Vascular endothelial growth factor receptor-1 promotes migration and invasion in pancreatic carcinoma cell lines

Abstract

BACKGROUND

Vascular endothelial growth factor receptor-1 (VEGFR-1) is one of three receptor tyrosine kinases for VEGF, a key regulator of angiogenesis in cancer. Although VEGFRs initially were believed to be expressed exclusively on endothelial cells (ECs), recent studies have demonstrated the presence of VEGFR-1 on non-EC types. The authors hypothesized that VEGFR-1 is present and functional in pancreatic carcinoma cells, contributing to the malignant phenotype.

METHODS

The authors assessed the expression of VEGFR-1 and its ligands in 11 pancreatic carcinoma cell lines by reverse-transcriptase–polymerase chain reaction, enzyme-linked immunosorbent assay, and/or Western blot analysis. The function of VEGFR-1 was evaluated by treating two representative cell lines with VEGF-B, a selective ligand for VEGFR-1, and/or a specific anti-VEGFR-1 antibody and assessing the effects on signaling, migration, invasion, and proliferation.

RESULTS

All 11 pancreatic carcinoma cell lines expressed VEGFR-1 mRNA and protein, as well as the VEGFR-1 ligands VEGF-A and VEGF-B. Two representative cell lines (L3.6 and Panc-1) exhibited VEGF-B-induced mitogen-activated protein kinase signaling. A VEGFR-1 neutralizing antibody abrogated signaling, confirming that the ligand effect was mediated through VEGFR-1. VEGFR-1 stimulation by VEGF-A or VEGF-B was found to promote migration in both cell lines. Panc-1 cells also demonstrated enhanced Matrigel invasion after VEGFR-1 stimulation. VEGFR-1-dependent migration and invasion were blocked by the VEGFR-1 neutralizing antibody. VEGFR-1 activation did not appear to enhance cell proliferation.

CONCLUSIONS

VEGFR-1 appears to be expressed ubiquitously in pancreatic carcinoma cell lines, in which it induces signaling and promotes migration and invasion. Overexpression of VEGF in tumors may activate tumor cells bearing VEGFR-1 via an autocrine pathway. Agents targeting VEGF or its receptors may have a dual inhibitory effect on tumor growth by suppressing both angiogenesis and tumor cell function. Cancer 2005. © 2005 American Cancer Society.

Vascular endothelial growth factor (VEGF) is a proangiogenic factor with roles in both physiologic and pathologic angiogenesis.1 Also known as VEGF-A, VEGF is the prototype member of a protein family that includes VEGF-B, VEGF-C, and VEGF-D, as well as placental growth factor (PlGF). VEGF's many effects include induction of endothelial cell (EC) proliferation and migration2 and an increase in microvascular permeability.3 VEGF interacts with two receptor tyrosine kinases, VEGFR-1 (Flt-1) and VEGFR-2 (KDR and the murine homologue Flk-1),4, 5 as well as with neuropilin-1 and neuropilin-2, which are believed to function as coreceptors for the VEGFRs.6–8 The majority of VEGF's effects are mediated through VEGFR-2; to our knowledge the role of VEGFR-1 is not as well defined.

VEGFR-1 was initially believed to play a relatively minor role in VEGF-mediated signal transduction because none of the traditional effects of VEGF (such as EC proliferation) were noted in response to stimulation of VEGFR-1.9 Furthermore, early investigations of the function of VEGFR-1 suggested that it was a decoy receptor for VEGF, serving as a negative regulator of VEGF.2, 10 Knockout studies targeting VEGFR-1 in mice, however, revealed that loss of VEGFR-1 function led to early embryonic death with disorganized blood vessels and EC overgrowth.11 This study implicated VEGFR-1 as a critical mediator of physiologic and developmental angiogenesis. Subsequent studies also supported functional roles for VEGFR-1 in the induction of monocyte migration,12, 13 recruitment of EC progenitors,14 adhesion of natural killer cells to ECs,15 and induction of hepatocyte growth factors from liver sinusoidal ECs.16

VEGFRs have traditionally been described as being expressed by ECs; however, recent studies have demonstrated that VEGFRs are also expressed in non-ECs. Given the current interest in VEGF as a target for cancer treatment and recent reports that anti-VEGF therapy enhances the effect of standard chemotherapy in patients with metastatic colorectal carcinoma,17 elucidating the possible roles of VEGFRs on tumor cells is of increasing importance. We hypothesized that VEGFR-1 is present and functional on pancreatic carcinoma cell lines. To test this hypothesis, reverse-transcriptase–polymerase chain reaction (RT-PCR) and/or Western blot analysis were utilized to demonstrate mRNA and/or protein expression of VEGFR-1 and its ligands in all pancreatic carcinoma cell lines tested. Stimulation with the VEGFR-1 ligand VEGF-B led to mitogen-activated protein kinase (MAPK) signaling, migration, and invasion, all of which were inhibited by a neutralizing VEGFR-1 antibody. These findings suggest that circulating VEGF family members may not only promote tumor angiogenesis by activating VEGFRs on ECs but also have an autocrine effect on tumor cells bearing VEGFRs.

