NF-κB activity blockade impairs the angiogenic potential of human pancreatic cancer cells

Authors

  • Henry Q. Xiong,

    1. Department of Gastrointestinal Medical Oncology, The University of Texas M. D. Anderson Cancer Center, Houston, TX, USA
    Search for more papers by this author
  • James L. Abbruzzese,

    1. Department of Gastrointestinal Medical Oncology, The University of Texas M. D. Anderson Cancer Center, Houston, TX, USA
    Search for more papers by this author
  • Edward Lin,

    1. Department of Gastrointestinal Medical Oncology, The University of Texas M. D. Anderson Cancer Center, Houston, TX, USA
    Search for more papers by this author
  • Liwei Wang,

    1. Department of Gastrointestinal Medical Oncology, The University of Texas M. D. Anderson Cancer Center, Houston, TX, USA
    Search for more papers by this author
  • Leizhen Zheng,

    1. Department of Gastrointestinal Medical Oncology, The University of Texas M. D. Anderson Cancer Center, Houston, TX, USA
    Search for more papers by this author
  • Keping Xie

    Corresponding author
    1. Department of Gastrointestinal Medical Oncology, The University of Texas M. D. Anderson Cancer Center, Houston, TX, USA
    2. Department of Cancer Biology, The University of Texas M. D. Anderson Cancer Center, Houston, TX, USA
    • Department of Gastrointestinal Medical Oncology, Unit 426, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, TX 77030
    Search for more papers by this author
    • Fax: +713-745-1163


Abstract

The effect of blockade of NF-κB activity on human pancreatic cancer angiogenesis was determined in an orthotopic xenograft model. Highly metastatic L3.3 human pancreatic cancer cells, which expressed an elevated level of constitutive NF-κB activity, were transfected with a mutated IκBα (IκBαM). After implantation in the pancreas of nude mice, parental (L3.3) and control vector-transfected (L3.3-Neo) cells produced rapidly growing tumors and liver metastases, whereas IκBαM-transfected (L3.3-IκBαM) cells had decreased tumorigenicity and metastatic potential. NF-κB signaling blockade significantly inhibited the in vitro and in vivo expression of the major proangiogenic molecules vascular endothelial growth factor and interleukin-8 and decreased tumor vascular formation. These events were correlated with retarded tumor growth and suppression of metastasis. Collectively, these data suggest that suppression of tumorigenicity and metastasis by NF-κB blockade is due to impaired angiogenic potential of tumor cells. © 2003 Wiley-Liss, Inc.

Metastatic pancreatic adenocarcinoma is a deadly disease, the aggressive nature of which is related to several abnormalities in growth factors and their receptors that affect the downstream signal transduction pathways involved in the control of growth and differentiation.1, 2 Other contributing molecular changes in pancreatic adenocarcinoma include activation of oncogenes and inactivation of tumor suppressor genes.3, 4 These perturbations confer a tremendous survival and growth advantage to pancreatic cancer cells. Understanding the biology and pathogenesis of pancreatic cancer is crucial to reversing the metastatic biology of this disease. As with other solid tumors, the growth and metastasis of pancreatic adenocarcinoma are dependent on angiogenesis.5 Of the numerous angiogenic factors discovered so far, vascular endothelial growth factor (VEGF)6, 7, 8 and, most recently, interleukin-8 (IL-8)9, 10 have been identified as key mediators of pancreatic tumor angiogenesis.

VEGF has been shown to induce the proliferation of endothelial cells, increase vascular permeability and induce the production of plasminogen activators by these cells.11, 12 IL-8, a chemoattractant cytokine, has been shown to attract and activate neutrophils in inflammatory regions and be angiogenic.13, 14 The mechanism of regulation of constitutive expression of these genes in malignant pancreatic cancer cells is not clear, however. Recent studies demonstrated that the pleiotropic transcription factor nuclear factor-κB (NF-κB) regulates the expression of multiple genes, including IL-8 and VEGF, in several types of cells.10, 14–17 NF-κB has been shown to be constitutively activated in pancreatic cancer cells;18, 19 NF-κB activity can also be elevated by several stress factors in human pancreatic cancer cells, such as hypoxia, acidosis, nitric oxide and proinflammatory cytokines.20 Accumulating evidence from in vitro and in vivo studies suggests an important role for NF-κB in the regulation of apoptosis, cell adhesion and oncogenesis.21, 22, 23 However, the critical role of constitutive NF-κB activity in pancreatic cancer angiogenesis has not been investigated.

