Expression and regulation of the novel vascular endothelial growth factor receptor neuropilin-1 by epidermal growth factor in human pancreatic carcinoma

Abstract

BACKGROUND

It was recently shown that neuropilin-1 (NRP-1), which was described originally as a receptor for the semaphorins/collapsins (ligands involved in neuronal guidance), is a coreceptor for vascular endothelial growth factor (VEGF) and increases the affinity of specific isoforms of VEGF to its receptor, VEGF-R2.

METHODS

The authors investigated the expression and regulation of NRP-1 in human pancreatic adenocarcinoma specimens and cell lines.

RESULTS

Immunohistochemical analysis revealed that NRP-1 was expressed in 12 of 12 human pancreatic adenocarcinoma specimens but was absent in nonmalignant pancreatic tissue. Northern blot analysis revealed NRP-1 mRNA expression in 8 of 11 human pancreatic adenocarcinoma cell lines. NRP-1 mRNA expression was increased by epidermal growth factor (EGF) but not by tumor necrosis factor α in several of the human pancreatic adenocarcinoma cell lines studied. Treating human Panc-48 adenocarcinoma cells with EGF activated Akt and Erk but not P-38. Blockade of the phosphatidylinositol-3 kinase (PI-3K)/Akt, mitogen-activated protein kinase (MAPK)/Erk, or P-38 pathways abrogated EGF-induced NRP-1 expression. Finally, EGF receptor blockade in vivo led to a decrease in NRP-1 expression in an orthotopic model of human pancreatic carcinoma.

CONCLUSIONS

NRP-1 is expressed in most human pancreatic adenocarcinomas and cell lines but not in nonmalignant pancreatic tissue. EGF regulates NRP-1 expression through the PI-3K/Akt and MAPK/Erk signaling pathways, and blockade of the EGF receptor is associated with decreased expression of NRP-1 in vivo. NRP-1 may act as a coreceptor for VEGF in pancreatic carcinoma, as it does in other tumor systems, thereby enhancing angiogenesis and the effect of VEGF on the growth of pancreatic adenocarcinoma. Cancer 2003;98:720–9. © 2003 American Cancer Society.

DOI 10.1002/cncr.11560

Vascular endothelial growth factor (VEGF), or vascular permeability factor, is the best characterized of the angiogenic factors. VEGF has been associated with increased angiogenesis and poor prognosis in patients with a variety of solid tumor types, including pancreatic carcinoma.1, 2 Members of the VEGF family mediate both vascular permeability and endothelial cell proliferation through three tyrosine kinase receptors: VEGF receptor 1 (VEGFR-1; Flt-1), VEGFR-2 (Flk-1/KDR), and VEGFR-3 (Flt-4).3 Although it was believed initially that VEGF affected only endothelial cells, recent evidence suggests that VEGF receptors may also be present on tumor cells.4, 5

Neuropilin-1 (NRP-1) was originally characterized as a receptor for semaphorins or collapsins, proteins involved in the development of the nervous system.6 Unlike the VEGF tyrosine kinase receptors, however, NRP-1 seems to be VEGF isoform specific and acts as a coreceptor for VEGF. It has been shown that NRP-1 binds the VEGF isoform VEGF₁₆₅ as well as other family members, such as VEGF-B, VEGF-E, and placental growth factor-2 (PGF-2).7–9 Although it was found originally that NRP-1 was expressed on endothelial cells, expression of NRP-1 has been described recently in prostate, breast, and melanoma cell lines and in several tumors in vivo.10–13 However, the expression of NRP-1 in gastrointestinal malignancies, including pancreatic adenocarcinoma, has not been characterized; and the regulation and control of NRP-1 expression in tumors have not been studied well. In this study, we assessed whether NRP-1 was expressed in human pancreatic adenocarcinoma tissue and/or in uninvolved human pancreatic tissue. We also investigated the cytokines and intracellular signal-transduction pathways that regulate NRP-1 expression in human pancreatic adenocarcinoma cell lines.

MATERIALS AND METHODS

Tissue Specimens

Human pancreatic adenocarcinoma and normal pancreatic specimens that were obtained from consenting patients at The University of Texas M. D. Anderson Cancer Center (M. D. Anderson) were frozen immediately after resection in optimal cutting temperature (OCT) solution (Miles Inc., Elkhart, IN) and stored at − 80 °C. Histopathologic diagnoses of pancreatic adenocarcinoma or normal pancreas were confirmed by the Department of Pathology at M. D. Anderson. None of the patients had received preoperative chemotherapy or radiation therapy.

