Enhanced Expression of Fibroblast Growth Factor Receptor 2 IIIc Promotes Human Pancreatic Cancer Cell Proliferation

Materials

Materials were purchased as follows: Immobilon P transfer membrane from Millipore (Yonezawa, Japan); FuGene HD transfection reagent from Roche Diagnostics (Mannheim, Germany); FastPure RNA kit from Takara Bio (Tokyo, Japan); high-capacity cDNA reverse transcription kit, TaqMan gene expression assays for FGFR-3 IIIb (Hs01005396_s1), FGFR-3 IIIc (Hs00997397_s1), FGFR4 (Hs01106908_s1), and 18S rRNA (Hs99999901_s1), Silencer Select custom-designed siRNA (s275290 and s275291), and Silencer negative control siRNA from Applied Biosystems (ABI; Life Technologies, Foster City, CA); TransIT-siQUEST transfection reagent from Mirus Bio (Madison, WI); M-PER mammalian protein extraction reagent and SuperSignal West Dura chemiluminescent substrates from Thermo Fisher Scientific (Rockford, IL); G418 (Geneticin; Gibco BRL, Grand Island, NY); Histofine simple stain MAX PO (R) kits, Histofine simple stain MAX PO (M) kits, and peroxidase-conjugated streptavidin from Nichirei (Tokyo, Japan); biotinylated anti-guinea pig IgG from Vector Laboratories (Burlingame, CA); HRP-conjugated goat anti-rabbit IgG secondary antibodies from American Qualex (San Clemente, CA); human serum from Lonza (Walkersville, MD); Zenon rabbit IgG labeling kit (Z-25351), Hoechst 33342 dye, and Click-iT EdU Pacific Blue flow cytometry assay kit from Invitrogen (Life Technologies, Carlsbad, CA); rabbit IgG (ab37415) from Abcam (Cambridge, UK); 35-μm filter, 8-μm pore size cell culture insert, and BioCoat Matrigel invasion chamber from BD Bioscience (Franklin Lakes, NJ); guinea pig polyclonal anti-swine insulin antibody, rabbit polyclonal anti-human IgG antibody, and mouse monoclonal anti-Ki-67 antibody from Dako (Santa Barbara, CA); rabbit polyclonal anti-extracellular signal-regulated kinase (ERK) 1 antibody (K-23) from Santa Cruz Biotechnology (Santa Cruz, CA); mouse monoclonal anti-human CD31 antibody from AbD Serotec (Kidlington, UK); phospho-p44/42 MAPK (p-ERK) rabbit monoclonal antibody from Cell Signaling Technology (Danvers, MA); recombinant human basic fibroblast growth factor (bFGF, FGF-2) from ReproCELL (Kanagawa, Japan); recombinant human EGF from Austral Biologicals (San Ramon, CA); ultra-low attachment surface plate from Corning (Lowell, MA); recombinant human FGF-7 from R&D Systems (Minneapolis, MN); WST-8 cell counting kit from Wako Pure Chemical Industries (Osaka, Japan); New Silane II slide glass and malinol mounting medium from Mutoh Chemical (Tokyo, Japan). All other chemicals and reagents were purchased from Sigma-Aldrich (St. Louis, MO).

Pancreatic Cancer Cell Lines

KLM-1, PANC-1, MIAPaCa-2, PK-1, and PK-8 PDAC cell lines were obtained from the Cell Resource Center for Biomedical Research, Institute of Development, Aging and Cancer, Tohoku University (Sendai, Japan); the Capan-1 pancreatic adenocarcinoma cell line was purchased from American Type Culture Collection. The cells were grown in the RPMI 1640 medium containing 10% fetal bovine serum (FBS) at 37°C under a humidified 5% CO₂ atmosphere. Capan-1 cells were incubated under the same conditions in RPMI 1640 medium with 15% FBS.

Patients and Tissues

Tissues from 117 patients with invasive PDAC were obtained for this study. These patients received treatment at Nippon Medical School Hospital (Tokyo, Japan) from 1995 to 2011. None of the patients received preoperative chemotherapy and radiotherapy. The patients were 66 men and 51 women; the median age was 66.9 years (range, 35 to 87 years). Clinicopathologic stage was determined according to the TNM classification system of the International Union Against Cancer (UICC) and was additionally characterized under the Japan Pancreas Society classification (Table 1). Fifty-nine patients were monitored after surgery; the median follow-up period was 15.6 months; 28 patients did not receive any postoperative chemotherapy and 31 patients received adjuvant chemotherapy after surgery. Of those receiving chemotherapy, 13 patients received tegafur-uracil and 12 patients received gemcitabine; 6 patients received tegafur-uracil and gemcitabine. Paraffin-embedded specimens were prepared for immunohistochemical analysis, as described previously.²⁷ This study was conducted in accordance with the principles embodied in the Declaration of Helsinki, 2008, and informed consent for the use of pancreatic tissues was obtained from each patient. Normal pancreatic tissues were obtained from Human Digestive Tissue Sets and Human Tissue Microarray 1 (both obtained from Novagen, Darmstadt, Germany).