NF-κB is an inducible dimeric transcription factor belonging to the Rel/NF-κB family of transcription factors whose prototype in most nonlymphoid cells is a heterodimer composed of the RelA (p65) and NF-κB1 (p50) subunits.24, 25, 26 NF-κB activation involves its release from a cytoplasm inhibitor of NF-κB (IκB); it subsequently translocates to the nucleus, where it binds to cognate sequences in the promoter region of multiple genes. Regulation of gene expression by NF-κB is controlled mainly by the inhibitory IκB proteins, which include IκBα.24, 26 Upon stimulation, IκBα is rapidly phosphorylated and degraded via the ubiquitin-proteasome pathway, permitting activation and nuclear importation of NF-κB. Researchers engineered dominant-negative mutant forms of IκBα (IκBαM) that could not be phosphorylated and degraded and thus constitutively repressed the nuclear translocation and DNA binding of NF-κB complexes in stably as well as transiently transfected cells.26, 27, 28, 29

In our study, we found that constitutive NF-κB activity directly correlated with the metastatic potential of pancreatic cancer cells, as highly metastatic L3.3 human pancreatic cancer cells had higher levels of NF-κB activity than other cells did. Blockade of the NF-κB activity via IκBαM transfection suppressed tumor growth and spontaneous metastasis. The antitumor activity correlated with decreased vascularization of pancreatic cancers growing in nude mice due at least partially to decreased VEGF and IL-8 expression. Our data demonstrate that blocking NF-κB activity suppresses the angiogenesis of human pancreatic cancer, suggesting that this transcription factor directly regulates the angiogenesis, growth and metastasis of this disease.

MATERIAL AND METHODS

Cell lines and reagents

The SG, FG and L3.3 variants were originally established from COLO357 human pancreatic carcinoma cell line30 by Vezeridis et al.31 The tumor cells were maintained in culture (5% CO2 and 95% air at 37°C) as adherent monolayers in minimal essential medium (MEM) supplemented with 10% fetal bovine serum, sodium pyruvate, nonessential amino acids, L-glutamine and vitamin solution (Flow Laboratories, Rockville, MD). All of the cultures were free of mycoplasma and pathogenic murine viruses. Also, all of the reagents used in the tissue cultures were free of endotoxins as determined using the Limulus amebocyte lysate assay (sensitivity limit, 0.125 ng/ml; Associates of Cape Cod, Falmouth, MA).

Animals

Male athymic BALB/c nude mice were purchased from the Animal Production Area of the National Cancer Institute, Frederick Cancer Research Facility (Frederick, MD). The mice were housed in laminar flow cabinets under specific pathogen-free conditions and used at 10 weeks of age. They were maintained according to institutional regulations in facilities approved by the Association for Assessment and Accreditation of Laboratory Animal Care and current regulations and standards of the U. S. Department of Agriculture, Department of Health and Human Services and National Institutes of Health.

ELISA for human IL-8 and VEGF expression

The level of IL-8 protein expression in culture supernatants was determined using a quantitative immunometric sandwich ELISA (Quantikine IL-8 and VEGF ELISA kits; R&D Systems, Minneapolis, MN). The absorbance of the samples was compared using the standard curve.8, 9

Northern blot analysis

Cellular mRNA was extracted from pancreatic cancer cells and tumor tissues using the FastTrack mRNA isolation kit (Invitrogen, San Diego, CA). The mRNA (2 μg) was separated electrophoretically on a 1% denaturing formaldehyde agarose gel, transferred to a GeneScreen nylon membrane (DuPont, Boston, MA) in 20× standard saline citrate and crosslinked using a UV-Stratalinker 1800 (Stratagene, La Jolla, CA). The cDNA probe used in the analysis was a 0.5 kb EcoRI cDNA fragment corresponding to human IL-8 and a 0.204 kb BamHI/EcoRI cDNA fragment corresponding to human VEGF/VPF.8, 9 The cDNA probes were labeled with [32P]-deoxycytidine triphosphate using a random labeling kit (Roche, Indianapolis, IN). The equivalence of the mRNA sample loading was monitored by hybridizing the same membrane filter with a β-actin cDNA probe.

Promoter reporters and dual luciferase assays

Luciferase reporters driven by either 2-copy wild-type (Wt-NF-κB) or mutant (Mut-NF-κB) NF-κB-responsive elements15, 18 were used in our study. pGL2-IL-8, a pGL2-basic reporter containing a full-length firefly luciferase gene under the control of an IL-8 promoter region flanking +44 to −1481 from pxp2-IL-8,14 and pGL2-VEGF, a pGL2-basic reporter containing a full-length firefly luciferase gene under the control of both a VEGF 5′-flanking region from +50 to −2274 and 3′-untranslated region from +1 to +1921,17 were used. Pancreatic cancer cells (1 × 105/well) growing in 6-well tissue culture plates were transfected with the indicated reporter plasmids using Lipofectin reagent (Life Technologies, Gaithersburg, MD). Normalization of transfection efficiency was performed using cotransfection with a pB-Actin-RL reporter containing a full-length Renilla luciferase gene (Promega, Madison, WI) under the control of a human β-actin promoter.32 Forty-eight hours after transfection, the cells were harvested in passive lysis buffer (Promega). The firefly and Renilla luciferase activities were quantified using the dual luciferase assay system (Promega). The fold luciferase activity was calculated relative to the luciferase activity of pGL2-basic in tumor cells, which were given a reference value of 1 as described previously.8, 32

