Award Number : DAMD 17 - 02 - 1 - 0342 TITLE : Ron in Breast Development and Cancer

RON is a member of the c-MET receptor tyrosine kinase family. Like c-MET, RON is expressed by a variety of epithelial-derived tumors and cancer cell lines and it is thought to play a functional role in tumorigenesis. To date, antagonists of RON activity have not been tested in vivo to validate RON as a potential cancer target. In this report, we used an antibody phage display library to generate IMC-41A10, a human immunoglobulin G1 (IgG1) antibody that binds with high affinity (ED50 = 0.15 nmol/L) to RON and effectively blocks interaction with its ligand, macrophage-stimulating protein (MSP; IC50 = 2 nmol/L). We found IMC-41A10 to be a potent inhibitor of receptor and downstream signaling, cell migration, and tumorigenesis. It antagonized MSP-induced phosphorylation of RON, mitogen-activated protein kinase (MAPK), and AKT in several cancer cell lines. In HT-29 colon, NCI-H292 lung, and BXPC-3 pancreatic cancer xenograft tumor models, IMC41A10 inhibited tumor growth by 50% to 60% as a single agent, and in BXPC-3 xenografts, it led to tumor regressions when combined with Erbitux. Western blot analyses of HT-29 and NCI-H292 xenograft tumors treated with IMC-41A10 revealed a decrease in MAPK phosphorylation compared with control IgG–treated tumors, suggesting that inhibition of MAPK activity may be required for the antitumor activity of IMC41A10. To our knowledge, this is the first demonstration that a RON antagonist and specifically an inhibitory antibody of RON negatively affects tumorigenesis. Another major contribution of this report is an extensive analysis of RON expression in f100 cancer cell lines and f300 patient tumor samples representing 10 major cancer types. Taken together, our results highlight the potential therapeutic usefulness of RON activity inhibition in human cancers. (Cancer Res 2006; 66(18): 9162-70) Introduction The macrophage-stimulating protein receptor (RON) belongs to the c-MET family of receptor tyrosine kinases (1–3). RON ligand, macrophage-stimulating protein (MSP), was originally found to modulate the function of certain macrophage by a variety of means. For example, addition of MSP not only induced shape changes, chemotaxis, macropinocytosis, and phagocytosis (4) but also inhibited lipopolysaccharide (LPS)-induced production of inflammatory mediators such as inducible nitric oxide (5), nitric oxide (6), prostaglandins, and cyclooxygenase-2 (7). Consistent with this notion that RON negatively regulates inflammation is the observation that activated macrophages from adult RON knockout mice produced increased levels of nitric oxide both in vitro and in vivo , thus rendering them more susceptible to LPS-induced endotoxic shock (8–10). Besides macrophage, RON has also been found expressed in several normal epithelial cells such as keratinocytes (11) where MSP was shown to phosphorylate RON and activate a number of signaling pathways that elicited cell adhesion/ motility and antiapoptotic and proliferative responses (12). As is the case with its better-known family member, c-MET, several lines of evidence suggest a role for RON in cancer. First, it is highly expressed in several epithelial tumors and cell lines of the colon (13), lung (14), breast (15, 16), stomach (17), ovary (18), pancreas (3), bladder (19), liver (20), and kidney (21, 22). RON is also found coexpressed in tumors with other growth factor receptors such as c-MET and epidermal growth factor receptor (EGFR; refs. 16, 19, 23, 24). Second, MSP and RON have been shown to influence the migration and invasion of cancer cells (2, 3, 25). Third, the oncogenic potential of RON has been shown on overexpression in cultured cell lines (26–28) and in transgenic mice (29, 30) where overexpression of RON led to a profound increase in proliferation and tumorigenesis, respectively. Downregulation of RON expression in RON-expressing cancer cell lines resulted in a reduction in proliferation (31). Similar oncogenic potential has also been observed for splice variants of RON identified in tumors and cell lines that generate receptors completely lacking the extracellular domain or that harbor deletions within this domain (26, 27, 32–35). Fourth, RON synergizes with known oncogenes. For example, the skin and breast tumorigenicity of transgenic mice overexpressing the polyoma middle T (36) and RAS (37) oncogenes, respectively, is repressed when these mice are crossed to RON knockout mice. Whereas several antagonists of c-MET have shown antitumor activity and thus validated c-MET as a cancer target, to date, antagonists of RON activity have not shown antitumor activity to validate RON as a potential cancer target. Consequently, through the screening of a human Fab phage display library, we developed IMC-41A10, a human IgG1 monoclonal antibody that binds with Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/). J.M. O’Toole and K.E. Rabenau contributed equally to this work. Requests for reprints: Daniel S. Pereira, Department of Tumor Biology, ImClone Systems, Inc., 180 Varick Street, New York, NY 10014. Phone: 646-638-5008; Fax: 212645-2054; E-mail: [email protected]. I2006 American Association for Cancer Research. doi:10.1158/0008-5472.CAN-06-0283 Cancer Res 2006; 66: (18). September 15, 2006 9162 www.aacrjournals.org Research Article high affinity to human RON and blocks MSP binding. In RONexpressing cancer cell lines, we show IMC-41A10 to be a potent inhibitor of receptor signaling, cell migration, and tumorigenesis. To our knowledge, this is the first demonstration of a RON antagonist, and specifically an inhibitory antibody of RON, that