MATERIALS AND METHODS

Materials

Antibodies for Western blot analysis were as follows: polyclonal rabbit antiphospho-Erk-1/2 MAPK (Thr202/Tyr204) (Cell Signaling Technology, Beverly, MA), polyclonal rabbit antiphospho-Akt (Ser473) (Cell Signaling Technology), polyclonal rabbit anti-MAPK (Oncogene Research Products, Cambridge, MA), polyclonal rabbit anti-Akt (Cell Signaling Technology), polyclonal goat anti-VEGFR-1 (Oncogene Research Products), monoclonal mouse anti-VEGF-B (R&D Systems, Minneapolis, MN), monoclonal mouse antivinculin (Sigma Chemical Company, St. Louis, MO), human immunoglobulin G (IgG) (Jackson ImmunoResearch, West Grove, PA), horseradish peroxidase (HRP)-conjugated sheep antimouse IgG (Amersham Biosciences, Piscataway, NJ), HRP-conjugated rabbit antigoat IgG (Jackson ImmunoResearch), and HRP-conjugated goat antirabbit IgG (Bio-Rad Laboratories, Hercules, CA). Recombinant human VEGF-A, VEGF-B, transforming growth factor-β1, tumor necrosis factor-α, and epidermal growth factor were obtained from R&D Systems. Humanized VEGFR-1 neutralizing antibody (18F1) was provided by ImClone Systems Incorporated (New York, NY). Latrunculin B was obtained from EMD Biosciences (San Diego, CA). Other chemicals utilized were purchased from Sigma Chemical Company unless otherwise indicated.

Cell Lines and Culture Conditions

Human umbilical vein endothelial cells (HUVECs) and pancreatic carcinoma cell lines (AsPC-1, BxPC3, CFPAC, HPAF2, MiaPaCa2, and Panc-1) were purchased from American Type Culture Collection (Manassas, VA). HS7665 and Panc-48 were kindly provided by D. J. McConkey, Ph.D., of The University of Texas M. D. Anderson Cancer Center in Houston. The L3.6pl, FG, and pancreatic carcinoma cell lines were kindly provided by I. J. Fidler, Ph.D., D.V.M., also of the The University of Texas M. D. Anderson Cancer Center. Cells were cultured and maintained in minimal essential medium (MEM) supplemented with 10% fetal bovine serum (FBS), 2 U/mL of a penicillin-streptomycin mixture (Flow Laboratories, Rockville, MD), vitamins (Life Technologies, Inc., Grand Island, NY), 1 mM of sodium pyruvate, 2 mM of L-glutamine, and nonessential amino acids and incubated in 5% carbon dioxide and 95% air at 37 °C. In vitro experiments were performed at 60–80% cell confluence, and cells were used at passages 3–15 after their receipt from the supplier.

Enzyme-Linked Immunoadsorbent Assay for VEGFR-1 and VEGFR-2

Pancreatic carcinoma cell lines (1 × 10⁶) were seeded into 10-cm dishes containing complete medium with 10% FBS. Conditioned medium was collected after 48 hours, and lysates were prepared according to the manufacturer's directions using lysis buffer from R&D Systems. Enzyme-linked immunosorbent assay (ELISA) was performed on the cell lysates for VEGFR-1 and VEGFR-2 using kits from R&D Systems according to the manufacturer's directions.