Stable transfection of pancreatic cancer cells with IκBαM and control vector

The cDNA plasmid pLXSN-IκBαM, which contains mutations at S32 and S36 of the NH2 terminus and a COOH-terminal PEST sequence mutation, was used.28 L3.3 cells (1 × 106) were transfected using 15 μl of Lipofectin reagent (Life Technologies) and 4 μg of pLXSN-IκBαM expression vector or control pLXSN vector. The transfections were carried out according to the manufacturer's instructions. Cells were selected using a standard medium containing G418 at 600 μg/ml. Fourteen days later, Neo-resistant colonies were isolated via trypsinization and established as subcultures.33

Western blot analysis

Cytosolic and nuclear protein was isolated from control and transfected pancreatic cancer cells. Soluble protein was separated on a 10% sodium dodecyl sulfate-polyacrylamide gel using electrophoresis and electrophoretically transferred onto an Immobilon-P transfer membrane (Millipore, Bedford, MA). Endogenous and mutant IκBα were probed with a polyclonal rabbit anti-human and -mouse IκBα (C-21) and NF-κB was probed with a polyclonal rabbit anti-human anti-p65 antibody (Santa Cruz Biotechnology, Santa Cruz, CA). Equal cytosolic and nuclear protein sample loading was monitored by hybridizing the same membrane filter with anti-β-actin and anti-Oct-1 antibodies, respectively. The probe proteins were detected using the Amersham enhanced chemiluminescence system according to the manufacturer's instructions (Amersham Biosciences, Piscataway, NJ).

EMSA

Nuclear protein extracts were prepared as described previously.18, 32 The NF-κB oligonucleotide probe was 5′-AGTTGAGGGACTTTCCCAGGC-3′. EMSA was also performed as described previously but with minor modifications.32 Five micrograms of nuclear extract protein and 30,000 cpm of an end-labeled double-stranded DNA probe were added to the mixture. The binding reaction was allowed to proceed for 25 min at 22°C. For supershift reactions, extracts were preincubated with an anti-p65 or -p50 antibody (Calbiochem-Novabiochem, San Diego, CA) for 20 min on ice.

Orthotopic implantation of tumor cells

For all in vivo experiments, tumor cells in exponential growth phase were harvested after brief exposure to a 0.25% trypsin/0.02% EDTA solution (w/v). The flask containing the cells was tapped sharply to dislodge them; MEM was then added, and the cell suspension was pipetted to obtain single-cell suspensions. The cells were washed, resuspended in Ca2+-and Mg2+-free Hanks' balanced salt solution (HBSS) and diluted to the desired cell number/inoculum. Cell viability was determined using trypan blue exclusion, and only single-cell suspensions of greater than 95% viability were used to determine the tumorigenic and metastatic potential in orthotopic xenograft models. In brief, nude mice were anesthetized with methoxyflurane and placed in the supine position. An upper midline abdominal incision was made, and the pancreas was exteriorized. Tumor cells (1 × 106 in 0.05 ml of HBSS) were then injected into the tail of the pancreas. The animals were killed 2 months after tumor-cell inoculation or when they became moribund. Tumors in the pancreas were harvested and weighed. In addition, each mouse's liver was fixed in Bouin's solution for 24 hr to differentiate the neoplastic lesions from the organ parenchyma. Metastases on the surface of liver were counted (double-blinded) with the aid of a dissecting microscope as described previously.9

Immunohistochemistry and quantitation of microvessel density

Pancreatic tumors harvested at autopsy were processed for immunostaining using anti-IL-8 (Biosource International, Camarillo, CA), anti-CD31/PECAM-1 (Pharmingen, San Diego, CA) and anti-VEGF (Santa Cruz Biotechnology) antibodies and appropriate peroxidase-conjugated secondary antibodies. The slides were examined using a bright-field microscope. A positive reaction was indicated by a reddish-brown precipitate in the cytoplasm. Negative controls were done using nonspecific IgG. The microvessel density of each tumor was analyzed using anti-CD31 immunostaining. Images were digitized using a Sony 3CD color video camera (Sony, Tokyo, Japan) and personal computer equipped with the Optimas image analysis software program (Optimas, Bothell, WA). For the quantification of MVD, 10 random 0.159 mm2 fields at 100× magnification were captured for each tumor, and microvessels were quantified according to the method described previously.9

Statistical analysis

The significance of the in vitro results was determined using Student's t-test (2-tailed). The significance of the in vivo metastasis results was determined using the Mann-Whitney U test.