negatively affects tumorigenesis. Another major contribution of this report is an extensive analysis of RON expression in f100 cancer cell lines and f300 patient tumor samples representing 10 and 6 major cancer types, respectively. Taken together, our results underscore the potential therapeutic usefulness of RON activity inhibition in human neoplasia. Materials and Methods Immunohistochemistry. The polyclonal RON (C-20) antibody from Santa Cruz Biotechnology (Santa Cruz, CA), which has previously been used to detect RON expression by immunohistochemistry (17), was used here to detect RON expression in breast (51 samples, IMH-364), lung (59 samples, IMH-305), prostate (40 samples, IMH-303), stomach (59 samples, IMH-316), and colon (59 samples, IMH-306) tumor tissue arrays purchased from Imgenex (Sorrento Valley, CA). Pancreatic tumor tissue arrays (54 samples, CC1401) were obtained from Cybrdi (Frederick, MD). These formalin-fixed paraffin-embedded arrays were treated and stained with 1.2 Ag/mL RON (C-20) antibody or rabbit IgG as a negative control (Jackson ImmunoResearch Laboratories, West Grove, PA) using the EnVision+ Rabbit Kit (DAKO, Carpinteria, CA) as described (38). Cell culture. Cell lines used in this study were obtained from the American Type Culture Collection (Manassas, VA) and grown exactly as recommended. Media used for their propagation were purchased from Life Technologies, Inc. (Grand Island, NY). Flow cytometry. To measure cell-surface expression of RON, cancer cell lines were grown to subconfluence, trypsinized, washed, pelleted (5 minutes, 1,400 rpm), and resuspended/filtered in 250 AL PBS + 5% fetal bovine serum (FBS). Five micrograms of RON antibody (IMC-41A10) were added to all tubes, except the control tube, and incubated for 30 minutes at 4jC. Cells were washed, pelleted (5 minutes, 1,400 rpm), and resuspended in 250 AL PBS + 5% FBS. One microgram of the secondary antibody, goat anti-human IgG-phycoerythrin (Jackson ImmunoResearch Laboratories), was added and left for 30 minutes at 4jC. Finally, cells were washed, pelleted, and resuspended in 500 AL PBS + 5% FBS for analysis on a Becton Dickinson (San Jose, CA) FACSVantage SE. Identification of human anti-RON Fab antibodies from a phage display library. A human Fab phage display library containing 3.7 ! 10 clones (39) was used to screen for and identify human anti-RON Fab antibodies following a previously described procedure (40) using RON-Fc (Sigma-Aldrich, St. Louis, MO) as the bait. Individual phage clones recovered after the second and the third round of selections were examined for binding to immobilized RON-Fc by ELISA. Plasmids of individual binders were then used to transform a nonsuppressor Escherichia coli host HB2151 for the expression of soluble Fab fragments. The soluble Fab proteins were purified from the bacteria periplasmic extracts by affinity chromatography using a Protein G column, following the protocol of the manufacturer (Amersham Pharmacia Biotech, Piscataway, NJ), and tested for binding to RON-Fc by ELISA. In MSP/RON blocking assay, Maxi-sorp 96-well microtiter plates (Nunc, Rochester, NY) were coated with MSP (1 Ag/mL ! 100 AL; R&D Systems, Minneapolis, MN) at room temperature for 1.5 hours. After washing the wells, they were blocked with 3% PBS/milk. Anti-RON antibodies (Fab or full IgG) were preincubated with RON-Fc (25 ng/well) at room temperature for 1 hour. The Fab/RON-Fc or IgG/RON-Fc mixtures were then added to the MSP-coated wells and allowed to incubate for 1.5 hours at room temperature. After several washes, a 1:1,000 dilution of the antihuman IgG, Fab-specific horseradish peroxidase (HRP)–conjugated antibody was added to the plates for 1.5 hours at room temperature to detect the anti-RON Fab or IgG that bound to RON but did not block the MSP/RON interaction. IMC-41A10 and IMC-42E12 are examples of antibodies that blocked and failed to block the MSP/RON interaction, respectively. ELISA to detect binding of IMC-41A10 to RON. Standard ELISA plates were coated overnight at 4jC with recombinant humanMSPR (R&D Systems) at 100 ng/well. The plateswerewashed oncewith 0.2%PBS/Tand blockedwith 3% milk (150 AL/well) for 1 hour at 37jC. Following a wash with 0.2% PBS/T, a 3 nmol/L solution of IMC-41A10 was prepared in 1 mL 3% milk and diluted 1:2 across plates. The plateswere incubated for 1 hour at room temperature on a rocker and washed five times with 0.2% PBS/T. The secondary antibody (goat anti-human Fc, HRP-conjugated) was added at 1:5,000 dilution and allowed to incubate for 1 hour at room temperature. Washing was done as above. One hundred microliters of substrate were added per well until a yellow color developed. The reaction was stopped with 50 AL of 1 N H2SO4 and the absorbance at 450 nm determined with a standard plate reader. To determine whether IMC-41A10 could cross-react with c-MET [hepatocyte growth factor receptor (HGFR)], an ELISA was done exactly as described above with the exception that a recombinant human HGFR (c-MET)/Fc chimeric protein (R&D Systems) was used to coat the plates and a mouse anti-human HGFR antibody (R&D Systems) was used as a positive control. A goat anti-mouse HRP antibody was used as the secondary antibody for the positive control. ELISA to detect IMC-41A10 blocking of MSP binding to RON. ELISA plates were coated with 100 ng/well carrier-free MSP (R&D Systems) overnight at 4jC on a rocker. The plate was washed once with 0.2% PBS/T and blocked for 2 hours at 37jC with 150 AL/well 3% milk. A 15 Ag/mL dilution of IMC-41A10 was prepared and serially diluted across another ELISA plate. To the IMC-41A10, we added 100 ng/well recombinant human MSPR (R&D