Quantitative Real-Time PCR of FGFR-2 IIIc in Pancreatic Cancer Cells

All pancreatic cancer cells were grown in RPMI 1640 medium supplemented with 10% or 15% FBS for 48 hours. Total RNA extraction was performed using a FastPure RNA kit. Next, cDNA synthesis was performed using a high-capacity cDNA reverse transcription kit according to the manufacturer's protocol. Quantitative real-time PCR (q-PCR) was performed using an ABI StepOnePlus system (Life Technologies). The real-time PCR primers used for FGFR-2 IIIc were nt1693-1716 (5′-GGATATCCTTTCACTCTGCATGGT-3′) and nt1770-1794 (5′-TGGAGTAAATGGCTATCTCCAGGTA-3′) of human FGFR-2 IIIc cDNA (102 bp; accession no. NM_000141.4). As TaqMan probes, 5′-CAGTTCTGCCAGCGCCTGGAAGA-3′ was used for FGFR-2 IIIc. PCR primers used for FGFR-1 IIIb were nt1991-2010 (5′-ACCAGTCTGCGTGGCTCACT-3′) and nt2036-2052 (5′-TGCCGGCCTCTCTTCCA-3′) of human FGFR-1 IIIb cDNA (62 bp; accession no. FJ809917.1). As TaqMan probes, 5′-CACCAGACCTGTGGCAA-3′ was used for FGFR-1 IIIb. PCR primers used for FGFR-1 IIIc were nt1295-1316 (5′-GGACTCTCCCATCACTCTGCAT-3′) and nt1381-1403 (5′-CCCCTGTGCAATAGATGATGATC-3′) of the human FGFR-1 IIIc cDNA (109 bp; accession no. M34186.1). As TaqMan probes, 5′-ACCGTTCTGGAAGCC-3′ was used for FGFR-1 IIIc. The PCR reaction mixture containing 2 μL of template cDNA, 10 μL of TaqMan fast universal PCR master mix and 1 μL of each of TaqMan gene expression assay was placed in a 96-well reaction plate. 18S rRNA, as the internal positive control, was amplified using a TaqMan gene expression assay. The optimized program involved denaturation at 95°C for 20 seconds, followed by 50 cycles of amplification (at 95°C for 1 second and at 60°C for 20 seconds) for FGFR-2 IIIc and 18S rRNA. Results were expressed as target/18S rRNA, as an internal standard concentration ratio. Gene expression measurements were performed in triplicate.

Rabbit Polyclonal Anti-Human FGFR-2 IIIc Antibody

To confirm the expression of FGFR-2 IIIc at the protein level in pancreatic cancer, specific antibody against human FGFR-2 IIIc was prepared, as described previously.²⁸ The anti-FGFR-2 IIIc antibody used was an affinity-purified rabbit polyclonal antibody raised against a peptide corresponding to amino acids AAGVNTTDKEIEVLYIR of the human FGFR-2 IIIc protein (accession no. NM_000141.4). This sequence is located at the carboxyl-terminal half of the Ig loop closest to the transmembrane region and is specific for FGFR-2 IIIc.

Western Blot Analysis of FGFR-2 IIIc in Pancreatic Cancer Cells

Protein extraction was performed according to the protocol involving the use of M-PER mammalian protein extraction reagent. Briefly, cultured pancreatic cancer cells were solubilized in M-PER reagent with protease inhibitor cocktail for mammalian tissues. Lysates were centrifuged for 10 minutes at 10,000 × g to pellet cell debris. The supernatants were collected and protein concentration was measured by the Bradford method. The cleared protein lysates were subjected to SDS-PAGE under reducing conditions, and the separated proteins were transferred to Immobilon P transfer membranes, which were then incubated for 16 hours at 4°C with the anti-FGFR-2 IIIc antibody. The membranes were washed and incubated with HRP-conjugated anti-rabbit IgG antibody for 60 minutes. After the wash, the blot was visualized by enhanced chemiluminescence. The membrane was reblotted with mouse monoclonal anti-β-actin antibody to confirm equal loading.

