CNTO 95, a fully human monoclonal antibody that inhibits αv integrins, has antitumor and antiangiogenic activity in vivo

Authors


Abstract

Integrins of the αv family, such as αvβ3 and αvβ5, are implicated in tumor-induced angiogenesis; but their role in tumor growth has not been fully explored. CNTO 95 is a fully human antibody that recognizes the αv family of integrins and is likely to be less immunogenic in humans compared to chimeric or humanized antibodies. CNTO 95 bound to purified αvβ3 and αvβ5 with a Kd of approximately 200 pM and to αv integrin–expressing human cells with a Kd of 1–24 nM. In vitro, CNTO 95 inhibited human melanoma cell adhesion, migration and invasion at doses ranging 7–20 nM. In a rat aortic ring sprouting assay, CNTO 95 (approx. 70 nM) completely inhibited sprouting. Using a human melanoma xenograft model in nude mice wherein CNTO 95 recognized αvβ3 and αvβ5 on human tumor cells but not mouse angiogenic integrins, CNTO 95 (10 mg/kg, 3 times/week) inhibited growth of human melanoma tumors in nude mice by approximately 80% (p = 0.0005), suggesting that CNTO 95 inhibited human tumor growth independently of its antiangiogenic activity. In a nude rat human xenograft model where CNTO 95 binds and blocks both tumor and host integrins, this antibody (10 mg/kg once/week) reduced final tumor weight by >99% (p < 0.0001). Based on these preclinical data, a dose-escalating phase I clinical trial in cancer patients has been initiated. To our knowledge, CNTO 95 is the first fully human MAb to αv integrins that has potent antitumor and antiangiogenic properties in in vivo preclinical models. © 2004 Wiley-Liss, Inc.

Angiogenesis, the formation of new blood vessels from preexisting capillaries, is fundamental to reproduction and development. This process is typically short in duration and tightly regulated through a balance between angiogenesis stimulators and inhibitors. However, when angiogenesis becomes deregulated, it can play a critical role in a number of pathologic processes, including cancer, cardiovascular diseases, chronic inflammation, diabetes, psoriasis, endometriosis and adiposity.1, 2

The progressive growth and metastasis of tumor cells is dependent on the ability of tumor cells to stimulate the formation of new blood vessels, to supply tumors with oxygen and essential nutrients. Without neovascularization, tumors rarely grow beyond 2–3 mm,3 and most persist in a relatively benign or dormant state and are often clinically undetectable.3, 4 At some point during this early growth phase, the tumor undergoes an “angiogenic switch”, where new microvessels are recruited from the surrounding vasculature. Tumor-associated neovascularization allows tumor cells to express their critical growth advantage and facilitates metastasis by establishing continuity with systemic circulation.5, 6 Results of clinical studies have supported the importance of angiogenesis in tumor growth and metastasis, and investigators have suggested that the degree of angiogenesis is a prognostic indicator for various tumor types, including primary breast carcinoma,7 laryngeal cancer8 and prostate cancer.9 Positive correlations between the degree of vascularity within the primary tumor and incidence of distant metastasis have also been reported in both melanoma10 and breast carcinoma.11 Identification of potential mediators of angiogenesis is an active area of research, and angiogenesis inhibitors have shown some potential in the treatment of cancer in both preclinical12, 13, 14 and clinical15 studies.

Certain integrin receptors are essential for angiogenesis. Integrins are heterodimeric transmembrane receptor proteins that consist of a diverse family of over 15 α and 8 β subunits that can heterodimerize in over 20 different combinations. Integrins facilitate the adhesion of stimulated endothelial cells to the ECM, trigger the secretion of ECM-rearranging proteases and propagate signaling events that promote the survival and differentiation of cells in the newly formed vasculature.16 One of the best-characterized integrins implicated in tumor-induced angiogenesis is αvβ3, which is highly expressed in human breast tumors but not readily detected in nonangiogenic benign breast tissue. The degree of integrin αvβ3 expression has been established as a prognostic indicator in breast carcinomas, and antibodies directed against αvβ3 disrupt intratumoral neovessels and inhibit neovascularization by inducing apoptotic signals.17 A humanized MAb to αvβ3, Vitaxin, is currently in clinical trials. One potential limitation of this MAb is that it blocks only αvβ3-mediated angiogenesis and does not block the angiogenic pathways of other integrin receptors. Other αv integrins, such as αvβ5, can also mediate tumor angiogenesis. There is evidence that αvβ3 and αvβ5 promote angiogenesis via distinct pathways, αvβ3 through bFGF and TNF-α and αvβ5 through VEGF and TGF-α.18 Both preclinical and clinical data suggest that αvβ3 and αvβ5 are logical targets for antitumor therapies and may provide therapeutic benefit in the treatment of solid tumors. CNTO 95 is a fully human MAb that recognizes the αv family of integrins, blocks both αvβ3 and αvβ5 and inhibits integrin-mediated tumor growth and angiogenesis in vitro and in vivo.

Abbreviations

AM, acetoxy methyl ester; bFGF, basic fibroblast growth factor; CNBr, cyanogen bromide; CPM, counts per minute; EBM, endothelial basal medium; ECM, extracellular matrix; HUVEC, human umbilical vein endothelial cell; MAb, monoclonal antibody; OTG, octyl-beta-D-glucopyranoside; PECAM, platelet endothelial cell adhesion molecule; TBS, TRIS-buffered saline; TGF-α, transforming growth factor-α; TNF-α, tumor necrosis factor-α; VEGF, vascular endothelial growth factor.