Systems). The IMC-41A10/rh-MSPR complex was allowed to form for 2 hours at room temperature on a rocker. Next, the MSP-coated ELISA plate was washed once with 0.2% PBS/T and 100 AL of the IMC-41A10/rh-MSPR complex were added per well. After a 1.5-hour incubation at room temperature, the plate was washed five times with 0.2% PBS/T and incubated for 1 hour at room temperature with a 1:2,000 dilution of anti–His-tag HRP antibody (Sigma), which recognizes a His tag on the recombinant human MSPR protein. The plate was washed five times with 0.2% PBS/T and 100-AL substrate was added per well until a yellow color developed. The reaction was stopped with 50 AL of 1 N H2SO4 and the absorbance at 450 nm determined with a standard plate reader. Analysis of MSP-induced phosphorylation of RON, mitogen-activated protein kinase, and AKT. Cells were grown in six-well plates or 10-cm dishes to f70% confluency and serum starved for 18 hours. For cells receiving antibody, 100 nmol/L 41A10 or 42E12 was added and incubated at 37jC for 1 hour. MSP (R&D Systems) was then added to the appropriate cells at 10 nmol/L and incubated for 30 minutes at 37jC. Cells were washed twice with ice-cold PBS and lysates prepared in ice-cold cell extraction buffer (Biosource International, Carlsbad, CA) plus fresh HALT protease inhibitor cocktail (Pierce, Rockford, IL) and 15 Ag of lysates were resolved by SDS-PAGE using 4% to 12% NuPAGE Novex Bis-Tris Gels with 1! MOPS running buffer. For Western blotting, proteins were transferred to a nitrocellulose membrane (Invitrogen, Carlsbad, CA) and blocked with 5% nonfat milk in TBS/0.1% Tween 20 (TBS-T) for 1 hour at room temperature on a rocker. After three 5-minute washes with TBS-T, the membrane was incubated with 1:1,000 rabbit polyclonal phospho-p44/42 MAPK (Thr/Tyr) antibody (Cell Signaling, Danvers, MA) diluted in 1% nonfat milk in TBS-T or 1:500 rabbit polyclonal anti-phospho-AKT (P-Ser) antibody (BD PharMingen, San Jose, CA) or 1:1,000 rabbit monoclonal phosphor-AKT (Ser) antibody (Cell Signaling). Following three 5-minute washes with TBS-T, a 1:5,000 dilution of anti-rabbit IgG, HRP-conjugated secondary antibody (Amersham) was added to the membrane for 1 hour at room temperature followed by three 5-minute TBS-T washes. Protein bands were detected with Amersham Enhanced Chemiluminescence Western Detection Kit. To determine expression levels of mitogen-activated protein kinase (MAPK), these membranes were stripped with Pierce RestoreWestern Blot Stripping Buffer (15 minutes at room temperature) and reprobed with a 1:1,000 dilution of p44/42 MAPK polyclonal rabbit antibody (Cell Signaling) or a 1:500 dilution of polyclonal rabbit anti-AKT1 antibody (BD PharMingen) or 1:2,000 rabbit polyclonal anti-AKTantibody (Cell Signaling). Neutralizing Antibody to RON Receptor Tyrosine Kinase www.aacrjournals.org 9163 Cancer Res 2006; 66: (18). September 15, 2006 RON-NIH3T3 cells were used to detect MSP-induced phosporylation of RON. The procedure used was exactly as described above with the exception that cells were serum starved for 2.5 hours and stimulated with or without 10 nmol/L MSP for 30 minutes at 37C. Before MSP stimulation, RON antibodies were added to cells for 1 hour at 100 nmol/L to assess blocking activity. To determine whether RON antibodies exhibited agonist activity on RON phosphorylation, antibody was added to cells in the absence of MSP stimulation. Cells were lysed and 30 Ag of lysate were resolved on a 4% to 12% gel. Western blot analysis was done with antiphosphotyrosine antibody, 4G10 (Upstate, Charlottesville, VA), as per recommendation of the manufacturer. Cell migration assay. To determine whether IMC-41A10 could block the migration of H596 cells induced by MSP, we used cell culture inserts containing porous translucent polyethylene terephthalate track–etched membranes (8.0 Am pore size; Becton Dickinson Falcon). Before the assay, the undersides of the porous membranes in the cell culture inserts were coated with collagen by placing them into a 24-well Falcon plate filled with 700 AL of Vitrogen-100 purified collagen solution (25 Ag/mL; Cohesion, Palo Alto, CA). The inserts were left for at least 1 hour at 4jC and placed in a new 24-well plate containing either 700-AL serum-free medium or 10% FBS. Next, 6 ! 10 viable cells that had been serum starved for 24 hours were rinsed twice with PBS and seeded into the upper chamber of the cell culture insert in 300 AL of serum-free medium. MSP was added to the lower chamber for 24 hours at 37jC to induce cell adhesion and migration through the collagen-coated porous membrane on the underside of the cell culture insert. Before the addition of MSP, IMC-41A10 was added to other wells to determine if it could inhibit the MSP-induced migration/invasion of H596 cells. IMC-42E12, which binds to RON but does not block MSP binding, was used as a negative control. At the conclusion of the assay, migrated cells present on the underside of the collagen-coated membrane were stained with H&E and visualized by bright field microscopy at !100 magnification. In vitro ‘‘wound’’ healing assay. HCA-7 cells were seeded at 2! 