We first analyzed three pancreatic cancer cell lines, BxPC-3, Capan-2, and SW1990, for the presence of integrin subunits, IL-1RI, uPA, and uPAR. In immunoblotting analysis, all three cell lines have expression of α₅, α₆, and α_v integrin subunits. All three cell lines lacked β₄ integrin subunits expression, while high expression of β₁ integrin subunit was observed. The high expression of IL-1RI was observed in three cell lines. The expression of uPA and uPAR was also observed in three cell lines (Figure 1).

In flow cytometric analysis, integrin surface expression was measured since the cells were fixed and not permeabilized prior to anti-integrin antibody incubation. The expression of α₅ and α_v integrin subunits remained unchanged in response to rIL-1α after 24 h, while the expression of α₆ and β₁ integrin subunits was enhanced in all pancreatic cancer cells (Table 1). In BxPC-3 cells, the expression of α₃ integrin subunit was enhanced in response to IL-α, while it remained unchanged in Capan-2 and SW1990 cell lines. The uPAR expression was also enhanced in response to rIL-1α after 24 h in all three cell lines. Enhancement of uPAR correlated with an increase in the cell surface expression of uPA in pancreatic cancer cells (Table 1).

We determined the proliferative response of three pancreatic cancer cell lines in response to rIL-1α for 24 h. Pancreatic cancer cells stimulated by rIL-1α showed 1.5–2.0-fold increase in proliferation compared to untreated conditions (Figure 2A and 2B). Inhibitory antibodies against α₆ and β₁ integrin inhibited the baseline and IL-1α-induced proliferation of all three pancreatic cancer cell lines (Figure 2A and 2B). No effect on proliferation was observed with control IgG (data not shown). These results indicate that α₆ and β₁ integrin subunits are involved in IL-1α-induced and normal proliferation of pancreatic cancer cells.

To determine whether IL-1α have any effect on the adhesion of pancreatic cancer cells, we investigated the adhesive response of pancreatic cancer cell lines to laminin, the putative ligand of the α₆β₁-integrin, in response to rIL-1α in pancreatic cancer cells. All three cell lines showed enhanced adhesion when stimulated by rIL-1α for 24 h (Figure 3). IL-1α-induced and basal adhesive response of these cell lines was suppressed by inhibitory antibodies against α₆ and β₁ integrin but not by control IgG.

The migration potential of pancreatic cancer cells stimulated with rIL-1α for 24 h was examined by using Matrigel-coated invasion chambers. All three pancreatic cancer cell lines showed significant enhancement of migration in the presence of IL-1α. The migration of pancreatic cancer cells correlated with an increase in cell surface bound uPA and uPAR (Table 1). IL-1α-induced migration of pancreatic cancer cells was inhibited by anti-α₆ integrin, anti-β₁ integrin, and anti-uPAR antibodies. Especially, each antibody inhibits the basal migration of these three cell lines (Table 2). Control IgG had no effect on the migration of pancreatic cancer cell lines. These data suggest that IL-1α-induced enhancement of α₆ and β₁ integrin and uPAR may have a role in enhancing the migration of pancreatic cancer cells.

To estimate whether IL-1α can activate Ras/ERK pathway to regulate proliferation, adhesion and migration of pancreatic cancer cells, we investigated the effect of IL-1α on this pathway. IL-1α enhanced the activation of Ras, as evidenced by the increased Ras-GTP levels in pancreatic cancer cells. Activation of Ras correlated with the phosphorylation of ERK. These results indicate that IL-1α may induce activation of ERK through a Ras-dependent pathway (Figure 4). To evaluate whether α₆β₁-integrin and uPAR affect activation of Ras and ERK, pancreatic cancer cells were treated for 30 min with the inhibitory antibodies before being exposed to rIL-1α for 30 min. Inhibition of α₆ and β₁ integrin subunits and uPAR signaling pathway inhibited activation of basal/IL-1α-induced Ras and phosphorylation of ERK (Figure 4). We performed the same examinations on all three cell lines. The results of three cell lines are very similar to the presented findings, therefore, we presented the data of SW1990 cell line. These results suggest that α₆β₁-integrin and uPAR expression have an important role in regulating IL-1α-induced activation of signaling pathways via IL-1RI. Detection of total ERK 1/2 levels served as a loading control.

