Specific tyrosine kinase receptors such as c-MET mediate epithelial-mesenchymal (EMT) transition, leading to phenotypic alterations associated with increased cell motility. It was hypothesized that RON, a tyrosine kinase receptor related to c-MET, would be expressed in human pancreatic cancer cells, induce EMT, and would thus serve as a target for therapy in a preclinical model.
RON expression in human pancreatic cancer specimens was assessed by immunohistochemistry. In pancreatic cancer cell lines, RON expression was assessed by reverse-transcriptase polymerase chain reaction (PCR) and Western blot analysis. The human pancreatic cancer cell line L3.6pl, with high RON expression, was exposed to macrophage stimulating protein (MSP), the RON ligand, and assessed for cell migration, invasion, and changes associated with EMT. Western blot analysis and immunofluorescent staining were used to assess alterations in protein expression and cellular location, respectively. A RON monoclonal antibody (MoAb) was used to block ligand-induced activation of RON.
Immunohistochemical staining revealed RON overexpression in 93% of human pancreatic cancer specimens relative to nonmalignant ductal tissue. RON mRNA and protein was expressed in 9 of 9 human pancreatic cancer cell lines. Treatment of L3.6pl cells with MSP increased Erk phosphorylation, cell migration, and invasion (P < .001). RON activation led to a decrease in membrane-bound E-cadherin in association with nuclear translocation of β-catenin. RON MoAb inhibited downstream signaling as well as cell migration and invasion. In nude mice, RON MoAb inhibited subcutaneous and orthotopic tumor growth by about 60%.
Despite systemic therapy, few patients with locally advanced or metastatic pancreatic cancer survive more than 2 years after diagnosis. Insights into the molecular mechanisms that mediate tumor progression and metastasis of pancreatic cancer may lead to the development of more effective antineoplastic therapies.
The tyrosine kinase receptors mediate multiple processes involved in tumor progression and metastasis, making them attractive targets for therapy. One such receptor, RON, is a member of the MET proto-oncogene family,1 and its ligand, macrophage stimulating protein (MSP), is secreted by the liver.2 RON activation mediates multiple signaling cascades involved in cell motility, adhesion, proliferation, and survival, including the c-Src, ras/mitogen-activated protein kinase (MAPK), phosphatidylinositol-3 kinase/Akt, and focal adhesion kinase pathways.1, 3–5
Epithelial-mesenchymal transition (EMT) is a process observed in embryonic development and tumorigenesis whereby cells lose epithelial characteristics and gain mesenchymal properties.6 EMT is characterized by loss of cellular adhesion and of normal cytoskeletal arrangements.6 Epithelial cells undergoing EMT exhibit mesenchymal features correlating with increased motility and invasion.6 EMT has also been implicated in tumor progression and metastasis formation and is associated with a worse clinical prognosis.7–9
In epithelial cells, activation of RON leads to increased cell invasion and migration, properties associated with EMT. Although RON has been studied in breast, colon, and ovarian cancers, where it has been shown to be overexpressed,10–12 there are no reports on the role of RON in pancreatic cancer. Determining the presence of RON and its function in pancreatic cancer may identify an important mediator of the aggressive behavior of pancreatic cancer and a potential target for molecular therapy. We hypothesized that the activation of RON, similar to c-MET, leads to EMT and that blockade of RON's activity with a monoclonal antibody (MoAb) inhibits tumor growth.