MATERIAL AND METHODS

Reagents

Bovine bFGF and human VEGF165 were obtained from R&D Systems (Minneapolis, MN). BIOCOAT cell culture inserts (pore size 8 μm) were purchased from Becton Dickinson (Bedford, MA). The calcein AM and Vybrant cell adhesion assay kit (V-13181) was purchased from Molecular Probes (Eugene, OR). Human plasminogen-free fibrinogen (VWF/Fn depleted) was purchased from Enzyme Research Labs (South Bend, IN). Bovine skin gelatin was purchased from Sigma (St. Louis, MO). Human vitronectin was purchased from Promega (Madison, WI), and type I collagen was from GIBCO BRL (Gaithersburg, MD). c7E3 IgG, which recognizes GPIIb/IIIa and αvβ3, and 10E5 (anti-GPIIb/IIIa MAb) were generated at Centocor.19 Anti-αvβ3 (LM609) and anti-αvβ5 (P1F6) were purchased from Chemicon (Temecula, CA). Anti-integrin αv, MAb 1978 IgG1, was purchased as ascites fluid (Chemicon). MAb 1978 IgG1 was purified with ImmunoPure Immobilized Protein A/G (catalogue number 20422; Pierce, Rockford, IL) following the manufacturer's instruction. Anti-integrin αv (Q-20, goat polyclonal IgG), anti-integrin β1 (4B7R, mouse MAb IgG1), anti-integrin β1 (N-20, goat polyclonal IgG), anti-integrin β3 (H-96, rabbit polyclonal IgG), anti-integrin β5 (H-96, rabbit polyclonal IgG) and anti-integrin β6 (H-110, rabbit polyclonal IgG) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA).

Cell lines

HUVECs were purchased from Clonetics (Walkersville, MD) and cultured in complete EBM (Clonetics) containing 10% FBS, long R IGF-1, ascorbic acid, hydrocortisone, human epidermal growth factor, human VEGF, gentamicin sulfate and amphotericin-B. Cells were grown at 37°C and 5% CO2, and medium was changed every 2–3 days. Cells were passaged when they reached 80% confluence. Passages 3–8 were used in all experiments. The A375.S2 human melanoma cell line and the monkey breast carcinoma cell line were obtained from the ATCC (Rockville, MD) and deemed free of mycoplasma and bacterial contaminants. Rabbit and rat endothelial cells were cultured from freshly excised aortas. Cells were cultured in DMEM supplemented with 10% FBS, 2 mM L-glutamine, 1 mM sodium pyruvate and 0.1 mM nonessential amino acids. Human melanoma M21 cells were obtained from Centocor Cell Biology Services, where this cell line was deemed free of mycoplasma and bacterial contamination. Tumor cells were cultured in RPMI medium supplemented with 10% FBS, 2 mM L-glutamine, 1 mM sodium pyruvate and 0.1 mM nonessential amino acids.20

Antibody generation

CNTO 95 was derived by immunizing mice that were transgenic for part of the human immunoglobulin repertoire (Medarex, Princeton, NJ). Applying conventional cell fusion technology, a hybridoma cell line was produced that secreted human IgG1 reactive with αvβ3 and αvβ5. cDNA molecules encoding the heavy chain and light chain variable regions were cloned from the hybridoma cell line. Cloned variable region genes were then inserted into plasmid expression vectors that provided coding sequences for human IgG1 and the resulting plasmids introduced into mouse myeloma cells by electroporation. Transfected cells secreting human IgG1 were designated CNTO 95.

Antibody specificity

The integrin specificity of CNTO 95 was determined by immunoaffinity chromatography. CNTO 95 was coupled to Sepharose and used to capture integrins from extracts of human placental tissue. Captured integrins were subsequently identified by Western blot.

Conjugation of CNTO 95 and MAb 1978 to CNBr-Sepharose.

CNTO 95 and MAb 1978 were dialyzed against coupling buffer (0.1 M sodium bicarbonate, pH 8.3) at 4°C overnight. CNBr-Sepharose was allowed to swell for 30 min in 1 mM HCl and washed with 1 mM HCl 8 times, followed by washing with coupling buffer 3 times. Antibodies were incubated with CNBr-Sepharose beads for 3.5 hr. Resin was blocked with 1 M ethanolamine (pH 9.0, Sigma) for 2 hr at room temperature. Blocked resin was then washed alternately with 50 mM TRIS, 1 M NaCl (pH 8.0) and 50 mM glycine, 1 M NaCl (pH 3.5). This cycle was repeated 6 times.

Protein extraction from human placenta.

Human placenta (approx. 300 g) was obtained with informed consent from an anonymous donor. Tissue was washed with ice-cold TBS to remove blood, and 600 ml of extraction buffer [TBS (pH 7.5), 1 mM CaCl2, 1 mM MnCl2, 100 mM OTG, plus EDTA-free protease inhibitor; Roche Applied Sciences] were added. Tissue was cut into small pieces with scissors, followed by homogenization with a blender. After homogenization, 17.52 g OTG were added to the homogenate, which was then agitated on a rotator at 4°C overnight. The tissue extract was centrifuged on a Sorvall (Wilmington, DE) RC5CPLUS centrifuge (rotor SLA 3000, at 16,900g) for 1 hr at 4°C. Supernatant was collected and stored at 4°C.

Affinity chromatography.

Placental extracts were incubated with CNTO 95–or MAb 1978-conjugated resin overnight at 4°C. Resin was then loaded to empty columns. Columns were washed with 10 ml of column wash buffer I [TRIS-HCl (pH 7.5), 1 mM CaCl2, 1 mM MnCl2, 0.1% NP-40], followed by 10 ml of column wash buffer II [10 mM Na acetate (pH 4.5), 100 mM NaCl, 1 mM CaCl2, 1 mM MnCl2, 0.1% OTG]. Integrin fractions were eluted with 10 ml of elution buffer [0.1 M glycine, 2% acetic acid (pH 3.1), 1 mM CaCl2, 1 mM MnCl2, 0.1% OTG]. Eluted material was concentrated and stored at 4°C for further analysis.

