SEARCH

SEARCH BY CITATION

Keywords:

  • pancreatic cancer;
  • angiogenesis;
  • CXCL8/IL-8;
  • CXCL12/ SDF-1α

Abstract

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

CXC-chemokines are involved in the chemotaxis of neutrophils, lymphocytes and monocytes. However, role of these chemokines in tumorigenesis, especially with regard to interaction between tumor and its microenvironment, has not been clearly elucidated. The purpose of this study was to analyze the co-operative role of CXCL8 and CXCL12 in the tumor-stromal interaction in pancreatic cancer (PaCa). Using enzyme-linked immunosorbent assay (ELISA) and reverse transcription polymerase chain reaction (RT-PCR), we initially confirmed the expression of ligands and receptors, respectively, of CXC-chemokines in PaCa and stromal cells. We examined the co-operative role of CXCL8 and CXCL12 in proliferation/invasion of PaCa and human umbilical vein endothelial cells (HUVECs), and in HUVEC tube-formations through tumor-stromal interaction by MTS, Matrigel invasion, and angiogenesis assays, respectively. We detected expression of CXCR4, but not CXCR2, in all PaCa cells and fibroblasts. PaCa cells secreted CXCL8, and fibroblast cells secreted CXCL12. CXCL8 production in PaCa was significantly enhanced by CXCL12, and CXCL12 production in fibroblasts was significantly enhanced by co-culturing with PaCa. CXCL8 enhanced proliferation/invasion of HUVECs but did not promote proliferation/invasion of PaCa. Both recombinant and PaCa-derived CXCL8 enhanced tube formation of HUVECs that were co-cultured with fibroblast cells. CXCL12 enhanced the proliferation/invasion of HUVECs and the invasion of PaCa cells but had no effect on tube formation of HUVEC. We showed that PaCa-derived CXCL8 and fibroblast-derived CXCL12 cooperatively induced angiogenesis in vitro by promoting HUVEC proliferation, invasion, and tube formation. Thus, corresponding receptors CXCR2 and CXCR4 are potential antiangiogenic and antimetastatic therapeutic targets in PaCa. © 2008 Wiley-Liss, Inc.

Pancreatic cancer (PaCa) is the fourth leading cause of cancer-related deaths in the United States, with about 32,000 newly diagnosed cases and an equal number of deaths occurring annually.1 The poor prognosis of PaCa is attributable to its tendency for late presentation, aggressive local invasion, early metastases, and poor response to chemotherapy.2 Thus, a better understanding of the fundamental nature of this cancer is needed to improve the clinical outcome.

Recently, there has been increasing evidence that chemokines have a role in tumor biology. Chemokines were first described as small peptides controlling cell migration, especially that of leukocytes during inflammation and immune response. Since then, a broad spectrum of biological activities has been described as chemokine-regulated tumorigenesis, which affects tumors and their microenvironment. The role of chemokines in tumor biology is important because these peptides may influence tumor growth, invasion, and metastasis.3–11 CXCL8 (interleukin-8 [IL-8]), one of the ELR (Glu-Leu-Arg) motif-positive (ELR+) CXC-chemokines, is secreted by leukocytes and tumor cells.12 CXCL8-mediated cellular response is affected by its two high-affinity cell surface receptors, G protein-couple receptors CXCR1 and CXCR2, which CXCL8 cross-links with and exerts biological function on.13 CXCL8 has diverse functions in immune surveillance, inflammation, and angiogenesis. Most primary and metastatic solid tumors, such as breast, uterine, prostate, colon and pancreatic carcinomas, melanoma, and glioblastoma, constitutively express IL-8 (CXCL8).12–18 A known function of CXCL8 in tumors is the enhancement of angiogenesis.15 Several studies have shown that tumor-derived CXCL8 directly modulates endothelial cell proliferation and migration, thus promoting angiogenesis.16, 17 On the other hand, stromal cell-derived factor-1 (SDF-1), now designated as CXCL12,19 is a homeostatic chemokine that signals through its receptor CXCR4,20 which in turn plays an important role in hematopoiesis and development and organization of the immune system. Initial studies on CXCR4 focused on its role in the pathogenesis of HIV infection.21 Subsequently, the correlation of hematopoietic malignancies with the CXCL12-CXCR4 axis was shown in many reports.22–24 Recently, several other nonhematopoietic neoplasias also have been described to express the CXCR4 chemokine receptor. Activation of CXCR4 by CXCL12 induces migration, angiogenesis, and/or survival of the neoplastic cells, including tumor cells from neuroblastoma cells,25 colorectal cancer,26 prostate cancer,27 melanoma,28 ovarian cancer29 and PaCa.30 These reports described the important role of these chemokines in tumor angiogenesis.