Immunohistochemistry

The same anti-FGFR-2 IIIc antibody used for Western blotting was also used for immunohistochemistry. Paraffin-embedded sections (3 μm thick) were subjected to immunostaining using a Histofine simple stain MAX PO (R) kit for FGFR-2 IIIc or a Histofine simple stain MAX PO (M) kit for Ki-67 and CD31. After deparaffinization, endogenous peroxidase activity was blocked by incubation for 30 minutes with 0.3% hydrogen peroxide in methanol for FGFR-2 IIIc and Ki-67 antibodies. The tissue sections were then incubated with the anti-FGFR-2 IIIc antibody (1:200 dilution), anti-Ki-67 antibody (1:100 dilution), or CD31 (1:50 in dilution) in phosphate-buffered saline (PBS) containing 1% bovine serum albumin (BSA) for 16 hours at 4°C. Bound antibodies were detected with the Histofine simple stain MAX PO (R) or (M) reagent, using diaminobenzidine tetrahydrochloride as the substrate. For insulin staining, guinea pig polyclonal anti-porcine insulin antibodies (1:1000 dilution), cross-reactive with human insulin, and a biotinylated goat anti-guinea pig IgG secondary antibody were used after incubation with 10% normal goat serum. Sections were then treated with streptavidin-peroxidase complex, using diaminobenzidine tetrahydrochloride as the substrate. The sections were then counterstained with Mayer's hematoxylin. Negative control tissue sections were prepared by omitting the primary antibody.

In Situ Hybridization

In situ hybridization was performed as previously reported.^5,29 Tissue sections were deparaffinized and incubated at room temperature for 20 minutes with 0.2 mol/L HCl and then at 37°C for 15 minutes with 100 μg/mL proteinase K. The sections were then postfixed for 5 minutes in PBS containing 4% paraformaldehyde, and incubated twice for 15 minutes each with PBS containing 2 mg/mL glycine at room temperature and once in 50% (v/v) formamide/2× standard saline citrate for 1 hour at 42°C before initiation of the hybridization reaction. The hybridization buffer contained 0.6 mol/L NaCl, 1 mmol/L EDTA, 10 mmol/L Tris-HCl (pH 7.6), 0.25% SDS, 200 mg/mL yeast tRNA, l× Denhardt's solution, 10% dextran sulfate, 40% formamide, and 500 ng/mL of the indicated digoxigenin-labeled riboprobe. Hybridization was performed in a moist chamber for 16 hours at 42°C. The sections were then washed sequentially with 2× standard saline citrate for 20 minutes at 42°C and 0.2× standard saline citrate for 20 minutes at 42°C. For immunological detection, a DIG nucleic acid detection kit (Roche Diagnostics) was used. The sections were washed briefly with buffer 1 (100 mmol/L Tris-HCl and 150 mmol/L NaCl, pH 7.5), incubated with 1% (w/v) blocking reagents in buffer 1 solution for 60 minutes at room temperature, and with alkaline-phosphatase-conjugated polyclonal sheep anti-digoxigenin Fab fragment containing 0.2% Tween 20 at l:2000 dilution for 60 minutes at room temperature. The sections were then washed twice for 15 minutes at room temperature with buffer 1 solution containing 0.2% Tween 20, equilibrated with buffer 3 solution (100 mmol/L Tris-HCl, 100 mmol/L NaCl, 50 mmol/L MgCl₂, pH 9.5) for 2 minutes, and incubated with a staining solution containing nitro blue tetrazolium and X-phosphate in a dark box for 1 hour. After the reaction was stopped with TE buffer (10 mmol/L Tris-HCl and 1 mmol/L EDTA, pH 8.0), the sections were mounted in an aqueous mounting medium.

Construction of FGFR-2 IIIc Expression Vector and Generation of Stably Transfected Clones

The full-length FGFR-2 IIIc cDNA fragment was ligated to the 3′ end of the human cytomegalovirus early promoter/enhancer in pIRES2-EGFP eukaryotic expression vector. Proper orientation of the insert was verified by DNA sequencing. Approximately 1 × 10⁵ cells/2 mL of KLM-1 cells were transfected with 3 μg of DNA using FuGene HD transfection reagent, and the cells were passaged and cultured with 600 μg/mL of G418. Independent colonies were isolated by ring cloning, transferred to microtiter wells, and expanded in 300 μg/mL of G418.

