer activity when CK6 antibody is combined with chemotherapy in SCLC and melanoma. Our findings support the continued evaluation of CK6 as potential therapy for c-Kit positive cancers.

Materials and Methods

Cell culture

Tumor cell lines were purchased from American Type Culture Collection (ATCC) except CMK and MO7e which were from Leibniz Institute DSMZ-German Collection of Microorganisms and Cell Cultures. All were cultured in the respective recommended media (Life Technologies) with serum (HyClone) at 37 °C in a humidified 5% CO₂ atmosphere.

Monoclonal antibody

c-Kit monoclonal antibody CK6 was derived by immunizing Medarex human immunoglobulin transgenic mice with recombinant human extracellular c-Kit protein (R&D Systems). The heavy and light chain variable region cDNA sequences of CK6 were cloned and fused in frame to the human immunoglobulin heavy chain gamma1 constant region of glutamine synthetase (GS) expression vector (Lonza Group Ltd). CK6 antibody was stably expressed in transfected CHO cells and purified by protein A affinity chromatography (POROS A; Life Technologies).

Solid phase binding and blocking assays

For c-Kit binding ELISA, recombinant human c-Kit protein (1 μg/mL) from R&D Systems was immobilized onto 96-well plate (100 μL/well) for 1 h at 4 °C. The plate was blocked with 2% milk in PBS/0.1%Tween-20 (PBS/T) for 1 h at room temperature. CK6 antibody at 297 nM starting concentration was titrated down 1:3, added to the plate (100 μL/well) and incubated for 1 h at room temperature. All plate washes were done with PBS/T. Goat anti-human IgG antibody HRP conjugate (Jackson ImmunoResearch Laboratories; 109-035-088) was added to each well at 1:6000 dilution in 2% milk/PBS/T and incubated for 1 h at room temperature. After washing the plate, color was developed by adding 100 μL TMB peroxidase substrate from KPL and stopped with 100 μL of 2 N H₂SO₄ stop solution (R&D Systems). Absorbance was measured at 450 nm using microplate reader (Molecular Devices) and data analysis performed using SoftMax Pro software.

For c-Kit blocking ELISA, recombinant human SCF (1 μg/mL) from PeproTech was coated onto 96-well plate (100 μL/well) for up to 2 h at 4 °C. Plate washes were done with PBS/T and then blocked with 2% milk/PBS/T for 1 h at 4 °C. Recombinant human c-Kit protein (50 μL of 1 μg/mL) was mixed with 100 μL CK6 antibody, at starting concentration of 660 nM and titrated down 1:3, and incubated for up to 2 h at room temperature. Antibody plus c-Kit protein mixture was added to the plate and incubated for 1 h at room temperature. Plates were washed with PBS/T. Goat anti-human c-Kit antibody AF332 (R&D Systems) was added to each well (100 μL of 1 μg/mL) and incubated at room temperature for 1 h. Plates were washed with PBS/T and secondary donkey anti-goat IgG antibody HRP conjugate (Jackson ImmunoResearch Laboratories; 705-035-003) added to each well at 1:8000 dilution and incubated for 1 h at room temperature. After washing the plate, 100 μL TMB peroxidase substrate from KPL was used to develop the color and stopped by adding 100 µL of 2 N H₂SO₄ stop solution. Absorbance was measured at 450 nm using Molecular Devices microplate reader and data analysis performed using SoftMax Pro software.

Biacore

The binding kinetics of CK6 antibody to c-Kit receptor protein was measured by surface plasmon resonance on a Biacore T200 instrument using Biacore reagents and software (GE Healthcare BioSciences). Recombinant extracellular human c-Kit protein at concentration of 2 µg/mL in 10 mM sodium acetate (pH 5.0) was immobilized at a density about 40 resonance units (RU) onto a CM5 sensor chip. CK6 antibody was injected at various concentrations and measurements obtained at 25 °C. Sensorgrams of concentrations ranging from 0.15 to 5.0 nM were evaluated using Biacore T200 software. Calculation of association (Ka) and dissociation (Kd) rate constants was based on a 1:1 Langmuir binding model fit. The equilibrium dissociation constant (K_D) or binding affinity constant was calculated from the ratio of kinetic rate constants Kd/Ka. The average K_D value from multiple independent runs was obtained.

Flow cytometry

To determine the extent of CK6 antibody binding to c-Kit receptor expressed on the cell surface, cells were rinsed in PBS and adherent cells detached using cell dissociation buffer (Life Technologies). Staining was performed with up to 2 × 10⁵ cells and 15 μg/mL CK6 antibody or human IgG as control (Jackson ImmunoResearch Laboratories; 009-000-003) in 5% FBS/PBS for 1 h at 4 °C. Cells were rinsed in PBS and goat anti-human IgG antibody Alexa-488 or PE conjugate (Jackson ImmunoResearch Laboratories; 109-546-088, 109-116-088) added for 1 h at 4 °C. After washing, signal was captured using an Accuri C6 HTFC system from BD Biosciences. Flow cytometry data analysis was performed with FlowJo software (Tree Star).