Tumor angiogenesis, the formation of new capillaries from the existing vascular network, is essential for tumor growth and metastasis. The progressive growth of malignant solid tumors is dependent on the development of new blood vessels that provide oxygen and nutrients to tumor cells.31 Angiogenesis is a complex multistep process involving extracellular matrix remodeling, endothelial cell migration and proliferation, and capillary tube formation.32 In tumors, these steps depend on the production of angiogenic factors by both tumor and stromal cells.33, 34 The new blood vessels embedded in the tumor provide a gateway for tumor cells to enter the circulation and to metastasize to distant organs such as the liver or lungs.

CXCL8 and CXCL12 could play an important role in tumor progression and angiogenesis in various tumors, but the co-operative functions of these chemokines have not been clearly elucidated. The purpose of this study was thus to determine the co-operative effects of CXCL8 and CXCL12 on PaCa angiogenesis, especially with regards to the interaction between the tumor and the microenvironment.

Material and methods

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Cell culture

PaCa cell lines BxPC-3, AsPC-1, MIA PaCa-2, Panc-1, Capan-2 and SW 1990 were obtained from the American Type Culture Collection (Rockville, MD). BxPC-3 and AsPC-1 cells were maintained in RPMI-1640 (Sigma Chemical, St. Louis, MO) supplemented with 10% fetal calf serum (FCS). MIA PaCa-2, Panc-1, Capan-2, and SW 1990 cells were maintained in Dulbecco's modified Eagle's medium (Sigma Chemical, St. Louis, MO) with high glucose and 10% FCS. Human umbilical vein endothelial cells (HUVECs) were obtained from Kurabo (Osaka, Japan) and maintained in HuMedia-EG2 medium supplemented with 2% FCS, 5 ng/mL basic fibroblast growth factor, 10 μg/mL heparin, 10 ng/mL epidermal growth factor, and 1 μg/mL hydrocortisone according to the supplier's instructions (Kurabo, Osaka, Japan). Fibroblasts were obtained from Lonza (Walkersville, MD) and maintained in FGM-2 medium supplemented with 2% FCS, 1 ng/mL basic fibroblast growth factor, and 1 μg/mL insulin according to the supplier's instructions (Lonza, Walkersville, MD). The human pancreatic ductal epithelium (HPDE) cells were a generous gift from Dr. Ming-Sound Tsao (University of Toronto, Ontario, Canada). HPDE cells were cultured in keratinocyte serum-free (KSF) medium supplied with 5 ng/mL EGF and 50 μg/mL bovine pituitary extract (Invitrogen, Carlsbad, CA). All cells were incubated at 37°C in a humidified atmosphere of 5% CO2 in air.

Reagents and antibodies

Recombinant human CXCL8 and CXCL12 were provided by R&D Systems (Minneapolis, MN). Neutralizing monoclonal anti-human CXCL8, CXCL12, and CXCR4 antibody (anti-CXCL8 Ab, anti-CXCL12 Ab and anti-CXCR4 Ab) were provided by R&D Systems (Minneapolis, MN).

RT-PCR analysis

Total RNA was prepared from all cell lines using an RNeasy Mini Kit by Qiagen (Valencia, CA). Reverse transcription polymerase chain reaction (RT-PCR) was performed according to the manufacturer's instructions using an OneStep RT-PCR Kit by Qiagen. For RT-PCR of CXCR2, we used CXCR2 primer pair kit from R&D Systems (PCR product size, 350 bp; GenBank Accession number, NM_001557), and for CXCR4 RT-PCR, we used the following pairs of forward and reverse primer sets: 5′-GAAGCTGTTGGCTGAAAAGG-3′ and 5′-GAGTCGATGCT GATCCCAAT-3′ (PCR product size, 345 bp; GenBank accession number, NM_003467). For CXCR2, PCR was performed for 35 cycles with denaturation at 94°C for 45 sec, annealing at 55°C for 45 sec, and extension at 72°C for 45 sec; for CXCR4, PCR was performed for 30 cycles with denaturation at 94°C for 30 sec, annealing at 54°C for 30 sec, and extension at 72°C for 60 sec. Amplified DNA fragments were resolved by electrophoresis on 1.2% agarose gels containing ethidium bromide.