Reagents and Chemicals

Recombinant human epidermal growth factor (EGF) and tumor necrosis factor α (TNF-α) were purchased from R&D Systems, Inc. (Minneapolis, MN). The antihuman EGF receptor (EGFR) monoclonal antibody C225 was kindly provided by Dan Hicklin (ImClone Systems, New York, NY). The mitogen-activated protein kinase (MAPK)/extracellular signal-regulated kinase (MEK)-1/2 inhibitor U0126 and the MEK1 inhibitor PD98059 were obtained from New England Biolabs Inc. (Beverly, MA). The phosphatidylinositol-3 kinase (PI-3K) inhibitor Wortmannin was purchased from Sigma Chemical Company (St. Louis, MO). The P-38 MAPK inhibitor SB203580 was purchased from Calbiochem (San Diego, CA).

Cell Lines

The human pancreatic adenocarcinoma cell lines Panc-1, CFPAC, ASPC-1, HS766T, HPAF-2, BX-PC3, and MIAPACA-2 were obtained from the American TypeCulture Collection (ATCC; Manassas, VA). The human pancreatic adenocarcinoma cell lines SG, FG, and L3.6pL were generous gifts of I. J. Fidler, D.V.M., Ph.D., of M. D. Anderson and were derived from COLO 357 cells, as described previously.14 The human pancreatic adenocarcinoma cell line Panc-48 was a generous gift of D. McConkey, Ph.D., of M. D. Anderson. Cell lines were cultured and maintained in minimal essential medium supplemented with 10% fetal bovine serum (FBS), 2 units/mL penicillin-streptomycin, vitamins, 1 mM sodium pyruvate, 2 mM L-glutamine, and nonessential amino acids at 37 °C in 5% CO₂ and 95% air.

Immunofluorescent Staining for NRP-1 and Cytokeratin-22 in Frozen Tissue Specimens

Tissue specimens frozen in OCT were sectioned (8–10 μm thick), mounted on positively charged Superfrost slides (Fisher Scientific, Houston, TX), and air dried for 30 minutes. Snap-frozen tissues were fixed in cold acetone (5 minutes), followed by 1:1 acetone:chloroform (5 minutes) and acetone (5 minutes), then washed with phosphate-buffered saline (PBS) 3 times for 3 minutes each. All samples were incubated with 3% hydrogen peroxide in methanol for 12 minutes at room temperature to block endogenous peroxidase. Sections were then washed 3 times for 3 minutes each with PBS, pH 7.5, then incubated for 20 minutes at room temperature in a protein-blocking solution consisting of PBS supplemented with 1% normal goat serum and 5% normal horse serum. The primary antibodies directed against NRP-1, polyclonal rabbit anti-NRP-1 at 1:100 dilution (Santa Cruz Biotechnologies, Santa Cruz, CA) or undiluted cytokeratin-22 (CK-22) (Fisher Scientific), were applied to the sections and incubated overnight at 4 °C. Sections were then rinsed 3 times for 3 minutes each in PBS and incubated for 10 minutes in protein-blocking solution. In a darkened room, the blocking solution was drained, and the samples were incubated with the fluorescein-conjugated secondary antibodies Alexa 594 goat antirabbit immunoglobulin G (IgG; CK-22; 1:400 dilution) (Molecular Probes, Eugene, OR) and Alexa 488 goat antimouse NRP-1 (1:400 dilution) (Molecular Probes) for 1 hour at room temperature; care was taken to avoid light exposure during the incubation. Samples were then washed 3 times with PBS and mounted with 4,6-diamidino-2-phenylindole fluorescent mounting media (Vector Laboratories, Burlingame, CA). Immunofluorescence microscopy was performed with a × 100–200 objective on an epifluorescence microscope equipped with narrow-band-pass excitation filters mounted in a filter wheel (Chroma Technology Corporation, Brattleboro, VT) to individually select for green, red, and blue fluorescence. Images were captured with a Hamamatsu C5810 camera (Hamamatsu Photonics K.K., Bridgewater, NJ) mounted on a Zeiss universal microscope (Carl Zeiss Inc., Thornwood, NY) using Optimas image-analysis software (Media Cybernetics, Silver Spring, MD) installed on a Pentium chip Compaq computer. Unaltered images were transferred to Adobe Photoshop (Adobe Systems, Mountain View, CA) for viewing and processing. The presence of CK-22 (epithelial cells) was identified by red fluorescence, and NRP-1 expression was detected by green fluorescence. The protocol for control specimens was similar except that the primary antibody was omitted.