RT–PCR

RT–PCR was performed as previously described.18 Briefly, total RNA was extracted from human pancreatic carcinoma cell lines using TriZOL reagent (Invitrogen Life Technologies, Carlsbad, CA) according to manufacturer's directions. cDNA synthesis was performed by heating 3 μg of total RNA with 0.5 μg of random primers and dNTP (final concentration of 0.5 mM) at 65 °C for 5 minutes. Reverse transcription was then performed at 42 °C for 1 hour with this mixture in a final volume of 20 μL containing 0.5 M of Tris-HCl (pH 8.0), 0.5 M of potassium chloride, 0.05 M of MgCl₂, 40 U of RNase inhibitor (Boehringer–Mannheim Biochemicals, Indianapolis, IN), 200 U of SuperScript II reverse transcriptase (Invitrogen), and 10 mM of dithiothreitol in RNase-free water. PCR amplification of VEGFR-1 and VEGFR-2, VEGF-B, and PlGF was performed on 2 μL of the synthesized cDNA under the following conditions: 95 °C for 5 minutes, 35 cycles of denaturing for 30 seconds at 95 °C, 1 minute of annealing at 57 °C, and 1 minute of extension at 72 °C, followed by a terminal 10-minute extension at 72 °C. Products were analyzed by electrophoresis of 10 μL of each PCR reaction mixture in a 1% agarose gel, and bands were visualized by staining with ethidium bromide.

The following primers were used. VEGFR-1; sense primer: 5′–TGGGACAGTAGAAAGGGCTT-3′ and anti-sense primer: 5′–GGTCCACTCCTTACACGACAA-3′; VEGFR-2: sense primer: 5′-CATCACATCCACTGGTATTGG-3′ and anti-sense primer: 5′-GCCAAGCTTGTACCATGTGAG-3′; VEGF-B19: sense primer: 5′-CCTTGACTGTGGAGCTCATG-3′ and anti-sense primer: 5′-TGTCTGGCTTCACAGCACTG-3′. RT-PCR for PlGF was performed using a kit from R&D Systems according to the manufacturer's directions.

Western Blot Hybridization for VEGFR-1

Eleven pancreatic carcinoma cell lines and HUVEC (positive control) were lysed with RIPA “B” protein lysis buffer (20 mM of sodium phosphate [pH 7.4], 150 mM of sodium chloride, 1% Triton X-100, 5 mM of ethylenediamine tetraacetic acid, 5 mM of phenylmethylsulfonyl fluoride, 1% of aprotinin, 1 μg/mL of leupeptin, and 500 μM of Na₃VO₄). Protein was quantitated by a commercially available modified Bradford assay (Bio-Rad Laboratories). Western blot samples were prepared by boiling 50 μg of protein with denaturing sample buffer. The protein was separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) on a 6% denaturing polyacrylamide gel and transferred to a polyvinylidene fluoride (PVDF) membrane (Millipore Corporation, Bedford, MA) as previously described.20 After blocking with 5% nonfat dry milk in Tris-buffered saline and 0.1% Tween 20 (TBS-T), the membranes were probed with goat anti-VEGFR-1 antibody at a 1:500 dilution in 5% milk and TBS-T overnight at 4 °C. The membranes then were washed and incubated with HRP-conjugated rabbit antigoat antibody at a 1:3000 dilution in 5% milk and TBS-T for 1 hour at room temperature. Membranes were washed, and protein bands were visualized using a commercially available enhanced chemiluminescence kit (ECL) (Amersham Biosciences). To verify accuracy of loading, membranes were washed in stripping solution (100 mM of 2-mercaptoethanol, 2% SDS, and 62.5 mM of Tris-HCL [pH 6.7]) for 30 minutes at 65 °C and reprobed with antivinculin antibody as a loading control.

Evaluation of VEGFR-1 Ligand Expression in Pancreatic Carcinoma Cell Lines

Western blot analysis for VEGF-B was also performed in 11 pancreatic carcinoma cell lines and HUVEC as described earlier. Briefly, 50-μg samples of protein from cell lysates were separated by 12% SDS-PAGE. The membranes were probed with monoclonal mouse anti-VEGF-B antibody at a 1:500 dilution in 5% milk and TBS-T. The membranes were washed and incubated with HRP-conjugated sheep antimouse antibody at a 1:3000 dilution in 5% milk and TBS-T.

ELISA for human VEGF-A was performed on conditioned medium and fresh medium from pancreatic carcinoma cell lines according to the manufacturer's directions using an human VEGF-A kit (R&D Systems).