MATERIALS AND METHODS
Immunohistochemistry for RON
Twenty-five formalin-fixed, paraffin-embedded pancreatic cancer specimens, obtained from the archives of the Department of Pathology at M. D. Anderson Cancer Center, were sectioned 5–8 μm in thickness and the sections were placed on poly-L-lysine-coated slides, dewaxed in xylol, passed through a decreasing series of alcohol concentrations, and rehydrated in distilled water. The specimens were then incubated in 1× Target Retrieval Solution (Dako, Carpinteria, CA) for 20 minutes after heating to 95–99°C in a vegetable steamer. Endogenous peroxidase was blocked with peroxidase block from an EnVision+ Rabbit Kit (Dako) for 5 minutes. After protein block for 1 hour the specimens were incubated with the anti-RON-C20 antibody for 1 hour at room temperature followed by labeled polymer from an EnVision+ Rabbit Kit for 30 minutes at room temperature. Specimens were developed with diaminobenzidine and hydrogen peroxide for 6 minutes and lightly stained with hematoxylin. Slides were analyzed by a single pathologist (S.R.H.) to confirm the presence of pancreatic adenocarcinoma and were graded based on the degree of pancreatic intraepithelial neoplasia (PIN) present.13
Cell Lines and Culture Conditions
The human pancreatic cancer cell lines AsPC-1, BxPC-3, HPAF-2, MIAPACA2, Panc-1, HS7665, and CFPAC-1 were obtained from the American Type Culture Collection (ATCC, Manassas, VA). The mouse fibroblast cell line NIH3T3 and the human gastric fibroblast cell line Hs677.St were also obtained from the ATCC. The human pancreatic cancer cell lines FG and L3.6pl were kindly provided by I. J. Fidler, DVM, PhD (University of Texas M. D. Anderson Cancer Center, Houston, TX). Cell lines were cultured in minimal essential medium (MEM) supplemented with 10% fetal bovine serum (FBS), penicillin-streptomycin, vitamins, sodium pyruvate, L-glutamine, and nonessential amino acids (Life Technologies, Grand Island, NY) at 37°C in 5% CO2 and 95% air. Cells were confirmed to be free of Mycoplasma.
Reagents and Antibodies
Recombinant human MSP was purchased from R&D Systems (Minneapolis, MN) and used at a concentration of 100 ng/mL in all experiments. The RON MoAb 41A10 was supplied by ImClone Systems (New York, NY). This antibody was validated by an enzyme-linked immunosorbent assay (ELISA) with recombinant RON protein that demonstrated the antibody's affinity for RON.14 Similarly, the antibody was shown not to crossreact with c-Met (data not shown).
Antibodies utilized were as follows: polyclonal rabbit anti-RON-C20 (Santa Cruz Biotechnology, Santa Cruz, CA), polyclonal rabbit anti-phosphorylated-p44/42 MAPK, polyclonal rabbit anti-p44/42 MAPK, polyclonal rabbit anti-phosphorylated-Akt, polyclonal rabbit anti-Akt, polyclonal rabbit anti β-catenin, polyclonal rabbit anti-phosphorylated-p38, polyclonal rabbit anti-p38 (all from Cell Signaling Technology, Beverly, MA), rabbit anti-actin (Sigma-Aldrich, St. Louis, MO), polyclonal mouse anti-E-cadherin, polyclonal rabbit anti-N-cadherin (both from Zymed Laboratories, San Francisco, CA), polyclonal goat anti-vimentin (Chemicon International, Temecula, CA), Alexa Fluor 488 goat antimouse, and Alexa Fluor 594 goat antirabbit (both from Molecular Probes, Eugene, OR).
Reverse-Transcriptase Polymerase Chain Reaction and Real-Time Quantitative PCR
mRNA was extracted from cells in 1 mL of Trizol reagent (Life Technologies) according to the manufacturer's protocol and subjected to reverse-transcriptase polymerase chain reaction (PCR). For cDNA synthesis, 3 μg of mRNA was mixed with 50 U of avian myeloblastosis virus reverse transcriptase (Life Technologies), 0.5 M Tris-HCl (pH 8.0), 0.5 M KCl, 0.05 M MgCl2, 2.5 mM dNTP, 40 U of RNase inhibitor (Boehringer Mannheim, Indianapolis, IN), and 0.5 μg of random primer to achieve a final volume of 20 μL. The cDNA synthesis reaction was performed for 1 hour at 37°C. A portion of this reaction mixture (5 μL) underwent PCR amplification in a reaction mixture(50 μL) that contained 1 μmol/L of each of 2 primers (sense and antisense), 1.5 mmol/L of MgCl2, 0.2 mmol/L of each of 4 deoxynucleotides, and 2.5 U of Taq polymerase (Promega, Madison, WI). PCR amplification of RON was performed under the following conditions: 95°C for 5 minutes; 35 cycles of 30 seconds of denaturing at 95°C and 30 seconds of annealing at 60°C; and 5 minutes of extension at 72°C. PCR products were analyzed by electrophoresis of 20 μL of each PCR reaction mixture in a 1.5% agarose gel, and bands were visualized by ethidium bromide staining. For negative controls, the fibroblast cell lines NIH3T3 (murine) and Hs677.St (human) were used.