Enzyme-linked immunosorbent assay

All cell lines were seeded at a density of 2 × 105 cells/mL into a 24-well plate containing medium with 10% FCS and cultured overnight. Medium was then exchanged, and cells were cultured for a further 48 hr. The culture media were then collected and microfuged at 1,500 rpm for 5 min to remove particles, and the supernatants frozen at −80°C until use in the enzyme-linked immunosorbent assay (ELISA). The concentrations of CXCL8 and CXCL12 were measured using an ELISA kit (R&D Systems) according to the manufacturer's instructions. To determine the effect of CXCL12 on CXCL8 production in PaCa cells, we stimulated Panc-1 cells with CXCL12 (20 ng/mL) cultured the cells for 48 hr, and then measured the concentration of CXCL8 using the aforementioned methods. To determine the synergistic effect of the tumor-stromal interaction, we cultured fibroblast cells (1 × 105 cells into 24 well-plates) with or without SW 1990 (5 × 104 cells into inserts with 0.4-μm pores [BD Biosciences, Franklin Lakes, NJ]) for 48 hr and then measured the CXCL12 concentration using the aforementioned methods. Also, to examine the effects of fibroblast-derived CXCL12 on CXCL8 production from PaCa, PaCa cell lines [Panc-1, AsPC-1 and Mia PaCa-2 cells (2 × 105 cells into 24-well plates)] were co-cultured with or without fibroblast cells (5 × 104 cells into inserts with 0.4-μm pores [BD Biosciences]) for 48 hr using a double chamber method, and subsequently CXCL8 concentrations were measured as mentioned earlier. Moreover, to determine the effects of anti-CXCR4 Ab (CXCR4 Ab) on enhanced CXCL8 production from PaCa by fibroblasts, PaCa cells were pretreated with CXCR4 Ab (15 μg/mL) for 1 hr, then co-cultured with fibroblast cells for 48 hr, and CXCL8 concentrations were measured. Each condition was assessed in 5 independent samples.

Proliferation assay

To confirm the effect of these chemokines on HUVECs, we initially performed the proliferation assay using the Celltiter 96 Aqueous One Solution Cell Proliferation Assay (MTS-assay) (Promega, Madison, WI) according to the manufacturer's instructions. Briefly, HUVECs were seeded at a density of 5 × 103 cells/100 μL in 96-well plates and allowed to adhere overnight. Then, cultures were re-fed with fresh media containing various concentrations of CXCL8 or CXCL12. After 48-hr incubation, 20 μl of CellTiter 96 Aqueous One Solution Reagent was added to each well and the trays were incubated for 3 hr at 37°C, after which absorbance was measured using a microplate reader with a test wave length of 490 nm. Each condition was assessed in 5 independent samples.

Invasion assay

In vitro invasion assay was performed using the BD Bio-Coat Matrigel invasion assay system (BD Biosciences) according to the manufacturer's instructions. Briefly, BxPC-3 cells (2 × 105 cells) or HUVECs (5 × 104 cells) were suspended in medium containing 2% FCS and seeded into the Matrigel precoated transwell chambers consisting of polycarbonate membranes with 8-μm pores. The transwell chambers were then placed into 24-well plates, into which we added basal medium only or basal medium containing various concentrations of CXCL8 or CXCL12. After incubating BxPC-3 cells for 24 hr and HUVECs for 16 hr, the upper surface of the transwell chambers was wiped with a cotton swab and the invading cells were fixed and stained with Diff-Quick stain. The number of invading cells was counted in 5 random microscopic fields (200×). To confirm whether fibroblast-derived CXCL12 caused an increase in the invasive potency of PaCa cells, we performed an invasion assay for BxPC-3 using a double-chamber method. Briefly, we co-cultured BxPC-3 cells (2 × 105 cells into transwell chambers) with fibroblast cells (1 × 105 cells into 24-well plates) blocking with or without anti-CXCL12 Ab (10 μg/mL) for 24 hr; then, invaded cells were counted in the same manner. Likewise, to investigate the interaction between endothelial cells and PaCa or fibroblast cells, we co-cultured HUVECs with PaCa (BxPC-3 or MIA PaCa-2) or with fibroblast cells blocking with or without anti-CXCL8 Ab (for PaCa) or anti-CXCL12 Ab (for fibroblasts). Each condition was assessed in triplicate.

Zymography

HUVEC cells were seeded at a density of 2 × 106 cells/3mL into 60-mm dish and cultured overnight. Medium were then exchanged and cells were cultured for a further 24 hr treated with or without recombinant CXCL12 (100 ng/mL). Using these supernatants and cell lysates, Zymography assay was performed. The samples were mixed with sample buffer without reducing agent or heating, and loaded into Zymogram gels (Invitrogen, Carlsbad, CA). After the electrophoresis with constant voltage, the gel was treated with Zymogram Renaturing Buffer (Invitrogen) and Zymogram Developing Buffer (Invitrogen) according to their instructions. Finally, the gel was stained with SimplyBlue Safestain (Invitrogen) according to the instruction. In the same way, the supernatants of AsPC-1, Panc-1 and BxPC-3 cells treated with or without CXCL12 were collected and Zymography analysis was performed as described earlier.