Immunoperoxidase Staining for NRP-1 in Frozen Tissue Specimens

Frozen specimens of normal pancreatic tissue and pancreatic adenocarcinoma were fixed and incubated with a primary NRP-1 antibody (polyclonal rabbit anti-NRP-1; 1:150 dilution; Santa Cruz Biotechnologies) overnight, as described earlier. The secondary antibody (peroxidase-conjugated goat antirabbit IgG [H+L]; Jackson Research Laboratories, Westgrove, PA) was then used at a 1:400 dilution. Sections were washed 3 times with PBS, rinsed with PBS and 0.1% Brij35 detergent, and incubated with stable diaminobenzidine substrate (Research Genetics, Huntsville, AL); during that incubation, the staining was monitored by brightfield microscopy. The reaction was halted by rinsing with double-distilled H₂O. The sections were counterstained with Gill no. 3 hematoxylin solution (Sigma Chemical Company), mounted with Universal Mount (Research Genetics), and analyzed by light microscopy. The protocol for control specimens was similar except that the primary antibody was omitted.

The Effects of EGF and TNF-α on the Expression of NRP-1 mRNA in Panc-48, Panc-1, and HPAF-2 Cells

To determine the effects of EGF and TNF-α on NRP-1 mRNA expression, Panc-48, Panc-1, and HPAF-2 cells were grown to subconfluence in standard medium, as described earlier, and the medium was changed to medium containing 5% FBS overnight. Cells were then incubated with EGF (100 ng/mL) or TNF-α (10–20 ng/mL) for 4–24 hours in medium containing 1% FBS. Total RNA was extracted, and NRP-1 mRNA expression was determined by Northern blot analysis.

The Effect of Increasing Doses of EGF on NRP-1 mRNA Expression in Panc-48 Cells

To determine the effect of increasing doses of EGF on NRP-1 mRNA expression, Panc-48 cells were grown to subconfluence in standard medium, as described earlier, and the medium was changed to medium containing 5% FBS overnight. Cells were then incubated with EGF (0–250 ng/mL) for 24 hours in medium containing 1% FBS. Total RNA was extracted, and NRP-1 mRNA expression was determined by Northern blot analysis.

The Effect of EGF on Erk-1/2, Akt, and P-38 Phosphorylation in Panc-48 Cells

To determine the effect of EGF on the protein levels and phosphorylation status of the signaling intermediates Erk-1/2, Akt, and P-38 MAPK, Panc-48 cells grown under the conditions described earlier were incubated with EGF (100 ng/mL) for 0 minutes, 10 minutes, 15 minutes, 30 minutes, or 60 minutes in medium containing 1% FBS and then lysed. Phosphorylated and total protein levels were determined by Western blot analyses.

The Effect of Inhibiting Erk-1/2, Akt, and P-38 MAPK on NRP-1 Induction by EGF

To determine the effect of inhibiting Erk-1/2, Akt, and P-38 MAPK on NRP-1 induction, Panc-48 cells grown under the conditions described earlier were treated with 50 μM PD98059 (MEK1 inhibitor), 10 μM U0126 (MEK-1/2 inhibitor), 200 nM Wortmannin (PI-3K inhibitor), or 25 μM SB203580 (P-38 MAPK inhibitor) for 1 hour in medium containing 1% FBS, followed by the addition of EGF (100 ng/mL) or medium containing 1% FBS alone, as a control, for 24 hours. Total RNA was extracted, and Northern blot analysis was performed.