RESULTS

Direct correlation of constitutive NF-κB activity with the metastatic potential of human pancreatic cells

In the first set of experiments, the constitutive NF-κB activity was determined in the COLO357 human pancreatic cancer cell system: COLO357, SG, FG and L3.3 cells. Both the NF-κB DNA-binding activity (electrophoretic mobility shift assay [EMSA]; Fig. 1a) and NF-κB promoter activity assay (Fig. 1b) showed increased NF-κB activity in highly metastatic L3.3 cells compared to poorly metastatic COLO357, SG and FG cells. A similar finding was observed in mouse pancreatic cancer cells, i.e., higher NF-κB DNA-binding activity in highly metastatic Panc02-H7 cells than that in poorly metastatic Panc02-H0 cells (Fig. 1c). Western blot experiments were performed using nuclear protein to confirm the quality/integrity of the nuclear extracts (Fig. 1d). These data suggest that constitutive NF-κB activity contributes to the increased metastatic potential of human pancreatic cancer.

Figure 1.

Constitutive NF-κB activity in human pancreatic cells differing in metastatic potential. (a) Nuclear protein was extracted from COLO357, SG, FG and L3.3 cells. The NF-κB DNA-binding activity was determined using EMSA. (b) Luciferase reporters driven by either wild-type (Wt-NF-κB) or mutant (Mut-NF-κB) NF-κB response elements were cotransfected with pB-Actin-RL into the pancreatic cancer cells listed above. The firefly and Renilla luciferase activity was quantified using the dual luciferase reporter assay system. The luciferase activity in test samples was compared to that in COLO357 cells transfected with Mut-NF-κB, which was assigned a value of 1. (c) The NF-κB DNA-binding activity also was determined by EMSA using nuclear protein extracts from Panc02-H0 (H0) and Panc02-H7 (H7) cells. (d) Western blot experiments were performed using nuclear protein to confirm the quality/integrity of the nuclear extracts using specific antibodies against p65 and Oct-1. This is one representative assay of 3 independent experiments.

Downregulation of constitutive NF-κB activity in metastatic human pancreatic cells via transfection of IκBαM

L3.3 cells were stably transfected with the IκB'3fM expression vector, which encodes a mutated IκBα. The expression of both endogenous and mutant IκBα was analyzed using Western blot analysis. As shown in Figure 2a, endogenous IκBα was detected in L3.3, control pLXSN-transfected (Neo) and IκBαM-transfected (IκBαM1 and IκBαM2) L3.3 cells, whereas exogenous mutant IκBα was detected only in IκBαM-transfected L3.3 cells.

Figure 2.

IκBαM transfection and NF-κB activity. Exogenous IκBαM expression. Cytosolic protein was extracted from L3.3 cells, L3.3 cells transfected with pLXSN (L3.3-Neo) and L3.3 cells transfected with pLXSN-IκBαM (L3.3-IκBαM1 and -IκBαM2). Western blot analysis was performed using a specific anti-IκBαantibody (a). NF-κB-binding activity. Nuclear protein was extracted from L3.3, L3.3-Neo and IκBαM-transfected (L3.3-IκBαM1, and -IκBαM2) cells. EMSA was performed using the NF-κB probe (b). This EMSA experiment was repeated using nuclear protein extracts from the same cells (c) and the quality/integrity of the nuclear extracts was confirmed by Western blot analysis using specific antibodies against p65 and Oct-1 (d). For NF-κB-Luc reporter activity, luciferase reporters driven by either wild-type (Wt-NF-κB) or mutant (Mut-NF-κB) NF-κB response elements were cotransfected with pB-Actin-RL into the pancreatic cancer cells listed above. The firefly and Renilla luciferase activity was quantified using the dual luciferase reporter assay system. The luciferase activity in test samples was compared to that in L3.3 cells transfected with Mut-NF-κB, which was assigned a value of 1 (e). This is 1 representative experiment of 2. Note that IκBαM transfection suppressed NF-κB activity (p < 0.01).