Real-time quantitative PCR (qRT-PCR) was performed for Snail and Twist, known transcription factors involved in EMT.15–18 cDNA synthesis was performed in the same manner as for PCR. cDNAs were diluted in 200 μL of diethylpyrocarbonate water and 5 μL of each reaction mixture was used in each 25-μL qRT-PCR reaction. Amplifications were performed in a GeneAmp 7000 sequence detector (Applied Biosystems, Foster City, CA) using a 2-step cycling parameter (95°C for 5 minutes followed by 40 cycles of 95°C for 15 seconds and 60°C for 120 seconds). PCR products were detected using SYBR Green I master mix (Applied Biosystems). Snail and Twist gene expression were normalized using reference primers for GAPDH. All reactions were performed in triplicate and the comparative cycle threshold methodology was used to determine relative changes in gene expression.19
For PCR, the following primers were used for RON: CATGGCATTTCATGGGCTGT and CAGACACTCAGTCCCATTGA; and for MSP: AATACCACCACTGCGGGCGT and TCAGTATCCACTGCTCCTTCA. For qRT-PCR the primers used for Snail were: CTAACTACAGCGAGCTGCAG and CTGTCAGATGAGGACAGTGG; for Twist: GGAGTCCGCAGTCTTACGAG and TCTGGAGGACCTGGTAGAGG; and for GAPDH: TCAAGGTCGGAGTCAACGGATTTGGT and CATGTGG CCCATGAGGTCCACCAC.
Assessing Morphologic Changes After Exposure to MSP
L3.6pl cells were grown in MEM with 10% FBS in a T25 flask. When cells reached 40% confluence, cells were placed in MEM with 1% FBS overnight. Cells were then exposed to MSP for 48 hours and digital images were obtained under light microscopy at ×10 magnification.
Western Blot Analysis
Cells were lysed in protein lysis buffer (20 mM sodium phosphate [pH 7.4], 150 mM sodium chloride, 1% Triton X-100, 5 mM EDTA, 5 mM phenylmethylsulfonyl fluoride, 1% aprotinin, 1 μg/mL leupeptin, and 500 μM Na3VO4). Proteins were subjected to electrophoresis on polyacrylamide gels and transferred to nylon membranes (Millipore, Billerica, MA) as previously described.20 After blocking with 5% milk in 0.1% Tween 20 in Tris-buffered saline, the membranes were probed with primary antibodies, washed, and treated with secondary antibodies labeled with horseradish peroxidase. Protein bands were visualized with a commercially available chemoluminescence kit (Amersham Biosciences, Piscataway, NJ).
Cells were grown on poly-L-lysine-coated glass coverslips in MEM with 10% FBS. At 40% confluence cells were placed in MEM with 1% FBS overnight. The following day cells were treated with MSP for 2, 6, or 24 hours and then fixed with acetone. After fixation cells were permeabilized by treatment with phosphate-buffered saline (PBS) containing 0.5% Triton X-100 and then blocked with 1% bovine serum albumin. Cells were incubated with the primary antibody for actin, E-cadherin, or β-catenin overnight at 4°C. The following day slides were washed, incubated with the corresponding Alexa Fluor secondary antibody, and stained with Hoechst 33342 (Molecular Probes) as a nuclear label. After the final washes and mounting, cells were examined using a laser scanning Olympus microscope.
Migration and Invasion Assays
Cell migration in response to MSP was assessed utilizing modified Boyden chambers according to the manufacturer's protocol (Becton Dickinson Labware, Bedford, MA). Briefly, 105 L3.6pl cells were incubated in MEM with 1% FBS on 8.0-μm pore size membrane inserts (Becton Dickinson Labware) in 24-well plates. Chemoattractants (MEM plus 1% FBS with or without MSP [100 ng/mL]) were placed in the bottom wells. In experiments using the RON MoAb the antibody was placed in the upper well 2 hours before the addition of MSP. At 6, 24, and 48 hours cells that had not migrated were removed from the top side of the inserts with a cotton swab. Cells that had migrated to the underside of the inserts were stained with Diff-Quik (Harleco, Gibbstown, NJ) and counted in 10 separate fields at ×100 magnification.