RNA Extraction and Northern Blot Analysis

Total RNA was harvested from subconfluent tumor cells in culture with the TRIZOL Reagent (Gibco-BRL, Grand Island, NY) according to the manufacturer's instructions. Northern blot analysis was performed as described previously.15 Briefly, 25 μg of RNA were fractionated on 1% denaturing formaldehyde-agarose gels and transferred to a Hybond-N+ positively charged nylon membrane (Amersham Life Science, Arlington Heights, IL) overnight by capillary elution. After ultraviolet (UV) cross-linking at 120,000 mJ/cm² with a UV Stratalinker 1800 (Stratagene, La Jolla, CA), the membranes were prehybridized at 65 °C for 3–4 hours in rapid hybridization buffer (Amersham Life Science). The membranes were then hybridized at 65 °C overnight with the cDNA probe for NRP-1 or glyceraldehyde 3-phosphate dehydrogenase (GAPDH). The probed membranes were then washed, and autoradiography was performed. The cDNA probes used were a human NRP-1 450 base pair cDNA probe derived from the reverse transcriptase-polymerase chain reaction product of PC-3 human prostate carcinoma cells (ATCC) using the primers (ACGATGAATGTGGCGATACT) 3′-5′ and (AGTGCATTCAAGGCTGTTGG) 5′-3′ and using a GAPDH probe purchased from ATCC. Probes were purified by agarose gel electrophoresis with the QIAEX gel-extraction kit (QIAGEN Inc., Chatsworth, CA). Each cDNA probe was radiolabeled with [α-³²P]deoxyribonucleotide triphosphate using the random-priming technique with the Rediprime labeling system (Amersham Life Science).

Western Blot Hybridization

Cells were rinsed twice with ice-cold PBS and lysed with protein lysis buffer (20 mM sodium phosphate, pH 7.4; 150 mM NaCl; 1% Triton X-100; 5 mM ethylenediamine tetraacetic acid; 5 mM phenylmethylsulfonyl fluoride; 1% aprotinin; 1 μg/mL leupeptin; and 500 μM Na₃VO₄). Aliquots (50 μg) of protein were subjected to electrophoresis on 8% polyacrylamide gels followed by electrotransfer to nitrocellulose membranes (Schleicher & Schleicher, Keene, NH). Membranes were blocked with 5% fat-free milk in 0.1% Tween-20 in PBS. The primary antibodies used in this study were 1:1000 dilutions of rabbit polyclonal antiphosphospecific p44/42 MAPK (Erk-1/2) antibody, antitotal Erk antibody, antiphosphospecific P-38 MAPK antibody, anti-P-38 MAPK antibody, antiphosphospecific Akt antibody, antitotal Akt antibody, and anti-β-actin antibody (all from New England Biolabs, Beverly, MA). The membranes were then washed and treated with the secondary antibody labeled with horseradish peroxidase (antirabbit immunoglobulin from donkey at 1:3000 dilution; Amersham Life Science). Protein bands were visualized by using a commercially available chemoluminescence kit (Amersham Life Science). Before reprobing, the membranes were washed with stripping solution (100 mM 2-mercaptoethanol; 2% sodium dodecyl sulfate; and 62.5 mM Tris-HCl, pH 6.7) for 30 minutes at 50 °C.

Immunofluorescent Staining for NRP-1 in Orthotopic L3.6pL Tumors

Tissues from a previous study16 were used to analyze the effect of EGFR inhibition on the expression of NRP-1 by orthotopic tumors. Briefly, L3.6pL human pancreatic adenocarcinoma cells were injected into the pancreas of male athymic mice. Seven days later, the mice were assigned randomly to undergo biweekly intraperitoneal injection of the chimeric anti-EGFR monoclonal antibody C225 (1 mg/injection; ImClone Systems) or saline (control), and tumors were harvested on Day 11 and immediately frozen in OCT compound.16 For these experiments, tumors were sectioned and immunofluorescently stained for NRP-1 as described above.

RESULTS

Immunohistochemical Staining of Human Pancreatic Adenocarcinomas for NRP-1

To determine whether NRP-1 expression is present in human pancreatic adenocarcinoma and/or normal human pancreas, immunoperoxidase staining was performed on 12 specimens of frozen human pancreatic adenocarcinoma and 10 specimens of frozen, uninvolved, nonmalignant human pancreas. NRP-1 was present in all 12 specimens of histologically confirmed pancreatic adenocarcinoma but in none of the nonmalignant pancreatic specimens (Fig. 1).

Figure 1. Representative sections showing hematoxylin and eosin (H&E) staining and immunohistochemical staining for neuropilin-1 (NRP-1) expression in specimens of nonmalignant human pancreas (n = 10 specimens) and in specimens of malignant human pancreatic adenocarcinoma (n = 12 specimens). Frozen sections were stained, and representative images were obtained at a magnification of × 100 (H&E; top row) or × 200 (NRP-1; bottom row) by light microscopy. NRP-1 production is indicated by reddish-brown staining.