Flow Cytometry of FGFR-2 IIIc

Anti-FGFR-2 IIIc antibody was labeled with allophycocyanin using a Zenon rabbit IgG labeling kit according to the manufacturer's protocol. Cells were incubated for 20 minutes at 4°C in PBS containing 10% human serum. Cells were then centrifuged briefly to remove the serum-containing medium and incubated (5 × 10⁵ cells/25 μL) with 1 μg of allophycocyanin-labeled anti-FGFR-2 IIIc antibody for 60 minutes in the dark at 4°C; 1 μg of propidium iodide was added, to label dead cells. Cells were washed with cold PBS containing 2% FBS, and the cell suspension was passed through a 35-μm filter. FGFR-2 IIIc expression was analyzed using a FACSAria II flow cytometer (BD Bioscience, Franklin Lakes, NJ). Isotype-matched rabbit IgG was used as a negative control.

Morphological Analysis of FGFR-2 IIIc-Transfected KLM-1 Cells

KLM-1 cells were cultured in 75-cm² flasks in RPMI 1640 medium containing 10% FBS for 48 hours. The cells were photographed using a Nikon Eclipse TE-2000-U system at ×200 magnification.

Cell Proliferation Assay

To monitor cell proliferation, nonradioactive cell proliferation assays were performed. Cells were plated at a density of 5 × 10³ cells per well in 96-well plates and grown overnight in the RPMI 1640 medium supplemented with 10% FBS. After 72 hours, cells were incubated with WST-8 cell counting reagent for 4 hours at 37°C and the optical density of the culture solution in the plate was measured using an enzyme-linked immunosorbent assay (ELISA) plate reader (Bio-Rad Laboratories, Hercules, CA) at 450 nm. Analysis was performed in triplicate.

Heterotopic and Orthotopic Implantation of FGFR-2 IIIc-Transfected KLM-1 Cells

To assess the effect of FGFR-2 IIIc expression on in vivo tumorigenicity, 1 × 10⁶ cells/animal were injected subcutaneously into 6-week-old male athymic (nude) mice (BALB/cA Jcl-nu/nu; CLEA Japan, Tokyo, Japan) (n = 6 per cell line). Tumor volume was calculated as a × b² × 0.5, where a is the longest diameter and b is the shortest diameter.³⁰ The tumors were removed and cut in half. Half of the tumor was fixed in 20% neutral-buffered formalin; the other half, used for orthotopic implantation, was cut into 2-mm² pieces. Under brief general inhalation anesthesia with isoflurane and using 7.0 Prolene sutures, KLM-1 fragments (2 mm²) were sutured onto the surface of the tail of the pancreas of 6-week-old male NOD/Shi-scid, IL-2γ^null (NOG; Central Institute for Experimental Animals, Kanagawa, Japan) and nude mice (n = 4 per cell line).³¹ The animals were monitored for 5 weeks.

Sphere Formation Assay

An important feature of stem cells and CSCs (or cancer initiating cells) is formation of cell spheres in floating culture.³² Therefore, to determine whether pancreatic cancer cells have CSC-like characteristics, we performed sphere formation assays. KLM-1 cells (1000/well) were plated in ultra-low attachment surface 24-well plates with serum-free medium supplemented with or without FGF-2 (10 ng/mL) and EGF (20 ng/mL). After 7 days, the number of spheres in each well was counted under a phase-contrast microscope (Nikon Eclipse TE2000-U). Analysis was performed in triplicate.

Flow Cytometry for Side Population Assay

Cells (1 × 10⁶/mL) were incubated at 37°C for 10 minutes, and then 5 μg/mL of Hoechst 33342 dye was added. Verapamil (30 μg/mL) was also added as a control. Cells were incubated at 37°C for 90 minutes, and then washed twice with cold PBS containing 2% FBS. Cells were suspended with PBS containing 1 mmol/L EDTA, 25 mmol/L HEPES pH 7.0, and 1% FBS. Propidium iodide (1 μg) was added to label dead cells. The cell suspension was passed through a 35-μm filter. The ratio of side population (SP) to major population (MP) was analyzed using a BD FACSAria II flow cytometer. Analysis was performed in triplicate.

