Publication of the International Union Against Cancer
Int. J. Cancer: 102, 592– 600 (2002) © 2002 Wiley-Liss, Inc.
VASCULAR ENDOTHELIAL GROWTH FACTOR MEDIATED ANGIOGENIC POTENTIAL OF PANCREATIC DUCTAL CARCINOMAS ENHANCED BY HYPOXIA: AN IN VITRO AND IN VIVO STUDY Bence SIPOS1,3*, Dirk WEBER1, Hendrik UNGEFROREN2, Holger KALTHOFF2, Andre ZU¨ HLSDORFF1, Claudia LUTHER2, Virag TO¨ RO¨ K3 and Gu¨nter KLO¨ PPEL1 1 Department of Pathology, University of Kiel, Kiel, Germany 2 Department of General Surgery, Molecular Oncology Division, University of Kiel, Kiel, Germany 3 2nd Department of Pathology, Semmelweis University, Budapest, Hungary Angiogenesis in pancreatic ductal adenocarcinomas depends on the presence of angiogenic factors such as vascular endothelial growth factor (VEGF) and basic fibroblast growth factor (bFGF) and is thought to be stimulated by hypoxia. We tested the angiogenic potential of 9 cell lines of pancreatic ductal carcinoma origin by screening mRNA and protein expression of VEGF and bFGF and the release of VEGF into culture medium under normoxic and hypoxic (5% or 0.2% O2) conditions. Angiogenic activity was determined using 2- and 3-D endothelial cell assays. Furthermore, VEGF expression and tumor vascularization were studied in human pancreatic carcinoma tissues from orthotopic xenografts and resection specimens. All cell lines expressed (mRNA, protein) and secreted VEGF, whereas bFGF was only found in 3 cell lines and was secreted into the medium in low concentrations. In addition to the dominant isoforms VEGF121,VEGF165 and VEGF189, 2 isoforms described recently, VEGF145 and VEGF183, were detected. Severe hypoxia (0.2% O2), but not moderate hypoxia (5% O2) raised VEGF mRNA expression and protein secretion in 7/9 and 5/9 cell lines, respectively. Conditioned media from 7/9, 6/9, 8/9 and 7/9 cell lines stimulated endothelial cell proliferation under normoxic (24 and 48 hr) or hypoxic (24 hr, 0.2% and 48 hr 5% O2) conditions, respectively. Conditioned media from 4/9 cell lines also induced capillary-like sprouting under normoxic conditions and from 6/9 under hypoxic (0.2% O2) conditions. In xenografted carcinoma tissues microvessel density was found not to be increased around areas of ischemic necrosis. In resected ductal carcinomas showing tumor necrosis VEGF expression and microvessel density were only increased in 3/12 and 2/13 cases, respectively. In conclusion, in vitro most pancreatic ductal carcinomas show a distinct VEGF related angiogenic potential, as demonstrated by 2- and 3-D endothelial cell proliferation, which may be promoted by severe hypoxia. Surprisingly, perinecrotic tumor areas, which are supposed to be hypoxic, only rarely showed the expected increase in microvessel density and VEGF expression. © 2002 Wiley-Liss, Inc. Key words: pancreatic carcinoma; angiogenesis; VEGF isoforms; hypoxia
The induction of angiogenesis by factors that stimulate the development of new blood vessels from the endothelium of the preexisting vasculature is a prerequisite for the growth of solid neoplasms. This also holds for pancreatic ductal adenocarcinoma (PDAC), one of the human cancers with the worst prognosis. Among the angiogenic factors that have been described in association with PDAC are vascular endothelial growth factor (VEGF),1–5 acidic and basic fibroblast growth factor (aFGF, bFGF),6 – 8 platelet derived endothelial cell growth factor (PDECGF),2,3 hepatocyte growth factor (HGF),9 –11 platelet derived growth factor (PDGF),12,13 epidermal growth factor (EGF)14,15 and pleiotrophin,16 however, only VEGF, PD-ECGF and bFGF were investigated for their angiogenic potential in these tumors. Because VEGF seems to be one of the main regulators of angiogenesis, most studies have focused on this angiogenic factor. In vitro, it was shown recently that PDAC cell lines constitutively overexpress VEGF and secrete the protein into the culture
medium. For this overexpression the Sp1 transcription factor binding site was basically responsible.17 Luo et al.18 showed that PDAC cell lines secrete functionally active VEGF using conditioned media of 2 PDAC cell lines that were able to stimulate the proliferation of human endothelial cells. There is a body of evidence indicating that VEGF not only acts as an angiogenic factor but also stimulates the proliferation of PDAC cells via KDR/flk1 receptor in autocrine manner.18,19 In vivo investigations on human PDACs demonstrated that 65–100% of them overexpress VEGF,1–5 whereas untransformed ductal epithelium did not express VEGF. These studies yielded controversial data about the correlation of VEGF expression and microvessel density (MVD), a marker of the angiogenic potential of tumors. In 4 studies MVD was found to correlate with VEGF expression in PDACs,2,5,20,21 1 study reported a moderate correlation,3 but in another1 no correlation was found. Hypoxia plays a role in many pathological processes, including tumor formation. The adaptation of tumor cells to hypoxia renders them able to escape apoptotic mechanisms and becoming angiogenic. It was shown that angiogenesis was enhanced around necrotic areas in solid tumors,22,23 probably due to a hypoxia stimulated increase in VEGF expression mediated by hypoxia inducible factor-1 (HIF1) via transcriptional activation.24 Very recently, it was demonstrated that HIF1␣ is overexpressed in PDAC cell lines without hypoxic stimulus in vitro.25 To further determine the angiogenic potential of PDAC, we investigated how hypoxia influenced the expression and secretion of VEGF and bFGF in 9 PDAC cell lines, using 2- and 3-D angiogenic assays modeling proliferation, migration and capillary formation in vitro. Furthermore, we examined VEGF expression and MVD in perinecrotic areas of xenografted and resected PDACs . MATERIAL AND METHODS