To localize the origin of the NRP-1 expressed, immunofluorescent staining was performed on the 12 frozen specimens of human pancreatic adenocarcinoma. The samples were immunostained for CK-22 (an epithelial cell marker), to identify the tumor epithelium, and for NRP-1 (double-stain technique). NRP-1 expression was present in all pancreatic adenocarcinoma specimens and was localized to the adenocarcinoma epithelium (Fig. 2).

Figure 2. Immunofluorescence analysis of neuropilin-1 (NRP-1) expression in specimens of human pancreatic adenocarcinoma. Frozen sections (n = 12 sections) were stained by immunofluorescence, and representative images were obtained at × 100 magnification. The expression of the epithelial cell marker cytokeratin-22 (CK-22) is indicated by the presence of red staining, and NRP-1 production is indicated by the presence of green staining. Coexpression of NRP-1 and CK-22 is shown as yellow staining.

Northern Blot Analysis of NRP-1 mRNA in Human Pancreatic Adenocarcinoma Cell Lines

Eleven human pancreatic adenocarcinoma cell lines were examined by Northern blotting for the expression of NRP-1. Eight of 11 cell lines (73%) expressed detectable, constitutive NRP-1 mRNA (Fig. 3) in various amounts. NRP-1 mRNA expression was greatest in the CFPAC-1 cell line and least (no constitutive expression detected) in the MiaPaca-2, HS766T, and FG cell lines.

Figure 3. Neuropilin-1 (NRP-1) mRNA expression in human pancreatic adenocarcinoma cell lines. Cells were grown to 80% confluence. Total RNA was harvested and, Northern blot analysis was used to quantify NRP-1 expression. GADPH: glyceraldehyde 3-phosphate dehydrogenase.

NRP-1 Induction by EGF in Human Pancreatic Adenocarcinoma Cells

It was recently shown that the cytokines EGF and TNF-α induce NRP-1 expression in endothelial cells and astrocytoma cells, respectively.17, 18 To examine the role of EGF and TNF-α in the regulation of NRP-1 in human pancreatic adenocarcinoma, we treated Panc-48 cells with EGF or TNF-α and subsequently analyzed NRP-1 mRNA levels. EGF increased NRP-1 mRNA expression, with maximum expression observed at 24 hours (Fig. 4A). In contrast, incubation of Panc-48 cells with TNF-α did not increase NRP-1 mRNA levels (Fig. 4A). These findings were confirmed in two additional human pancreatic adenocarcinoma cell lines; NRP-1 mRNA expression was increased in Panc-1 and HPAF-2 cells in response to EGF, but not in response to TNF-α (Fig. 4B). Extending the period of incubation with TNF-α or increasing the dose to 20 ng/mL also failed to increase NRP-1 mRNA levels in all three cell lines (data not shown).

Figure 4. The effect of epidermal growth factor (EGF) and tumor necrosis factor α (TNF-α) on neuropilin-1 (NRP-1) mRNA induction in human pancreatic adenocarcinoma cell lines. Panc-48, Panc-1, and HPAF-2 cells were grown to 80% confluence and treated with EGF (100 ng/mL) or TNF-α (10 ng/mL) for 24 hours. Total RNA was then harvested, and Northern blot analysis was performed for NRP-1 analysis. 18S indicates ethidium bromide staining for 18S RNA that was used as a loading control. GADPH: glyceraldehyde 3-phosphate dehydrogenase.

NRP-1 Induction by EGF in Human Pancreatic Cells Is Dose Dependent

To determine the effects of different concentrations of EGF on NRP-1 mRNA expression, we treated Panc-48 cells with escalating doses of EGF (0–250 ng/mL) for 24 hours and subsequently analyzed NRP-1 mRNA levels. NRP-1 expression levels increased with increasing concentrations of EGF, with maximum mRNA levels present with an EGF dose of 100 ng/mL, as determined by densitometric analysis (Fig. 5).

Figure 5. Effect of increasing doses of epidermal growth factor (EGF) on neuropilin-1 (NRP-1) mRNA expression in Panc-48 human adenocarcinoma cells. Cells were grown to 80% confluence and treated with varying doses of EGF, as indicated, for 24 hours. Total RNA was then harvested for Northern blot analysis of NRP-1 mRNA expression. GADPH: glyceraldehyde 3-phosphate dehydrogenase.