Western blot.

Proteins in the integrin fraction were separated by electrophoresis on 4–12% SDS polyacrylamide gels and transferred to nitrocellulose filters. Filters were blocked with 5% nonfat dry milk in TBS containing 0.05% Tween-20 (wash buffer) at room temperature for 1 hr and then incubated with anti-integrin antibodies. After thorough washing, filters were incubated with appropriate peroxidase-conjugated secondary antibodies (1:20,000 dilution). Antigen–antibody complexes were visualized using the SuperSignal West Pico Chemiluminescent Substrate kit (Pierce, Rockford, IL).21

Antibody affinity

The binding affinity of CNTO 95 was determined using purified αvβ3, αvβ5, GPIIb/IIIa and α5β1 proteins as well as cells that express these receptors. The binding affinity of CNTO 95 to purified integrin receptors was assessed as described.19

Antibody (125I-labeled CNTO 95) binding to cells was measured with cells in suspension as described.19 Bound 125I antibody was separated from unbound 125I antibody by washing 3 times with excess cold medium and centrifugation. Nonspecific binding was determined by incubating cells with unlabeled antibody (200-fold molar excess over 125I-labeled antibody) at each concentration. Specific CPM was calculated by subtracting nonspecific binding CPM from total bound CPM. The receptor number per cell was calculated as follows:

equation image

Antibody binding to purified integrins was also determined by ELISA, as described.19

Cell adhesion, migration and proliferation

Adhesion assay.

Microtiter plates (Linbro-Titertek; ICN Biomedicals, Costa Mesa, CA) were coated at 4°C overnight with vitronectin (1 μg/ml), gelatin (0.1%), fibrinogen (100 μg/ml), type I collagen (10 μg/ml) or fibronectin (10 μg/ml). Fibrin-coated microtiter wells were formed by thrombin treatment (1 U/ml) of fibrinogen. A calcein-based assay was used to quantify cell adhesion, as previously described.20 Adhesion was measured in a fluorescence plate reader (Fluoroskan; Tecan, Research Triangle Park, NC) at 485–538 nm. Cell adhesion to BSA-coated wells served as a negative control. Isotype-matched antibodies served as a negative control.

Cell migration assay.

Cell migration assays were performed in 24-transwell chambers with a polystyrene membrane (6.5 mm diameter, 10 μm thickness and 8 μm pore size) as described previously.20 Extent of cell migration was determined by light microscopy, and images were analyzed using the Phase 3 image analysis software (Phase 3 Image Systems, Glen Mills, PA); data were analyzed as previously described.20 The software analyzes the total area occupied by stained cells on the bottom side of the filter, and this is directly proportional to the extent of cell migration.

Cell proliferation assay.

Subconfluent HUVECs or A375.S2 melanoma cells were trypsinized, washed and resuspended in complete medium. HUVECs (5,000) or A375.S2 cells (8,000) were added to each well of 96-well plates precoated with vitronectin (1 μg/ml). Cells were allowed to attach for 2 hr, medium was aspirated, wells were washed once with PBS and 100 μl of medium (serum-free M199 or 2% serum M199) containing 20 ng/ml bovine bFGF-2 and various concentrations of CNTO 95 were added to each well of HUVECs. This level of bFGF was found to be saturating in dose-finding studies (not shown). Medium (DMEM + 2% FBS) containing various concentrations of CNTO 95 was added to A375.S2 cells. Plates were incubated at 37°C for 48 hr (HUVECs) or 72 hr (A375.S2). Extent of cell proliferation was determined by the ATP kit (Packard, Meriden, CT). Luminescence intensity was measured for the ATP assay in a TopCount reader (Packard).

Microvessel sprouting

Rat aortic ring assay.

The rat aortic ring assay was performed as described by Nicosia and Ottinetti22 with slight modifications.23 Agarose wells (1.5%, 10 mm diameter) prepared in 100 × 15 mm tissue culture dishes were filled with M199 medium containing rat tail type 1 collagen (2.8 mg/ml, Becton Dickinson), NaHCO3 (28 mM) and either CNTO 95 (10 μg/ml), nonspecific control murine IgG (20 μg/ml) or BSA (20 μg/ml). Rat aortic ring sections (1 mm) were placed on top of collagen gels within the agarose rings; wells were filled with collagen solution and incubated at 37°C. After the collagen had gelled, collagen–aortic ring sandwiches were transferred to 12-well plates containing 1 ml EBM-2, BSA (0.1%), bFGF (10 ng/ml), penicillin (100 U/ml), streptomycin (100 μg/ml), amphotericin B (0.25 μg/ml) (all from Clonetics) and either CNTO 95 (increasing concentrations), nonspecific control mIgG (20 μg/ml) or BSA (20 μg/ml). Plates were maintained in a 37°C tissue culture incubator with media changes every other day. After 10 days, microvessel sprouts originating from each aortic ring were counted microscopically using the Phase 3 Image Analysis System.

Rat and monkey Matrigel assays

We determined whether CNTO 95 could functionally block αvβ3 and αvβ5 integrins in a nonhuman primate model, a nude rat model of growth factor–induced angiogenesis and a nude rat model of tumor cell–induced angiogenesis. The monkey study was performed at Charles River Laboratories (Worcester, MA) on young adult female cynomulgus monkeys (species Mucaca fascicularis). Nude female rats (5–7 weeks old) were obtained from Harlan (Indianapolis, IN). Recombinant human bFGF was obtained from R&D Systems. Matrigel, prepared from the Engelbreth-Holm-Swarm tumor, was obtained from Becton Dickinson.