A similar protocol was used to assess cell invasion in response to MSP except that the membrane insert was coated with Matrigel and cells were treated for 24, 48, and 72 hours.
Animal Studies With RON MoAb
Male athymic nude mice were obtained from the National Cancer Institute Frederick Animal Production Area (Frederick, MD) and were acclimated for 2 weeks. All animal studies were conducted under guidelines approved by the Animal Care and Use Committee of the University of Texas M. D. Anderson Cancer Center. Under sterile conditions, L3.6pl cells (106) in Hank balanced salt solution (HBSS) (0.1 mL) were injected subcutaneously into the right flank of each mouse. Four days after mice were injected, 10 mice were randomly assigned to receive 1) PBS (control) by twice weekly intraperitoneal (IP) injections or 2) 41A10 (1 mg/mouse) by twice weekly IP injections. The mice were killed on Day 18, when tumors in the control group exceeded 1.5 cm in longest diameter. Tumors were excised and weighed and tumor sections were placed in 10% buffered formalin for paraffin fixation or in optimal cutting temperature compound (Miles, Elkhart, IN) with freezing in liquid nitrogen for frozen tissue sections. The tumor volume was calculated using the formula: tumor volume = ab2/2, in which a is the longest diameter and b is the shortest diameter of the tumor.
The results of our initial study using a subcutaneous tumor model were validated with an orthotopic pancreatic cancer experiment. L3.6pl cells (5 × 105 cells) in HBSS (0.05 mL) were injected directly into the tail of the pancreas through a left flank incision.21 Treatment groups and protocols were similar to the experiment described above.
Immunohistochemistry for Apoptotic Cells
Terminal deoxynucleotidyl transferase-mediated nick end-labeling (TUNEL) was performed on paraffin-embedded tumor sections using a commercially available apoptosis detection kit (Promega). Sections were examined using a Zeiss photomicroscope (Carl Zeiss, Thornwood, NY) equipped with a 3-chip charge-coupled color camera (DXC-960 MD; Sony, Tokyo, Japan). The images were analyzed using Optimas image analysis software (v. 5.2; Bothell, WA).
The differences between the means of the treatment groups were examined using InStat statistical software (GraphPad Software, San Diego, CA) using the Mann-Whitney U-test. Statistical analysis was performed using InStat statistical software (v. 2.03; GraphPad Software). Data were tested for outliers using the Grubb test (http://www.graphpad.com) for all experimental groups to avoid bias. Outliers were then removed from the analysis. P ≤ .05 was considered statistically significant.
RON Expressed in Human Pancreatic Cancer Specimens
Immunohistochemical analysis revealed that 24 (96%) of the 25 human pancreatic cancer specimens expressed RON. All slides evaluated had both normal pancreatic ductal tissue and pancreatic carcinoma present. One specimen had no invasive carcinoma, but PIN was present.13 The 1 tumor specimen that did not demonstrate RON staining had extensive areas of necrosis that potentially compromised the integrity of the tumor cells.
RON was expressed in only 14 (56%) of 25 nonmalignant pancreatic ductal tissue specimens. The pancreatic cancer tissue overexpressed RON relative to the matched nonmalignant pancreatic ductal tissue in 13 (93%) of the 14 cases in which RON was expressed in the nonmalignant tissue. Among the 11 cases in which RON expression was not observed in the normal tissue, RON was detected in the cancer tissue in 10 cases. Overall, RON expression was higher in the pancreatic neoplastic epithelium than in nonmalignant ducts in 23 cases (93%).
The RON staining intensity subjectively increased from normal tissue to PIN to invasive cancer (Fig. 1). The pattern of staining was heterogeneous, with both cytoplasmic and membranous staining evident. Ten tumors displayed both a membranous and cytoplasmic staining pattern, whereas the remaining 13 had only cytoplasmic staining for RON. Although the location of staining was heterogeneous, a consistent apical vesicle-like distribution was demonstrated, possibly indicating a lysosomal location. In the 8 specimens with PIN a similar vesicle-like pattern was observed. In general, the invasive cancer stained more intensely than the PIN, which stained more intensely than the normal tissue.