The clinical features of 42 patients with invasive ductal adenocarcinoma of the pancreas were evaluated (Table 3). The mean age of all patients was 64.3 ± 9.1 years (range: 45 – 81 years). None had received prior chemotherapy or radiation therapy. Some patients received postoperative therapy; however, there was no difference in outcome among the various treatment modalities. The pT, pN, and pM categories were determined according to the TNM classification 19. The M category was determined from the intraoperative findings, chest and bone radiography, ultrasonography, computed tomography, and laboratory tests reflecting bone, chest, and liver metastasis. Ten patients had synchronous or heterochronous liver metastasis. The tissue specimens were obtained at pancreatoduodenectomy (n = 35) and distal pancreatectomy (n = 7).

Immunohistochemical expression of α₆ and β₁ integrin subunits and uPAR was evaluated in invasive ductal adenocarcinomas (n = 42) and duct cells in non-cancerous region of pancreas (n = 42). In 20 cancerous regions, the α₆ integrin subunit and uPAR were strongly to moderately expressed (Figure 5A and 5B), whereas the α₆ integrin subunit was not expressed or expressed weakly and uPAR was absent in non-cancerous region of the pancreas (Figure 5C and 5D). Weak-to-strong expression of the β₁ integrin subunit was observed in both malignant and non-cancerous region, and there was no trend in β₁ integrin subunit expression. A significant association was found between strong expression of α₆ integrin subunit and uPAR in pancreatic cancer specimens (p = 0.011, Table 3). The α₆ integrin subunit expression was significantly higher in malignant regions compared to non-cancerous regions of the pancreas (Table 4).

There was a significant association between strong co-expression of α₆ integrin subunit with uPAR and the presence of liver metastasis (p = 0.019), lymph node metastasis (p = 0.045), and retroperitoneal invasion of pancreatic cancer (p = 0.045). No significant correlation was found between co-expression of α₆ integrin subunit with uPAR and gender of patients, age, TNM stage, tumor location, cancer cell differentiation, intrapancreatic nerve invasion, lymphatic system invasion, and venous system invasion (Table 3). We also carried out a multivariate analysis of survival using the Cox's proportional hazards regression model, including each of the pathological parameters and strong expression of α₆ integrin subunit and uPAR (Table 5). The strong expression of α₆ integrin subunit (hazard ratio = 0.547, p = 0.026) and uPAR (hazard ratio = 0.491, p = 0.006) can be independent prognostic factors.

At the time of analysis, the median follow-up time for patients was 17.9 months (range; 1.3 – 89.6 months) after surgery. Twenty-six patients died of pancreatic cancer, and three patients died of other diseases. There was a statistically significant association between cases with and without strong expression of α₆ integrin in poor prognosis of patients with invasive ductal adenocarcinoma of the pancreas (Figure 6A). In addition, a statistically significant association was also detected between cases with and without strong uPAR expression (Figure 6B).

The human pancreatic cancer cell lines, BxPC-3, Capan-2, and SW1990, were from the American Type Culture Collection (Rockville, MD). The BxPC-3 cells were maintained in RPMI 1640 (Gibco BRL, Eggenstein, Germany) supplemented with 10% fetal calf serum (FCS). SW1990 and Capan-2 cells were maintained in Dulbecco modified Eagle medium (Gibco BRL) with high glucose and 10% FCS. All cells were incubated at 37°C in a humidified atmosphere of 5% CO₂ in air.

Human pancreatic tissues were obtained in Department of Gastroenterological Surgery, Nagoya City University Hospital with patients' or their relatives' informed consent. Tissue samples were fixed in 10% formalin and then embedded in paraffin. Immunohistochemical studies on tumor-free pancreatic tissue were performed using non-cancerous regions of tumor-containing pancreas.

Recombinant human IL-1α (rIL-1α) was provided by Gibco BRL. The monoclonal antibodies (mAbs) used included anti-β₁ (P5D2), anti-α₆ (GoH3), anti-α_v (AV1), and anti-β₄ (439-9B) from Chemicon International, Inc. (Temecula, CA, USA); anti-IL-1RI (35730) from Genzyme/Techne; anti-uPA-specific antibody (#3471) and uPAR specific antibody (#3936) from American Diagnostica (Temecula, CA, USA); anti-phospho-ERK 1/2 (Thr 202/Tyr 204), anti-ERK 1 (C-16), and anti-ERK 2 (C-14) from Santa Cruz Biotechnology (Santa Cruz, CA, USA).