Liquid Matrigel was maintained at 4°C. Angiogenesis assays were performed as described.20 For growth factor–induced angiogenesis, human bFGF (5 μg/ml) was added to the Matrigel solution and allowed to mix thoroughly overnight. Matrigel was then mixed with antibodies or control solutions and kept on ice. For tumor cell–induced angiogenesis, human melanoma M21 cells were harvested with trypsin-EDTA, centrifuged, washed twice with DMEM and resuspended in ice-cold DMEM. M21 cells were gently added to the Matrigel solution at a final concentration of 0.5 × 106/ml. The tumor cell–Matrigel solution was gently mixed and stored on ice until it was injected into nude rats.

Monkeys were injected at each site s.c. with 2 ml of Matrigel solution, while rats were injected with 1 ml of Matrigel each. In the tumor cell–induced study, rats were injected at 2 sites with 1 ml of the ice-cold tumor cell–Matrigel solution. Gel formation was confirmed after injection. Animals received test article via i.v. or i.p. bolus injection. At the end of each study, animals were killed and Matrigel was harvested from the injection sites. Matrigel implants were weighed, photographed and graded for angiogenesis using the Phase 3 Image Analysis System. To measure the total area of neovessels, photomicrographs were taken from both the top surface and the bottom surface of each Matrigel plug at ×2 magnification on an inverted phase contrast microscope. Vessel length and number of vessels per field were calculated using the tracing function within the Phase 3 Image System. The mean value from all fields was calculated for each Matrigel plug, and the mean vessel number and vessel length for each test group were calculated.

As a third method to measure angiogenesis, immunohistochemistry was performed on 10 μm serial cryostat sections cut from frozen Matrigel plugs. Sections were immediately fixed in cold acetone (5 min) and air-dried. Sections were washed 3 times in PBS to remove frozen mounting media, blocked for 1 hr with 5% mouse serum and 5% goat serum in PBS and rinsed in PBS. Sections were then blocked with avidin–biotin solution (X0590; Dako, Carpinteria, CA) for 10 min. After washing, endogenous peroxidase was quenched by incubation in 3% hydrogen peroxide for 10 min. Then, sections were incubated for 60 min with primary antibody (mouse antihuman PECAM, 10 μg/ml; Pharmingen, Bedford, MA) diluted with antibody diluent solution (S3022, Dako). Immunoreactive sites were detected using a Dako kit and sections counterstained with hematoxylin. Irrelevant mouse IgG1 was used as a negative control in all cases. Photomicrographs were taken from all slides at ×20 magnification; each entire section was photographed. Vessel density per field was calculated using the Phase 3 Image System software. Vessel density was quantitated by measuring the percentage of cross-sectional area of each Matrigel section occupied by stained microvessels. The mean value for each slide was calculated and the mean vessel density for each group determined.

Tumor growth studies

For mouse studies, female nude mice, aged 4–5 weeks, were purchased from Charles River Laboratories (Wilmington, MA) and maintained according to NIH standards. Twenty mice were inoculated s.c. with A375.S2 cells (3 × 106) in the flank region (day 0). On day 3, mice were randomly divided into 2 groups. One group was injected i.p. with CNTO 95 (10 mg/kg in PBS), while the other group received vehicle. Dosing was continued 3 times a week thereafter until day 26. Tumors were measured by calipers twice a week and tumor volumes calculated by the formula (length × width2)/2. Body weights were also recorded weekly.

For rat studies, female nude rats, aged 6–7 weeks, were purchased from Harlan. Twenty rats were inoculated s.c. with A375.S2 cells (3 × 106) in the flank region (day 1). On day 4, rats were randomly assigned to 2 groups. One group was injected i.v. with CNTO 95 (10 mg/kg in PBS), while the other group received an isotype-matched control IgG (10 mg/kg). Dosing was continued weekly thereafter until day 46 (total of 6 doses). Tumors were measured by calipers twice a week and tumor volumes calculated by the formula (length × width2)/2. Body weights were also recorded weekly.

Statistical analysis

Statistical comparison of group mean values (i.e., tumor volumes, vessel counts, cell numbers) was performed using Student's 2-tailed t-test.

RESULTS

CNTO 95 is a fully human IgG1 MAb generated by immunizing mice transgenic for part of the human immunoglobulin repertoire. To define the specificity of CNTO 95, human placental extract was loaded onto a CNTO 95 affinity column. Parallel affinity purification was done with a commercially available antibody to human αv (MAb 1978). Figure 1a shows that bound proteins eluted from both the CNTO 95 column and the anti-αv MAb 1978 column contained 2 major proteins, one with apparent m.w. of 130 kDa and the other of approximately 100 kDa. Integrins typically migrate at these weights under reducing conditions.

Figure 1.

Binding of CNTO 95 to multiple αv integrins. Extracts of human placenta were exposed to Sepharose resin conjugated to either CNTO 95 or MAb 1978 (anti-αv). (a) On a gel stained with simple blue, proteins eluted from the CNTO 95 column (lane 1) and the MAb 1978 column (lane 2) contained 2 major bands, one with apparent m.w. of 130 kDa and the other of approximately 100 kDa. (b) Bound proteins eluted from the CNTO 95 column were separated by reducing SDS-PAGE and analyzed by Western blot using antibodies specific to αv, β1, β3, β5 and β6 integrin subunits. (c) Western blot of eluted proteins using antibodies against integrins α5 and αIIb. Lanes 1 and 3 contain CNTO 95–purified proteins. Purified integrin α5β1 (lane 2) and crude placental extract (lane 4) served as positive controls. Lanes 1 and 2 were probed with anti-α5 antibody and lanes 3 and 4 with anti-αIIb antibody.