RON Expressed in Human Pancreatic Cancer Cell Lines
Reverse-transcriptase PCR analysis revealed RON mRNA expression in all 9 pancreatic cancer cell lines (Fig. 2A) (the MIAPACA2 cell line demonstrated only minimal RON mRNA expression). Western blot analysis confirmed the presence of RON in all 9 pancreatic cell lines (Fig. 2B). The highest expression of RON was observed in L3.6pl cells. Conversely, Panc-1 cells had minimal RON expression by Western blot analysis, similar to the findings by PCR analysis.
RON-Activated Downstream Signaling Mediators
To assess the downstream signaling mediators activated through RON, Western blot analysis was performed on L3.6pl cells exposed to MSP for various times. MSP led to marked phosphorylation of Erk (Thr202/Tyr204), with peak activation at 60 minutes (Fig. 3A). MSP also increased Akt (Ser473) phosphorylation after 60 minutes of MSP exposure. P38 phosphorylation was not induced by MSP in L3.6pl cells (data not shown). Pretreatment of cells with RON MoAb inhibited MSP induction of Erk phosphorylation in L3.6pl cells (Fig. 3B).
RON-Induced Morphologic Changes Associated With EMT
EMT is characterized by both a loss of cell-cell contact and a change in cell shape from a polarized cell to one that is more spindle-shaped. We assessed the morphologic changes in L3.6pl cells after MSP exposure for 48 hours. MSP led to altered cell morphology, with increased spindle-shaped cells representing loss of polarity, increased intercellular separation, and increased pseudopodia formation (Fig. 4). These morphologic alterations were confirmed by immunofluorescent staining for actin. In contrast to the highly organized actin filaments in the control cells, the cells treated with MSP demonstrate reorganization of the actin filaments with the presence of both pseudopodia and the formation of stress fibers (Fig. 4). Similar results were obtained in a second pancreatic cancer cell line (FG) (data not shown).
RON Activation Enhanced Pancreatic Cancer Cell Migration and Invasion
MSP led to an increase in migration of L3.6pl cells at all times evaluated (6, 24, and 48 hours). A 4-fold increase in migration was demonstrated after 48 hours of MSP treatment (P < .001) (data not shown).
Similar to the findings with cell migration, MSP led to an increase in invasion by L3.6pl cells at all times evaluated (24, 48, and 72 hours). A greater than 4-fold increase in invasion was demonstrated after 24 hours of MSP treatment (P < .001) (data not shown).
The migration and invasion assays were repeated with the addition of the RON MoAb to block RON activation before exposure of cells to MSP. Pretreatment with RON MoAb inhibited migration by 30% (P< .001) and invasion by 80% (P < .001) at 24 hours (Fig. 5). Similar findings were demonstrated in the FG and CFPAC-1 cell lines (data not shown).
RON Activation Altered Expression of E-cadherin
Because EMT is characterized by loss of both cellular adhesion and normal cytoskeletal arrangements, we first attempted to demonstrate decreased expression of E-cadherin with MSP treatment. As shown by Western blot analysis, exposure of L3.6pl cells to MSP for 72 hours reduced expression of E-cadherin (Fig. 6A). In contrast, MSP increased the expression of vimentin relative to control cells at 72 hours (Fig. 6A), with minimal effect on N-cadherin.
To further explore the altered expression of E-cadherin, immunofluorescent staining was performed at 2, 6, and 24 hours after exposure to MSP to assess the cellular localization of E-cadherin. In the control cells, E-cadherin was associated with the cell membrane in regions of cell-cell contact. In contrast, MSP decreased membrane-associated E-cadherin (Fig. 6B), and this loss of membrane association increased in a time-dependent manner. Maximal changes were demonstrated at 24 hours.
Recently, the transcription factors Snail and Twist have been shown to down-regulate the expression of E-cadherin.15–18 Exposure of L3.6pl cells to MSP for 24 hours induced an 18-fold increase in induction of Snail and a 2.5-fold increase in induction of Twist as determined by qRT-PCR (data not shown).
RON Activation Led to β-Catenin Nuclear Translocation
Western blot analysis of nuclear protein extracts demonstrated that MSP treatment increased nuclear levels of β-catenin in a time-dependent fashion (Fig. 6C). The nuclear translocation of β-catenin was confirmed by immunofluorescent staining of cells (Fig. 6D). In the untreated L3.6pl cells, β-catenin was associated with the cell membrane in regions of cell-cell contact, similar to the results observed for E-cadherin. After exposure to MSP for 24 hours the membrane-associated β-catenin decreased, with translocation of β-catenin to the nucleus.