Cell culture The human PDAC cell lines Capan-2, HPAF-2, Hs700t, Hs-766t were purchased from American Type Culture Collection (Manassas, VA), while A818-4 and PT45P1 were established by our Grant sponsor: Hungarian E¨otv¨os Fellowship; Grant sponsor: OTKA; Grant number: F026017; Grant sponsor: FKFP; Grant number: 1235/96; Grant sponsor: Werner and Klara Kreitz Foundation. *Correspondence to: Department of Pathology, University of Kiel, Michaelisstr. 11, 24105 Kiel, Germany. Fax: ⫹49-431-597-3462. E-mail:
[email protected] Received 4 April 2002; Revised 12 July 2002; Accepted 27 August 2002 DOI 10.1002/ijc.10753
ANGIOGENESIS IN PANCREATIC CARCINOMA
group. Colo357 was provided by R. Morgan (Denver, CO). Panc89 (T3M4) was a gift from T. Okabe (Japan); PancTu-I was donated by M. von Bu¨ low (Mainz, Germany).26 Cells were grown to subconfluence in RPMI 1640 with Glutamax (GIBCO, Karlsruhe, Germany) containing 10% FCS (PAA Laboratories, Co¨ lbe, Germany). After washing, the medium was replaced with RPMI 1640 containing 0.5% FCS with or without heparin (100 g/ml, Sigma, Deisenhofen, Germany) and incubated in a humidified incubator under normoxic (5% CO2, 95% air), moderately hypoxic (5% O2, 10% CO2, 85% NO2) and severely hypoxic (0.2% O2, 18% CO2, 82% N2) conditions. A hypoxic milieu was generated in anaerobic jars (Merck, Darmstadt, Germany) with Anaerocult C and A (Merck, Darmstadt, Germany) reagents, respectively. After 24 hr (normoxic-severely hypoxic pairs) or 48 hr (normoxic-moderately hypoxic pairs) conditioned media were collected, centrifuged, filtered through a 0.22 m pore, low protein binding filter (Millipore) and stored at ⫺80°C. Cells were trypsinized and counted in a hemocytometer, centrifuged and the pellets stored at ⫺80°C. Every experiment was carried out in duplicate and in 2 independent series. Conditioned media (CM) were tested for mycoplasma contamination with a polymerase chain reaction-based test (Takara Biomedical, Shiga, Japan), and only negative samples were further processed. Human umbilical vein endothelial cells (HUVECs) and dermal microvascular endothelial cells (DMVECs) were purchased from Promocell (Heidelberg, Germany). The cells were cultured in Endothelial Cell Growth Medium (C-39210, Promocell, Heidelberg) and Endothelial Cell Growth Medium MV (C-39220, Promocell, Heidelberg) containing 2% and 5% FCS, respectively. Cells between the third and fifth passage were used for the experiments. RNA extraction and cDNA synthesis Total cellular RNA from pancreatic cancer cell lines was isolated using the RNA extraction reagent RNAzol (Tel-Test, Inc., Friendswood, TX). The quality of the RNA samples was determined by electrophoresis through denaturing agarose gels. For cDNA synthesis, 5 g of total RNA from each cell line was denatured together with oligo(dT)15 primer (50 pmol) for 15 min at 70°C. After this the sample was chilled on ice for 5 min, poly(A)⫹-RNA was reverse transcribed at 42°C for 60 min in RT solution [50 mM Tris-HCl, pH 8.3, 40 mM potassium chloride, 8 mM magnesium chloride, 0.5 mM each dNTP, 225 g/ml BSA, 5 mM dithiothreitol, 20 U of RNAsin and 20 U of AMV reverse transcriptase] with a total volume of 25 l. All reagents were provided by a cDNA synthesis kit (Promega Biotech, Madison, WI). The cDNA was incubated at 95°C for 5 min to inactivate the reverse transcriptase and stored at ⫺80°C until analysis. RT-PCR for bFGF For the detection of bFGF mRNA the forward and reverse primers were 5⬘-GGG ACC ATG GCA GCC GGG AG-3⬘ and 5⬘TCA GCT CTT AGC AGA CAT TGG AAG-3⬘. The PCR conditions were as follows: initial denaturation at 95°C for 4 min, amplification at 95°C, 57°C, 72°C each for 1 min for 30 cycles, final extension at 72°C for 10 min. PCR was carried out using 0.5 U of Taq DNA polymerase (Life Technologies, Eggenstein, Germany) with 10 pmol of each primer (bFGF or GAPDH, respectively), 200 M each dNTP, 2 mM MgCl2, 10 mM Tris-HCl, pH 8.3, 50 mM KCl (Life Technologies) and water in a total volume of 25 l. As an internal control the housekeeping gene GAPDH mRNA was also amplified under the same conditions. Two microliters of each PCR reaction were diluted with a 10⫻ gel loading buffer (Life Technologies) and dH2O to a volume of 10 . Eight l aliquots of each were loaded on 5% polyacrylamide/urea gels and run for 2 hr at 400 V. Bands were subsequently visualized by silver staining.