To identify which integrins were retained by the CNTO 95 affinity column, Western blot analysis was performed on the eluted protein fraction. Purified integrins were separated by SDS-PAGE under reducing conditions and blotted to nitrocellulose, and strips of nitrocellulose were probed with antibodies against specific integrin subunits. As shown Figure 1b, positive bands were detected for integrins αv, β1, β3, β5 and β6, suggesting that these integrin subunits are components of integrin complexes bound by CNTO 95. No bands were detected by antibodies against the α5 or αIIb subunits (Fig. 1c), suggesting that CNTO 95 does not recognize α5β1 or αIIbβ3 integrins, both of which are abundantly expressed in placenta. Since integrins are heterodimers of α and β subunits, these results suggest that CNTO 95 is a pan-αv antibody, recognizing αvβ1, αvβ3, αvβ5 and αvβ6 integrins.

The binding affinity of CNTO 95 was determined using purified αvβ3 and αvβ5 proteins and cells that express these receptors. Antibody binding to purified αvβ3 and αvβ5 was dose-dependent and saturable. Mean Kd values of CNTO 95 for αvβ3 and αvβ5 were 2.1 ± 1.33 × 10−10 M and 2.5 ± 1.04 × 10−10 M, respectively (Fig. 2a,b). Binding studies confirmed that CNTO 95 did not cross-react with GPIIbIIIa (αIIbβ3) or α5β1, even at doses as high as 400 μg/ml (Fig. 2c,d), whereas maximum binding of the positive control antibodies was achieved at concentrations of 1.0 μg/ml and lower. CNTO 95 bound with high affinity to human and nonhuman primate cells and with lower affinity to cells isolated from rabbits and rats (Table I).

Figure 2.

Binding of CNTO 95 to αvβ3, αvβ5, αIIbβ3 and α5β1 integrins. (a,b) Binding of CNTO 95 to αvβ3 and αvβ5 was detected using 125I-labeled antibody. Graphs indicate specific binding. (c) c7E3 IgG (anti-αIIbβ3/αvβ3) was used as a positive control for binding to GPIIb/IIIa. Human IgG was an isotype-matched negative control antibody. Bound antibodies were detected by ELISA. Note that no binding of CNTO 95 to GPIIb/IIIa was observed at 400 μg/ml. (d) CNTO 95 did not bind to α5β1. MAb 2253Z was a positive control antibody directed to β1 integrin that dose-dependently recognized α5β1. c7E3 IgG was used as an isotype-matched negative control for CNTO 95. Neither c7E3 IgG nor CNTO 95 bound to α5β1 integrin. Each data point is the mean of triplicate determinations (±SD).

Table I. Binding Affinity of CNTO 95
CellEC50 (μg/ml)1Kd (nM)CNTO 95 binding sites/cell
  • 1

    Mean effective binding concentration.

  • 2

    Rhesus monkey mammary carcinoma cells.

  • 3

    Freshly cultured aortic endothelial cells.

A375.S2 (human melanoma)0.85.2120,000
HT29 (human colon carcinoma)0.21.381,000
M21 (human melanoma)3.624240,000
CMTT22.819200,000
Cynomulgus30.85.144,000
Rabbit320.8140Not saturable
Rat332.8220Not saturable

Inhibition of cell adhesion, migration and proliferation

Cell adhesion and migration are important processes involved in tumor growth and angiogenesis; therefore, the ability of CNTO 95 to inhibit integrin-mediated adhesion and migration of HUVECs and human melanoma (A375.S2) cells was studied. CNTO 95 dose-dependently inhibited the adhesion of HUVECs and A375.S2 cells to vitronectin, fibrinogen, gelatin and fibrin, all ligands for αvβ3 and αvβ5, indicating that it blocked the function of these integrins (Fig. 3). CNTO 95 did not inhibit cell adhesion to type I collagen or fibronectin, indicating that it did not block β1 integrin–mediated cell adhesion. To determine the relative contribution of different αv integrins to cell adhesion, labeled HUVECs and A375.S2 cells were allowed to adhere to vitronectin in the absence or presence of integrin-specific antibodies. As seen in Figure 4, blockage of αvβ3 by LM609 or of αvβ5 by P1F6 at saturating concentrations20 partially inhibited adhesion of both HUVECs and melanoma cells to vitronectin. In contrast, blockage of cell adhesion by CNTO 95 alone was almost complete, similar to that seen with the combination of LM609 and P1F6. A similar pattern of inhibition was observed in endothelial cell migration assays with CNTO 95, LM609 and P1F6 (Fig. 5), suggesting that αvβ3 and αvβ5 are the predominant integrins at work in this system.

Figure 3.

Adhesion of HUVECs and A375.S2 melanoma cells to matrix proteins. Adhesion assay was performed as described in Material and Methods. Cell adhesion to BSA-coated wells served as a negative control. Extent of cell adhesion in the presence of various concentrations of antibody was plotted as a percent of cell adhesion in the absence of antibody that was considered as 100%. Each data point is the mean of triplicate determinations (±SD) and representative of at least 3 experiments.

Figure 4.