HUVEC tube formation assay for angiogenesis by co-culturing PaCa, HUVEC and fibroblasts

To investigate the influence of CXCL8 on tube formation by HUVECs, we co-cultured HUVECs and fibroblast cells in basal medium containing different concentrations of CXCL8 using an angiogenesis kit (Kurabo) according to the manufacturer's protocols.35, 36 Briefly, HUVECs and fibroblasts were co-cultured in 24-well plates with basal medium only (control) or basal medium containing CXCL8 (1 or 10 ng/mL). Media were changed every 3 days, and HUVECs and fibroblasts were co-cultured for a total of 11 days. HUVECs were then stained with anti-CD31 antibody according to the manufacturer's protocols. The tube formation area was measured quantitatively over 15 different fields for each condition using an image analyzer (Kurabo). To investigate the influence of PaCa on tube formation by HUVECs, we co-cultured PaCa cell lines (BxPC-3 or MIA PaCa-2), HUVECs, and fibroblasts using a double-chamber method in 24-well plates. BxPC-3 or MIA PaCa-2 cells (1 × 104 cells) were seeded into transwell chambers consisting of polycarbonate membranes with 0.45-μm pores (Kurabo) and allowed to adhere overnight. The chambers were then placed into the HUVEC/Fibroblasts co-culture system and exchanged on Day 7. Cells were cultured for totally 11 days, and HUVEC tube formation was evaluated as described earlier. The assay allowed us to evaluate angiogenesis quantitatively and to examine tumor-stromal cell interactions. Using the same method, we assessed the effects of anti-IL-8 Ab (10 μg/mL) on HUVEC tube formation in the presence of PaCa cells. Each condition was assessed in triplicate.

HUVEC tube formation assay for angiogenesis in vitro

To confirm the effect of CXCL12 on tube formation in HUVECs, we performed an angiogenesis assay on Matrigel (BD Biosciences). For reconstitution of a basement membrane, Matrigel was diluted 2-fold with cold DMEM (without FCS) and added to the 24-well tissue culture plate (250 μL/well) at 4°C. The 24-well plate was incubated for 2 hr in a 37°C cell culture incubator to allow the Matrigel to solidify. HUVECs were trypsinized, counted, resuspended in basal medium, and added on top of the reconstructed basement membrane (5 × 104 cells/well) in the absence or presence of various concentrations of CXCL12 or anti-CXCL12 Ab (10 μg/mL). Cells were incubated for 16 hr to allow formation of capillary-like structures. Endotubes were quantified by counting 9 random fields/sample under the microscope (100×). Each condition was assessed in triplicate.

Statistical analysis

Differences in the mean of 2 samples were analyzed by an unpaired t test. Multiple group comparisons were performed by one-way analysis of variance (ANOVA) with a post hoc test for subsequent individual group comparisons. A p value < 0.05 was considered statistically significant. Mean values and standard deviations (SDs) were calculated for experiments performed in triplicate (or more).

Results

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

RT-PCR for CXCR2 and CXCR4 mRNA

Our RT-PCR experiments revealed that none of the cell lines expressed CXCR2 mRNA but that all these cell lines and fibroblasts did express CXCR4 mRNA (Fig. 1a).

thumbnail image

Figure 1. Expression of chemokine receptors and ligands in PaCa and stromal cells. (a) Detection of CXCR2 and CXCR4 mRNA in PaCa and stromal cells. PCR products were subjected to 1.2 % agarose gel electrophoresis and stained with ethidium bromide. (b) Quantification of CXCL8 and CXCL12 proteins in culture medium of PaCa and stromal cells. All cells were cultured for 48 hr, and then the concentrations of both chemokines were measured by ELISA. (c) Change in chemokine production by tumor-stromal interaction. Alteration of CXCL8 production from PaCa (Panc-1) by stimulation with CXCL12 (20 ng/mL) (left) and alternation of CXCL12 production in fibroblasts by co-culturing with PaCa (SW 1990) (right). Values are expressed as mean ± SD. *, p < 0.01.

Download figure to PowerPoint

ELISA for measurement of CXCL8 and CXCL12 protein secretion

Using ELISA, we measured secreted levels of CXCL8 and CXCL12 in PaCa cell lines and stromal cells. On the basis of the assay results for CXCL8, we classified the PaCa cell lines into 2 groups: those cell lines secreting high levels of CXCL8 (BxPC-3 and SW 1990) and those secreting low levels of it (MIA PaCA-2, Panc-1, Capan-2 and AsPC-1). Levels of CXCL12 secreted by PaCa cell lines were undetectable by ELISA. Fibroblast cells secreted high levels of CXCL12 (20 ng/mL) in culture medium (Fig. 1b). CXCL8 production in PaCa cells was significantly increased by cellular stimulation with recombinant CXCL12. Likewise, secretion levels of CXCL12 by fibroblast cells were increased by co-culturing with PaCa cells (Fig. 1c). CXCL8 production from PaCa cells was significantly enhanced by co-culturing with fibroblasts (2.34, 3.17 and 6.91 folds in Panc-1, AsPC-1 and MIA PaCa-2, respectively). Moreover, the enhanced CXCL8 production by co-culturing with fibroblast cells was significantly inhibited by the addition of CXCR4 Ab (Fig. 2).