Effect of VEGFR-1 Stimulation on Signaling

Cells were grown to 60–70% cell confluence and incubated overnight in medium containing 5% FBS. The medium was changed to medium containing 1% FBS for 1 hour. Cytokines then were added (VEGF-A at 10 ng/mL or VEGF-B at 50 ng/mL) for 3 minutes, 10 minutes, or 30 minutes. Cells were lysed with RIPA “B” buffer, and 50-μg protein samples were subjected to Western blot analysis by SDS-PAGE on a denaturing 10% gel. Activated signaling pathways were identified by incubation with polyclonal rabbit antiphospho-Erk-1/2 antibody at a 1:1000 dilution in 5% milk and TBS-T. To verify accuracy of loading, membranes were stripped as described earlier and reprobed with polyclonal rabbit anti-MAPK antibody at a 1:1000 dilution in 5% milk and TBS-T. Detection by ECL was as described above. Akt and JNK/SAPK signaling were assessed by stripping the membranes and reprobing with antiphospho-Akt and antiphospho-JNK antibodies in a similar fashion.

To assess the effect of a VEGFR-1 neutralizing antibody, 18F1, on VEGFR-1-dependent signaling, cells were plated as described earlier. After the medium was changed to 1% FBS for 1 hour, 18F1 was added at 20 μg/mL. Control cells received no antibody. After incubation for an additional hour, cytokines were added as described earlier, and Western blot analysis was performed to identify signaling intermediates.

Effect of VEGFR-1 Activation on Migration and Invasion

Evaluation of cell migration utilized control inserts with 8.0-μm pores (BD Biosciences Discovery Labware, Bedford, MA) in 24-well plates. After detaching the cells in 10% trypsin and counting, cells were diluted to 60,000 cells/mL (Panc-1) or 100,000 cells/mL (L3.6) in medium containing 1% FBS. A total of 500 μL of cells (30,000 Panc-1 cells or 50,000 L3.6 cells) then were placed in the top chamber of the insert above medium containing 10% FBS with or without cytokines (VEGF-A at 10 ng/mL or VEGF-B at 50 ng/mL). After 6 hours (Panc-1) or 24 hours (L3.6pl), the cells on the upper surface of the membrane were removed with a cotton swab and the migrated cells on the underside of the membrane were fixed and stained using Diff-Quik (American Scientific Products, McGraw Park, IL) and counted in 10 random fields at ×100 magnification. Results are expressed as the means ± the standard errors of the mean (SEM).

To confirm that activation of VEGFR-1 was the pathway involved in the above studies, the VEGFR-1 was blocked with 18F1 prior to the addition of cytokines. For these studies, cells were preincubated overnight with 18F1 (20 μg/mL) prior to being detached and counted. Cells also were preincubated with 18F1 (20 μg/mL) for 10 minutes prior to transfer to the top level of the insert. Studies of invasion were performed as described earlier, except the membranes utilized were Matrigel™-coated invasion chambers (BD Biosciences, San Jose, CA) that were prehydrated in serum-free medium. Invasion studies with Panc-1 were performed for 18 hours prior to the fixing and staining of cells.

Effect of VEGFR-1 Activation on Cell Proliferation and Chemosensitivity

L3.6 or Panc-1 cells (3 × 10³) were plated in 96-well plates. After cell attachment for 24 hours, the medium was changed to complete medium containing 10% FBS containing cytokines (VEGF-A at 10 ng/mL or VEGF-B at 50 ng/mL) or antibody (18F1 at 20 μg/mL or human IgG at 20 μg/mL) for 10 minutes. Medium containing various concentrations of gemcitabine (reconstituted in phosphate-buffered saline) then was added, and cells were incubated for 24–48 hours. To detect surviving and proliferating cells, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) was added to a final concentration of 0.5 mg/mL and the cells were incubated for 90 minutes. Medium and MTT were removed, dimethyl sulfoxide was added for 1 minute, and absorption was read at 570 nanometers (nm). Proliferation in response to cytokines was assessed by comparing the absorption of treated cells with that of untreated cells. Cell survival was calculated as the percentage of cells surviving compared with control untreated cells.

Effect of VEGFR-1 Activation on Colony Formation in Soft Agar

One milliliter of complete medium containing 10% FBS and 1% SeaPlaque® low-melting-point agarose (Cambrex Corporation, East Rutherford, NJ) was plated in each well of six-well plates. Panc-1 cells were plated and treated with or without 18F1 (20 μg/mL) or control human IgG (20 μg/mL) for 1 hour. Cytokines (none/control, VEGF-A at 10 ng/mL, or VEGF-B at 50 ng/mL) then were added for 4 hours. Cells were trypsinized, and 500 cells/mL were plated in 1 mL of medium containing 0.5% agarose and the appropriate antibodies and/or cytokines on top of the 1% agarose base layer. Additional medium containing cytokines and/or antibodies was added biweekly to keep the agarose hydrated. Cells were allowed to grow for 20 days, after which colonies greater than 100 μm were counted using light microscopy.