EGF Induction of the PI-3K and MAPK Signal-Transduction Pathways

Treatment of Panc-48 cells with EGF resulted in phosphorylation of both Erk-1/2 (the MAPK pathway) and Akt (the PI-3K pathway) (Fig. 6). In contrast, EGF treatment had no effect on the levels of phosphorylated P-38 MAPK (Fig. 6).

Figure 6. The effect of epidermal growth factor (EGF) on intracellular signaling pathways in Panc-48 human adenocarcinoma cells. Cells were grown to 80% confluence, treated with EGF (100 ng/mL) for the indicated times, and harvested for Western blot analysis of the indicated signaling intermediates.

EGF Induction of NRP-1 Is Regulated by the PI-3K, Erk MAPK, and P-38 MAPK Pathways

To investigate the roles of specific signal-transduction pathways on constitutive NRP-1 expression in human pancreatic adenocarcinoma cells, we treated Panc-48 adenocarcinoma cells with the PI-3K inhibitor Wortmannin, the MEK inhibitors U0126 and PD98059, or the P-38 MAPK inhibitor SB203580. Inhibition of the MEK pathway blocked constitutive NRP-1 mRNA expression (Fig. 7A). In contrast, inhibition of the PI-3K or P-38 pathways had no effect on constitutive NRP-1 mRNA expression (Fig. 7A). To investigate the roles of these signaling pathways on EGF-induced NRP-1 expression in human pancreatic adenocarcinoma cells, we pretreated Panc-48 adenocarcinoma cells with the same signaling pathway inhibitors followed by the addition of EGF. Inhibition of any of these three pathways abrogated EGF-induced NRP-1 mRNA expression, as determined by Northern blot analysis (Fig. 7B).

Figure 7. Northern blot analysis showing inhibition of epidermal growth factor (EGF)-induced neuropilin-1 (NRP-1) mRNA expression by signaling inhibitors. Panc-48 cells were grown to 80% confluence and were treated with 200 nM Wortmannin (Wt) (a phosphatidylinositol-3 kinase inhibitor), 10 μM U0126 (a mitogen-activated protein kinase [MAPK]/extracellular signal-regulated kinase [MEK]/Erk-1/2 inhibitor), 50 μM PD98059 (a MEK1 inhibitor), or 25 μM SB203580 (a P-38 MAPK inhibitor). One hour later, medium (control) (A) or EGF (100 ng/mL) (B) was added, and the mixture was incubated for 24 hours. Total RNA was then harvested for analysis of NRP-1 mRNA expression. GADPH: glyceraldehyde 3-phosphate dehydrogenase.

In Vivo Inhibition of NRP-1 Protein Production by C225 in an Orthotopic Model of Human Pancreatic Carcinoma

Bruns et al. showed that blockade of the EGFR with the monoclonal antibody C225 resulted in regression of L3.6pL human pancreatic adenocarcinoma growing in an orthotopic nude mice model.16 This blockade was also associated with decreased levels of VEGF and interleukin 8 (IL-8) 11 days after the start of treatment and decreased microvessel density in the tumors 18 days after the start of treatment.16 When we examined the same tumors harvested at 11 days for NRP-1 expression by immunohistochemistry, we also found that blockade of the EGFR with C225 led to a visible decrease in NRP-1 production (Fig. 8) in sections of histologically confirmed adenocarcinoma.

Figure 8. Representative sections showing immunofluorescent staining for neuropilin-1 (NRP-1) production in tumor sections of L3.6pL human pancreatic adenocarcinoma cells growing in the pancreas of untreated nude mice (n = 9 mice; top row) or in nude mice treated with the epidermal growth factor receptor antibody C225 (n = 9 mice; bottom row). Frozen sections of histologically confirmed pancreatic adenocarcinoma were stained for NRP-1, and representative images were obtained at a magnification of × 100 under fluorescent microscopy. NRP-1 production is indicated by green fluorescence.