593
RT-PCR quantification of VEGF mRNA The PCR oligonucleotide primer sequences for VEGF were chosen to include the entire coding sequence (nucleotides [nts] 17–592, Gene bank accession number x62568) VEGF-sense (nts 16 –37): CATGAACTTTCTGCTGTCTTGG, VEGF-antisense (nts 629 – 649): CCTGGTGAGAGATCTGGTTCC. Thermal cycling was carried out with Taq polymerase (Life Technologies) as follows: denaturation at 95°C for 1 min, annealing at 65°C for 1 min, and extension at 72°C for 1 min. Every third cycle the annealing temperature was reduced by 2°C until a temperature of 55°C was reached, at which 24 cycles were carried out. PCR products were analyzed by electrophoresis through 1.5% agarose gels and viewed under UV light after ethidium bromide staining. To identify the various mRNA isoforms, the lower 4 of 5 specific amplification products obtained were excised from the gel, electroeluted using the Biotrap device (Schleicher & Schu¨ ll, Dassel, Germany) and purified by pheno/chloroform-extraction. After ethanol precipitation, PCR products were subcloned into the pDrive vector (Qiagen, Hilden, Germany) and sequenced on both strands. To detect the VEGF 206 and 189 isoforms, we synthesized an 18-mer (GGTATAAGTCCTGGAGCGT) that corresponded to the 6 amino acids present only in VEGF 206 and 18927 and used this together with the VEGF antisense primer to amplify a calculated region of 552 base pairs. Discrimination between VEGF 206 and 189 was accomplished by RT-PCR with VEGF-sense and an antisense primer (TCCAGGGCATTAGACAGCAGC) that binds to a nucleotide sequence unique to VEGF 206. For quantification of VEGF mRNA a competitive approach was chosen. Multiple reactions were run in parallel with identical amounts of cDNA (corresponding to 100 ng of total RNA) but different amounts of standard DNA. The standard DNA was constructed by PCR using the VEGF-antisense primer and a hybrid sense primer that contained nts 16 –37 followed by nts 273–292 (CATGAACTTTCTGCTGTCTTGG-GCTGCAATGACGAGGGCCTG). Reamplification of this product with VEGF-sense and VEGF-antisense primers yielded the 399 bp standard fragment. The induction of VEGF mRNA upon incubation of cells under various oxygen pressures was assessed from those reactions that showed an equal intensity of ethidium bromide staining of the approx. 500 bp amplification product (corresponding to VEGF121) and the standard. Immunoblot assay For preparation of protein extracts, cells and tumor tissues from a xenograft model were mechanically disrupted with micro-pellet pestles and lysed in cold lysis buffer (0.1% SDS, 1% NP40, 0.5 % sodium deoxycholate in PBS) for 30 min on ice. The lysate was cleared by centrifugation and protein concentration was determined with a BCA Protein Assay (Pierce Chemical Company, IL). Eighty g protein of each lysate were loaded on a standard 15% SDS polyacrylamide gel under reducing conditions followed by blotting onto a PVDF membrane (Immobilon-P, Millipore, Eschborn, Germany). After blocking with low-fat milk overnight, VEGF was detected using polyclonal rabbit anti-VEGF antibody (Santa Cruz, A-20, sc-152) in a concentration of 2 g/ml. The
FIGURE 1 – RT-PCR analysis of VEGF isoforms in pancreatic carcinoma cell lines under normoxic conditions.
594
SIPOS ET AL.
FIGURE 2 – Immunoblot analysis of whole cell lysates of pancreatic carcinoma cell lines under reducing conditions detected by anti-VEGF antibody (A20, Santa Cruz). Human recombinant VEGF121 and VEGF165 (R&D Systems) served as positive control. Antibody competition was done by coincubation with 10⫻ concentrated rhVEGF121 and VEGF165 protein before staining.
FIGURE 3 – Quantification of VEGF mRNA from the HPAF-2 cell line under normoxic and severely hypoxic conditions by competitive RT-PCR. The numbers indicate the dilution factor of the internal standard in the PCR at which the staining intensity of the VEGF121 amplification product and that of the internal standard were comparable.
secondary antibody was horseradish-peroxidase conjugated goat anti-rabbit IgG from Zymed (San Francisco, CA). The blots were developed with Western blotting detection reagent ECL (Amersham Pharmacia Biotech, Buckinghamshire, UK) and autoradiographed onto a BIOMAX MR film (EASTMAN KODAK Company, Rochester, NY). As a control of the specificity of the anti-VEGF antibody, it was preincubated with recombinant human VEGF121 and VEGF165 (R&D Systems, Wiesbaden, Germany) for 1 hr. ELISA of conditioned media The VEGF and bFGF concentrations of CM were determined with Quantikine ELISA Kits (R&D Systems) according to the manufacturers instructions. Each CM was tested in duplicate in 2 separate measurements using a Microplate Reader (MWG Biotech, Ebersberg, Germany). Results were calculated with SoftMAx Pro software (Molecular Devices, Sunnyvale, CA). Endothelial cell proliferation assay DMVECs and HUVECs were plated overnight in endothelial cell growth medium in 96-well culture plates at a density of 3 ⫻ 103/well. Before seeding the plates were coated with human collagen-I (Becton Dickinson) at a concentration of ⬃1 g/cm.2. The
medium was replaced by serum free endothelial cell medium (PromoCell, Heidelberg, Germany) and incubated for 24 hr. To test the CM, cells were overlaid with 100 l of medium containing 50% CM and 50% endothelial cell medium with 0.5 % FCS and incubated for 72 hr. One hour before incubation, antiVEGF neutralizing antibody (2 g/ml, R&D Systems, Wiesbaden, Germany) was added to 1 set of the CM. Every CM was tested 6-fold in 3 independent series with DMVECs and in 1 series with HUVECs. Cell proliferation was determined with a tetrazolium salt based proliferation assay (Cell Proliferation Kit II, Roche Diagnostics, Mannheim, Germany). In 1 of the DMVEC series, the cells were pulsed with 2 Ci/well of [3H]thymidine (Amersham Pharmacia Biotech, Freiburg, Germany) for the last 6 hr of incubation. The cells were harvested, and the incorporated radioactivity was counted in a Micro-Betaplate liquid scintillation counter (Wallac, Gaithersburg, MD). As a control, endothelial cells from each series were stimulated with human recombinant VEGF (R&D Systems) in 1, 3, 7 and 12 ng/ml concentration with and without neutralizing anti-VEGF antibody under identical conditions. Three-dimensional in vitro angiogenesis assay The capability of CM to induce capillary-like sprout formation in vitro was tested in a 3-D HUVEC spheroid assay described previously.28 Briefly, HUVECs were seeded in 96-well roundbottom plates in endothelial cell growth medium containing 20% methylcellulose (Sigma) and incubated overnight, after which they were embedded in collagen gel. Collagen stock solution was prepared from 8 vol rat tail collagen solution (Serva, Heidelberg, Germany) with 1 vol 10⫻ EBSS (Gibco BRL, Eggenstein, Germany) and the pH was adjusted to 7.4 with 0.1 M NaOH. The stock solution was mixed 1:1 with endothelial cell basal medium containing 20% FCS and 20% methylcellulose and the spheroids were seeded in this solution in 24-well plates (0.5 ml/well). After 15 min incubation at 37°C the collagen was polymerized. CM (0.5 ml) with or without preincubation with neutralizing anti-VEGF antibody (R&D Systems) were pipetted on top of the gels. After 36 hr incubation, the sprouts of each intact spheroid were counted (⬃50/ well) with an Olympus microscope. Each CM of the severely hypoxic-normoxic conditions was tested twice in 2 independent series. For controls, the spheroids were incubated with RPMI 1640 only, 2 ng/ml and 5 ng/ml recombinant human VEGF (R&D Systems, Wiesbaden, Germany).
