Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Angiogenesis, the process by which new capillary blood vessels originate from pre-existing vessels, is an absolute requirement for the establishment of a vascular supply in both normal and neoplastic tissues. During angiogenesis, previously quiescent endothelial cells are stimulated to degrade their basement membrane and to invade the surrounding stroma, initially as solid endothelial cell cords. Later, these cords develop a lumen and deposit a new basement membrane, thus resulting in functional new capillaries. Several molecules affect one or more endothelial cell functions involved in these processes. Many are polypeptide growth factors, among which acidic and basic fibroblast growth factors (bFGF) and vascular endothelial growth factor/vascular permeability factor (VEGF/VPF; referred as VEGF in this paper) are the best characterized. FGFs and VEGF are mitogenic for endothelial cells and stimulate endothelial cell migration and pericellular proteolysis (for review, see Gerwins et al., 2000).

VEGF is an endothelial cell-specific mitogen that acts through two tyrosine kinase receptors, VEGFR-1/Flt and VEGFR-2/Flk-1/KDR, whose expression is restricted to endothelial cells, monocytes, and hematopoietic precursors. Disruption of either the VEGFR-1 or the VEGFR-2 gene results in lethal embryonic vascular and/or hematopoietic abnormalities. Similar defects have been described in mice lacking a single copy of the VEGF gene, which indicates a crucial dose-dependent requirement for VEGF during early development (for review, see Ferrara, 2001). VEGF-C, a protein with structural homology to VEGF, was the first VEGFR-3/Flt-4 ligand to be described (for review, see Veikkola et al., 2000). VEGFR-3 is expressed widely in embryonic endothelium, but in postnatal life becomes restricted to lymphatic endothelial cells and some venules (Kaipainen et al., 1995). Based on its expression profile and binding to VEGFR-3, VEGF-C has been implicated in the development of the lymphatic system, and tissue-specific overexpression of VEGF-C in the mouse results in a specific lymphangiogenic effect (Jeltsch et al., 1997; Mandriota et al., 2001). However, the role of VEGF-C/VEGFR-3 is not restricted to lymphangiogenesis. First, targeted inactivation of the VEGFR-3 gene resulted in embryonic lethal vascular defects, before the emergence of lymphatic vessels (Dumont et al., 1998). Second, VEGFR-3 is re-expressed in blood vessels of some tumors (Valtola et al., 1999; Kubo et al., 2000). Third, VEGF-C can induce angiogenesis under certain circumstances (Cao et al., 1998; Witzenbichler et al., 1998). Interestingly, both VEGF and VEGF-C have been reported to synergize with bFGF in the induction of angiogenesis in vitro (Pepper et al., 1992a, 1998).

The observation that VEGFR expression correlates well with tumor angiogenesis points to VEGFR tyrosine kinase activities as potential targets for anti-angiogenic tumor therapies. Because VEGFR-2 is the main signal transducing receptor for VEGF, much effort has gone into the development of VEGFR-2 tyrosine kinase inhibitors. Low molecular weight compounds of different chemical classes (Fong et al., 1999; Wood et al., 2000) have been developed that are potent inhibitors of VEGFR tyrosine kinases. A series of 1-anilino-phthalazines have been synthesized that not only inhibit VEGFR-2 but also VEGFR-1 and VEGFR-3 and that do not affect FGFR tyrosine kinase activity in the same dose range (Bold et al., 2000a). These compounds completely block VEGF activity in in vivo and in vitro models of angiogenesis (Bold et al., 2000a; Wood et al., 2000). They also inhibit tumor angiogenesis, tumor growth, and the formation of metastasis in rodent tumor models (Wood et al., 2000).

In the present studies, we assessed the effects of several 1-anilino-phthalazines on VEGF- and bFGF-induced angiogenesis in vivo and in vitro. Surprisingly, we observed inhibition of angiogenesis induced by either cytokine, despite the fact that 1-anilino-phthalazines specifically block the activity of VEGFR-2 without effecting FGFR-1, -3, or -4. This prompted us to assess whether bFGF-induced in vitro angiogenesis may be dependent on endogenous VEGF and/or VEGF-C. Using a variety of intra- and extracellular antagonists, we were able to demonstrate that bFGF-induced invasion is indeed partially dependent on endogenous VEGF and VEGF-C.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Reagents. The VEGFR-2 inhibitors PTK787/ZK222584, AAC789/ZK202650, and AAD777/ZK202664 come from a series of anilino-phthalazines discovered and synthesized by Novartis Pharma AG (Basel, Switzerland) as part of a joint research collaboration with Schering AG (Berlin, Germany). The discovery, structure, and profile of these compounds is described in detail elsewhere (Bold et al., 2000a, 2000b; Wood et al., 2000; Manley et al., 2001). Recombinant human bFGF (155 amino acid form) was provided by Dr P. Sarmientos (Farmitalia Carlo Erba, Milan, Italy). Recombinant human VEGF (165-amino acid homodimeric isoform) was purchased from Peprotech Inc. (Rocky Hill, NJ). Recombinant human VEGF-C (NC) was provided by Dr. K. Alitalo (Molecular/Cancer Biology Laboratory, Helsinki, Finland). VEGFR-1-IgG (Flt-IgG) and CD4-IgG were provided by Dr. N. Ferrara (Genentech Inc., South San Francisco, CA). VEGFR-3-IgG (Flt-4-IgG) was provided by Dr. W. I. Wood (Genentech). Anti-human VEGFR-2 antibodies (p1C11 and 6.64) and an isotype control for p1C11 (C225) have been described previously (Witte et al., 1998; Zhu et al., 1998).