Tumorigenicity of SP and MP Cells in KLM-1 Cells

To examine the tumor formation ratio of SP and MP cells in KLM-1 cells, we sorted each cells by flow cytometry, as described above. A total number of 10⁵, 10⁴, 10³, or 10² cells of SP and MP fractions were subcutaneously injected into 6-week-old male NOD.CB17-Prkdc^scid/J (NOD/SCID) mice (Charles River Laboratories Japan, Kanagawa, Japan). Tumor formation was determined after 5 weeks.

Cell Cycle Flow Cytometry

Cell cycle analysis was performed using a Click-iT EdU Pacific Blue flow cytometry assay kit according to the manufacturer's protocol. In brief, 10 μmol/L of 5-ethynyl-2′-deoxyuridine (EdU) was added into culture medium, and cells were incubated for 60 minutes at 37°C. Cells were fixed with 4% paraformaldehyde for 60 minutes, and then the EdU was labeled with Pacific Blue. 7-Aminoactinomycin D (7-AAD) was added for measuring DNA content and cell cycle distribution, and cell cycle analysis was performed using a BD FACSAria II flow cytometer.

Cell Signaling Pathway of FGFR-2 IIIc-Transfected Pancreatic Cancer

Cells were seeded in 60 mm-dish (2.5 × 10⁵ cells) and grown in RPMI 1640 medium supplemented with 10% FBS for 24 hours. The cells were then washed with serum-free medium and cultured with same medium for 24 hours, and recombinant FGF-2 (100 ng/mL) and/or heparin (1 μg/mL) was added to the plates. Protein was then extracted and used for the Western blot analyses using rabbit monoclonal anti-p-ERK (1:1000) and rabbit polyclonal anti-ERK antibodies (1:1000), as described above.

Single-Cell Movement Assay

Cells (5 × 10³ per well) were seeded onto a four-well glass-bottom dish and grown for 24 hours, and the medium was then changed to RPMI 1640 supplemented with 10% FBS, 100 ng/mL FGF-2, or 100 ng/mL FGF-7. Cell movement in the presence of an anti-FGFR-2 IIIc antibody or control IgG was then monitored for 24 hours using a digital Nikon Eclipse TE 2000-E motorized inverted microscope, acquiring images every 5 minutes. The total distance that individual cells traveled during 24 hours was determined using MetaMorph software version 7.6 (Universal Imaging, Marlow, UK).

Transfection of FGFR-2 IIIc siRNA

Down-regulation of FGFR-2 IIIc expression in PANC-1 cells was induced using siRNA. Two different types of custom-designed siRNA against specific IIIc region of FGFR-2 IIIc were purchased, and the sense sequence was 5′-GGAAUGUAACUUUUGAGGATT-3′ (s275290) and 5′-GUGCUUGGCGGGUAAUCCUTT-3′ (s275291) (underlining denotes common sequences). The cells were plated at a density of 1 × 10⁵ cells in a 35-mm dish and transfected with 5 nmol/L siRNAs for FGFR-2 IIIc and Silencer negative control siRNA for control using TransIT-siQUEST according to the manufacturer's protocol. To confirm the effective transfection of siRNA in PANC-1 cells, total RNA was prepared at 24 hours after the transfection and FGFR-2 IIIc mRNA levels were determined by q-PCR.

Effects of Anti-FGFR-2 IIIc Antibody on Cell Migration and Invasion

Cell migration was assessed using modified Boyden chambers with uncoated inserts (6.4 mm in diameter; 8-μm pores). Serum-free RPMI 1640 medium (500 μL) was placed in each upper chamber and RPMI 1640 medium with 10% FBS (750 μL) was into placed in each lower chamber. PANC-1 cells were then plated on the inner surface of the inserts at a density of 1 × 10⁵ cells per insert in the presence of either the anti-FGFR-2 IIIc antibody or control IgG (100 μg/mL), followed by incubation at 37°C in a humidified 5% CO₂ atmosphere. After 6 hours, the cells that moved to the outer surface of the inserts were fixed, stained with a Diff-Quik staining kit (Fisher Scientific, Pittsburgh, PA), and counted in five high-power fields (20× objective). All assays were performed in triplicate. Cell invasion assays were performed using modified Boyden chambers in which the inner surfaces of the inserts were coated with Matrigel. Cell counting was performed as described above.

Statistical Analysis

Results for cell proliferation or tumor volume are presented as means ± SE. Data were compared between different groups using Student's t-test. Whenever indicated, the χ² test and Fisher's exact test were used to analyze the correlation between FGFR-2 IIIc expression and clinicopathologic features. Cumulative survival rates were calculated by the Kaplan-Meier method, and the significance of differences in survival rate was analyzed by the log-rank test. Data were compared between multiple groups using one-way analysis of variance and then analyzed using a post hoc test. P < 0.05 was considered significant in all analyses. Computations were performed using the StatView J version 5.0 software package (SAS Institute, Cary, NC).