thumbnail image

Figure 2. Effects on CXCL8 production from PaCa by co-culture with fibroblasts. The alteration of CXCL8 production from PaCa (Panc-1, AsPC-1 and MIA PaCa-2) by co-culturing with fibroblast (FB) was evaluated by ELISA as described in “Materials and methods” section. Also, to examine the role of fibroblast-derived CXCL12 on enhanced CXCL8 production from PaCa cell lines, PaCa cells were pretreated with anti-CXCR4 Ab (Ab) for 1 h and ELISA experiments were performed as described in “Materials and methods” section. FB; fibroblast only, control; PaCa cells only, with FB; PaCa cells co-cultured with fibroblast, with FB + Ab; PaCa cells co-cultured with fibroblast treated with anti-CXCR4 Ab. Values are expressed as mean ± SD. *p < 0.01.

Download figure to PowerPoint

Effects of CXCL8 and CXCL12 on proliferation of PaCa cells, fibroblasts and HUVECs

Having confirmed the expression of CXCL8 in PaCa cell lines and of CXCL12 in fibroblasts, we used the MTS assay to determine the effect of these chemokines on cell growth. Neither CXCL8 nor CXCL12 enhanced the growth of PaCa cells (BxPC-3, AsPC-1, MIA PaCa-2 and Panc-1) (data not shown). To determine whether tumor cell-derived CXCL8 affected the proliferation of PaCa cells in an autocrine manner, we also performed an MTS assay using neutralizing anti-CXCL8 antibody (CXCL8 Ab). However, owing to the blockade of CXCL8 by CXCL8 Ab, there was no significant inhibition of growth in PaCa cells (data not shown). Also, neither CXCL8 nor CXCL12 promoted proliferation of fibroblasts (data not shown). HUVEC proliferation, however, was significantly enhanced by stimulation with both CXCL8 and CXCL12 (p < 0.01) (Fig. 3).

thumbnail image

Figure 3. Effects of CXCL8 and CXCL12 on proliferation of HUVECs. Cell proliferation was assessed using the MTS-assay as described in Material and Methods. Multiple comparisons were performed by one-way ANOVA followed by the Student-Newman-Keuls test (SNK test). Values are expressed as mean ± SD. *p < 0.01 and **p < 0.05 vs. control.

Download figure to PowerPoint

CXCL12, but not CXCL8, enhances PaCa invasiveness

Initially, we used the Matrigel double-chamber invasion assay to determine whether recombinant CXCL8 and CXCL12 modulated the invasiveness of PaCa cells. The invasive behavior of BxPC-3 was significantly enhanced by stimulation with recombinant CXCL12 in a dose-dependent manner (p < 0.01). The invasive ability of BxPC-3 was also significantly enhanced by co-culturing with fibroblasts in the lower chamber (p < 0.01); enhancement of invasiveness in BxPC-3 was inhibited by preincubating these cells with neutralizing anti-CXCL12 antibody (p < 0.01). CXCL8 did not affect the invasiveness of PaCa cells (Figs. 4a and 4b).

thumbnail image

Figure 4. Effects of CXCL8 and CXCL12 on invasiveness of PaCa and HUVECs. Both PaCa invasiveness and HUVEC invasiveness were assessed by the BD Bio-Coat Matrigel invasion assay system (BD Biosciences) as described in “Material and methods” section. (a, b) PaCa invasion assay. PaCa (BxPC-3) invasion assay was performed under the condition of basal medium containing various concentrations of CXCL8 or CXCL12, or co-culture with fibroblast treated with or without anti-CXCL12 Ab (10 μg/mL). Invading cells were fixed and stained with Diff-Quick stain. The invading cells were counted in 5 random microscopic fields (×200). Multiple comparisons were performed by one-way ANOVA followed by the SNK test. Bars indicate SD, *p < 0.01 compared with control. (b1) control; (b2) cultured with CXCL8 (10 ng/mL); (b3) cultured with CXCL12 (10 ng/mL); (b4) co-cultured with fibroblasts; (b5) co-cultured with fibroblasts treated with anti-CXCL12 Ab (10 μg/mL). (c, d) HUVEC invasion assay. HUVEC invasion assay was performed under the condition of basal medium only or basal medium containing various concentrations of CXCL8 or CXCL12. To assess the interaction between HUVECs and PaCa or fibroblasts, HUVECs were co-cultured with PaCa (MIA PaCa-2 or BxPC-3) treated with or without anti-CXCL8 Ab or co-cultured with fibroblasts treated with or without anti-CXCL12 Ab. The invading cells were stained with Diff-Quick stain and counted in 5 random microscopic fields (×200). Multiple comparisons were performed by one-way ANOVA followed by the SNK test. Bars indicate SD, *p < 0.01 compared with control. (d1) control; (d2) cultured with CXCL8 (10 ng/mL); (d3) cultured with CXCL12 (10 ng/mL); (d4) co-cultured with BxPC-3; (d5) co-cultured with BxPC-3 treated with anti-CXCL8 Ab (10 μg/mL); (d6) co-cultured with MIA PaCa-2; (d7) co-cultured with MIA PaCa-2 treated with anti-CXCL8 Ab (10 μg/mL); (d8) co-cultured with fibroblast; (d9) co-cultured with fibroblast treated with anti-CXCL12 Ab (10 μg/mL). [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