To determine whether expression of IκBαM led to blockade of the nuclear translocation of NF-κB, EMSA was performed using nuclear extracts of L3.3, L3.3-Neo, and IκBαM-transfected L3.3 cells. As shown in Figure 2b, constitutive NF-κB-binding activity was present in L3.3 cells. The expression of IκBαM significantly inhibited the NF-κB activity in IκBαM-transfected cells (IκBαM1 and IκBαM2) but not in the control cells. Western blot experiments were performed using nuclear protein to confirm the quality/integrity of the nuclear extracts (Fig. 2d). The specificity of the observed bandshift was verified using anti-p65 and -p50 antibodies (supershift), indicating that NF-κB complexes contained both p50 and p65 components (data not shown).

Furthermore, we confirmed the suppressive effect of IκBαM transfection on the constitutive level of NF-κB activity using an NF-κB-dependent luciferase reporter activity assay. A Wt-NF-κB (wild-type) or Mut-NF-κB (mutant) reporter15, 18 was transiently transfected into the cells. As shown in Figure 2e, the constitutive NF-κB reporter activity was significantly decreased in IκBαM-transfected IκBαM1 and IκBαM2 cells, which was consistent with the EMSA results (Fig. 2b,c). Therefore, we showed that L3.3 cells constitutively express NF-κB activity, which can be blocked by IκBαM transfection.

Suppression of the tumorigenicity and metastasis of human pancreatic cancer cells by NF-κB blockade

To evaluate the effects of NF-κB activity on the tumorigenicity and metastasis of human pancreatic cancer cells, L3.3, L3.3-Neo and L3.3-IκBαM cells were injected into the pancreas of nude mice (1 × 106/mouse). The mice were killed 2 months after tumor-cell inoculation, and the tumor incidence and liver metastasis rate were recorded (Table I). In contrast with L3.3 and L3.3-Neo cells, L3.3-IκM (IκBαM1 and IκBαM2) cells produced smaller tumors with a significantly decreased incidence of spontaneous liver metastasis. Therefore, inhibition of NF-κB activity by IκBαM transfection suppressed both tumorigenicity and metastasis of human pancreatic cancer cells.

Table I. Suppression of tumor growth and metastasis of L3.3 human pancreatic cancer cells by transfection of the IκBαM expression vector
Cell linePrimary pancreatic tumorLiver metastasis
IncidenceMedian weight (range)IncidenceMedian number (range)
  • 1

    L3.3-P, L3.3-Neo, L3.3-IκBαM1, and L3.3-IκBαM2 cells (1 × 106) were injected into the pancreas of nude mice. The animals were killed 2 months after tumor-cell inoculation or when they became moribund. Primary tumors in the pancreas were harvested and weighed. In addition, each mouse's liver was fixed in Bouin's solution for 24 h to differentiate the neoplastic lesions from the organ parenchyma. Metastases on the surface of liver were counted (double-blinded) with the aid of a dissecting microscope as described in “Materials and Methods”.

  • a

    P < 0.001, Mann-Whitney U test.

L3.310/10890 (670–1,100)8/1026 (0–67)
L3.3-Neo10/10990 (760–1,200)7/1019 (0–53)
L3.3-IκBαM15/10180 (0–420)a1/100 (0–4)1
L3.3-IκBαM26/10330 (0–540)a2/100 (0–3)1

In vitro growth of human pancreatic cells transfected with IκBαM

We next tested whether stable transfection of L3.3 cells with IκBαM affected cell growth and proliferation in vitro. L3.3, L3.3-Neo and IκBαM-transfected (IκBαM1 and IκBαM2) cells were seeded into 96-well plates (5 × 103 cells/well) for 48 hr in 10% CMEM. In vitro cell growth and proliferation were measured using a methylthiotetrazole (MTT) assay and [3H]TdR incorporation. The in vitro rates of growth of all of the lines were very similar (Fig. 3a,b), suggesting that the stable expression of IκBαM did not alter the in vitro growth of L3.3 cells.

Figure 3.

In vitro growth of human pancreatic cells transfected with an IκB dominant-negative expression vector. (a) MTT assay. A total of 5 × 103 L3.3, L3.3-Neo, L3.3-IκBαM1 and L3.3-IκBαM2 cells were cultured in vitro for 48 hr. The number of metabolically active cells was determined using the MTT assay. (b) Uptake of [3H]TdR. Cells were plated in 96-well plates for 24 hr; 0.5 μCi/ml [3H]TdR was then added. [3H]TdR incorporation was measured 12 hr later. No significant differences in proliferation were found in IκBαM-transfected and control cells. This is 1 representative experiment of 3.

Altered expression of apoptosis-regulatory molecules in human pancreatic cells transfected with IκBαM

The effect of NF-κB inhibition on the expression of apoptotic (both proapoptotic and antiapoptotic) molecules, including Bcl-2 and Bax, was determined in IκBαM-transfected L3.3 cells. As shown in Figure 4, there was increased Bcl-2 expression but decreased Bax expression in IκBαM-transfected cells when compared to that in control cells. These results suggested that overexpression of antiapoptotic molecules coupled with decreased expression of proapoptotic molecules may help tumor cells survive by overcoming the effect of decreased NF-κB activity.