Adhesion of HUVECs and A375.S2 cells to vitronectin. Extent of cell adhesion in the presence of various anti-integrin MAbs was plotted as a percent of cell adhesion in the absence of antibody (100%). Antibodies were used at 10 μg/ml; at this concentration, LM609 and P1F6 completely blocked αvβ3 and αvβ5, respectively. Each data point is the mean of triplicate determinations (±SD).

Figure 5.

Migration of HUVECs toward vitronectin (Vn) in the presence of bFGF. The undersides of migration chamber filters were coated with vitronectin, and the assay was performed as described in Material and Methods. Cells were allowed to migrate for 6 hr. Each data point is the mean of 3 transwell filters (±SD).

Integrins αvβ3 and αvβ5 have also been identified as factors involved in cell proliferation. Dose-dependent inhibition of cell proliferation was observed with CNTO 95 when endothelial cells were stimulated with bFGF in the absence or presence of 2% FBS on vitronectin-coated plates (Fig. 6a). No effect of CNTO 95 was observed on quiescent (unstimulated) endothelial cells, suggesting that blockade of αvβ3 and αvβ5 is important only for stimulated endothelial cells but not for quiescent endothelial cells. In addition, CNTO 95 reduced the proliferation of human A375.S2 melanoma cells (Fig. 6b), suggesting that integrin blockage directly affects both endothelial and tumor cells.

Figure 6.

Inhibition of endothelial and tumor cell proliferation by CNTO 95. Cell proliferation assays were performed as described in Material and Methods. CNTO 95 inhibited bFGF and bFGF + 2% serum-stimulated proliferation of human endothelial cells (a), as well as proliferation of the human melanoma cell line A375.S2 (b). Data are plotted as luminescence intensity, which is directly proportional to cell number. Each point represents the mean of triplicate determinations (±SD).

Inhibition of angiogenesis in vitro

A microvessel sprouting assay was used to test the in vitro antiangiogenic activity of CNTO 95. In this assay, freshly excised rat aortas were cultured in 3-D fibrin or collagen gels. In the presence of various angiogenic factors, such as bFGF, the aorta sprouts microcapillaries after a few days. This process is a representation of angiogenesis as it involves endothelial cell adhesion, migration, invasion and proliferation. As shown in Figure 7, CNTO 95 inhibited sprouting of microvessels from aortas excised from rats in a dose-dependent manner These results demonstrate that CNTO 95 is an inhibitor of angiogenesis in vitro.

Figure 7.

Inhibition of microvessel sprouting from rat aorta explants by CNTO 95. (a) Representative digital microscopic images of rat aortic rings treated with nonspecific control human IgG (20 μg/ml) or CNTO 95 (20 μg/ml). (b) Effect of CNTO 95 on vascular sprouting number from rat aortic rings. BSA (20 μg/ml) and irrelevant human IgG (20 μg/ml) were used as negative controls. Data points represent one rat aorta, with mean values for each group indicated by lines. p values were determined by comparing antibody-treated groups to the BSA-treated group using a 2-tailed unpaired t-test

Inhibition of angiogenesis in nude rats and cynomulgus monkeys

Inclusion of human bFGF in Matrigel implants in rats resulted in increased angiogenesis as measured by vessel length and number (Fig. 8). Systemic treatment of rats with CNTO 95 significantly inhibited bFGF-stimulated increases in vessel length and total vessel number within Matrigel implants, as measured by visual inspection (p < 0.001). Inhibition of angiogenesis by CNTO 95 was dose-dependent, with a dose of 1 mg/kg being active in this model (Fig. 8c,d).

Figure 8.

Inhibition of bFGF-induced angiogenesis by CNTO 95. Appearance of Matrigel under microscopy (×2). (a) bFGF (5 μg/ml) + C57 (10 mg/kg, i.v.). (b) bFGF (5 μg/ml) + CNTO 95 (10 mg/kg, i.v.). Number (c) and length (d) of blood vessels in Matrigel from nude rats on day 7. Vessel number and length were assessed by microscopy (×2) aided by image analysis software (Phase 3 Image System). Each point represents average view from one Matrigel sample (2 plugs/animal); line represents group mean. A 2-tailed unpaired t-test indicated p < 0.0001 for all 4 CNTO 95 dose groups compared to the control IgG group.

Injection of Matrigel s.c. in cynomulgus monkeys provides a nonhuman primate model of angiogenesis. Inclusion of human bFGF in Matrigel implants resulted in increased angiogenesis, as measured by vessel length, number and vessel density. Systemic treatment of monkeys with CNTO 95 significantly inhibited bFGF-stimulated increases in vessel length and total vessel number within Matrigel implants, as measured by image analysis quantification (Fig. 9). Systemic treatment of monkeys with CNTO 95 also reduced the bFGF-stimulated increase in microvessel density within Matrigel implants, as measured by immunostaining for CD31 expression and image analysis (Fig. 10).

Figure 9.

Inhibition of bFGF-induced angiogenesis. Length (a) and number (b) of blood vessels in monkey Matrigel plugs on day 7. Each point represents average per field from one Matrigel sample (4 Matrigel plugs/animal). Horizontal lines indicate mean. Matrigel alone; bFGF-PBS, 5 μg/ml bFGF; bFGF-control IgG, 5 μg/ml bFGF, 10 mg/kg control IgG i.v.; bFGF-CNTO 95, 5 μg/ml bFGF, 10 mg/kg CNTO 95 i.v. A 2-tailed unpaired t-test indicated p < 0.001 for Matrigel alone and bFGF-CNTO 95 groups compared to the bFGF-PBS and bFGF-control IgG groups. There was no difference between the bFGF-CNTO 95 and Matrigel alone groups (p > 0.05) [n = 2 animals (8 implants) per group except in Matrigel alone (n = 1 animal)].

Figure 10.

