Volume 123, Issue 1, Pages 301-313 (July 2002) IIIc isoform of fibroblast growth factor receptor 1 is overexpressed in human pancreatic cancer and enhances tumorigenicity of hamster ductal cells  Marko Kornmann, Toshiyuki Ishiwata, Kei Matsuda, Martha E. Lopez, Kimi Fukahi, Goro Asano, Hans G. Beger, Murray Korc  Gastroenterology  Volume 123, Issue 1, Pages 301-313 (July 2002) DOI: 10.1053/gast.2002.34174 Copyright © 2002 American Gastroenterological Association Terms and Conditions

Fig. 1 Analysis of FGFR-1 isoform expression. (A) Diagram of the structure of FGFR-1 III isoforms, modified from Johnson and Williams.18 In case of FGFR-1 IIIc, the second half of the Ig-like domain III (solid) corresponds to the exon 7 sequence of the FGFR-1 gene, and, in the case of IIIb, the second half of the Ig domain III (open) corresponds to the exon 6 sequence of the FGFR-1 gene.19 The IIIa isoform is devoid of any signaling capacity. (B and C) Ribonuclease protection assay distinguishing the IIIc and IIIb isoforms of FGFR-1. Hybridization with the FGFR-1 probe (330 nt) resulted in protected bands of 301 nt for the IIIc and of 156 nt for the IIIb isoform as indicated at right. (B) Human pancreatic cancer cell lines: DNA ladder (lane 1), total RNA (10 μg) from ASPC-1 (lane 2), CAPAN-1 (lane 3), COLO-357 (lane 4), HS776T (lane 5), Mia PaCa-2 (lane 6), PANC-1 (lane 7), and T3M4 (lane 8) cells, yeast tRNA (10 μg; lane 9), yeast tRNA without RNAse (lane 10). (C) Pancreatic tissues: total RNA (20 μg) from 4 normal (lanes 1–4) and 6 cancerous (lanes 5–10) pancreatic tissues, yeast tRNA (20 μg; lane 11), yeast tRNA without RNAse (lane 12). Gastroenterology 2002 123, 301-313DOI: (10.1053/gast.2002.34174) Copyright © 2002 American Gastroenterological Association Terms and Conditions

Fig. 2 Expression of FGFR-1 IIIc in human pancreatic cancer tissues. (A and B) In situ hybridization analysis revealed a moderate to strong FGFR-1 IIIc mRNA signal in most of the duct-like cancer cells and a faint to moderate signal in some fibroblasts surrounding the cancer cells. Arrow denotes the area of higher magnification in B. Serial sections with a sense probe for FGFR-1 IIIc did not show any specific signal (C). [Original magnification: (A) 200×, (B and C) 400×.] Gastroenterology 2002 123, 301-313DOI: (10.1053/gast.2002.34174) Copyright © 2002 American Gastroenterological Association Terms and Conditions

Fig. 3 Comparison of the nucleotide and amino acid sequences of FGFR-1 isoforms. The nucleotide and derived amino acid sequences of human FGFR-1 IIIc (147 nt) and IIIb (153 nt) cDNAs are shown. Boldface letters indicate shared amino acids. Gastroenterology 2002 123, 301-313DOI: (10.1053/gast.2002.34174) Copyright © 2002 American Gastroenterological Association Terms and Conditions

Fig. 4 Characterization of FGFR-1 expression in TAKA-1 cells. For immunoblotting (upper panel), total cell lysates (20 μg/lane) from the indicated cells were subjected to 8% SDS-PAGE, transferred to an Immobilon-P membrane, and blotted with a highly specific anti-FGFR-1 antibody (1:1000 or 500 ng/mL). Location of molecular weight markers is shown at left. For Northern blot analysis (middle and lower panels), total RNA (20 μg/lane) from the indicated cells was hybridized with an [α-32P]dCTP-labeled FGFR-1 cDNA probe. Location of rRNA is indicated at left. A 7S cDNA probe was used as loading control (lower panel). Exposure times were 12 hours for FGFR-1 and 1 hour for 7S. Gastroenterology 2002 123, 301-313DOI: (10.1053/gast.2002.34174) Copyright © 2002 American Gastroenterological Association Terms and Conditions

Fig. 5 Characterization of anchorage-dependent growth. Cells (5000/well) were seeded in 96-well plates and incubated for 48 hours in complete medium before initiation of the MTT assay. Results are expressed as growth above basal growth of parental TAKA-1 cells and are the means ± SEM from quadruplicate determinations from 5 separate experiments. *P < 0.008 and **P < 0.0001 compared with parental and sham-transfected TAKA-1 cells, respectively. Gastroenterology 2002 123, 301-313DOI: (10.1053/gast.2002.34174) Copyright © 2002 American Gastroenterological Association Terms and Conditions