10 in a 4-mL suspension in a 60-mm tissue culture dish and allowed to grow to confluence, refreshing the medium every 3 to 4 days. When 41A10 or anti– keyhole limpet hemocyanin (KLH) antibodies were used, the cells were preincubated with the antibodies for 1 hour before wound infliction. At the start of the assay, a small wound was inflicted with a plastic pipette tip. The cells were incubated for 24 hours with and without the presence of MSP plus anti-KLH or anti-RON (41A10) antibodies. The cells were imaged on a Zeiss microscope immediately after wounding (time 0) and at 16 and 24 hours. Tumor xenograft models. To establish tumor xenograft models with which to test the antitumor activity of IMC-41A10, 5 million HT-29, NCIH292, and BXPC-3 cells were mixed with Matrigel (Collaborative Research Biochemicals, Bedford, MA) and s.c. injected into the left flank of 56-weekold female athymic (nu/nu) mice (Charles River Laboratories, Wilmington, MA). Tumors were allowed to reach 150 to 300 mm in size and mice were randomized into groups of 12 animals each. Mice were treated by i.p. injection every 3 days with control antibody (human IgG) or IMC-41A10 antibody at a dose of 40 mg/kg or Erbitux at a dose of 40 mg/kg. Treatment of animals was continued for the duration of the study. Tumors were measured twice each week with a caliper and tumor volumes calculated using the following formula: (k/6 (w1 ! w2 ! w2)), where w1 represents the largest tumor diameter, and w2 represents the smallest tumor. Analysis of MAPK activity in HT-29 and NCI-H292 tumor xenografts. HT-29 and NCI-H292 tumor xenografts were established as described above with the exception that six mice were used for IMC-41A10 treatment and six used for control IgG treatment. Once tumors reached f300 mm, mice were injected once with an antibody dose of 40 mg/kg. After 24 hours, mice were sacrificed and tumors excised. Tumor lysates were prepared and subjected to Western blot analysis. To determine whether IMC-41A10 had an effect on MAPK activity, phosphorylated and total MAPK levels were detected as described above and subjected to densitometric analysis. The densitometric values for the total MAPK bands were normalized to 20,000 and, in turn, the densitometric values of the phosphorylated MAPK bands were proportionately adjusted. The adjusted phosphorylated MAPK values for the six IMC-41A10 mice were averaged and compared with those of the six control IgG mice. Statistical significance between the two groups was determined by Student’s t test. Results RON is expressed in several cancerous tissues and cell lines. We did an extensive analysis of RON expression in several cancer tissues by tumor tissue arrays as well as in cell lines by flow cytometry and Western blot analyses. In addition, through access to the Gene Logic Ascenta database, we were able to assess RON expression profile in a variety of cancer tissues as well as in the NCI-60 cancer cell line panel. Table 1 is a summary of RON expression in human cancer as determined by immunohistochemistry of Imgenex tumor tissue arrays. RON was expressed in 100%, 93%, 92%, 73%, 69%, and 65% of breast, lung, prostate, gastric, pancreas, and colon tumors, respectively. Table 1 also shows the mean intensity of RON expression in the tumor epithelial cells for each of these cancer types. Figure 1 is a montage of tumor tissue sections from these tumor tissue arrays that exhibit a mean epithelial cell intensity of RON expression similar to that listed in Table 1. Patient information for the tumor tissue sections shown in Fig. 1 is given in the figure legend. With respect to cancer cell lines, Table 2 lists those that are positive and negative for RON expression on the basis of flow cytometry, Western blotting, and Ascenta data (Gene Logic). Of the f100 cell lines tested that cover 10 different cancer types, approximately half exhibited RON expression. Interestingly, none of the leukemic and renal cancer cell lines expressed RON by these analyses. Identification and characterization of IMC-41A10 as an antibody that binds to RON and blocks its interaction with MSP. To determine whether RON plays a functional role in tumor growth, we generated a neutralizing monoclonal antibody. For this task, a human Fab antibody/phage display library was used to isolate fully human high-affinity antibodies capable of binding the RON receptor and blocking ligand binding (data not shown). To begin this process, an Fc fusion protein of the extracellular portion of the RON receptor was used to screen the Fab phage display Table 1. RON expression in patient tumor samples as determined by immunohistochemistry analysis of tumor tissue arrays Cancer tissue %Positive* Mean intensity c Breast 100% (47 of 47) 166 (20-300) Lung 93% (50 of 54) 142 (15-300) Prostate 92% (35 of 38) 164 (20-300) Gastric 73% (38 of 52) 172 (20-300) Pancreas 69% (36 of 52) 100 (30-200) Colon 65% (36 of 55) 54 (15-225) *The percentage of RON-expressing tumor samples on the tumor tissue array is listed. In parentheses, the numerator represents the number of RON-expressing tumor cores and the denominator equals the total number of tumor cores present on the tumor tissue array. cMean intensity represents tumor epithelial staining intensity in RONpositive samples. This value is obtained by multiplying a relative intensity score (0-3) by the percentage of epithelial cells present in the sample core. See Materials and Methods for experimental details. Cancer Research Cancer Res 2006; 66: (18). September 15, 2006 9164 www.aacrjournals.org library, which contains f10 naı̈ve recombinant Fab fragments. Following three rounds of selection, binding Fab fragments were identified. Four of these were subsequently found to efficiently block the interaction of MSP ligand with RON. Following the cloning of these Fabs into mammalian expression vectors harboring the fully human IgG1 backbone, one antibody, IMC41A10, was chosen for further studies. To quantitate the binding and blocking potential of IMC-41A10 to RON, ELISA assays were established. ELISA data revealed a receptor binding ED50 of 0.15 nmol/L and a receptor blocking IC50 of 2 nmol/L for IMC-41A10 (Fig. 2A and B). Because RON is a member