Both CXCL8 and CXCL12 enhance HUVEC invasion in a dose-dependent manner

As expression of CXCR2 (the receptor of CXCL8) in vascular endothelial cells has been already reported,37, 38 we reconfirmed the effect of CXCL8 on the invasiveness of HUVECs using the Matrigel double-chamber invasion assay. CXCL8 enhanced HUVECs' invasiveness in a dose-dependent manner (p < 0.01). Furthermore, having grouped PaCa cell lines according to their CXCL8 secretion, into high- and low-CXCL8 secretion groups, we examined how PaCa cells with different CXCL8 secretion potential influence the invasive ability of HUVECs. Co-culturing with BxPC-3 cells (high CXCL8 secretion group) significantly enhanced the invasive ability of HUVECs compared with these cells' invasiveness when co-cultured with MIA PaCa-2 cells (low-CXCL8 secretion group) (p < 0.01). The BxPC-3-enhanced invasive ability of HUVECs was significantly inhibited by treatment with anti-CXCL8 Ab (p < 0.01). Also, recombinant CXCL12 enhanced the invasive ability of HUVECs in a dose-dependent manner. The invasive ability of HUVECs was also significantly enhanced by co-culturing with fibroblasts; that enhancement was inhibited by the addition of anti-CXCL12 Ab (p < 0.01) (Figs. 4c and 4d).

Effects on MMP-2 and MMP-9 activities in HUVEC or PaCa cells following treatment with CXCL12

As CXCL12 enhanced both HUVEC and PaCa invasion, we performed zymography experiments to determine MMP-2 and MMP-9 activities in HUVEC and PaCa upon stimulation with CXCL12 (Supp. Info. Fig. 2). We observed for the first time that HUVEC has a basal MMP-9 activity (in supernatant), not significantly enhanced by CXCL12 (Supp. Info. Fig. 2A). We also performed similar experiments with PaCa cells. Interestingly, MMP-9 activity of PaCa cells is not enhanced by CXCL12. However, Panc-1 cells showed a modest increase in MMP-2 activity upon CXCL12 stimulation (Supp. Info. Fig. 2B).

Effect of chemokines on HUVEC tube formation

Since CXCL8 stimulated endothelial cell migration, we next examined this cytokine's effect on tube formation by HUVECs. HUVEC tube formation was significantly enhanced by treatment with recombinant CXCL8. Moreover, the high-CXCL8-secreting BxPC-3 cells significantly enhanced HUVEC tube formation; that enhancement was inhibited by treatment with anti-CXCL8 Ab (Figs. 5a–5c). On the other hand, recombinant CXCL12 did not significantly affect HUVEC tube formation (Figs. 5d and 5e).