595
ANGIOGENESIS IN PANCREATIC CARCINOMA
FIGURE 4 – VEGF concentration of conditioned media (CM) of pancreatic carcinoma cells under normoxic and severely hypoxic conditions measured by ELISA. The rows below the diagram indicate the effect of the CM on the proliferation of dermal microvascular cells (upper row) and on the capillary-like sprouting of human umbilical vein endothelial cells (lower row). ⫺, no significant effect, ⫹, significant stimulation, ⫹1, additional significant stimulation by the anoxic CM. The proliferation assays were carried out in sextuplicate in 3 independent experiments. The capillary-like sprouting assays were done in 2 independent series assessing 30 –50 spheroids in each series.
Orthotopic xenograft model PancTu-I and Panc89 cells suspended in RPMI (106 cells/20 l) were inoculated into the pancreas of SCID-beige mice (n ⫽ 10). Solid tumors had grown after 35 days (PancTu-I) and 45 days (Panc89), at which time the mice were sacrificed. The tumors were removed and immediately snap frozen in a mixture of isopentane and dry ice. The mean volume of the tumors was 1,250 mm3 (PancTu-I) and 1,310 mm3 (Panc89). All animal experiments were carried out in accordance with the ethical guidelines for animal experiments of the State of Schleswig-Holstein. Human PDAC tissues Human PDAC were surgically removed, formalin fixed and processed for routine histological examination. Hematoxylin and eosin stained sections of 97 consecutively resected PDACs were evaluated for the presence of necrotic areas. Nineteen carcinomas showed necrotic areas, and of these 19 cases 13 had a sufficient number of tumor blocks. From each of the 13 3–5 large blocks were processed for further immunohistochemical evaluations. Immunohistochemistry and determination of microvessel density Five micron thin cryosections of the xenografted tumors were cut, air dried and acetone fixed. After blocking with 5% rabbit serum, sections were incubated with rat anti-mouse CD31 antibody (Immunotech) and the reaction product was detected by the peroxidase-antiperoxidase method using anti-rat and APAAP antibodies (DAKO, Hamburg, Germany). The color reaction was developed with New Fuchsin and the sections were counterstained with hemalum. Formalin fixed sections of the human PDACs were blocked with 0.03 % H2O2 and 4% low fat milk powder solution. For the VEGF stain, sections were heat-pretreated in citrate buffer in a pressure cooker for 3 min. A mouse monoclonal anti-VEGF antibody (C1, Santa Cruz, CA) was applied, followed by avidin-biotin-peroxidase detection with the Vectastain-ABC Kit (Vector Laboratories, Burlingame, CA). The staining was carried out with an immunostainer (DAKO, Hamburg, Germany). For endothelial cell labeling, sections were incubated with anti-human CD34 antibody (Pharmingen-Becton Dickinson, Heidelberg, Germany) and the reaction was detected by the APAAP method using anti-mouse and APAAP antibodies (DAKO, Hamburg, Germany). The MVD of the xenografted tumors and human PDACs was determined by the hot spot method.29 Large fields (1 mm2) with vascular hot spots directly lining necrotic areas and in non necrotic carcinoma tissues were chosen and counted with the help of an ocular grid. Four to 11 hot spots from each zone were evaluated in
each tumor. For statistical evaluation 3 spots with the highest MVD were used. VEGF expression was evaluated by a semiquantitative method as follows: the staining intensity of the tumor cells was scored from 0 –3 and the proportion of the positive cells was expressed in percent. The final score for a 1 mm2 field was calculated as product of the staining quality and quantity. Statistical analysis The significance of the results was analyzed by means of the impaired t-test using Microsoft Excel Software. RESULTS
Expression of VEGF in PDAC cell lines All PDAC cell lines expressed VEGF at both the RNA and protein level. In all samples tested, the PCR amplification yielded 5 fragments ranging in size from 500 –700 bp (Fig. 1). Subcloning of the lower 4 fragments and sequencing on both strands showed that these various mRNA isoforms corresponded to VEGF121, VEGF145, VEGF165, VEGF183, respectively. Using oligonucleotide primers specific for VEGF189/206 and VEGF206, the mRNA isoform corresponding to the PCR product of the highest molecular weight was identified as VEGF189. Immunoblot analysis of the cell lysates showed 3 groups of bands at 17–22 kD, 24 –27 kD, 30 –31 kD. These correspond to the VEGF isoforms detected by RT-PCR (Fig. 2). In the original study by Houck et al.27 a 15 kD band of VEGF121 was also demonstrated, which we did not detect. All pancreatic carcinoma cell lines secreted VEGF into the CM as detected 1–2 days after confluence. Concentrations ranged from 0.07–15.5 ng/ml/106 cells (Fig. 4). Effect of moderate and severe hypoxia on VEGF expression in vitro Severe hypoxia caused a 2–50-fold increase in VEGF121 mRNA expression in all cell lines (Fig. 3 and data not shown). All splice variants were proportionately expressed. In 7 of 9 cell lines an increase in the smaller isoforms (VEGF121, VEGF145 or VEGF165) was observed at the protein level (Fig. 2). Increased VEGF release upon severe hypoxia was detected in 5 of 9 cell lines (Fig. 4). Moderate hypoxia had no effect on VEGF expression at the mRNA or protein level. In vitro angiogenic activity of PDAC cell lines The stimulatory effects of 36 CM that were obtained from 9 PDAC cell lines cultured under 4 conditions were tested. Under
596
SIPOS ET AL.
FIGURE 5 – Three-dimensional capillary-like sprout formation assay with human umbilical vein endothelial cell spheroids stimulated by CM of pancreatic carcinoma cell lines cultured under severely hypoxic conditions without or with neutralizing anti-VEGF antibody. Insert: examples of HUVEC spheroids after 36 hr incubation with (a) RPMI 1640, (b) 5 ng/ml rhVEGF165, (c) CM of Panc89 cell line or (d) CM of Panc89 with 2 g/ml neutralizing anti-VEGF antibody.