of the c-MET receptor family, we wanted to determine whether IMC-41A10 could also bind c-MET. Using an ELISA-based binding assay with immobilized soluble extracellular c-MET protein, we showed that IMC-41A10 could not bind c-MET (Fig. 2C). IMC-41A10 significantly inhibits MSP-dependent phosphorylation of RON and downstream signaling. Once IMC-41A10 was shown to function in solid-phase receptor binding and blocking assays, we evaluated its ability to block the activity of the RON receptor and the downstream effectors, MAPK and AKT, in cellbased assays. To assess the ability of IMC-41A10 to inhibit RON phosphorylation, we used NIH3T3 cells overexpressing the recombinant wild-type RON protein (Fig. 3A). When these cells were serum starved and stimulated with MSP, tyrosine phosphorylation of RON was readily detected and could be significantly inhibited (f75%) by the addition of IMC-41A10. In contrast, no inhibition of RON phosphorylation was noted with IMC-42E12 (an antibody that binds to RON but does not block MSP binding to RON). It is important to note that when 41A10 was added to the RON-NIH3T3 cells in the absence of MSP (Fig. 3A, lane 3), no induction of RON phosphorylation was observed, indicating that IMC-41A10 is devoid of agonist activity on the RON receptor. Because the MSP/RON pathway is known to affect MAPK and AKT activity, we were interested in evaluating the potential of IMC41A10 to inhibit the phosphorylation and subsequent activation of MAPK (Fig. 3B). Using cell lines that represent a number of different cancers [i.e., HT-29 (colon), NCI-H292 (lung), HCC-1937 (breast), BXPC-3 (pancreas), DU145 (prostate), and AGS (gastric)], we determined that IMC-41A10 was able to completely inhibit MSP-induced phosphorylation of MAPK in every case and of AKT in HT-29, DU145, and AGS cells. Following MSP treatment, AKT phosphorylation did not occur in NCI-H292, BXPC-3 and HCC-1937 cells. It is also interesting to note that for many of the aforementioned cell lines, constitutive phosphorylation of MAPK and AKT was noted in the absence of exogenously added MSP. IMC-41A10 significantly inhibits cell migration. Like HGF and c-MET, MSP and RON have been implicated in the migration and invasion of certain macrophage and epithelial cells. Previously, MSP was shown to induce migration of the H596 lung cancer cell line in vitro (14). We modified this assay (Fig. 4A) and tested the Table 2. Flow cytometric, Western blot, and Gene Logic analyses of RON expression in cancer cell lines Tumor type Positive cell lines Negative cell lines Colon DiFi, HCT-8, HCT-15, HCT-116 HCC-2998, Colo205, Colo-201 DLD-1, GEO, T84, HT-29, KM12 Sw480, Sw620 SNU-C1, LoVo, CaCo-2 Lung NCI-H1650, NCI-H292, NCI-H1975, NCI-H358, NCI-H1666, NCI-H226, NCI-H441, NCI-H1650, NCI-H727, NCI-H596, NCI-H322M, EKVX, A549 NCI-H460, NCI-H23, NCI-H522 Hop-62, Hop-92 Pancreas BXPC-3, Capan-2, HPAC, HPAFII L3#7.p1, Hs766.T, ASPC-1, CFPAC-1 MIA-PACA-2 Breast T47D, HCC-1937, MDA-MB-468, BT20, DU4475 MDA-MB-231, MDA-MB-435, MCF-7, BT-474, SKBR3, Hs578T Ovary SKOV3, OVCAR3, OVCAR5, MDAH-2774, IGOV-1, OV90 OVCAR8, CaCOV3, CaCOV4 Prostate PC-3, DU145, LnCAP 22RV.1 Stomach NCI-N87, AGS SNU-1, SNU-16, KKVR Liver HepG2 SNU-182, SNU-449, SNU-475 SNU-398, Hep3B, PLC/PRF/5 Leukemia EOL, JM-1, HEL, Jurkat, TF-1 U937, HL-60 Kidney A498, Caki-1, SKRC29, 786-0, ACHN, RXF-393, SN12C, TK-10, UO-31 NOTE: Of the cell lines present in the NCI-60 cancer cell line panel, expression analysis was collected from Gene Logic’s Ascenta database, which contains Affymetrix expression data for the NCI-60 cancer cell line panel. These data were accessed through an institutional subscription to the Ascenta database. Figure 1. Montage of immunohistochemistry images from the tumor tissue array analysis of RON expression described in Table 1. The images shown in this figure were chosen to match the mean expression intensity of RON reported for each cancer type in Table 1. Patient data corresponding to the tumor images are as follows: Breast (infiltrating ductal carcinoma, stage T4b), Lung (adenosquamous carcinoma, stage IIIA, T2N2M0), Prostate (adenocarcinoma, stage III, T3N0M0), Gastric (stage IB, T2N0M0, well differentiated), Pancreas (ductal adenoma, stage III), and Colon (sigmoid adenocarcinoma, stage II, T3N0M0, moderately differentiated). Neutralizing Antibody to RON Receptor Tyrosine Kinase www.aacrjournals.org 9165 Cancer Res 2006; 66: (18). September 15, 2006 ability of IMC-41A10 to inhibit MSP-induced migration of H596 cells. In comparison with IMC-42E12, IMC-41A10 showed >90% inhibition of MSP-mediated cell migration (Fig. 4B). These data are shown graphically in Supplementary Fig. S1. Using an in vitro wound healing assay, we also showed that IMC-41A10 could significantly inhibit (f75%) the ability of HCA7 colon cancer cells to migrate in response to MSP and close a ‘‘wound’’ created in their monolayer (Fig. 4C). Taken together, these data confirm that MSPdependent tumor cell migration can be inhibited by antagonizing MSP binding to RON. IMC-41A10 possesses antitumor activity. Because IMC-41A10 successfully abrogated RON signaling and cell migration/invasion, we were interested in determining whether it could inhibit tumorigenicity. To begin to test the antitumor activity of IMC41A10, murine tumor xenograft models were established using HT-29 colon, NCI-H292 lung, and BXPC-3 pancreatic cancer cells. These cells were s.c. injected into nude mice and tumors allowed to establish to a size of f250 mm before being randomized into treatment groups. IMC-41A10 and a control human IgG were subsequently injected i.p. at a dose of 40 mg/kg every 3 days. In comparison with