Evaluation of Density-Dependent VEGFR-1 Expression

Panc-1 or L3.6pl cells were plated at 0.125 × 10⁶ cells, 0.25 × 10⁶ cells, 0.5 × 10⁶ cells, 1.0× 10⁶ cells, 2.5 × 10⁶ cells, or 5 × 10⁶ cells per 10-cm dish. After growth overnight, the cells were lysed in RIPA “B,” and 50-μg protein samples were utilized for Western blot analysis as described earlier. Membranes then were stripped and reprobed with antivinculin antibody as a loading control.

Statistical Analyses

All statistical analyses were performed with InStat Statistical Software (V2.03; GraphPad Software, San Diego, CA) with P values of less than 0.05 considered to be statistically significant.

RESULTS

Pancreatic Carcinoma Cell Lines Express VEGFR-1 but Not VEGFR-2

Screening by RT-PCR revealed strong VEGFR-1 mRNA expression in 5 of 11 pancreatic carcinoma cell lines tested, with weak expression found in 6 of 11 pancreatic carcinoma cell lines (data not shown). ELISA of pancreatic carcinoma cell lysates demonstrated VEGFR-1 protein levels between 3–366 pg/10⁶ cells in 6 of 9 pancreatic carcinoma cell lines tested. However, Western blot analysis confirmed VEGFR-1 protein expression in 11 of 11 pancreatic carcinoma cell lines (Fig. 1A). Because VEGF-A is also a ligand for VEGFR-2, we evaluated pancreatic carcinoma cells for VEGFR-2 expression. VEGFR-2 mRNA was detected in only 1 of the 11 pancreatic carcinoma cell lines by RT-PCR (Fig. 1B); however, we were unable to detect VEGFR-2 protein expression in this cell line (CFPAC) using ELISA.

Figure 1. Vascular endothelial growth factor receptor (VEGFR) expression in pancreatic carcinoma cell lines. (A) Western blot analysis demonstrating VEGFR-1 (180 kilodaltons [kD]) expression in 11 of 11 cell lines. Lysate from human umbilical vein endothelial cells (HUVECs) was included as a positive control. Vinculin is shown as a loading control. (B) Reverse transcriptase–polymerase chain reaction analysis of 11 pancreatic carcinoma cell lines demonstrated only minimal expression of VEGFR-2 in 1 of 11 pancreatic carcinoma cell lines (note that VEGFR-2 protein was not detectable in this cell line using enzyme-linked immunosorbent assay). HUVEC was utilized as a positive control. bp: base pairs.

Pancreatic Carcinoma Cell Lines Express VEGFR-1 Ligands

Having demonstrated the presence of VEGFR-1 on pancreatic carcinoma cells, we sought to determine whether these cells also produced the ligands for VEGFR-1. VEGFR-1 binds several of the VEGF family members. VEGF-A binds to both VEGFR-1 and VEGFR-2, whereas VEGF-B21 and PlGF13 are unique ligands for VEGFR-1. Western blot analysis for VEGF-B demonstrated expression of this ligand in 11 of 11 pancreatic carcinoma cell lines (Fig. 2). Expression of PlGF was demonstrated in selected cell lines (L3.6pl and Panc-1) by RT-PCR (data not shown). VEGF-B expression also was verified in the conditioned medium of these two cell lines by Western blot analysis (data not shown). VEGF-A expression was verified by ELISA in all pancreatic carcinoma cell lines tested. Levels of VEGF-A in conditioned medium ranged between 7–20 μg/10⁶ cells. No human VEGF-A was detectable in standard culture medium containing 10% FBS.

Figure 2. Vascular endothelial growth factor receptor-B (VEGF-B) protein expression in 11 pancreatic carcinoma cell lines. Western blot analysis of lysates from pancreatic cell lines demonstrated expression of VEGF-B₁₆₇ (36 kilodaltons [kD]) in 11 of 11 pancreatic carcinoma cell lines. Lysate from human umbilical vein endothelial cells (HUVECs) was included as a positive control.