RON MoAb Affected Growth of Subcutaneous Pancreatic Tumors
Twice weekly IP therapy with RON MoAb decreased the volume of subcutaneously implanted pancreatic cancers by about 60% relative to control treatment with PBS (Fig. 7A). Tumor weight changes (data not shown) were similar to tumor volume changes. The tumors treated with RON MoAb demonstrated a 1.3-fold increase in apoptosis compared with the control group by TUNEL staining (P = .03) (data not shown).
Effect of RON MoAb on the Growth of Orthotopic Pancreatic Tumors
Similar results were obtained in the orthotopic pancreatic tumor model. RON MoAb decreased the volume of the orthotopic tumors by about 40% compared with control tumors treated with PBS (P = .01) (data not shown). Tumor weight changes (data not shown) were similar to tumor volume changes.
EMT is a morphologic alteration in epithelial cells that generates a more motile phenotype similar to that characteristic of mesenchymal cells. Because essential steps in tumor progression and metastasis are dependent on tumor cell motility, EMT represents a fundamental process in the progression of cancer. Growth factors such as epidermal growth factor and hepatocyte growth factor have been implicated in mediating tumor progression and growth, and these same factors mediate EMT.22, 23
Our studies demonstrated 1) the presence of increased RON expression in pancreatic neoplasia (both invasive cancer and PIN); 2) RON's ability to induce EMT in pancreatic cancer cells; and 3) the ability of a RON MoAb to inhibit tumor cell signaling, migration, invasion, and growth. We demonstrated that RON activation induced EMT and increased migration and invasion of pancreatic cancer cells. Exposure of pancreatic cancer cells to the RON ligand MSP led to loss of membrane-bound E-cadherin in association with nuclear translocation of β-catenin. The loss of membrane-bound E-cadherin corresponded to morphologic changes in the cancer cells, with loss of cell-cell contact and increased motility. These alterations in protein expression and cellular activity stimulated by RON activation are consistent with EMT.
Immunohistochemical staining revealed that RON was overexpressed in 93% of the pancreatic cancer specimens relative to matched nonmalignant pancreatic tissues. PIN (the precursor to invasive pancreatic cancer) also demonstrated high RON expression, suggesting that RON is overexpressed early in the progression of malignancy.
Investigations in breast, colon, and ovarian cancer demonstrated RON overexpression and increased cellular motility similar to our results in pancreatic cancer.10–12 To our knowledge, this is the first report demonstrating the presence of RON in pancreatic cancer and also in premalignant PIN.
The finding that RON exhibits a functional role in pancreatic cancer by inducing EMT combined with its high expression in pancreatic cancer specimens and cell lines suggests that RON represents a potential therapeutic target. Using the RON MoAb, we inhibited MSP-induced downstream signaling and migration and invasion of cancer cells in vitro. In vivo, the RON MoAb inhibited tumor growth by about 60% in a subcutaneous model of pancreatic cancer. These results were confirmed in an orthotopic model of pancreatic cancer, where the RON MoAb led to about a 40% inhibition of tumor growth. No metastasis were observed in any of the treatment groups, possibly due to terminating the experiment only after 18 days (based on the rapid local growth of the implanted tumors, and to remain in compliance with animal protocols). The ability of the RON MoAb to inhibit downstream signaling mediators such as pERK may help to explain the antibody's mechanism of action. pERK mediates pathways related to invasion, migration, cell growth, and survival. The loss of RON's ability to regulate pERK may contribute to the increased cellular apoptosis and decreased tumor growth observed in the in vivo experiments with the RON MoAb.
The ability of RON to induce EMT has widespread implications relating to the role of RON in tumor progression and metastasis. In pancreatic cancer, RON activation may induce early invasion as well as enhance metastasis formation through EMT. Improved understanding of EMT may lead to identification of potential molecular targets for cancer therapy as well as increase our knowledge of critical processes in cancer progression. Ultimately, RON may serve as a molecular target for pancreatic cancer therapy owing to its high expression and function in cancer cells.