control IgG-injected mice, IMC-41A10 treatment led to a 58%, 50%, and 50% inhibition of HT-29, NCI-H292, and BXPC-3 tumor growth, respectively (Fig. 5A-C). It should be noted that no differences in tumor necrosis were observed in control IgG versus IMC-41A10-treated tumors (see Supplementary Fig. S2 for examples of HT-29 tumor xenografts). To determine whether IMC-41A10 could be more efficacious in HT-29 xenografts, we initiated a subcutaneous prophylactic murine tumor xenograft model where IMC-41A10 was administered at 40 and 80 mg/kg every 3 days commencing 1 day after HT-29 cells were injected (Supplementary Fig. S3). IMC-41A10 did not exhibit increased efficacy in this model versus the established subcutaneous HT-29 tumor xenograft model in Fig. 5A . IMC-41A10 inhibits phosphorylation of MAPK in HT-29 and NCI-H292 tumor xenografts. Because IMC-41A10 was shown to inhibit MAPK phosphorylation in cultured HT-29 and NCI-H292 cells (Fig. 3), we were interested in confirming whether MAPK signaling was also inhibited in xenograft tumors of IMC-41A10treated mice. After treating established HT-29 and NCI-H292 tumors with a single 40 mg/kg dose of IMC-41A10, tumor lysates were prepared after 24 and 72 hours, respectively, to quantitate MAPK phosphorylation levels by Western blot analysis. We found that IMC-41A10 treatment of HT-29 and NCI-H292 xenografts resulted in a 35% and 28% mean decrease in phosphorylated MAPK levels compared with control IgG-treated tumors (Fig. 6). IMC41A10-treated HT-29 tumors 2, 4, and 5 as well as NCI-H292-treated tumors 1, 2, and 4 showed phosphorylated MAPK levels below that of any of the IgG-treated tumors. With the exception of the IMC41A10-treated NCI-H292 tumor 3, none of the IMC-41A10-treated tumors exhibited higher phosphorylated MAPK levels than the highest control IgG-treated tumors, a finding which also shows that IMC-41A10 is devoid of agonist activity. Taken together, these data show that treatment of tumor xenografts with IMC-41A10 led to a significant inhibition of tumor growth via inhibition of RONdependent signaling in tumor cells. Discussion A substantial body of evidence exists to support a functional role for c-MET in human cancer (2). In addition, several c-MET receptor antagonists have shown antitumor activity in animal models, thereby implicating c-MET as a potential target for therapeutic intervention (2). In comparison with c-MET, RON has not been as extensively studied. Within the last few years, however, data have accumulated to suggest that not only is RON expressed in several cancers but it also may play a functional role in tumor formation. Figure 2. Solid-phase binding and blocking characteristics of IMC-41A10. A, ELISA to determine the ED50 of IMC-41A10 binding to immobilized rh RON protein. B, ELISA to determine the IC50 of IMC-41A10 needed to block the interaction of recombinant human MSP to immobilized recombinant human RON protein. C, ELISA to determine whether IMC-41A10 can bind to immobilized recombinant human c-MET protein. An antihuman HGFR (c-MET) antibody is used as a positive control. Cancer Research Cancer Res 2006; 66: (18). September 15, 2006 9166 www.aacrjournals.org Although RON antagonists have been reported (41–44), to our knowledge, none have been tested for antitumor activity to validate RON as a potential cancer target. Consequently, we developed IMC-41A10, a fully human IgG1 antibody that binds to human RON with high affinity (ED50 = 0.15 nmol/L) and blocks interaction with its ligand, MSP (IC50 = 2 nmol/L). IMC-41A10 does not bind to c-MET and possesses no agonist activity. Whereas other neutralizing antibodies have been reported for RON (42), IMC-41A10 represents the first RON antibody to show antitumor activity. As a single agent, IMC-41A10 showed 50% to 60% inhibition of established tumor growth in three s.c. xenograft models using HT-29 colon, NCI-H292 lung, and BXPC-3 pancreatic cancer cells. We hypothesize that the reason IMC-41A10 treatment alone did not lead to tumor growth stabilization or regression in these three cell lines is because pathways in addition to RON are required for their tumorigenicity. Consistent with this notion is the fact that when IMC-41A10 treatment was given in combination with Erbitux (an anti-EGFR antibody) in BXPC-3 xenografts, tumor regression was observed in 5 of 12 mice. In contrast, although IMC-41A10 and Erbitux significantly inhibited BXPC-3 tumor growth, neither antibody, on its own, led to tumor growth stabilization or regression. In the future, it will be important to determine whether enhanced tumor inhibition will result following combination treatment with other antibodies or targeted therapies as well Figure 3. Inhibition of MSP-dependent receptor and MAPK phosphorylation. A, effect of IMC-41A10 on MSP-induced phosphorylation of RON. RON-NIH3T3 cells were serum starved (lane 1) and stimulated with MSP (lanes 2, 4 , and 5). Lanes 3 and 4, effect of adding IMC-41A10 in the absence and presence of MSP, respectively. Lane 5, effect of adding IMC-42E12, an antibody that binds to RON but does not block MSP binding, in the presence of MSP. B, effect of IMC-41A10 on MSP-induced phosphorylation of MAPK in cell lines representing six types of cancer. These lines were serum starved (lane 1) and stimulated with MSP (lane 2). Lanes 3 and 4, effect of adding IMC-41A10 in comparison with IMC-42E12, respectively. Figure 4. Ability of IMC-41A10 to inhibit cell migration. A, schematic of cell migration assay. i, cells are placed in the upper chamber of a cell culture insert and lowered into a well of a 24-well plate containing serum-free medium. The cell culture insert contains a porous