FIGURE 6 – Vascular hot spot in PancTu-I tumor (a) in nonnecrotic area and (b) in perinecrotic area (anti-mouse CD31 immunostain, APAAP detection, New Fuchsin chromogen). N, necrotic area.
normoxic conditions (24 hr and 48 hr) 6/9 and 8/9 CM, respectively were able to stimulate the proliferation of human endothelial cells; under severely hypoxic conditions (24 hr) 8/9 and under moderately hypoxic conditions (48 hr) 7/9. The stimulatory effect could be blocked with neutralizing anti-VEGF antibody in all but 1 case (28/29, 97%). We tested the severely hypoxic-normoxic CM pairs as to their capability to induce capillary-like sprouting of HUVECs. Four of 9 CM were able to induce sprouting in the normoxic setting and 6/9 in the severely hypoxic setting. Coincubation with neutralizing anti-VEGF antibody abrogated the induction of sprout formation in all cases (10/10) (Fig. 5). Expression of bFGF in PDAC cells bFGF mRNA was detected in only 3 (Hs766t, Colo357 and Panc89) of the 9 cell lines tested. Low levels of bFGF (10 –150 pg/ml), that could not be increased by heparin treatment, were detected in CM of the same 3 cell lines (data not shown).
VEGF expression and MVD in xenografted PDAC cell lines and resected PDACs Xenografted Panc89 and PancTu-I cell lines30 formed solid tumors in the pancreas of SCID mice. Immunoblot analyses of tissue lysates of the xenografted tumors showed the expression of all VEGF isoforms that were detected in vitro (data not shown). All tumors displayed central necrotic areas. There was no difference in MVD when hot spots from perinecrotic (PancTu-I: 62.1 ⫾ 12.4; Panc89: 91.3 ⫾ 21.6) and non necrotic (PancTu-I: 61.9 ⫾ 13.8: Panc89: 78.5 ⫾ 27.8) tumor areas were compared (Fig. 6). Nineteen of 97 human PDACs, contained large necrotic areas. These were more common in poorly differentiated carcinomas (11/39, 28.9%) than in moderately (7/45, 15%) and well differentiated ones (1/12, 8.3%). The largest necrotic area was observed in 1 undifferentiated carcinoma. Thirteen of these PDACs were immunohistochemically screened for VEGF expression and MVD. VEGF expression was found in tumor cells, stromal cells and
ANGIOGENESIS IN PANCREATIC CARCINOMA
occasionally macrophages. In peritumoral pancreatic tissue, acini at the periphery of lobules and islets were also positive. Overexpression of VEGF in perinecrotic areas was found in 3 tumors (23%). In only 2 of them (15%) was there a significant increase in the perinecrotic MVD (Figs. 7,8). Enhanced staining intensity was usually found in the invading zone of the tumors . DISCUSSION
Our study confirms and extends the observation that most PDAC cell lines show considerable, though variable, constitutive expression and secretion of VEGF and its isoform, including VEGF145 and VEGF183 mRNA transcripts, which had not been recognized previously in PDAC cell lines. In contrast to VEGF, bFGF showed only low level expression in a few PDAC cell lines. The secreted VEGF was found to stimulate proliferation and capillary-like sprouting of endothelial cells, and severe hypoxia promoted the in vitro VEGF-induced angiogenesis. In perinecrotic PDAC tissue, neither VEGF expression nor MVD were found to be consistently increased. PDACs are known to express VEGF in vitro and in vivo, the rate ranging from 56 –100%. We found constitutive VEGF mRNA expression and protein production in all 9 PDAC-derived cell lines tested. Their capacity for VEGF production differed greatly, however, with the highest values being present in Hs700t, Panc89 and PT45P1, and the lowest in Capan-2 and PancTu-I. As there was only a weak correlation between mRNA and protein levels of the detected VEGF isoforms, we assume that there are translational or posttranslational mechanism regulating VEGF expression. To test the relevance of our in vitro data on the distribution patterns of VEGF isoform expression to the in vivo situation, we xenografted a high and a low VEGF producing cell line, Panc89 and PancTu-I, and obtained results that were quite compatible with the in vitro data.
597
Until recently, VEGF was known to have 4 mRNA splice variants resulting in VEGF121,VEGF165,VEGF189 and VEGF206. The 2 smaller isoforms are soluble with no or moderate heparin binding capacity, whereas VEGF189 and VEGF206 are heparin binding and sequestered in the extracellular matrix.27,31 They also differ in their angiogenic/tumorigenic properties and to some extent in their receptor affinity. In a mouse mammary carcinoma model, VEGF121 was reported to be more angiogenic and tumorigenic than VEGF165 or VEGF189.32 VEGF121 selectively binds flk-1, one of the VEGF receptors.33 Moreover, neuropilin-1, the third VEGF receptor, is specific for VEGF165.34 This suggests that VEGF isoforms may have distinct biological functions. Surprisingly, we detected 2 other VEGF mRNA transcripts, VEGF145 and VEGF183, which had so far not been identified in PDACs, but were found to be expressed in endometrial carcinoma cell lines (VEGF145)35 and in heart tissue (VEGF183).36 VEGF145 is generated by alternative splicing at the same region as VEGF165, but contains exon 6a instead of exon 7. It has a receptor affinity only to KDR/flk1 but not to the 2 other VEGF receptors and has considerable heparin binding capacity.37 Because its angiogenicity has been demonstrated in vivo,37 it is likely that it also contributes to the angiogenic activity of PDAC. The structural similarity between VEGF183 and VEGF189 suggests an analogous biological activity of the 2 isoforms.36 The angiogenic effect of VEGF results in the stepwise formation of new vessels. Of these steps proliferation, migration and tube formation can be assessed in vitro.38 Luo et al.18 already showed that it is predominantly VEGF that determines the direct in vitro angiogenic activity of PDAC cells. Using a proliferation and a 3D capillary-like sprout formation assay, in which these 3 steps in angiogenesis can be reproducibly quantified, we confirmed and extended these results. We found increased proliferation upon incubation with the CM of 6 of 9 and 8 of 9 PDAC cell lines cultured for 24 hr and 48 hr under normoxic conditions. To prove
FIGURE 7 – Upper diagram: VEGF expression in human pancreatic carcinomas containing necrotic areas. The staining intensity of the cells and the proportion of cells stained were estimated semiquantitatively for each field, resulting in a score of 0 –3. Lower diagram: intratumoral microvessel density (MVD) of human pancreatic carcinomas. MVD was determined by the hot spot method in non necrotic and directly in perinecrotic areas. Black columns indicate significant differences (p ⬍ 0.05).