DISCUSSION

NRP-1 was identified initially as a 130–140 kDa cell-surface glycoprotein expressed in the developing Xenopus laevis nervous system. Subsequently, it was discovered that this transmembrane glycoprotein served as a receptor for the semaphorins or collapsins, a large family of secreted and transmembrane proteins that serve as repulsive guidance signals in axonal and neuronal development.8, 12, 19, 20 NRP-1 apparently has a role in vasculogenesis and angiogenesis as well. NRP-1 is expressed in the skeletal and cardiovascular systems in developing embryos.12, 21 NRP-1 knockout mice suffer from insufficient and delayed vascularization, leading to embryonic death.22–24 Over-expression of NRP-1 in transgenic mice also is lethal due to hemorrhage in the head and neck, formation of excess blood vessels, and malformations of the heart.21 In addition to its expression in embryonic tissue, it has been found that NRP-1 is expressed in adult endothelial cells and in a variety of other tissues (including lung, heart, liver, kidney, pancreas, and placenta) as well as in osteoblasts and bone marrow stromal cells.7, 25

Although the specific functions of NRP-1 in vessel development and angiogenesis are not known fully, it has been shown that NRP-1 acts as a coreceptor for VEGF₁₆₅.7–9 The binding of VEGF₁₆₅ to NRP-1 appears to be mediated by amino acids that reside at the carboxyl-terminal part of the exon 7-encoded peptide of VEGF₁₆₅.7 In contrast, the binding of VEGF₁₆₅ to VEGFR-1 (Flt-1) and VEGFR-2 (KDR) occurs through exon 4,7 thus enabling VEGF₁₆₅ to bind to both NRP-1 and VEGFR-1 or VEGFR-2 simultaneously. For example, in vitro inhibition of VEGF₁₆₅ binding to NRP-1 in endothelial cells decreased its binding to VEGFR-2 and decreased subsequent mitogenic activity.26 In agreement, cotransfection of NRP-1 into VEGFR-2-expressing endothelial cells enhanced the binding of VEGF₁₆₅ to VEGFR-2 and enhanced subsequent mitogenic and chemotactic activity compared with cells that expressed VEGFR-2 alone.7, 19 Endothelial cells that expressed NRP-1 alone but not express VEGFR-2 did not respond to any VEGF isoform, suggesting that NRP-1 is not a signaling receptor for chemotaxis in and of itself but, rather, acts as a coreceptor for VEGFR-2, enhancing the angiogenic activity of VEGF.7

Other evidence that suggests a role for NRP-1 in tumor growth and angiogenesis includes the description of NRP-1 expression in prostate, breast, and colon carcinoma cell lines and in several tumors in vivo, including prostate, pituitary, and squamous cell carcinomas.8 Overexpression of NRP-1 in rat prostate carcinoma cells resulted in increased tumor cell size in vivo as well as increased microvessel density, increased endothelial cell proliferation, dilated blood vessels, and less tumor cell apoptosis along with higher levels of VEGF protein.27 In other studies, NRP-1 expression appeared to correlate with the metastatic potential of prostate carcinoma cells and with advanced stage and grade.11

To our knowledge, this is first study to describe the expression of the novel VEGF receptor NRP-1 in human pancreatic adenocarcinoma. NRP-1 was expressed in all human pancreatic adenocarcinoma specimens and in most pancreatic adenocarcinoma cell lines tested but not in nonmalignant human pancreatic tissue. This finding suggests that the production of NRP-1 is associated with the development of pancreatic adenocarcinoma, perhaps by enhancing VEGF binding and VEGF effects in this system. Although pancreatic adenocarcinoma generally is considered a hypovascular tumor, several lines of evidence suggest that angiogenesis may be important in its development and growth. Early studies showed that pancreatic carcinoma cells express a wide range of proangiogenic factors, including VEGF, that seem to be important in the angiogenesis and growth of primary and metastatic pancreatic carcinoma.28 Expression of VEGF in human pancreatic carcinoma has been associated with increased microvessel density,29–32 with worse prognosis, and with higher grade pancreatic tumors.30, 31 Previous studies from our laboratory demonstrated that antibodies to VEGFR-2 can decrease the growth of pancreatic tumors and the formation of metastases, both effects that are secondary to a decrease in angiogenesis.33

After finding that NRP-1 was expressed in pancreatic carcinoma, we sought to identify the potential regulatory factors that govern its expression. It has been shown that EGF regulates a variety of different angiogenic and regulatory factors, including VEGF, in several tumor types.34, 35 It was also shown that EGF regulates NRP-1 expression in astrocytoma cells17; and other studies have shown that inhibitors of EGFR, alone or in combination with gemcitabine, decreased the growth of pancreatic adenocarcinoma in animal models and were associated with decreases in microvessel density and VEGF production.16, 36 Other investigators have shown that adenoviral infection of pancreatic adenocarcinoma cells with a truncated, inactive form of EGFR led to attenuated EGF dependent mitogenic signal transduction, decreased EGF dependent cell growth, and increased sensitivity to chemotherapeutic agents.37, 38 Our current finding that EGF increased NRP-1 mRNA expression in several pancreatic adenocarcinoma cell lines further supports a putative role for EGF in the angiogenesis and growth of pancreatic carcinoma. It was reported previously that TNF-α, an inflammatory cytokine, increased VEGFR-2 and NRP-1 expression in human vascular endothelial cells.18 In the current study, however, TNF-α failed to increase NRP-1 expression in any of the three pancreatic adenocarcinoma cell lines tested.