membrane with collagen coated on its underside. After 24 hours, fresh serum-free medium containing MSPF antibody is added to the lower chamber (ii ) and cells migrate through the porous membrane and collagen and adhere to the underside of the collagen (iii ). B, H&E staining of H596 lung cancer cells that have adhered to and migrated through to the underside of the collagen in the presence of IMC-41A10 and IMC-42E12, an antibody that binds to RON but does not block MSP binding. C, in vitro wound healing assay. A scratch was made in a monolayer of HCA7 colon cancer cells. IMC-41A10 was added to determine whether it could inhibit the ability of MSP to induce the migration of cells to fill the wound. Neutralizing Antibody to RON Receptor Tyrosine Kinase www.aacrjournals.org 9167 Cancer Res 2006; 66: (18). September 15, 2006 as with various forms of chemotherapy and radiation. Testing the efficacy of IMC-41A10 in other subcutaneous as well as orthotopic tumor xenograft models will also be of interest. Interestingly, IMC-41A10 inhibition of RON activity and tumorigenicity in HT-29 cells is consistent with the work of Xu et al. (31) who showed that repression of RON expression in HT-29 cells by RNA interference caused a decrease in cell proliferation and an increase in apoptosis. HT-29 cells have been shown to express constitutively active splice variants of RON with small deletions in the extracellular portion that possess enhanced oncogenic potential (26–28). Whereas we have not determined whether IMC-41A10 can interact with these constitutively active oncogenic RON variants, it is noteworthy that IMC-41A10 showed strong antitumor activity in HT-29 cells that express wild-type RON in addition to the splice variants. Despite the fact that IMC-41A10 cross-reacts with murine RON (data not shown), no adverse effects were noted during treatment of tumor-bearing mice in this study. Thus, it seems that toxicity is not associated with inhibition of murine RON activity. This is an important finding given the phenotype of RON knockout mice and the role that RON is thought to play in normal inflammatory cells as a potential negative regulator of inflammation (8, 9). To investigate the antitumor activity of IMC-41A10, we considered ligand-dependent and -independent mechanisms. With regard to the former, we evaluated whether IMC-41A10 could modulate MSP-dependent MAPK and AKT activation as well as cell migration. Like HGF, it is well established that MSP (HGF-like protein) affects cell migration/invasion. Moreover, it is believed that cancer cells expressing c-MET or RON are prone to tissue invasion and metastasis (2, 3, 25). Using two in vitro cell migration assays, we showed IMC-41A10 to be a potent inhibitor of MSP-induced cell migration in H596 lung and HCA7 colon cancer cells. In the future, it will be of interest to determine whether IMC-41A10 will show efficacy in in vivo tumor metastasis models using these and other cell lines. MAPK and AKT are known to be key downstream effector molecules governing a variety of cellular pathways respectively associated with cell growth and survival. Because MAPK and AKT have been shown to be modulated by RON receptor signaling (12), we were keen to discover whether IMC-41A10 could inhibit their phosphorylation and subsequent activation. We found IMC-41A10 to be a potent inhibitor of MSP-induced phosphorylation of MAPK in vitro in HT-29 cells as well as in cultured NCI-H292, DU145, AGS, BXPC-3, and HCC-1937 cells, which represent lung, prostate, gastric, pancreatic, and breast cancer, respectively. IMC-41A10 also proved to be a potent inhibitor of AKT phosphorylation in HT-29, DU145, and AGS cells. In HT-29 and NCI-H292 tumor xenografts treated with IMC-41A10 for 24 and 72 hours, respectively, we observed a 35% and 28% inhibition of MAPK phosphorylation following a single treatment. Because IMC-41A10 treatment of HT-29 and NCI-H292 tumor xenografts occurred for several weeks, it is reasonable to assume that the cumulative effect of IMC-41A10 treatment on MAPK activity would be to significantly and negatively affect tumor growth. Because HT-29 and NCI-H292 cells do not produce MSP (Supplementary Fig. S4) but are sensitive to MSP stimulation in vitro , our data suggest not only that in vivo activation of MAPK in these cell lines is mediated by paracrine stimulation of the RON receptor with murine MSP but also that IMC-41A10 effectively inhibits the murine MSP-RON interaction. Taken together, the inhibitory effects of IMC-41A10 on MAPK and AKT phosphorylation would suggest not only a negative effect on cell growth and survival but may also be predictive of the antitumor potential of IMC-41A10 treatment in cancer cell lines where it modulates phosphorylation of MAPK and AKT. With respect to MSP-independent mechanisms of tumor inhibition, we did experiments to test whether IMC-41A10 treatment led to (a) Figure 5. IMC-41A10 inhibits the growth of tumor xenografts in nude mice. HT-29 colon (A), NCI-H292 lung (B ), and BXPC-3 pancreatic (C ) cancer cells were injected s.c. into nude mice and allowed to grow to f250 mm. Groups of 12 mice each were treated i.p. with 40 mg/kg of control human IgG or IMC-41A10 every 3 days. BXPC-3 xenograft tumor-bearing mice were also i.p. treated with Erbitux or the combination of Erbitux and IMC-41A10, each administered at 40 mg/kg every 3 days. Tumor size was measured with a caliper at regular intervals. Bars, SE. Statistical significance was determined by Student’s t test. Cancer Research Cancer Res 2006; 66: (18). September 15, 2006 9168 www.aacrjournals.org RON receptor down-regulation (Supplementary Fig. S5A and B); (b) repression of vascular