598
SIPOS ET AL.
FIGURE 8 – VEGF expression and microvessel density (MVD) in human pancreatic carcinomas. Carcinoma from Patient 12 with upregulation of VEGF and elevation of MVD in a perinecrotic area. (a) VEGF expression in tumor tissue. (b) VEGF expression in a perinecrotic area (immunostain with peroxidase detection, DAB chromogen). (c) MVD in tumor tissue. (d) MVD in a perinecrotic area (CD34 immunostain, APAAP detection, New Fuchsin chromogen). Carcinoma from Patient 11 with no upregulation of VEGF and no elevation of MVD in a perinecrotic area. (e) VEGF expression in tumor tissue. (f) VEGF expression in a perinecrotic area (immunostain with peroxidase detection, DAB chromogen). (g) MVD in tumor tissue. (h) MVD in a perinecrotic area (CD34 immunostain, APAAP detection, New Fuchsin chromogen). N, necrotic area.
that it is mainly VEGF that is responsible for the stimulatory effect of the CM tested, we used a neutralizing antibody against VEGF, which abrogated endothelial cell proliferation in almost all cases. Migration and capillary-like sprout formation tested in HUVEC spheroids were induced in 62% of the cell lines and inhibited by anti-VEGF in all cases. VEGF concentration in the CM and the stimulatory effect on sprouting correlated very well because 9 of 10 CM with VEGF levels of more than 2 ng/ml stimulated sprout formation, whereas only 1 of 8 CM with lower VEGF concentrations had an effect on sprouting. Hypoxia and its metabolic consequences, hypoglycemia and acidosis, may develop in solid tumor areas with diminished blood supply and are thought to upregulate the expression of angiogenic factors in tumor cells, in particular VEGF,24 mediated by HIF1␣.39,40 In PDAC cell lines moderate hypoxia (5% O2) failed to elevate VEGF mRNA and protein levels. Severe hypoxia (0.2% O2), however, led to an increase in VEGF mRNA expression in all
cell lines, stimulated VEGF production in 7/9 cell lines and enhanced VEGF secretion in 5/9 cell lines. It also exerted a clear-cut effect on the proliferation and sprouting of endothelial cells in two-thirds of the PDAC cell lines. To test these in vitro findings in vivo, we studied immunohistochemical VEGF expression and MVD in tumor tissues surrounding necrotic foci in xenografted and also in resected PDACs. Because perinecrotic tumor areas have a decreased oxygen tension, as was demonstrated in experimental studies,41– 44 they should theoretically display increased VEGF expression and MVD. We obtained little evidence for this assumption. No increased MVD was detected in perinecrotic areas of the 2 xenografted tumors, and of the 13 resected PDACs that contained necrotic areas only 2 tumors showed overexpression of VEGF and also displayed increased MVD. These results are surprising, but may be explained in several ways. First, it is conceivable that ischemic tumor necrosis is due to a sudden thrombotic occlusion of a tumor vessel.
599
ANGIOGENESIS IN PANCREATIC CARCINOMA
In this case it is possible that the perinecrotic tissue becomes temporarily severely hypoxic and later largely recovers. Second, a sustained reduction of the oxygen tension in tissues surrounding ischemic tumor necrosis could slow down the oxygen consumption of perinecrotic tumor cells, rendering them less sensitive to hypoxia. Third, in studies concerning HIF1␣ it has been found that HIF1␣ transfected PDAC cell lines expressing high levels of VEGF failed to increase microvessel density in nude mouse tumors.25 Conversely, loss of HIF1␣ did not alter tumor vascularization.45 Taken together, these considerations indicate that the in vivo situation is probably too complex to allow the in vitro findings to be directly transferred to the mechanisms of angiogenesis in perinecrotic tumor tissue. bFGF seems to play a role in the angiogenesis associated with carcinomas of the bladder, ovary and esophagus.46 – 48 Although it is expressed in more than half of PDACs and also influences the proliferation of PDAC cells by autocrine-paracrine mechanisms,49,50 it obviously is of minor importance for angiogenesis in PDAC.21 Our findings confirm this notion because bFGF expres-
sion was very low and only found in one-third of the PDAC cell lines. In summary, our data indicate that the angiogenic potential of PDAC cells mainly depends on their ability to produce and secrete VEGF. In vitro, this capacity can be stimulated by severe hypoxia. Whether this is also true in PDAC tissues remains an open question, as the findings are not yet conclusive, most likely because of the complexity and variability of the in vivo situation. ACKNOWLEDGEMENTS
We would like to thank M. Pacena, A. Paulus and S. Vollbehr for their excellent technical assistance and S. Haye, M. Riechmann and R. Zavatzka for their outstanding assistance with the mouse experiments. The 3-D endothelial spheroid assay was established with the kind support of T. Korff and H. Augustin (Freiburg, Germany). Our study was supported by the Hungarian Eo¨ tvo¨ s Fellowship (B.S.), by OTKA grant (F026017), FKFP grant (1235/ 96) and by the Werner and Klara Kreitz Foundation.
REFERENCES
1. 2.
3.
4. 5.
6.
7.
8.
9. 10. 11. 12.
13.
14. 15. 16.