Fig. 6 Effects of FGFR-1 IIIc expression on tyrosine phosphorylation. (A) Immunoblot with the PY20 antiphosphotyrosine antibody. Lysates (20 μg/lane) from exponentially growing cells were subjected to 8% SDS-PAGE and immunoblotted with PY20 antibodies (1:5000). The membrane was stripped and reprobed. Proteins that comigrated on these immunoblots with the tyrosine-phosphorylated bands are indicated at right. Migration of the molecular weight markers is indicated at left. (B) Equivalent loading of lanes was confirmed by immunoblotting with anti-PI 3-kinase antibodies (1:1000). (C) Immunoprecipitation of FGFR-1. FGFR-1 was immunoprecipitated (IP) in cell lysates (1 mg in 1 mL) with 2 μg of the anti-FGFR-1 antibody and subsequently immunoblotted (IB) with the PY20 antibody (1:5000). (D) Pull-down (P) of tyrosine phosphorylated Shc with Grb2-GST fusion protein. Cell lysates (1 mg in 1 mL) were incubated with 10 μL of Grb2-GST fusion protein. Immunoblotting (IB) of captured complexes was carried out with anti-Shc antibodies (1:500). (E) Immunoprecipitation of PLCγ. PLCγ was immunoprecipitated from cell lysates (1.5 mg in 1 mL) with 4 μg of the anti-PLCγ antibody and subsequently immunoblotted with the PY20 antibody (1:5000). (F) Precipitation with p13suc1. Immunoblot analysis with the PY20 antibody (1:5000) of p13suc1 precipitates from the indicated exponentially growing cells identified the p89 in A as phosphorylated FRS2. Gastroenterology 2002 123, 301-313DOI: (10.1053/gast.2002.34174) Copyright © 2002 American Gastroenterological Association Terms and Conditions

Fig. 7 MAPK activity in TAKA-1 cells and effects of PD98059. Exponentially growing cells were incubated without or with 5 μmol/L (upper panel) or 20 μmol/L (lower panel) of the MAPK inhibitor PD98059 for 2 hours. After cell lysis, immunoblotting was carried out with specific Anti-Active MAPK antibodies (1:4000). After stripping, membranes were reprobed with anti-ERK-2 antibodies (1:2000) recognizing active and inactive (pan) ERK-2 to confirm equivalent loading of the lanes. Gastroenterology 2002 123, 301-313DOI: (10.1053/gast.2002.34174) Copyright © 2002 American Gastroenterological Association Terms and Conditions

Fig. 8 Characterization of anchorage-independent growth by soft agar assay. Cells (10,000/well) were seeded in 6-well plates in complete medium containing 0.3% agar. After 21 days, colonies were stained with MTT and counted using an inverted light microscope. Data are expressed as mean colony number (±SD) of triplicate determinations for each cell line from a representative of 3 experiments. *P ≤ 0.009 and **P ≤ 0.0003 compared with TAKA-1 or F9 cells. Gastroenterology 2002 123, 301-313DOI: (10.1053/gast.2002.34174) Copyright © 2002 American Gastroenterological Association Terms and Conditions

Fig. 9 Tumor formation in nude mice. Cells (1 × 106) were injected SC at each of 2 sites in athymic (nude) mice. (A) Tumor size was measured externally every 2 weeks in 2 dimensions and tumor volume determined by the equation vol = (l × w2) × 0.5, where vol is volume, l is length, and w is width of the tumor. Values are means ± SE (n = 6); F9 (sham, ▴), F15 (●), F16 (■). (B) After 8 weeks, no tumors were visible at sites that had been injected with cells from clone F9 (arrows, left mouse). In contrast, sites that had been injected with cells of clone F15 or F16 (outlined by arrowheads, right mouse) developed visible tumors of at least 1-cm diameter within 8 weeks. Gastroenterology 2002 123, 301-313DOI: (10.1053/gast.2002.34174) Copyright © 2002 American Gastroenterological Association Terms and Conditions

Fig. 10 FGFR-1 immunohistochemistry in human pancreatic cancer and in xenograft tumors. (A) FGFR-1 immunohistochemistry in human pancreatic cancer, using the polyclonal anti-FGFR-1 antibody (1:16,000) that was used for immunoprecipitation. A higher magnification (A, inset) revealed a strong cytoplasmic and nuclear staining signal in most cancer cells and a moderate signal in some of the surrounding stromal cells (arrowheads). (B) The xenograft tumors consisted of areas in which the cancer cells exhibited strong FGFR-1 immunoreactivity (periphery of panel), moderate immunoreactivity (arrow, depicted in inset), and weak immunoreactivity (arrowheads). (C) Longitudinal sectioning revealed that the cells were arranged in cord-like structures that exhibited regions of strong FGFR-1 immunoreactivity and regions of weak immunoreactivity (arrowheads). (D) High-power view of C reveals clusters of cancer cells exhibiting intense membranous and moderate cytoplasmic FGFR-1 signal (outlined by arrowheads). (E) Immunohistochemistry of serial sections using a Ki-67 antibody revealed the presence of strong nuclear staining in the majority of the cells forming the clusters in D (outlined by arrowheads). In contrast, cells that exhibited weak FGFR-1 immunoreactivity did not exhibit Ki-67 immunoreactivity (outlined by arrows). [Original magnification: (A) 80×, (B) 50×, (C, E, insets) 200×, (D) 500×.] Gastroenterology 2002 123, 301-313DOI: (10.1053/gast.2002.34174) Copyright © 2002 American Gastroenterological Association Terms and Conditions