endothelial growth factor secretion (Supplementary Fig. S6A); (c) induction of apoptosis (Supplementary Fig. S6B); and (d) induction of complement-dependent cytotoxic or antibody-dependent cell-mediated cytotoxic activity (Supplementary Fig. S7A and B). Whereas results from these experiments were conclusive, they were also negative and did not suggest an MSP-independent mechanism of action for the antitumor activity of IMC-41A10. With respect to the inability of IMC-41A10 to elicit complement-dependent cytotoxic and antibody-dependent cell-mediated cytotoxic activities on cell lines tested, we emphasize that it is possible that IMC-41A10 (a human IgG1 antibody) could elicit such activity in vivo . Furthermore, we stress that other MSP-independent mechanisms may contribute to the antitumor efficacy of IMC-41A10 but await further experimentation. Whereas some reports exist that provide evidence for RON expression in human tumors, the number of patients was not always sufficient and methods used to examine expression were varied. In this report, we conducted an extensive systematic analysis of RON expression in f100 cancer cell lines representing 10 cancer types and in 6 different tumor types using commercial tumor tissue arrays harboring nearly 300 individual tumor samples. RON was expressed in approximately half of the cell lines and was expressed by at least one cell line per cancer type with the exception of the leukemic and renal cancer cell lines tested. With respect to the tumor tissue array results, RON was well expressed in every cancer type (breast, 100%; lung, 93%; prostate, 92%; gastric, 73%; pancreas, 69%; and colon, 65%). We observed no significant correlation between the extent of RON expression and the stage or severity of disease. These results make a significant contribution to the expression profile of RON in a variety of tumor tissues and cancer cell lines. In conclusion, our data show that not only is RON well expressed in several cancers but it is also an ‘‘inhibitable’’ target that contributes to tumorigenesis. Our findings also underscore the potential of using an antibody to antagonize RON activity for the treatment of human cancers such as colon, lung, pancreas, and probably other cancers. Acknowledgments Received 1/24/2006; revised 6/7/2006; accepted 7/12/2006. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. We thank Cellestan for pathologic analyses of our tumor tissue array data; Chris Damoci, Deevi Dhanvanthri, and Jim Tonra for animal work; and Anthony Kayas, Mary Amasia, Xenia Jimenez, and Dave Sargent for assistance with antibody production, purification, and characterization studies. Dedication: ILU-MdLSP-RIP. Figure 6. IMC-41A10 inhibits phosphorylation of MAPK in HT-29 and NCI-H292 tumor xenografts. HT-29 and NCI-H292 cells were s.c. injected into 24 nude mice and allowed to reach f300 mm. For each cell line, six mice were treated with IMC-41A10 and six with control IgG. Twenty-four hours (HT-29) and 72 hours (NCI-H292) after a single injection with antibody at a dose of 40 mg/kg, tumors were excised and lysates prepared for Western blot analysis to detect MAPK and phosphorylated MAPK levels. Densitometric values for total MAPK bands were normalized to 20,000 and, in turn, the densitometric values of the phosphorylated MAPK bands were proportionately adjusted. The adjusted phosphorylated MAPK values for IMC-41A10-treated mice were averaged and compared with those of the control IgG-treated mice. Statistical significance between the two groups was determined by Student’s t test. References 1. Ronsin C, Muscatelli F, Mattei MG, Breathnach R. A novel putative receptor protein tyrosine kinase of the met family. Oncogene 1993;8:1195–202. 2. Christensen JG, Burrows J, Salgia R. c-Met as a target for human cancer and characterization of inhibitors for therapeutic intervention. Cancer Lett 2005;225:1–26. 3. Camp ER, Liu W, Fan F, Yang A, Somcio R, Ellis LM. RON, a tyrosine kinase receptor involved in tumor progression and metastasis. Ann Surg Oncol 2005;12:273–81. 4. Leonard EJ, Danilkovitch A. Macrophage stimulating protein. Adv Cancer Res 2000;77:139–67. 5. Chen YQ, Fisher JH, Wang MH. Activation of the RON receptor tyrosine kinase inhibits inducible nitric oxide synthase (iNOS) expression by murine peritoneal exudate macrophages: phosphatidylinositol-3 kinase is required for RON-mediated inhibition of iNOS expression. J Immunol 1998;161:4950–9. 6. Wang MH, Fung HL, Chen YQ. Regulation of the RON receptor tyrosine kinase expression in macrophages: blocking the RON gene transcription by endotoxininduced nitric oxide. J Immunol 2000;164:3815–21. 7. Zhou YQ, Chen YQ, Fisher JH, Wang MH. Activation of the RON receptor tyrosine kinase by macrophagestimulating protein inhibits inducible cyclooxygenase-2 expression in murine macrophages. J Biol Chem 2002; 277:38104–10. 8. Leonis MA, Toney-Earley K, Degen SJ, Waltz SE. Deletion of the Ron receptor tyrosine kinase domain in mice provides protection from endotoxin-induced acute liver failure. Hepatology 2002;36:1053–60. 9. Wang MH, Zhou YQ, Chen YQ. Macrophage-stimulating protein and RON receptor tyrosine kinase: potential regulators of macrophage inflammatory activities. Scand J Immunol 2002;56:545–53. 10. Chen YQ, Zhou YQ, Wang MH. Activation of the RON receptor tyrosine kinase protects murine macrophages from apoptotic death induced by bacterial lipopolysaccharide. J Leukoc Biol 2002;71:359–66. 11. Wang MH, Dlugosz AA, Sun Y, Suda T, Skeel A, Leonard EJ. Macrophage-stimulating protein induces Neutralizing Antibody to RON Receptor Tyrosine Kinase www.aacrjournals.org 9169 Cancer Res 2006; 66: (18). September 15, 2006