Inhibition of bFGF-induced angiogenesis by CNTO 95. Microvessels within Matrigel implants in monkeys on day 7 were visualized by immunostaining with anti-PECAM antibody, followed by Dako alkaline phosphatase kit (red color). (a) Matrigel alone. (b) bFGF-PBS. (c) bFGF control IgG. (d) bFGF-CNTO 95. Photomicrographs taken at ×20 magnification; arrows indicate stained vessels. (e) Vessel density within Matrigel implants in monkeys. Percentage of Matrigel cross-sectional area occupied by vessels was calculated using computer-assisted image analysis. Each point represents the average density per Matrigel sample (4 samples/animal). A 2-tailed unpaired t-test indicated p < 0.01 for Matrigel alone and bFGF-CNTO 95 groups compared to the bFGF-HBSS and bFGF-control IgG groups. There was no difference between the bFGF-CNTO 95 and Matrigel alone groups (p > 0.05).

In the tumor cell–induced angiogenesis model, human melanoma M21 cells were mixed with Matrigel and injected s.c. into nude rats, as described in Material and Methods. A single dose of 10 mg/kg CNTO 95 inhibited tumor cell–induced angiogenesis. No difference in inhibitory activity was observed when CNTO 95 was administered as a single i.v. dose or when it was mixed with the Matrigel–tumor cell suspension prior to injection into rats. Inhibition of angiogenesis by CNTO 95 was demonstrated by decreased length and number of microvessels (Fig. 11) and decreased hemoglobin content in Matrigel plugs (data not shown).

Figure 11.

Inhibition of angiogenesis by CNTO 95 in a tumor cell–induced angiogenesis model in nude rats. Mean length (a) and number (b) of microvessels. Each point represents average per view from one Matrigel plug, and the line is the mean from 10 plugs (2 plugs/per animal). Inside refers to CNTO 95 (40 μg/ml) that was included in the Matrigel solution prior to injection. IV refers to CNTO 95 (10 mg/kg) injected i.v. as a bolus after Matrigel solution was injected into rats. Inside+IV refers to CNTO 95 mixed with Matrigel solution and injected i.v. A 2-tailed unpaired t-test indicated p < 0.001 for all 3 CNTO 95 groups compared to the control IgG group. The 3 CNTO 95 groups were not statistically different from each other.

Inhibition of tumor growth in mice

To determine the antitumor efficacy of CNTO 95 in vivo, a human A375.S2 melanoma xenograft tumor model was established in nude mice. Mice were treated with CNTO 95 (10 mg/kg) 3 times per week by i.p. injection, starting 3 days after tumor inoculation. As shown in Figure 12, dosing with CNTO 95 inhibited growth of human melanoma tumors in nude mice. At day 26, CNTO 95 inhibited tumor growth by approximately 80% compared to tumors from control-treated animals. In this model, CNTO 95 does not interact with host angiogenic vessels since it does not bind mouse integrins, suggesting that blockade of human tumor-expressed integrins alone can inhibit tumor growth in mice independently of antiangiogenic effects.

Figure 12.

Inhibition of human melanoma tumor growth in nude mice by CNTO 95. Mice were inoculated s.c. with A375.S2 cells (3 × 106), and therapy with CNTO 95 or control was initiated 3 days later. Mice were treated with CNTO 95 or vehicle 3 times per week at a dose of 10 mg/kg i.p. Each data point is the mean tumor volume from 10 tumor-bearing animals (±SEM). CNTO 95 given 3 times per week significantly inhibited growth of tumors compared to control-treated animals at day 26 (p = 0.0005).

Inhibition of tumor growth in rats

To determine the antitumor efficacy of CNTO 95 in another xenograft animal model, an A375.S2 human melanoma model was developed in female nude rats. In this model, CNTO 95 is capable of blocking both rat angiogenic integrins and human tumor cell–expressed integrins. Weekly treatment of tumor-bearing nude rats with CNTO 95 at 10 mg/kg reduced tumor growth compared to the isotype-matched human IgG control MAb (Fig. 13). By day 46, treatment with CNTO 95 resulted in significant reduction in final tumor size compared to control-treated nude rats (p = 0.0007).

Figure 13.

Inhibition of human melanoma tumor growth in nude rats by CNTO 95. Rats were inoculated s.c. with A375.S2 cells (3 × 106), and therapy with CNTO 95 or control was initiated 3 days later. Rats were treated with CNTO 95 or vehicle once per week at a dose of 10 mg/kg i.v. Each data point is the mean tumor volume from 9 tumor-bearing animals (±SEM).

DISCUSSION

We have developed a unique fully human MAb, CNTO 95, which binds members of the αv family of integrins and neutralizes the biologic effects of the integrin receptors αvβ3 and αvβ5 in vitro and in vivo. CNTO 95 inhibited adhesion, migration, proliferation and invasion of both tumor and endothelial cells in vitro and demonstrated that binding and blocking multiple αv integrin receptors was more effective than blocking of a single integrin alone. In addition, CNTO 95 inhibited angiogenesis and tumor growth in vivo. Growth of human melanoma tumors was significantly reduced by blockage of tumor cell integrins in the mouse model or by combined blockage of tumor cell and host angiogenic integrins in the rat model, highlighting the potential importance of targeting multiple cellular targets for antitumor efficacy.