FGFR Isoform Expression in Pancreatic Cancer Cell Lines

To determine whether cultured human pancreatic cancer cell lines express FGFR-2 IIIc, and to compare any such expression with the expression of other FGFR isoforms, q-PCR was performed using RNA from six human pancreatic cancer cell lines. FGFR-2 IIIc mRNA was expressed in all six cell lines. FGFR-2 IIIc mRNA levels were highest in PANC-1 cells, intermediate in MIA PaCa-2 cells, and relatively low in KLM-1, Capan-1, PK-1, and PK-8 cells (Figure 1A). The FGFR-2 IIIc level was 106-fold higher in PANC-1 cells than in Capan-1 cells, which demonstrates the extent of variability of FGFR-2 IIIc mRNA levels across the different cell lines. PANC-1 and MIA PaCa-2 cells (which express high levels of FGFR-2 IIIc) expressed low FGFR-2 IIIb, whereas KLM-1, Capan-1, PK-1, and PK-8 cells (which express low FGFR-2 IIIc) expressed higher FGFR-2 IIIb levels. A similar tendency was observed in the expression levels of FGFR-1 and FGFR-3 isoforms in the pancreatic cancer cell lines (Figure 1A).

Next, immunoblotting with an anti-FGFR-2 IIIc antibody was performed, using protein lysates from the same six pancreatic cancer cell lines. FGFR-2 IIIc was expressed at the protein level in all the cell lines (Figure 1B). FGFR-2 IIIc protein was most abundant in PANC-1 cells and least abundant in KLM-1, Capan-1, PK-1, and PK-8 cells (Figure 1B). Analysis by flow cytometry revealed similar expression patterns for FGFR-2 IIIc at the cell surface of the PDAC cells as determined by Western blotting (Figure 1C). Overall, there was a close correlation between FGFR-2 IIIc mRNA and protein levels in all six pancreatic cancer cell lines.

FGFR-2 IIIc Expression in Pancreatic Cancer Tissues

Immunohistochemical and in situ hybridization analysis of PDAC samples was performed next, to determine whether and in which cell type FGFR-2 IIIc is expressed in human pancreatic cancer tissues. In the normal pancreas, weak FGFR-2 IIIc immunoreactivity was detected in the endocrine islets, but not in ductal cells or capillary endothelial cells (Figure 2A), and strong insulin immunoreactivity was present in the islets (Figure 2B). Moderate FGFR-2 IIIc immunoreactivity was evident in the vascular smooth muscle cells and endothelial cells of relatively large vessels in the peripancreatic tissues (Figure 2C). As expected, the vascular endothelial cells were positive for CD31 (Figure 2D). By contrast to the epithelial cells in the normal pancreas, there was strong FGFR-2 IIIc immunoreactivity in the cancer cells in 84 of 117 (71.8%) PDAC samples (Figure 3A and Table 1). Moderate FGFR-2 IIIc immunoreactivity was also evident in the fibroblasts adjacent to cancer cells (Figure 3A). Both the cancer cells and the fibroblasts exhibited a moderately strong FGFR-2 IIIc mRNA in in situ hybridization (Figure 3B). A sense probe used as negative control did not yield a signal (Figure 3C).

FGFR-2 IIIc expression in the cancer cells did not correlate with the clinicopathological factors (Table 1). The overall 2-year survival rate for all 59 cases of PDAC was 16.9%, and the overall survival rates of the FGFR-2 IIIc-positive group tended to be lower, but this difference was not statistically significant (P = 0.1571) (Figure 3D). Time to development of liver metastasis after resection was significantly shorter in the FGFR-2 IIIc-positive group than in the FGFR-2 IIIc-negative group (P = 0.0071) (Figure 3E).