VEGFR-1 Induces Signaling in Pancreatic Carcinoma Cell Lines

Having demonstrated that pancreatic carcinoma cell lines express VEGFR-1 and its ligands, we examined the functionality of VEGFR-1 on these cells by stimulation with VEGFR-1 ligands and assessment of signaling activation. We chose two cell lines that were well characterized in our laboratory for further studies. VEGF-A induced Erk-1/2 phosphorylation in both L3.6p1 and Panc-1 cells (data not shown). This VEGF-induced signaling cascade was mediated via activation of VEGFR-1 because none of the cell lines in our studies expressed VEGFR-2. Nevertheless, we confirmed that VEGFR-1 mediated signaling by utilizing the VEGFR-1-specific ligand VEGF-B. Activated Erk-1/2 signaling was again observed in both L3.6pl and Panc-1 cells, with maximal stimulation at 3 minutes that decreased to baseline levels by 30 minutes (Fig. 3). No activation of Akt or JNK/SAPK signaling was detected (data not shown).

Figure 3. Vascular endothelial growth factor receptor-1 (VEGFR-1)-induced signaling. (A) L3.6pl and (B) Panc-1 pancreatic carcinoma cells were stimulated with VEGF-B (50 ng/mL) for 3 minutes, 10 minutes, or 30 minutes. Western blot analysis for phosphorylated Erk-1/2 demonstrated increased phosphorylation at 3 minutes; this decreased to the baseline level by 30 minutes. The treatment of cells for 1 hour with the neutralizing VEGFR-1 antibody 18F1 (20 μg/mL) prior to treatment with VEGF-B abrogated the increase in Erk-1/2 phosphorylation. Total Erk-1/2 levels are shown as a loading control.

To verify further the specificity of VEGFR-1-induced signaling, the VEGFR-1 neutralizing antibody 18F1 was utilized. Preincubation of cells with 18F1 blocked VEGF-B-induced Erk-1/2 phosphorylation (Fig. 3), confirming that the signaling was VEGFR-1 dependent.

VEGFR-1 Promotes Migration in Pancreatic Carcinoma Cell Lines

Having determined that VEGFR-1 is present and capable of inducing signaling in pancreatic carcinoma cell lines, we assessed its effects on other functions involved in tumor progression and metastasis. In a modified Boyden chamber assay, a twofold increase in the migration of Panc-1 cells (P < 0.0001) was observed in response to VEGF-A and VEGF-B after 6 hours (Fig. 4). The addition of the neutralizing VEGFR-1 antibody 18F1 was found to inhibit this increase in migration, confirming that the process was VEGFR-1 dependent. A similar increase in migration was observed after L3.6pl cells were treated with either VEGF-A or VEGF-B. The addition of 18F1 also abrogated the enhanced migration in L3.6pl cells (data not shown), confirming that the VEGFR-1–dependent migration was not unique to Panc-1 cells.

Figure 4. Vascular endothelial growth factor receptor-1 (VEGFR-1)-dependent migration of Panc-1 cells. Panc-1 cell migration was assessed using a modified Boyden chamber. Medium containing 10% fetal bovine serum with or without VEGF-A or VEGF-B was used as a chemoattractant in the bottom chamber, and trans-well migration was assessed after 6 hours. Both VEGF-A and VEGF-B induced at least a twofold increase in migration compared with the control. The increased migration was blocked by pretreatment with the neutralizing VEGFR-1 antibody 18F1. Data shown are the means ± the standard error of the mean of three separate experiments. Representative photomicrographs of migrated cells at a ×100 magnification are shown. *P < 0.0001. HPF: high-power field.

VEGFR-1 Promotes Invasion in Panc-1 Cells

Because the activation of VEGFR-1 appeared to promote the migration of pancreatic carcinoma cells, we assessed the effect of VEGFR-1 on invasion. In Matrigel-coated Boyden chambers, Panc-1 cells exhibited a five-fold increase in invasion after treatment with VEGF-A or VEGF-B compared with control (10% MEM) (P < 0.0001 at 18 hours) (Fig. 5). This increased invasion also was inhibited by pretreatment with 18F1, confirming that the process was VEGFR-1 dependent. L3.6pl cells were not found to invade, regardless of the chemoattractant utilized (data not shown); therefore, the effects of activation or blockade of VEGFR-1 on invasion could not be evaluated in this cell line.

Figure 5. Vascular endothelial growth factor receptor-1 (VEGFR-1)-dependent invasion by Panc-1 cells. Panc-1 cell invasion was assessed using modified Boyden chambers coated with Matrigel. Complete minimal essential medium with 10% fetal bovine serum with or without VEGF-A or VEGF-B was utilized as a chemoattractant in the bottom chamber, and invasion was assessed after 18 hours. Both VEGF-A and VEGF-B induced a fivefold increase in invasion compared with the control. Pretreatment with the neutralizing VEGFR-1 antibody 18F1 abrogated this effect. Representative photomicrographs of invading cells at ×100 migration are shown. Similar results were obtained in two separate experiments. *P < 0.0001. HPF: high-power field.