Western blot analysis for SCF induced signaling

Cells were serum-starved overnight at a density of 1 × 10⁶cells/ml and the following day either treated in presence or absence of CK6 or control human IgG at varying doses for 2 h at 37 °C. Human SCF (1 μg/mL) stimulation was performed for 15 min at 37 °C. Cells were rinsed in PBS and lysates prepared in 1× cell lysis buffer (Cell Signaling Technology) containing Halt Protease and Phosphatase Inhibitor cocktails (Thermo Fisher Scientific). Lysate analysis was performed on Novex Bis-Tris gels and protein transferred onto nitrocellulose membrane with iBlot system (Life Technologies). Blots were incubated with primary antibody dilutions: 1:4000 phospho-cKit antibody (Life Technologies; 44-496G), 1:1000 phospho-p44/p42 MAPK and 1:1000 phospho-AKT antibodies (Cell Signaling Technology; 9101, 9271, 9275), and 1:5000 glyceraldehyde-3-phosphate dehydrogenase (GAPDH) antibody (Life Technologies; AM4300). Secondary antibody-HRP conjugates were from GE Healthcare BioSciences. Enhanced chemiluminescence (ECL) reagents (GE Healthcare BioSciences) were used to detect protein band signal. All blot images were captured with an ImageQuant LAS 4000 digital imaging system from GE Healthcare BioSciences.

Proliferation assay

MO7e cells were plated onto 96-well plate (5 × 10³ cells/well) and grown in 2% FBS containing RPMI media without GM-CSF overnight. CK6 antibody or control human IgG were added (330 nM starting concentration diluted 1:2 down) and incubated for 2 h at 37 °C. Cells were then stimulated in the presence or absence of human SCF (100 ng/mL) for 72 h. CellTiter-Glo Assay (Promega) was used to measure cell viability as luminescence signal readout according to manufacturer’s instructions. Multiple independent experiments were run.

Soft agar growth assay

For cell growth testing in three-dimensional agarose assay, NCI-H69 cells were grown to subconfluency in complete media. CK6 antibody or control human IgG were added to the cells and incubated for 1 h at 37 °C before adding 0.4% agarose (Sigma-Aldrich). This top agarose mixture was overlaid onto 12-well plate surface coated with bottom agarose (0.5% agarose containing either CK6 antibody or control human IgG) and incubated at 37 °C until colonies appeared (up to 29 d). Cells were fed media and antibody once every week. Colonies were visualized by staining cells with MTT agent (Sigma-Aldrich) and counted using GelCount technology from Oxford Optronix. The percentage growth change relative to control human IgG was calculated as an average from three separate experiments.

Tumor xenograft studies in vivo

Tumor cells (5 × 10⁶ NCI-H526 or 1 × 10⁷ Malme-3M) were mixed with 50% Matrigel and injected subcutaneously into left flank of female athymic nu/nu mice obtained from Charles River Laboratories. When tumors reached approximately 250–300 mm³ in size, mice were randomized into groups of 12–15 animals each and treatment begun by IP injection with 0.9% USP Saline, CK6 antibody, and/or chemotherapy. Cisplatin, etoposide, and dacarbazine were purchased from Sigma-Aldrich. CK6 was dosed at 40 mg/kg 3 times weekly, cisplatin 5 mg/kg/etoposide 40 mg/kg every 7 d for 3 wk, dacarbazine 80 mg/kg on day 2, 3, and 4. Combinations of CK6 and chemotherapy followed the above dosing and schedule. Tumor size was measured twice weekly with a caliper and tumor volumes calculated using the formula Volume (mm³) = (π/6) l × w². Tumor growth was calculated as T/C% (ratio of relative mean tumor volume of treated group/relative mean tumor volume of control group). Statistical analysis of tumor growth data was performed using RM ANOVA (SigmaPlot; Systat Software). All experiments and procedures were approved by an Internal Animal Care and Use Committee and performed in accordance with the United States Department of Agriculture and the National Institute of Health policies regarding the humane care and use of laboratory animals.

Immunohistochemistry

Formalin-fixed paraffin-embedded (FFPE) human tumor tissue microarrays (TMAs) were from US Biomax. Arrayed samples included SCLC, NSCLC and non-neoplastic lung. Polyclonal rabbit anti-human CD117 antibody A4502 from Dako was used to determine the level of c-Kit protein. TMA slides were baked at 60 °C, deparaffinized and rehydrated. Antigen retrieval was performed using Dako Target Retrieval Solution, pH9 (S2367) and processed for staining using Dako Autostainer platform. For visualization, Dako AEC Substrate Chromogen (K3464) was used. Normal human skin containing mast cells and melanocytes served as the positive control for the assay. c-Kit immunoreactivity analysis was performed by the pathologist with immunolabeling of the cores scored as weak (1+), moderate (2+), and strong (3+). The percentage of c-Kit positive samples on the TMAs was calculated as ratio of positive sample number / total sample number. Images were captured using Aperio ScanScope XT.

Glossary

Abbreviations:

c-Kit

stem cell factor receptor

SCF

stem cell factor

SCLC

small cell lung carcinoma

MAPK

mitogen-activated protein kinase

AKT

protein kinase B

AML

acute myeloid leukemia

GIST

gastrointestinal stromal tumor

MFI

mean fluorescence intensity

FFPE

formalin-fixed paraffin-embedded

TMAs

tumor tissue microarrays

NSCLC

non-small cell lung carcinoma

GAPDH

glyceraldehyde-3-phosphate dehydrogenase