Stable FGFR-2 IIIc-Transfected KLM-1 Cells

To examine the role of FGFR-2 IIIc in pancreatic cancer cells, we prepared FGFR-2 IIIc-overexpressing pancreatic cancer cells and mock-transfected cells as controls. A full-length FGFR-2 IIIc cDNA was subcloned into pIRES2-EGFP vector and stably transfected into KLM-1 cells, which express a low level of FGFR-2 IIIc. Capan-1 cells were not suitable for stable transfection, because, although they expressed the lowest level of FGFR-2 IIIc, these cells grow very slowly. FGFR-2 IIIc mRNA levels in mock-transfected cells and FGFR-2 IIIc-transfected KLM-1 cells were examined by q-PCR in relation to the 18S rRNA housekeeping gene. FGFR-2 IIIc/18S rRNA levels were very high in FGFR-2 IIIc clones 1 and 6 and low in the mock clones (Figure 4A). There was a parallel but relatively small decrease in IIIb mRNA levels in the FGFR-2 IIIc-transfected clones (Figure 4A). Moreover, as analyzed with flow cytometry using the anti-FGFR-2 IIIc antibody, high levels of FGFR-2 IIIc protein were evident in FGFR-2 IIIc stably transfected KLM-1 cells (Figure 4B), indicating that FGFR-2 IIIc was abundant on the cell surface of these clones. FGFR-2 IIIc transfection did not alter FGFR-3 IIIb or FGFR4 levels (see Supplemental Figure S1 at http://ajp.amjpathol.org), and the levels of the other FGFRs were exceedingly low (data not shown).

Effects of FGFR-2 IIIc on Cell Morphology and Proliferation

FGFR-2 IIIc-transfected KLM-1 cells and the corresponding mock-transfected cells exhibited similar morphologies. However, the number of mitotic cells was greater in the FGFR-2 IIIc-transfected KLM-1 cells than in mock-transfected cells (Figure 4C). Moreover, FGFR-2 IIIc stably transfected KLM-1 cells demonstrated increased cell proliferation, compared with mock cells (Figure 4D).

Effects of FGFR-2 IIIc on Subcutaneous Tumor Formation in Nude Mice

To determine whether FGFR-2 IIIc modulated the in vivo proliferation of pancreatic cancer cells, FGFR-2 IIIc-transfected KLM-1 cells and mock-transfected cells were injected subcutaneously in nude mice. Compared with two different mock clones, FGFR-2 IIIc-transfected KLM-1 clones 1 and 6 formed significantly larger tumors (Figure 5A). There was strong FGFR-2 IIIc immunoreactivity in the FGFR-2 IIIc-6-derived tumors and weak immunostaining in the mock-3-derived tumors (Figure 5B). FGFR-2 IIIc immunoreactivity was not detected in stromal cells in the nude mice. Moreover, there was a marked increase in Ki-67 immunoreactivity in the FGFR-2 IIIc-6-derived tumors, compared with the mock-3-derived tumors, and the calculated Ki-67 labeling index in the subcutaneous tumor was twofold higher in FGFR-2 IIIc-transfected KLM-1 cells, compared with mock-transfected cells (Figure 5B).

Effects of FGFR-2 IIIc on Orthotopic Tumor Formation and Liver Metastasis in NOG Mice

We next performed orthotopic implantation of FGFR-2 IIIc-transfected KLM-1 cells into NOG mice. FGFR-2 IIIc-transfected KLM-1 cells formed larger intrapancreatic tumors, compared with mock-transfected cells, and the tumors were significantly heavier (Figure 5C). In addition, in this model, liver metastases were detected in all four tumors derived from FGFR-2 IIIc-transfected cells, but only in two of the four mock-transfected cell-derived tumors. The number of liver metastases was also increased in FGFR-2 IIIc-transfected cells, compared with mock-transfected cells, and the livers of FGFR-2 IIIc-expressing tumors were significantly heavier (Figure 5D).

Orthotopic implantation was also performed in nude mice, which have greater immunological responses than NOG mice. In this model, the mean intrapancreatic tumor weight of the four FGFR-2 IIIc-6-derived tumors was 0.804 ± 0.275 g. By contrast, the mean weight of the four mock-3-derived tumors was 0.357 ± 0.023 g. Moreover, liver metastases were observed in only one mouse implanted with FGFR-2 IIIc-6 expressing tumor fragments, and in none of the four mice implanted with mock-3 expressing tumor fragments.

Effects of FGFR-2 IIIc on Sphere Formation and Side Population

To determine whether FGFR-2 IIIc contributed to CSC-like features, we performed sphere formation and SP assays in FGFR-2 IIIc-overexpressing pancreatic cancer cells. In the sphere formation assay, FGFR-2 IIIc-transfected cells formed larger (Figure 6A) and more numerous (Figure 6B) spheres, compared with mock-transfected cells. Next, the proportion of SP fraction in pancreatic cancer cells was determined by flow cytometry (Figure 6C). FGFR-2 IIIc-transfected cells (FGFR-2 IIIc-1 and -6) also exhibited increased SP fractions, compared with mock and wild (parental) cells (Figure 6D). Subcutaneous injection of SP and MP cells of KLM-1 cells into NOD/SCID mice, revealed that the tumors formed when either 1.0 × 10⁴ or 1.0 × 10⁵ cells were injected, but not when fewer cells were injected (Table 2). Moreover, the tumor formation ratio was always higher in SP cells than in MP cells (Table 2).