Figure 4.

Expression of apoptosis-regulatory proteins. Cytosolic protein was extracted from L3.3 cells, L3.3 cells transfected with pLXSN (L3.3-Neo) and L3.3 cells transfected with pLXSN-IκBαM (L3.3-IκBαM1 and -IκBαM2). Western blot analysis was performed using specific anti-Bcl-2 and -Bax antibodies (Santa Cruz). Quantitative analysis was performed by densitometry, standardized to b-actin and expressed as fold difference from that of the control L3.3 cells, which was assigned a value of 1.0.

Downregulation of multiple proangiogenic molecules in metastatic human pancreatic cells transfected with IκBαM

The effect of NF-κB inhibition on the expression of proangiogenic molecules, including VEGF and IL-8, in IκBαM-transfected L3.3 cells was studied. Expression of the VEGF and IL-8 genes in IκBαM-transfected cells was further determined at both the mRNA level using Northern blot analysis and protein level using quantitative IL-8 and VEGF enzyme-linked immunosorbent assay (ELISA). As shown in Figure 5a, there was a significant decrease in VEGF and IL-8 mRNA expression in IκBαM1 and IκBα2 cells compared to that in control cells. Also, the IκBαM-transfected L3.3 cells consistently secreted VEGF and IL-8 into the culture supernatant at significantly decreased levels (Fig. 5b). These results suggested that NF-κB blockade transcriptionally represses the expression of VEGF and IL-8.

Figure 5.

Expression of proangiogenic molecules. (a) ELISA. The amount of IL-8 and VEGF protein in the culture supernatants as determined using ELISA was expressed as pg/106 cells/48 hr. (b) Northern blot analysis. Cellular mRNA was extracted, Northern blot analysis was performed using IL-8 and VEGF cDNA probes and equal loading of mRNA was monitored by hybridizing the same membrane filter with GAPDH and β-actin cDNA probes. (c) IL-8 and VEGF promoter activity. pGL2-IL-8 and pGL2-VEGF reporters were cotransfected with a pB-Actin-RL (Renilla) reporter into the pancreatic cancer cells. pGL2-basic luciferase reporters were used as negative controls. The luciferase activity was determined as described in Figure 1b. The fold increase in luciferase activity was calculated relative to the luciferase activity of pGL2-basic in L3.3 cells and assigned a value of 1. Note that IκBαM transfection suppressed IL-8 and VEGF protein production (p < 0.01).

Next, the promoter activity of the VEGF and IL-8 genes in IκBαM-transfected and control cells was analyzed. Consistent with the decreased NF-κB promoter activity, the promoter activity of VEGF and IL-8 was significantly suppressed in IκBαM-transfected L3.3 cells compared to that in parental L3.3 and L3.3-Neo cells (Fig. 5c).

Decreased tumor angiogenesis correlated with the inhibition of angiogenic molecule expression in vivo

To determine whether NF-κB activity blockade leads to impaired angiogenic potential in vivo, the expression of VEGF and IL-8 protein in vivo was evaluated using immunohistochemistry. As shown in Figure 6, staining for VEGF and IL-8 was observed in L3.3 and L3.3-Neo tumors but significantly decreased in IκBαM-transfected tumors. Furthermore, Northern blot analysis has confirmed the lower IL-8 and VEGF mRNA expression in IκBαM-transfected tumors than that in control tumors (Fig. 7a). Thus, expression of IκBαM in pancreatic cancer cells inhibited constitutive activation of NF-κB and subsequently suppressed expression of the VEGF and IL-8 genes in vivo as well as in vitro. The use of specific neutralizing antibodies against IL-8 and VEGF was sufficient to inhibit tumor angiogenesis (Fig. 7b), growth and metastasis (Table II).

Figure 6.

In vivo angiogenesis of tumor cells having altered NF-κB activity. L3.3, L3.3-Neo and L3.3-IκBαM cells were injected into the pancreas of nude mice. Similar-sized pancreatic tumors were harvested and processed for immunohistochemical analysis. The blood vessels were counted using a light microscope after immunostaining of tissue sections with anti-CD31 antibodies. VEGF and IL-8 immunostaining was performed as described in Materials and Methods. Note that tumors formed by IκBαM-transfected cells had lower levels of activated VEGF and IL-8 expression and were less vascularized than tumors produced by L3.3 or Neo-transfected cells were.

Figure 7.