Ellis LM, Takahashi Y, Fenoglio CJ, Cleary KR, Bucana CD, Evans DB. Vessel counts and vascular endothelial growth factor expression in pancreatic adenocarcinoma. Eur J Cancer 1998;34:337– 40. Fujimoto K, Hosotani R, Wada M, Lee JU, Koshiba T, Miyamoto Y, Tsuji S, Nakajima S, Doi R, Imamura M. Expression of two angiogenic factors, vascular endothelial growth factor and platelet-derived endothelial cell growth factor in human pancreatic cancer, and its relationship to angiogenesis. Eur J Cancer 1998;34:1439 – 47. Ikeda N, Adachi M, Taki T, Huang C, Hashida H, Takabayashi A, Sho M, Nakajima Y, Kanehiro H, Hisanaga M, Nakano H, Miyake M. Prognostic significance of angiogenesis in human pancreatic cancer. Br J Cancer 1999;79:1553– 63. Itakura J, Ishiwata T, Shen B, Kornmann M, Korc M. Concomitant over-expression of vascular endothelial growth factor and its receptors in pancreatic cancer. Int J Cancer 2000;85:27–34. Itakura J, Ishiwata T, Friess H, Fujii H, Matsumoto Y, Buchler MW, Korc M. Enhanced expression of vascular endothelial growth factor in human pancreatic cancer correlates with local disease progression. Clin Cancer Res 1997;3:1309 –16. Ohta T, Yamamoto M, Numata M, Iseki S, Tsukioka Y, Miyashita T, Kayahara M, Nagakawa T, Miyazaki I, Nishikawa K, Yoshitake Y. Expression of basic fibroblast growth factor and its receptor in human pancreatic carcinomas. Br J Cancer 1995;72:824 –31. Yamanaka Y, Friess H, Buchler M, Beger HG, Uchida E, Onda M, Kobrin MS, Korc M. Overexpression of acidic and basic fibroblast growth factors in human pancreatic cancer correlates with advanced tumor stage. Cancer Res 1993;53:5289 –96. Yamazaki K, Nagao T, Yamaguchi T, Saisho H, Kondo Y. Expression of basic fibroblast growth factor (FGF-2)-associated with tumour proliferation in human pancreatic carcinoma. Virchows Arch 1997; 431:95–101. Ebert M, Yokoyama M, Friess H, Buchler MW, Korc M. Coexpression of the c-met proto-oncogene and hepatocyte growth factor in human pancreatic cancer. Cancer Res 1994;54:5775– 8. Furukawa T, Duguid WP, Kobari M, Matsuno S, Tsao MS. Hepatocyte growth factor and Met receptor expression in human pancreatic carcinogenesis. Am J Pathol 1995;147:889 –95. Kiehne K, Herzig KH, F¨olsch UR. c-met expression in pancreatic cancer and effects of hepatocyte growth factor on pancreatic cancer cell growth. Pancreas 1997;15:35– 40. Kalthoff H, Roeder C, Humburg I, Thiele HG, Greten H, Schmiegel W. Modulation of platelet-derived growth factor A- and B-chain/c-sis mRNA by tumor necrosis factor and other agents in adenocarcinoma cells. Oncogene 1991;6:1015–21. Ebert M, Yokoyama M, Friess H, Kobrin MS, Buchler MW, Korc M. Induction of platelet-derived growth factor A and B chains and over-expression of their receptors in human pancreatic cancer. Int J Cancer 1995;62:529 –35. Barton CM, Hall PA, Hughes CM, Gullick WJ, Lemoine NR. Transforming growth factor alpha and epidermal growth factor in human pancreatic cancer. J Pathol 1991;163:111– 6. Lemoine NR, Hughes CM, Barton CM, Poulsom R, Jeffery RE, Klo¨ ppel G, Hall PA, Gullick WJ. The epidermal growth factor receptor in human pancreatic cancer. J Pathol 1992;166:7–12. Weber D, Klomp HJ, Czubayko F, Wellstein A, Juhl H. Pleiotrophin can be rate-limiting for pancreatic cancer cell growth. Cancer Res 2000;60:5384 –288.
17. Shi Q, Le X, Abbruzzese JL, Peng Z, Qian CN, Tang H, Xiong Q, Wang B, Li XC, Xie K. Constitutive Sp1 activity is essential for differential constitutive expression of vascular endothelial growth factor in human pancreatic adenocarcinoma. Cancer Res 2001;61: 4143–54. 18. Luo J, Guo P, Matsuda K, Truong N, Lee A, Chun C, Cheng SY, Korc M. Pancreatic cancer cell-derived vascular endothelial growth factor is biologically active in vitro and enhances tumorigenicity in vivo. Int J Cancer 2001;92:361–9. 19. von Marschall Z, Cramer T, Hocker M, Burde R, Plath T, Schirner M, Heidenreich R, Breier G, Riecken EO, Wiedenmann B, Rosewicz S. De novo expression of vascular endothelial growth factor in human pancreatic cancer: evidence for an autocrine mitogenic loop. Gastroenterology 2000;119:1358 –72. 20. Seo Y, Baba H, Fukuda T, Takashima M, Sugimachi K. High expression of vascular endothelial growth factor is associated with liver metastasis and a poor prognosis for patients with ductal pancreatic adenocarcinoma. Cancer 2000;88:2239 – 45. 21. Fujioka S, Yoshida K, Yanagisawa S, Kawakami M, Aoki T, Yamazaki Y. Angiogenesis in pancreatic carcinoma. Thymidine phosphorylase expression in stromal cells and intratumoral microvessel density as independent predictors of overall and relapse-free survival. Cancer 2001;92:1788 –97. 22. Shweiki D, Itin A, Soffer D, Keshet E. Vascular endothelial growth factor induced by hypoxia may mediate hypoxia-initiated angiogenesis. Nature 1992;359:843–5. 23. Shweiki D, Neeman M, Itin A, Keshet E. Induction of vascular endothelial growth factor expression by hypoxia and by glucose deficiency in multicell spheroids: implications for tumor angiogenesis. Proc Natl Acad Sci USA 1995;92:768 –72. 24. Forsythe JA, Jiang BH, Iyer NV, Agani F, Leung SW, Koos RD, Semenza GL. Activation of vascular endothelial growth factor gene transcription by hypoxia-inducible factor 1. Mol Cell Biol 1996;16: 4604 –13. 25. Akakura N, Kobayashi M, Horiuchi I, Suzuki A, Wang J, Chen J, Niizeki H, Kawamura K, Hosokawa M, Asaka M. Constitutive expression of hypoxia-inducible factor-1␣ renders pancreatic cancer cells resistant to apoptosis induced by hypoxia and nutrient deprivation. Cancer Res 2001;61:6548 –54. 26. Moore PS, Sipos B, Orlandini S, Sorio C, Real FX, Lemoine NR, Gress T, Bassi C, Klo¨ ppel G, Kalthoff H, Ungefroren H, Lo¨ hr M, et al. Genetic profile of 22 pancreatic carcinoma cell lines. Analysis of K-ras, p53, p16 and DPC4/Smad4. Virchows Arch 2001;439:798 – 802. 27. Houck KA, Ferrara N, Winer J, Cachianes G, Li B, Leung DW. The vascular endothelial growth factor family: identification of a fourth molecular species and characterization of alternative splicing of RNA. Mol Endocrinol 1991;5:1806 –14. 28. Korff T, Augustin HG. Integration of endothelial cells in multicellular spheroids prevents apoptosis and induces differentiation. J Cell Biol 1998;143:1341–52. 29. Weidner N. Current pathologic methods for measuring intratumoral microvessel density within breast carcinoma and other solid tumors. Breast Cancer Res Treat 1995;36:169 – 80. 30. Alves F, Borchers U, Padge B, Augustin H, Nebendahl K, Klo¨ ppel G, Tietze LF. Inhibitory effect of a matrix metalloproteinase inhibitor on growth and spread of human pancreatic ductal adenocarcinoma eval-
600
31.