Mitogenic signaling through FGFRs often involves activation of the MAPK pathway. We therefore sought to determine whether FGF-2-mediated activation of MAPK was altered as a consequence of changes in the levels of FGFR-2 IIIc. FGF-2 caused a rapid increase in p-ERK levels in both mock-2 cells and FGFR-2 IIIc-6 transfected PANC-1 cells. However, the phosphorylation levels of ERK were markedly higher at 10 and 20 minutes after FGF-2 addition in FGFR-2 IIIc-transfected KLM-1 cells (Figure 6E). By contrast, the number of apoptotic cells (sub G0/G1 phase) was decreased in FGFR-2 IIIc-transfected cells (2.12% and 2.80%), compared with wild (4.24%) and mock-transfected cells (3.57% and 4.25%) (Figure 6F).

Transient Transfection of FGFR-2 IIIc in Capan-1 Cells

To confirm our findings with KLM-1 cells, we next assessed the consequences of transient transfection of FGFR-2 IIIc in Capan-1 cells, because these cells exhibited the lowest expression levels of FGFR-2 IIIc mRNA. FGFR-2 IIIc mRNA (see Supplemental Figure S2A at http://ajp.amjpathol.org) and protein levels (see Supplemental Figure S2B at http://ajp.amjpathol.org) were markedly increased 72 hours after transfection. FGFR-2 IIIc-transfected Capan-1 cells exhibited increased cell proliferation, compared with wild and mock-transfected cells (see Supplemental Figure S2C at http://ajp.amjpathol.org). Moreover, FGF-2 and FGF-7 enhanced the migration of Capan-1 cells, but this effect was statistically greater in FGFR-2 IIIc-transfected cells, compared with mock-transfected cells (see Supplemental Figure S2D at http://ajp.amjpathol.org).

Effects of FGFR-2 IIIc Down-Regulation on PANC-1 Cells

To further assess the importance of FGFR-2 IIIc in the modulation of pancreatic cancer cell function, we next sought to examine the consequences of down-regulating FGFR-2 IIIc in PANC-1 cells, which expressed the highest levels of FGFR-2 IIIc among all our cell lines. Accordingly, two different sequences of siRNA against FGFR-2 IIIc mRNA (s275290 and s275291) were used to induce the down-regulation of FGFR-2 IIIc. Both siRNAs markedly decreased FGFR-2 IIIc mRNA levels, compared with those in control siRNA-transfected PANC-1 cells, at 24 hours after transfection (Figure 7A). By flow cytometry, there were fewer FGFR-2 IIIc-expressing cells after transfection with FGFR-2 IIIc siRNA (1.85% and 1.06%), compared with cells transfected with negative siRNA (7.37%) (Figure 7B). Basal proliferation rates of PANC-1 cells transfected with either FGFR-2 IIIc siRNAs were significantly lower than proliferation in control siRNA-transfected cells (Figure 7C). Compared with negative siRNA-transfected cells, FGFR-2 IIIc siRNA-transfected PANC-1 cells showed decreased ERK phosphorylation in response to FGF-2 (Figure 7D).

Next, we sought to determine whether the anti-FGFR-2 IIIc antibody would inhibit the proliferation, migration, and invasion of PANC-1 cells. Incubation of PANC-1 cells with the anti-FGFR-2 IIIc antibody exerted a dose-dependent inhibitory effect on the PANC-1 cell proliferation, maximal inhibition occurring at a concentration of 100 μg/mL, whereas the control rabbit IgG was without effect (Figure 8A). The anti-FGFR-2 IIIc antibody also inhibited the migration of PANC-1 cells (Figure 8B), but did not alter their invasion (Figure 8C). Time-lapse analysis indicated that the anti-FGFR-2 IIIc antibody markedly inhibited cell motility, compared with control IgG (Figure 8D). Proliferation and migration of KLM-1 cells, which express very low levels of FGFR-2 IIIc, were not inhibited by the addition of FGFR-2 IIIc antibodies (Figure 8, A and D).