The receptor for EGF is a transmembrane glycoprotein with an external binding domain and an intracellular tyrosine kinase domain. After ligand binding, EGFR initiates a kinase cascade that activates members of the Ras/Raf/Erk MAPK family as well as the PI-3K/Akt pathway in specific cell systems.39 In this study, treatment of Panc-48 human pancreatic adenocarcinoma cells with EGF led to the phosphorylation and, thus, the activation of both Erk (MAPK pathway) and Akt (PI-3K pathway). However, treatment with EGF did not have an effect on activation of P-38 MAPK, which our results showed was present constitutively. These findings suggest that, in this pancreatic adenocarcinoma system, the P-38 MAPK pathway may not be as important as the Erk MAPK and PI-3K intracellular signaling pathways in terms of EGFR function. These results also support the view that binding of EGF to EGFR leads to activation of specific signal-transduction pathways rather than leading to a generalized cell response.

To our knowledge, the intracellular signaling pathways that govern NRP-1 induction have not been elucidated fully, although a study in astrocytoma cells suggested that the ras-GTP (MAPK) signaling pathway may regulate NRP-1 expression. Consistent with those results is our finding that both constitutive NRP-1 expression and EGF-induced NRP-1 expression appear to be regulated, at least in part, by the Erk MAPK signaling pathway in pancreatic adenocarcinoma. In contrast, the PI-3K/Akt signaling pathway appears to be involved in EGF-induced NRP-1 expression, which has not been described previously, but not in constitutive NRP-1 expression.

It is noteworthy that the treatment of Panc-48 cells with SB203580, a specific inhibitor of P-38 MAPK, decreased NRP-1 mRNA expression in response to EGF, even though EGF itself did not lead to an increase in phosphorylated (activated) P-38. Although one possible explanation is that the P-38 MAPK pathway may regulate constitutive NRP-1 expression, this does not appear to be the case, as shown in Figure 7A. Another possibility is the cross-talk by P-38 with other signal-transduction pathways.

Finally, the results of this study demonstrated that, in an established orthotopic model of human pancreatic carcinoma in which treatment with the antihuman EGFR antibody C225 inhibited tumor growth and metastasis significantly,16 NRP-1 production also was decreased significantly compared with tumors from untreated mice and was associated with regression of the tumor. Furthermore, this abrogation in NRP-1 production was associated temporally with the previously reported reduced VEGF and IL-8 production and preceded the decrease in microvessel density, which reportedly was observed 7 days later. These results lend further support to the concept that EGFR-targeted therapy may have a significant antitumor effect in vivo at least in part through its antiangiogenic mechanisms affecting the VEGF/VEGFR system.

The current study findings suggest that the novel VEGF receptor NRP-1 is associated with human pancreatic adenocarcinoma but not with nonmalignant pancreatic tissue. Although the exact mechanisms by which NRP-1 may regulate tumorigenesis and angiogenesis remain unclear, EGF well may play an important role, lending further support to the potential involvement and potential mechanism by which EGF may be involved in the growth of pancreatic carcinoma. Regulation of NRP-1 expression appears to involve, at least in part, EGF dependent signaling pathways that also have been shown to be important in modulating the production of other angiogenic factors, including VEGF. In addition, the signaling pathways that appear to regulate EGF-induced NRP-1 expression are different from the pathways that regulate constitutive NRP-1 expression. Further in vitro studies, particularly those involving the NRP-1 promoter as well as other factors regulating NRP-1 expression, in addition to in vivo studies investigating tumor growth and survival will be necessary to determine both the mechanism and the role of NRP-1 and its regulatory mechanisms in the pathogenesis of human pancreatic adenocarcinoma.

Acknowledgements

The authors thank Donna Reynolds for assistance with immunohistochemistry; Guido Sclabas, M.D., for assistance with obtaining human pancreatic specimens; and Christine Wogan from M. D. Anderson's Department of Scientific Publications for editorial assistance.