32.
33.
34.
35.
36. 37.
38. 39. 40.
SIPOS ET AL.
uated in an orthotopic severe combined immunodeficient (SCID) mouse model 31. Cancer Lett 2001;165:161–70. Kim KJ, Li B, Houck K, Winer J, Ferrara N. The vascular endothelial growth factor proteins: identification of biologically relevant regions by neutralizing monoclonal antibodies. Growth Factors 1992;7:53– 64. Zhang HT, Scott PA, Morbidelli L, Peak S, Moore J, Turley H, Harris AL, Ziche M, Bicknell R. The 121 amino acid isoform of vascular endothelial growth factor is more strongly tumorigenic than other splice variants in vivo. Br J Cancer 2000;83:63– 8. Gitay-Goren H, Cohen T, Tessler S, Soker S, Gengrinovitch S, Rockwell P, Klagsbrun M, Levi BZ, Neufeld G. Selective binding of VEGF121 to one of the three vascular endothelial growth factor receptors of vascular endothelial cells. J Biol Chem 1996;271:5519 – 23. Soker S, Takashima S, Miao HQ, Neufeld G, Klagsbrun M. Neuropilin-1 is expressed by endothelial and tumor cells as an isoformspecific receptor for vascular endothelial growth factor. Cell 1998;92: 735– 45. Charnock-Jones DS, Sharkey AM, Rajput-Williams J, Burch D, Schofield JP, Fountain SA, Boocock CA, Smith SK. Identification and localization of alternately spliced mRNAs for vascular endothelial growth factor in human uterus and estrogen regulation in endometrial carcinoma cell lines. Biol Reprod 1993;48:1120 – 8. Lei J, Jiang A, Pei D. Identification and characterization of a new splicing variant of vascular endothelial growth factor: VEGF183. Biochim Biophys Acta 1998;1443:400 – 6. Poltorak Z, Cohen T, Sivan R, Kandelis Y, Spira G, Vlodavsky I, Keshet E, Neufeld G. VEGF145, a secreted vascular endothelial growth factor isoform that binds to extracellular matrix. J Biol Chem 1997;272:7151– 8. Auerbach R, Auerbach W, Polakowski I. Assays for angiogenesis: a review. Pharmacol Ther 1991;51:1–11. O’Rourke JF, Dachs GU, Gleadle JM, Maxwell PH, Pugh CW, Stratford IJ, Wood SM, Ratcliffe PJ. Hypoxia response elements. Oncol Res 1997;9:327–32. Maxwell PH, Dachs GU, Gleadle JM, Nicholls LG, Harris AL, Stratford IJ, Hankinson O, Pugh CW, Ratcliffe PJ. Hypoxia-inducible factor-1 modulates gene expression in solid tumors and influences
41.
42.
43.
44.
45. 46.
47.
48.
49. 50.
both angiogenesis and tumor growth. Proc Natl Acad Sci USA 1997; 94:8104 –9. De Jaeger K, Merlo FM, Kavanagh MC, Fyles AW, Hedley D, Hill RP. Heterogeneity of tumor oxygenation: relationship to tumor necrosis, tumor size, and metastasis. Int J Radiat Oncol Biol Phys 1998;42:717–21. Beasley NJ, Wykoff CC, Watson PH, Leek R, Turley H, Gatter K, Pastorek J, Cox GJ, Ratcliffe P, Harris AL. Carbonic anhydrase IX, an endogenous hypoxia marker, expression in head and neck squamous cell carcinoma and its relationship to hypoxia, necrosis, and microvessel density. Cancer Res 2001;61:5262–7. Jenkins WT, Evans SM, Koch CJ. Hypoxia and necrosis in rat 9L glioma and Morris 7777 hepatoma tumors: comparative measurements using EF5 binding and the Eppendorf needle electrode. Int J Radiat Oncol Biol Phys 2000;46:1005–17. Olive PL, Aquino-Parsons C, MacPhail SH, Liao SY, Raleigh JA, Lerman MI, Stanbridge EJ. Carbonic anhydrase 9 as an endogenous marker for hypoxic cells in cervical cancer. Cancer Res 2001;61: 8924 –9. Ryan HE, Poloni M, McNulty W, Elson D, Gassmann M, Arbeit JM, Johnson RS. Hypoxia-inducible factor-1␣ is a positive factor in solid tumor growth. Cancer Res 2000;60:4010 –5. Jouanneau J, Plouet J, Moens G, Thiery JP. FGF-2 and FGF-1 expressed in rat bladder carcinoma cells have similar angiogenic potential but different tumorigenic properties in vivo. Oncogene 1997; 14:671– 6. Di Blasio AM, Cremonesi L, Vigano P, Ferrari M, Gospodarowicz D, Vignali M, Jaffe RB. Basic fibroblast growth factor and its receptor messenger ribonucleic acids are expressed in human ovarian epithelial neoplasms. Am J Obstet Gynecol 1993;169:1517–23. Li Z, Shimada Y, Uchida S, Maeda M, Kawabe A, Mori A, Itami A, Kano M, Watanabe G, Imamura M. TGF-␣ as well as VEGF, PD-ECGF and bFGF contribute to angiogenesis of esophageal squamous cell carcinoma. Int J Oncol 2000;17:453– 60. Leung HY, Gullick WJ, Lemoine NR. Expression and functional activity of fibroblast growth factors and their receptors in human pancreatic cancer. Int J Cancer 1994;59:667–75. Wagner M, Lopez ME, Cahn M, Korc M. Suppression of fibroblast growth factor receptor signaling inhibits pancreatic cancer growth in vitro and in vivo. Gastroenterology 1998;114:798 – 807.