thumbnail image

Figure 5. Effects of both chemokines on angiogenesis. (a) Effects of CXCL8 on tube formation by HUVEC. After incubation of the HUVEC/fibroblasts co-culture system (Kurabo) in the presence or absence of CXCL8 for 11 days, HUVECs were stained with anti-CD31 antibody. Tube formation area was measured quantitatively using an image analyzer. Multiple comparisons were performed by one-way ANOVA followed by the SNK test. Bars indicate SD, *p < 0.01 compared with control. (b) Effects of PaCa cells with different concentrations of CXCL8 production on tube formation by HUVECs. BxPC-3 or MIA PaCa-2 cells were cultured with HUVECs and fibroblasts using a double chamber treated without anti-CXCL8 Ab (white columns) or with anti-CXCL8 Ab (black columns). The tube formation by HUVEC for each condition was measured using an image analyzer. Multiple comparisons were performed by one-way ANOVA followed by an SNK test. Bars indicate SD, *p < 0.01 compared with treatment with anti-CXCL8 Ab and MIA PaCa-2 cells. (c) Tube formation assay for angiogenesis by co-culturing PaCa, HUVEC, and fibroblasts. (c1) control; (c2) cultured with CXCL8 (1 ng/mL); (c3) cultured with CXCL8 (10 ng/mL); (c4) co-cultured with BxPC-3; (c5) co-cultured with BxPC-3 treated with anti-CXCL8 Ab (10 μg/mL); (c6) co-cultured with MIA PaCa-2; (c7) co-cultured with MIA PaCa-2 treated with anti-CXCL8 Ab (10 μg/mL) (×40). (d) Effect of CXCL12 on HUVEC tube formation. Tube formation assay for angiogenesis on Matrigel was performed as described in “Material and methods” section. The number of endotubes was additionally quantified by counting 9 random fields/sample under the microscope (×100). (e) Tube formation assay for angiogenesis on Matrigel. (e1) control, (e2) cultured with CXCL12 (10 ng/mL); (e3) cultured with CXCL12 (100 ng/mL); (e4) cultured with anti-CXCL12 Ab (10 μg/mL).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Our study demonstrates the co-operative roles of CXCL8 and CXCL12 in angiogenesis with regards to the interaction between PaCa and stromal cells. CXCL12 significantly enhanced CXCL8 production in PaCa cells and CXCL12 production in fibroblasts was increased by co-culturing with PaCa cells. These results suggest that interaction between PaCa and stromal cells co-operatively regulates the production of these chemokines so that angiogenesis is enhanced. In angiogenesis, complex and diverse cellular actions are implicated, such as extracellular matrix degradation, proliferation and migration of endothelial cells, and morphologic differentiation of endothelial cells to form tubes.32 Each process is regulated by the balance between stimulatory and inhibitory signals such as growth factors, integrins and chemokines.39 In tumors, these signals are produced not only by tumors but also by stromal cells.40, 41 Thus, our finding that both CXCL8 and CXCL12 were upregulated by the interaction between PaCa and fibroblast cells have important biological implications.

One of the salient features of our study was that fibroblasts secreted high levels of CXCL12 compared to that of PaCa cells. Interestingly, secreted CXCL12 levels in fibroblasts were significantly enhanced by co-culturing with PaCa cells. This suggests paracrine factors secreted by PaCa cells enhanced CXCL12 production. In osteoblasts, these factors include IL-1β, platelet derived growth factor-BB (PDGF-BB), vascular endothelial growth factor-A (VEGF-A) and TNF-α.42 However, not much is known about regulation of CXCL12 production by PaCa-associated fibroblasts. A previous study noted increased migration of AsPC-1 cells in vitro by recombinant CXCL12 and of CFPAC-1 PaCa cells by co-culturing with MRC-9 fibroblasts cells, which were significantly inhibited by a CXCR4 antagonist, T22.43 In our study, we observed striking cooperative interaction between PaCa cells-derived CXCL8 and fibroblasts-derived CXCL12 in promoting invasion and angiogenesis. These findings indicate that the interaction between PaCa and stromal cells could produce an important cytokine network regulating invasion and angiogenesis of PaCa.

Chemokines are chemo-attractant cytokines that regulate motogenic activity of leukocytes and other cell types, including tumor and stromal cells.44 However, little is known about cooperative interactions between tumor and stromal cells to promote migration/invasion of PaCa cells. An important observation of our study was that CXCL12 enhanced PaCa cell invasiveness through CXCR4 in a paracrine manner, which was significantly inhibited by the blockade of CXCL12 signaling. Recently, the molecular mechanism by which CXC12-CXCR4 signaling axis can promote PaCa tumorigenesis was reported.45 In this study, Urrutia R et al. showed that CXCR4 mediated signaling significantly enhanced ERK activation through epidermal growth factor receptor (EGFR) transphosphorylation in PaCa cells and promoted HUVEC tube formation in vitro. Our results further extend this observation by demonstrating cooperative interactions between CXCL8 and CXCL12 in promoting PaCa associated angiogenesis. Also, we showed that fibroblast-derived CXL12 potently enhanced CXCL8 production by PaCa cells, thus potentiating HUVEC tube formation in vitro. Collectively, these observations highlight the fact that CXCR2 and CXCR4, corresponding receptors for CXCL8 and CXCL12 respectively, are potential therapeutic targets in PaCa.

We examined the effects of both cytokines on cellular proliferation of PaCa, fibroblasts, and HUVECs. Neither CXCL8 nor CXCL12 enhanced PaCa or fibroblast cell proliferation, but either chemokine did significantly enhanced HUVEC proliferation dose dependently. Because a previous report suggested that endogenous CXCL8 might play a role in PaCa proliferation in an autocrine manner,46 we examined the alteration of proliferation potency in PaCa cells by neutralizing anti-CXCL8 antibody but found no significant difference (data was not shown). Thus, with regard to cellular proliferation, it appears that these CXC-chemokines affect only vascular endothelial cells. This could be due to the fact that multiple growth factors including EGF, IGF, PDGF and VEGF provide substantial survival and proliferation advantage to PaCa cells and fibroblasts. Thus, chemokines including CXCL8 and CXC12 have distinct biological functions in the process of tumorigenesis.