VEGFR-1 Does Not Promote Proliferation or Survival in Pancreatic Carcinoma Cell Lines

Because VEGF-A is known to induce the proliferation of ECs through VEGFR-2, we assessed the effect of VEGF-B on the proliferation of Panc-1 and L3.6p1 cells. In contrast to the effects of VEGF-A on ECs, neither VEGF-A nor VEGF-B was found to induce the proliferation of Panc-1 or L3.6pl cells (data not shown). Blockade of VEGFR-1 using 18F1 also did not appear to affect proliferation.

Some of our previous studies of VEGFs in pancreatic carcinoma cell lines had demonstrated expression of the VEGF coreceptor neuropilin-1 and its promotion of chemoresistance and anchorage-independent growth in pancreatic carcinoma cells (unpublished data). We hypothesized that VEGFR-1 has similar effects in pancreatic carcinoma cells. Because clinical studies have shown that gemcitabine is effective therapy (albeit minimally) in patients with pancreatic carcinoma,22 we utilized this agent for in vitro chemosensitivity assays. However, in L3.6pl and Panc-1 cells, neither the activation of VEGFR-1 by the addition of ligands nor the blockade of VEGFR-1 by the addition of a neutralizing antibody appeared to affect the cells' sensitivity to gemcitabine using the MTT assay or their capability for anchorage-independent growth, as assessed by soft agar colony formation (data not shown).

VEGFR-1 Expression is Cell Density Dependent in Pancreatic Carcinoma Cell Lines

During the course of our studies of VEGFR-1 in pancreatic carcinoma cell lines, an association between VEGFR-1 protein expression level and cell density was observed. To confirm this observation, pancreatic carcinoma cell lines were cultured at different densities, ranging from no cell-cell contact to 100% confluency, and VEGFR-1 protein expression was assessed using Western blot analysis. At low confluency, cells expressed low levels of VEGFR-1 protein, but VEGFR-1 levels were found to increase in a density-dependent fashion in both Panc-1 cells (Fig. 6) and L3.6pl cells (data not shown).

Figure 6. Density-dependent regulation of vascular endothelial growth factor receptor-1 (VEGFR-1). Plating of 0.125 × 10⁶ Panc-1 cells per 10-cm dish (1.6 × 10³ cells/cm²) to 5 × 10⁶ Panc-1 cells per 10-cm dish (63.7 × 10³ cells/cm²) induced a corresponding density-dependent increase in VEGFR-1 protein expression. Lysate from human umbilical vein endothelial cells (HUVECs) was included as a positive control for VEGFR-1. Vinculin is shown as a loading control.

DISCUSSION

In the current study, we demonstrated protein expression of VEGFR-1 and its ligands, VEGF-A, VEGF-B, and PlGF, in all pancreatic carcinoma cell lines tested. Stimulation of two representative pancreatic carcinoma cell lines led to activation of MAPK signaling, migration, and invasion but not to cell proliferation or survival. VEGFR-1 protein levels also were noted to be cell density dependent, possibly explaining some of the variability in the expression levels noted between RT-PCR, ELISA, and Western blot analysis.

To our knowledge, VEGFR-1 has been studied less frequently than VEGFR-2, in part because VEGFR-1 does not mediate the traditional VEGF-associated functions in ECs. The current study findings are in concordance with the literature regarding ECs because VEGFR-1 did not appear to promote pancreatic tumor cell proliferation or survival, characteristics typically associated with VEGFR-2 activation. The increased tumor cell migration and invasion we observed with VEGFR-1 stimulation is also in keeping with studies of VEGFR-1 in other cell types, such as monocytes.12, 13 It is interesting to note that others have described an endogenous soluble VEGFR-1 as a component of the extracellular matrix, in which it appears to facilitate EC migration and attachment.23 This finding is of interest because tumor-expressed VEGFR-1 may promote the migration and invasion of tumor cells and also facilitate VEGFR-1-dependent EC chemotaxis and adhesion, thereby promoting tumor angiogenesis. However, this hypothesis remains to be studied.