In vitro results demonstrate that CNTO 95 has potent antiangiogenic properties, inhibiting endothelial cell adhesion, proliferation, migration and capillary sprouting. In addition, CNTO 95 blocked angiogenesis stimulated by both bFGF and M21 melanoma cells in the rat Matrigel model and by bFGF in a primate angiogenesis model. The majority of literature surrounding the role of integrins in angiogenesis and tumor growth has focused on αvβ3. During the angiogenic process, αvβ3 is upregulated on the surface of activated endothelial cells, which in turn helps these cells to migrate, proliferate and invade the tumor.16, 18, 24 However, there is also evidence that αvβ5, one of the most widely expressed members of the integrin family,25 may play an important role in angiogenesis, distinct from αvβ3. In 2 separate models of angiogenesis, Friedlander et al.18 demonstrated that LM609, a function-blocking antibody to αvβ3, selectively inhibited angiogenesis stimulated by bFGF. In contrast P1F6, an antibody that blocks αvβ5, preferentially inhibited angiogenesis stimulated by VEGF, suggesting the 2 integrins participate in angiogenesis through distinct pathways. We found that antibody blockage of either αvβ3 by LM609 or αvβ5 by P1F6 reduced endothelial cell adhesion and migration in vitro but that the 2 agents had to be combined to reach the level of inhibition achieved by CNTO 95. These observations clearly illustrate the potential benefit of targeting multiple integrin pathways to inhibit angiogenesis.

The ability to block multiple pathways may also prove critical to effectively inhibit angiogenesis and tumor growth in vivo. As mentioned above, CNTO 95 displayed potent antiangiogenic effects in both a rodent model and a novel nonhuman primate model in cynomulgus monkeys. In a classic study, it was shown that antibody antagonism of αvβ3 by LM609 inhibited tumor growth through its antiangiogenic effect.26 In that study, αvβ3-negative human melanoma cells were injected into full-thickness human skin grafted onto SCID mice. The resulting tumors induced a human angiogenic response that enhanced the growth of tumor cells in an orthotopic microenvironment. Administration of LM609 significantly inhibited growth of αvβ3-negative tumors by blocking the growth of human blood vessels. The fact that LM609 recognizes only αvβ3 may present a limitation to its efficacy in vivo as we have found that dual blockage of αvβ3 and αvβ5 more completely inhibited endothelial cell functions in vitro. Synthetic dual integrin inhibitors have been created and have demonstrated antiangiogenic and antitumor efficacy in vivo. For example, the peptidomimetic compound S247 has demonstrated antiangiogenic and antimetastatic effects against colon tumor xenografts in mice dosed with 70 mg/kg daily.27 Similar results were reported for the nonpeptide RGD mimetic SCH 221153, which inhibited human melanoma tumor growth and angiogenesis in mice when dosed twice daily at 20 and 50 mg/kg.28 These observations point to the potential importance of dual integrin antagonism in inhibiting tumor growth and angiogenesis.

In addition to blocking integrins on angiogenic endothelium, the ability to inhibit integrin function on tumor cells themselves may prove critical in halting the growth of some tumors. A number of αv integrins have been suggested to play critical roles in tumor cell biology. For example, changes in expression of αvβ3 and αvβ5 have been associated with the progression, growth and dissemination of melanomas.29, 30, 31 Integrins αvβ1, αvβ6 and αvβ8 are the other known αv integrins; and the biology of these receptors is less characterized. αvβ1 acts as an adhesion receptor for vitronectin and fibronectin in tumor cells lacking αvβ3,32 and there is increasing evidence pointing to a major role for αvβ6 in the progression of oral squamous cell carcinoma.33, 34 Thus, it is possible that the recognition of multiple αv integrins by CNTO 95 may broaden its applicability to multiple tumor types with different integrin expression patterns. In the present study, CNTO 95 exerted its inhibitory effects primarily through integrins αvβ3 and αvβ5, based on in vitro evidence demonstrating complete blockage of cell adhesion to vitronectin by blockage of these 2 integrins and the fact that expression of αvβ6 could not be detected on A375.S2 cells (not shown). In our mouse xenograft model where CNTO 95 does not cross-react with host integrins, treatment with CNTO 95 significantly inhibited the growth of αvβ3/β5-positive melanoma tumors.

When CNTO 95 was able to bind and block integrins on both tumor cells and angiogenic endothelium, it caused a marked reduction in final tumor size at the end of the study. Together these data suggest that, through combined blockade of αvβ3 and αvβ5 integrins on tumor and endothelial cells, CNTO 95 may have multiple mechanisms of action that contribute to its observed antitumor efficacy in animal models.

One of the most important features of CNTO 95 is its fully human nature. The only other antibody-based integrin antagonist in development for cancer treatment is the humanized form of the murine anti-αvβ3 antibody LM609 (Vitaxin).35, 36 A dose-escalating phase I study in cancer patients demonstrated that it was safe,21 and similar findings have been reported for cilengitide, a peptide antagonist of αvβ3 and αvβ5.37 These early clinical results suggest that antagonism of αv integrins is a safe therapeutic approach. In a nonclinical toxicology evaluation in cynomulgus monkeys, CNTO 95 at 50 mg/kg weekly for 9 weeks was well tolerated. A phase I clinical study (first-in-human safety assessment) of CNTO 95 is currently ongoing in patients with solid tumors. Because it is fully human, CNTO 95 may carry the advantage of being less likely to cause immune responses in patients. Furthermore, because CNTO 95 is able to bind not only αvβ3 and αvβ5 but also other αv integrins, such as αvβ6 and αvβ1, it has the potential to inhibit multiple integrin-mediated events. The potential benefits of CNTO 95 interaction with these less characterized integrins is being explored. In conclusion, CNTO 95 is a promising agent that can inhibit tumor growth and angiogenesis.

Acknowledgements

We thank Ms. B. Heiser-Neas and Ms. L. Mylett for help in the preparation of the manuscript.

Ancillary