The cells were lysed in lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM CaCl₂, 1% Triton X-100, 0.1% SDS, 0.1% Nonidet P-40, 2 mM PMSF, 1 mM vanadate, 5 μg/ml Trasylol, 10 μM Pepstatin A and 10 μM leupeptin). Following a low-speed spin (500 rpm, 5 min) to pellet nuclei and cell debris, the supernatant fraction was further centrifuged (100,000 g, 30 min), and the crude plasma membranes obtained in the pellet were re-suspended in 20 mM Tris-HCl (pH 7.4). Protein concentrations were determined with a BCA protein assay kit (Pierce, Rockford, IL, USA). The amounts of samples were 50 μg per each lane. Western blot analyses were performed following SDS-PAGE. The lysates were separated by 10% SDS-polyacrylamide gel electrophoresis, transferred to polyvinylidene difluoride membranes (Immobilon PVDF; Nihon Millipore Ltd., Tokyo, Japan) and immunoblotted with each antibody.

Flow-cytometric analysis was performed using FACScan (Becton Dickinson Immunocytometry Systems, Mountain View, CA, USA). The indirect immunofluorescence method was applied to stain the cancer cells with various monoclonal antibodies as the primary antibody (stained for 30 min at room temperature), followed by the addition of the secondary antibody conjugated fluorescein isothiocyanate (Dako, Glostrup, Denmark). Results are expressed as mean fluorescence intensity for triplicate determinations.

Pancreatic cancer cell proliferation was determined using the MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide dye reduction method] assay and cell count. In MTT assay, pancreatic cancer cells were seeded at a density of 2 × 10³ cells/100 μl into 96-well plates and allowed to adhere overnight. Culture media were replaced, and the cells then cultured in medium alone (control) or in medium with/without 10 ng/ml of rIL-1α. After 24 h of incubation, cells were cultured for 4 h with the metabolic substrate tetrazolium salt MTT at a final concentration of 0.5 mg/ml. Formazan was detected spectorphotometically at 540 nm with a multiwell spectrophotometer (ELISA Reader; Biotek Instruments, Burlington, VT, USA).

In cell count, pancreatic cancer cells were seeded at a density of 2 × 10⁵ cells on 35 mm well in media containing 10% FCS. After 24 h, cells were starved with 0.5% FCS for another 24 hours. Culture media was replaced to the fresh serum free media, and added rIL-1α at concentration of 10 ng/ml. After 24 h incubation, cells were washed once with phosphate-buffered saline (PBS), trypsinized, and centrifuged for 3 min at 1,500 rpm. The cell pellet was re-suspended in 2 ml of PBS and cells were counted using a light microscope.

Before the stimulating experiments with IL-1α were attempted, the lowest effective concentration was determined using rIL-1α at concentrations of 0.1 ng/ml, 0.5 ng/ml, 1.0 ng/ml, 10 ng/ml, and 100 ng/ml. A concentration of 10 ng/ml was determined to be the lowest effective concentration for stimulating experiments (data not shown). In some experiments, 0.5 μg/ml anti-α₆ integrin or anti-β₁ integrin mAbs was added to the cancer cells for 24 h. Before the blocking experiments were attempted, the lowest effective antibody concentration was determined using antibodies at concentrations of 0.1 μg/ml, 0.25 μg/ml, 0.5 μg/ml, 0.75 μg/ml, and 1.0 μg/ml. A concentration of 0.5 μg/ml was determined to be the lowest effective concentration for blocking experiments (data not shown). Experiments were performed in triplicate and repeated three times.

Adhesion assay was performed as described previously with some modifications 8. 24-well plates coated with laminin, the putative ligand of the α₆β₁-integrin, were obtained from Becton-Dickinson Labware (Franklin Lakes, NJ, USA). Briefly, cancer cells were incubated for 24 hours with/without rIL-1α (10 ng/ml) and then added (2 × 10⁵ cells/well) to each well and incubated at 37°C for 30 min. The wells were then washed three times with PBS to remove unattached cells. In some experiments, 0.5 μg/ml anti-α₆ or anti-β₁ integrin antibodies were added to the cancer cells for >30 min prior to addition of rIL-1α.