To summarize findings from our study, we propose the following model (Fig. 6). This model depicts the cooperative interactions between PaCa cells, endothelial cells, and fibroblasts with respect to the biological effects of chemokines. We showed that PaCa cell-derived CXCL8 and fibroblast-derived CXCL12 promote HUVEC proliferation, migration/invasion, and tube formation. Also, CXCL12 enhanced CXCL8 production by PaCa cells in a paracrine manner. Our studies were not designed to dissect detailed chemokine receptors-mediated signaling pathways in these cell types. We also ignored the possibility that growth factors including EGF and VEGF can cooperate with chemokines including CXCL8 and CXCL12 to promote migration/invasion and angiogenesis in PaCa. Further studies will identify key signaling molecules downstream of CXCR2 and CXCR4 that are critical mediators of PaCa tumorigenesis. Furthermore, understanding regulation of CXCL12 production in fibroblasts by growth factors or cytokines will lead to the development of novel therapeutic targets. Collectively our data provide theoretical justification for clinical testing of drugs targeting CXCR4 and CXCR2 in addition to other known biological targets, to block metastasis and angiogenesis in PaCa.

thumbnail image

Figure 6. CXCL8 and CXCL12 cooperatively promote invasion and angiogenesis of PaCa. PaCa cells secrete CXCL8. Fibroblast-derived CXCL12 enhanced CXCL8 production in PaCa cells. CXCL8 from PaCa upregulated proliferation, invasion, and tube formation by HUVECs. Fibroblast cells produced CXCL12; this production was increased by co-culturing with PaCa cells. CXCL12 upregulated not only PaCa invasion but also proliferation and invasion of HUVECs. In this way, CXCL8 and CXCL12 cooperatively promote invasion and angiogenesis of PaCa.

Download figure to PowerPoint

Acknowledgements

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

The authors like to thank Ms. Diane Hackett from the Department of Scientific Publications for carefully reviewing this manuscript. This work was supported in part by AGA Mentors Research Scholar Award (to S.G.), The University of Texas MD Anderson Cancer Center Physician Scientist Program Award (to S.G.), and NIH grant (Cancer Center Support Grant to M.D. Anderson Cancer Center).

References

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
IJC_24040_sm_suppfigure1.tif2130KSupporting Figure 1: (A) Effects of recombinant CXCL12 on CXCL8 production from AsPC-1 PaCa cell line. To determine the effects of CXCL12 on CXCL8 production from a distinct PaCa cell line, AsPC-1 cells were seeded at a density of 2 × 105 cells/mL into a 24-well plate containing medium with 10% FCS and cultured overnight. Medium was then exchanged, and cells were cultured for a further 48 h treated with or without CXCL12 (20 ng/mL). The concentration of CXCL8 was measured using ELISA kit as described in Materials and Methods. Values are expressed as mean ± SD. **, P < 0.05. (B) Effects of recombinant CXCL8 on CXCL12 production from fibroblast cells. To determine the effects of CXCL8 on CXCL12 production from fibroblast cells, fibroblast cells were cultured treated with or without CXCL8 (20 ng/mL) for 48 h and the concentration of CXCL12 in the supernatants was measured by ELISA kit in the same way. Values are expressed as mean ± SD. n. s., no significant difference. (C) To determine the effects of co-culture with a distinct PaCa cell line on CXCL12 production from fibroblast cells, we co-cultured fibroblasts with AsPC-1 cells using a double chamber method (AsPC-1; 5×104 cells into inserts with 0.4-μm pores [BD Biosciences], fibroblast; 2×105 cells into 24-well plates). After 48 h incubation, the concentration of CXCL12 in supernatant was measured by ELISA kit in the same way. Values are expressed as mean ± SD. **, P < 0.05
IJC_24040_sm_suppfigure2.tif1501KSupporting Figure 2: Effects on MMP-2 and MMP-9 activities in HUVEC or PaCa cells following treatment with CXCL12. (A) Zymography analysis of HUVEC. HUVEC cells were seeded at a density of 2 × 106 cells/3mL into 60 mm dish and cultured overnight. Medium were then exchanged and cells were cultured for a further 24 h treated with or without recombinant CXCL12 (100 ng/ml). Using these supernatants and cell lysates, Zymography assay was performed. The samples were mixed with sample buffer without reducing agent or heating, and loaded into Zymogram gels (Invitrogen, Carlsbad, CA). After the electrophoresis with constant voltage, the gel was treated with Zymogram Renaturing Buffer (Invitrogen) and Zymogram Developing Buffer (Invitrogen) according to their instructions. Finally the gel was stained with SimplyBlue Safestain (Invitrogen) according to the instruction. (B) Zymography analysis of PaCa cells. In the same way, the supernatants of AsPC-1, Panc-1, and BxPC-3 cells treated with or without CXCL12 were collected and Zymography analysis was performed as described above.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.