Evaluation of angiogenic molecule expression and microvessel formation in the tumors growing in mice. (a) Northern blot analysis. Cellular mRNA was extracted from tumor specimens obtained from the experiment of Table I, Northern blot analysis was performed using IL-8 and VEGF cDNA probes and equal loading of mRNA was monitored by hybridizing the same membrane filter with □□□□-actin cDNA probes. (b) Frozen sections were prepared using similar-sized pancreatic tumors harvested from the experiments of Tables I and II. The blood vessels were counted using a light microscope after immunostaining the sections with anti-CD31 antibodies.

Table II. Suppression of tumor growth and metastasis of L3.3 human pancreatic cancer cells by neutralizing antibodies against VEGF and IL-81
TreatmentPrimary pancreatic tumorLiver metastasis
IncidenceMedian weight (range)IncidenceMedian no. (range)
  • 1

    L3.3-P cells (1 × 106) were injected into the pancreas of nude mice. Three weeks after tumor injection, anti-VEGF and -IL-8 antibodies (R & D, Systems, Minneapolis, MN) or IgG (100 μg/mouse) were injected i.p. once every 5 days for 30 days. The animals were killed 2 months after tumor-cell inoculation or when they became moribund. Primary tumors in the pancreas were harvested and weighed. In addition, each mouse's liver was fixed in Bouin's solution for 24 h to differentiate the neoplastic lesions from the organ parenchyma. Metastases on the surface of liver were counted (double-blinded) with the aid of a dissecting microscope as described in Materials and Methods.

  • a

    P < 0.001, Mann-Whitney U test.

Control10/10820 (550–1050)10/1016 (6–43)
Control IgG10/10768 (470–960)9/1012 (0–35)
Anti-IL-88/10256 (0–405)a6/104 (0–12)a
Anti-VEGF7/10186 (0–390)a5/106 (0–9)a
Anti-IL-8/-VEGF5/10110 (0–253)a2/100 (0–5)a

Finally, we sought to determine whether the decreased NF-κB activity and subsequent decrease in VEGF and IL-8 production led to suppression of tumor angiogenesis. Tumor-associated neovascularization (as indicated by MVD) was determined using immunohistochemistry with anti-CD31 antibodies. As shown in Figure 6, tumors formed by control cells were highly vascularized, whereas tumors formed by IκBαM-transfected IκBαM1 and IκBαM2 cells had decreased vascular density (Fig. 7b). These studies indicated that tumor-associated neovascularization correlated directly with NF-κB activity, tumorigenicity and metastasis of human pancreatic cells.

DISCUSSION

Our results demonstrated that there was an increased level of constitutive NF-κB activity in metastatic pancreatic cancer cells and that blockade of NF-κB activity by IκBαM transfection inhibited the cells' tumorigenic and metastatic properties. The antitumor activity was associated with suppression of angiogenesis, as IκBαM transfection resulted in significant downregulation of the expression of 2 key angiogenic molecules, VEGF and IL-8, in both cultured tumor cells and tumor cells growing in the pancreas of nude mice. The decreased expression of VEGF and IL-8 in vivo directly correlated with decreased neovascularization. Therefore, our results provided direct evidence of the involvement of NF-κB in the molecular control of angiogenesis and metastasis of pancreatic cancer.

The aggressive growth and metastasis of pancreatic cancer depend on angiogenesis, which is mediated by angiogenic factors derived from both tumor and stromal cells.5 Pancreatic cells secrete a variety of proangiogenic molecules, including VEGF and IL-8, which have been identified as key mediators of pancreatic cancer angiogenesis.6, 7, 8, 9, 10 Specifically, elevated expression of VEGF has been reported in human pancreatic cancer biopsy specimens.7, 27 Increasing evidence suggests that increased VEGF expression is attributed to a plethora of external regulatory factors. Major stimulators of VEGF expression include hypoxia and acidosis,17, 34 which occur frequently in diverse types of expanding tumors, particularly in regions surrounding necrotic areas. The major role of hypoxia in the angiogenesis switch has been clearly demonstrated in a mouse model.35 It is clear that hypoxia-mediated upregulation of VEGF mainly involves the HIF-1 pathway.36 However, in many types of tumors, elevated VEGF production can often be detected in tumor cells located in the extreme periphery of the tumor, where there is no apparent hypoxia.37 Moreover, VEGF promoter analyses have revealed many other potential transcription factor-binding sites, such as AP-1, AP-2, Egr-1, Sp1 and NF-κB,38, 39, 40, 41, 42 suggesting that multiple signal transduction pathways may be involved in VEGF transcription regulation. For example, activation of NF-κB by many growth factors and cytokines may contribute to elevated VEGF expression in tumor tissues.43, 44, 45 These observations are consistent with numerous recent findings indicating that exogenous factors such as hormones, cytokines and growth factors modulate VEGF expression and then angiogenesis.46