The presence of VEGFRs on tumor cells has recently been recognized in several tumor types. VEGFR-1 in particular has been identified at the mRNA and/or protein level in a variety of solid tumors, including ovarian carcinoma,24 prostate carcinoma,25 nonsmall cell lung carcinoma,26 gastric carcinoma,27 and melanoma.28 Other authors also have identified VEGFR-1 and/or VEGFR-2 in pancreatic carcinoma specimens and/or cell lines,29–31 although the functional portion of these studies focused on the role of VEGFR-2, concluding that that receptor was responsible for the VEGF-induced proliferation observed in selected pancreatic carcinoma cell lines.29, 30 Consistent with our findings, these studies also demonstrated the presence of VEGFR-1 in neoplastic, but not in nonneoplastic, pancreatic specimens by Northern blot analysis.30 However, using Northern blot analysis, it is not possible to determine the cell type of origin of the protein. Although immunohistochemistry may help clarify this issue, to our knowledge currently available antibodies are not sensitive enough to detect relatively low expression levels in various cell types within tumors. These studies also concurred that VEGFR-1 is present at the RNA level in various pancreatic carcinoma cell lines, including Panc-1 cells, although to our knowledge the current study is the first to demonstrate the function of this receptor in pancreatic carcinoma cell lines. Because several of the cell lines studied by others overlap with those assessed in the current study, it is unclear why some of their cell lines appeared to express VEGFR-2 but those in the current study did not. To our knowledge, the mechanism by which pancreatic carcinoma cells acquire expression of VEGFRs is unknown, but the inherent genetic instability of neoplastic cells and/or the dedifferentiation that enables them to express these receptors may account for these differences. In addition, our finding that cell density affects VEGFR-1 expression may help to explain some of the discrepancies noted between the studies and, in particular, the difficulty experienced by other authors in correlating mRNA and protein expression for this receptor.

Others have shown that cell density regulates the expression of certain proteins, including actin and stat3.32, 33 One postulated mechanism by which cell density regulates gene expression is through the treadmilling of actin, yielding differing proportions of filamentous and soluble actin that in turn lead to gene regulation.33 We tested the role of actin polymerization in pancreatic carcinoma cells by adding the actin polymerizing inhibitor latrunculin B (EMD Biosciences, San Diego, CA, San Diego, CA) to pancreatic carcinoma cells plated at different densities in culture; the previously observed density-dependent increase in VEGFR-1 expression was not found to be affected (data not shown), suggesting that other mechanisms are responsible. Other studies also have shown that conditioned medium from hypoxic tumor cells increases VEGFR-1 mRNA in ECs.34 Because it is possible that increasing cell confluency stresses the cells in a manner similar to hypoxia, we tested the hypothesis that confluent cells release a soluble factor responsible for increasing VEGFR-1 expression. However, incubating low-density cells with conditioned medium from high-density cells did not appear to increase VEGFR-1 levels (data not shown). Other authors also have shown that mechanical denudation of ECs leads to up-regulation of VEGFR-1 through the transcription factor early growth response protein-1.35 It may be that disruption of adhesion by trypsin and/or those factors related to increased cell density (adhesion molecules or extracellular matrix) influence the cell density-dependent VEGFR-1 expression.

The presence and function of VEGFRs on tumor cells are of importance given the current attention to VEGF-targeted therapy for cancer.36 An association between the overexpression of VEGF and tumor progression and/or poor prognosis has been documented in several types of malignancy, including colon carcinoma,37, 38 gastric carcinoma,39, 40 pancreatic carcinoma,41, 42 breast carcinoma,43 prostate carcinoma,44 and melanoma.45 To our knowledge the research focus to date has been on anti-VEGF therapy as a means of inhibiting tumor angiogenesis by decreasing the effect of VEGF on ECs, but VEGF-targeted therapy also may have a secondary antitumor effect on the tumor cells themselves if functional VEGFRs are present. In light of reports correlating increased levels of the VEGFR-1 ligands VEGF-B46–49 and/or PlGF50, 51 with tumor presence or progression, the possibility that these ligands also may be relevant targets for treatment must be assessed.

The results of the current study demonstrate that VEGFR-1 is present on pancreatic carcinoma cells and is capable of increasing MAPK signaling, migration, and invasion. Moreover, these effects may be regulated in an autocrine fashion because pancreatic carcinoma cells also produce the ligands for VEGFR-1. In Phase II trials, anti-VEGF therapy has been shown to be of potential benefit in patients with metastatic or locally advanced pancreatic carcinoma (unpublished data), which may be due in part to the targeting of both the tumor cells and ECs. A better understanding of the role of VEGFR-1 in tumor cells will facilitate the development of targeted therapies and predictors of efficacy.

Acknowledgements

The authors thank Melissa G. Burkett, Department of Scientific Publications, and Rita Hernandez, Department of Surgical Oncology, for editorial assistance.