The migration response of pancreatic cancer cells in response to IL-1α was determined by using Matrigel-coated invasion chambers (Becton and Dickinson, USA). Cancer cells were added (1 × 10⁵ cells/well) to the inner chamber of a cell culture insert and incubated at 37°C for 24 h, either with culture media containing 10 ng/ml rIL-1α or with culture media containing 10 ng/ml rIL-1α and 0.5 μg/ml anti-α₆ integrin, anti-β₁ integrin, or anti-uPAR antibodies. Complete medium containing 20% fetal bovine serum served as a chemo-attractant in the lower chamber. To quantitate migration, the filters were fixed in 70 % ethanol for 30 min and stained with Giemsa. Cells were removed from the upper surface of the filters by rubbing gently with a cotton-tipped applicator. Cells that had migrated through the membrane were counted in five random microscope fields of the lower filter surface.

The activation state of Ras was determined using the Ras Activation Assay Kit provided by Upstate (Lake Placid, NY, USA). Briefly, pancreatic cancer cells were serum starved for 24 h, and then incubated in serum-free medium with/without rIL-1α (10 ng/ml) for 30 min. Cells were harvested and lysed in lysis buffer (100 mM HEPES, pH 7.5, 200 mM NaCl, 1% Nonidet P-40, 10 mM MgCl₂, 5 mM EDTA and 10% glycerol), and supernatant prepared by centrifugation for 5 min at 4°C at 14,000 g. Ras-GTP from various treated lysates was "pulled down" using the GST fusion protein corresponding to human Ras binding domain of Raf-1 bound to agarose. The presence of Ras-GTP was detected by Western blotting using anti-Ras antibody (Upstate).

Pancreatic tissues were studied using the labeled streptavidin biotin method 4041. Specimens were sectioned at 3.5-μm thick and deparaffinized. After rinsing in phosphate-buffered saline (pH 7.2), 10% bovine serum (Wako, Osaka, Japan) was applied for 10 min to block nonspecific binding. Sections were then incubated with anti-α₆ integrin (overnight at 4°C), anti-β₁ integrin (over night at 4°C), or anti-uPAR (60 min at 37°C) mAbs as primary antibodies. After rinsing in phosphate-buffered saline, sections were treated with biotinylated anti-mouse immunoglobulin (Ig) (Dako, Copenhagen, Denmark) for 10 min. After rinsing in phosphate-buffered saline, sections were treated with horseradish peroxidase-labeled streptavidin (Dako, Copenhagen, Denmark) for 10 minutes. The peroxidase reaction was visualized by incubating the sections with 0.02 % 3,3'-diaminobenzidine tetrahydrochloride in 0.05 M Tris buffer (pH 7.6) with 0.01 % hydrogen peroxide, followed by hematoxylin counterstaining. Negative control sections were prepared using normal mouse IgG instead of primary antibody.

Two observers (H.S. and H.F.) independently evaluated the immunostaining results. The concordance ratio was > 90%. Differences of opinion were resolved by reaching a consensus with the assistance of a third evaluator (Y.M.). The intensity of tissue staining was graded semiquantitatively on a 4-point scale (-, +, ++, and +++). Likewise, the proportion of cells stained was assessed on a 4-point scale (1, 0–15%; 2, 15–50%; 3, 50–85%; 4, 85–100% cells stained). To evaluate immunohistochemical findings from pancreatic cancer tissues, cases were classed in strongly staining (Group S) and weakly staining groups (Group W) by intensity and proportion of immunostaining. Immunostaining of intensity more than +++ or a staining area was more than 3 for α₆ integrin subunit, β₁ integrin subunit, or uPAR was defined as Group S.

Statistical comparisons were made using the Student's t test for paired observations or by one-way ANOVA for multiple comparisons. The Mann-Whitney U test was used to compare the immunohistochemical characteristics. Differences between Kaplan-Meier survival curves based on Immunohistochemical analysis were tested with the Wilcoxon test. Multiple survival analysis was calculated according to Cox's proportional hazards model. Statistical significance was indicated by p < 0.05. Data are presented as mean ± standard deviations (s.d.). Each experiment was repeated three times and was carried out in triplicate.