Conversely, many pancreatic cancer cells can constitutively express VEGF in vitro without any apparent external stimulation.8 Previous studies indicated that most pancreatic cancer cells exhibited constitutive activation of NF-κB.18, 47 Also, several previous findings suggested that NF-κB regulates VEGF expression.17, 33 More recently, it was shown that VEGF induction by ultraviolet (UV)-irradiated skin-derived cell lines can be blocked through treatment using NF-κB decoy oligodeoxynucleotides, which inhibits NF-κB activity.48 This has been substantiated by the identification of putative NF-κB-binding sites on 3′ regulatory regions of the VEGF gene.17 Our study clearly demonstrated that constitutive activation of NF-κB contributes, at least in part, to constitutive VEGF overexpression in human pancreatic cancer cells. Furthermore, the significant decrease in VEGF and IL-8 promoter activity found in IκBαM-transfected cells suggested that NF-κB regulates constitutive expression of these genes at least in part at the transcriptional level. Although the upstream signal pathway leading to constitutive NF-κB activation remains to be elucidated, several lines of evidence point to the roles of mutations of oncogenes and/or tumor suppressor genes.24, 25, 26 In fact, loss or inactivation of tumor suppressor genes, such as p53, and activation of oncogenes, such as Ras, are associated with VEGF overexpression.49, 50

As a pleiotropic transcription factor, NF-κB also controls the expression of many other genes important to tumor growth and metastasis. It is well established that NF-κB is essential for both inducible and constitutive IL-8 expression.10, 14, 18 In fact, human pancreatic cancer cells secrete the cytokine IL-8 in both a constitutive and inducible manner. Because IL-8 may promote tumor growth and metastasis,10 primarily through its ability to act as an angiogenic factor,13 constitutive NF-κB activity drives the constitutive overexpression of VEGF and IL-8 and contributes to the elevated angiogenic phenotype and aggressive biology of human pancreatic cancer. However, many other mechanisms may also contribute to the antitumor activity of NF-κB blockade. For example, it has been shown that NF-κB blockade results in inhibition of cell adhesion,51 proinflammatory cytokine production52 and plasminogen activator uPA19 or MMP-9,15, 16 which are important for tumor angiogenesis, growth and metastasis. NF-κB activation is also involved in the regulation of cell proliferation and apoptosis.28, 53, 54 However, consistent with previous reports showing that stable inhibition of NF-κB in cancer cells by stable transfection of IκBαM does not inhibit cell growth in vitro,15, 16, 18, 19 no discernible differences in the in vitro growth rate were found between IκBαM-transfected and control cells. Our results, together with those of several previous studies, strongly indicate that blockade of NF-κB activity also leads to a decrease in angiogenesis and metastatic potential without a decrease in the in vitro growth rate.15, 16 Interestingly, we found that IκBαM-transfected cells had an elevated level of antiapoptotic Bcl-2 protein expression but a decreased level of proapoptotic Bax protein expression. Although many other molecules may be involved in regulating pancreatic cancer cells survival/apoptosis, it is possible that imbalanced expression of apoptosis-regulatory proteins directly promotes in vitro survival of IκBαM-transfected cells, whereas the angiogenic potential of tumor cells is not crucial to in vitro survival of tumor cells. In sharp contrast, both the survival and angiogenic phenotype are essential for tumor growth and metastasis. Therefore, in this experimental system, suppression of tumor growth and metastasis by NF-κB blockade may be mainly due to impaired angiogenic potential of tumor cells.

In summary, human pancreatic cancer cells have high levels of constitutive NF-κB activity, and many cytokines and growth factors, which are present in the pancreatic cancer microenvironment, can activate NF-κB. NF-κB activation in pancreatic cancer cells may provide a growth advantage via multiple mechanisms, including increased angiogenesis and invasion ability as well as suppression of the apoptotic response. In our study, suppression of NF-κB activity decreased angiogenesis and hence retarded tumor growth and metastasis through the downregulation of multiple angiogenic molecules, including VEGF and IL-8, suggesting a critical role for NF-κB in angiogenesis, growth and metastasis. Targeting NF-κB may therefore be a potential approach for controlling the growth and metastasis of human pancreatic cancer.

Acknowledgements

We thank Dr. P. Chiao for the pLXSN-IκBαM expression construct, Dr. B. Su for the 2x NF-κB reporter construct and Drs. N. Mukaida and K. Matsushima for the pxp2-IL8-1481 promoter. The authors are grateful to Mr. D. Norwood for editorial comments and Ms. J. King for assistance in the preparation of the manuscript.

Ancillary