Cell Culture. Human VEGFR-2/KDR-transfected CHO cells were obtained from the Institute of Molecular Medicine, Tumor Biology Center (Freiburg, Germany). Human umbilical vein endothelial (HUVE) cells (Promo Cell Nr. C-12250, BioConcept AG, Switzerland) were cultivated in endothelial cell growth medium (Promo Cell Nr. C-22110). Bovine adrenal cortex-derived microvascular endothelial (BME) cells (Furie et al., 1984) were grown in minimal essential medium, -modification (MEM; Invitrogen, Carlsbad, CA), supplemented with 5% heat-inactivated donor calf serum (DCS) (Flow Laboratories, Baar, Switzerland), penicillin (110 U/ml), and streptomycin (110 µg/ml). Bovine aortic endothelial (BAE) cells, isolated from scrapings of adult bovine thoracic aortas and cloned by limiting dilution as previously described (Pepper et al., 1992b), were cultured in low glucose Dulbecco's modified minimal essential medium (DMEM) (Invitrogen) supplemented with 10% DCS and antibiotics.

VEGF-Receptor Tyrosine Kinase Assays. In vitro kinase assays were performed in 96-well plates as a filter binding assay using recombinant GST-fused kinase domains expressed in baculovirus and purified over glutathione-Sepharose. [-³³P]ATP was used as the phosphate donor and poly-(Glu:Tyr 4:1) peptide (Sigma, Buchs, Switzerland) was used as the acceptor. Recombinant GST-fusion proteins were diluted in 20 mM Tris-HCl, pH 7.5, containing 1 to 3 mM MnCl₂, 3 to 10 mM Mg Cl₂, 0.25 mg/ml PEG 20000, and 1 mM DTT, according to their specific activity. Each GST-fused kinase was incubated under optimized buffer conditions (20 mM Tris-HCl buffer, pH 7.5, 1-3 mM MnCl₂, 3-10 mM MgCl₂, 3-8 µg/ml poly-(Glu:Tyr 4:1), 0.25 mg/ml PEG 20000, 8 µM ATP, 10 µM sodium vanadate, 1 mM DTT, and 0.2 µCi [-³³P]ATP) in a total volume of 30 µl in the presence or absence of the test substance for 10 min at room temperature. The reaction was stopped by adding 10 µl of 250 mM EDTA. Using a 96-well filter system, half of the volume (20 µl) was transferred onto an Immobilon-PVDF membrane (Millipore, Bedford, MA). The membrane was washed extensively in 0.5% H₃PO₄ and then soaked in ethanol. After drying, Microscint cocktail (Packard, Meriden, CT) was added, and scintillation counting was performed. IC₅₀ values in these and all assays described below were calculated by linear regression analysis of the percentage inhibition.

Kinase Selectivity Assays. To determine the enzyme selectivity profile of the compounds, their effects on the kinase activity of human VEGFR-1, human and mouse VEGFR-2, human VEGFR-3, human PDGFR-, human PDGFR-, human FGFR-1, human FGFR-3, human FGFR-4, and human Tie-2 were measured using the same substrate and procedure as described above for VEGF receptor kinases.

In Vitro Cell Proliferation Assays. HUVE cell proliferation was determined by measuring the incorporation of the pyrimidine analog bromodeoxyuracil (BrdU) into DNA using the Biotrak Cell Proliferation Elisa System V.2 (Amersham Pharmacia Biotech AB, Uppsala, Sweden). Cells were seeded into 1.5% gelatin-coated 96-well plates (5 × 10³ cells per well) and incubated in endothelial cell growth medium containing 5% FCS for 24 h. The medium was replaced with essential basic medium (1.5% FCS), and the cells were incubated for another 24 h. Essential basic medium was then replaced with fresh medium containing either 50 ng/ml VEGF or 0.5 ng/ml bFGF. Inhibitors were added just before addition of growth factors. The cells were incubated for a further 24 h before adding the BrdU labeling solution. Twenty four hours later, the labeling solution was removed, the cells were fixed, and the incorporated BrdU was visualized with a peroxidase-labeled anti-BrdU antibody and TMB substrate, as described by the manufacturer.

VEGFR-2 Phosphorylation Assays. Human VEGFR-2-transfected CHO cells were seeded into 6-well plates and grown to about 80% confluency. Inhibitor was added in serial dilutions and the cells incubated for 2 h at 37°C in medium without FCS. VEGF (20 ng/ml) was then added. After a 5-min incubation at 37°C, the cells were washed twice with ice-cold phosphate-buffered saline and lysed (50 mM Tris-HCl, pH 7.4, 150 mM sodium chloride, 5 mM EDTA, 1 mM EGTA, 1% NP-40, 2 mM sodium ortho-vanadate, 1 mM phenylmethylsulfonyl fluoride, 50 mg/ml aprotinin, and 80 mg/ml leupeptin). Nuclei were removed by centrifugation for 10 min at 4°C. Protein concentrations of the lysates were determined using BSA as a standard. Microtiter plates were coated with a monoclonal antibody to human VEGFR-2/KDR (Mab 1495.12.14, Novartis), which served as a capture antibody. Cell lysates (20 µg of protein per well) were added in triplicate together with PY-20(AP), an alkaline phosphatase-labeled anti-phosphotyrosine antibody (Transduction Laboratories, Lexington, KY). After an overnight incubation at 4°C, the bound PY-20(AP) was detected by chemiluminescence with a luminescent alkaline phosphatase substrate (TROPIX, Bedford, MA).

In Vivo Growth Factor-Induced Angiogenesis Model. Porous Teflon chambers (volume, 0.5 ml) filled with 0.8% w/v agar-containing heparin (20 U/ml) with or without VEGF (2 µg/ml) or bFGF (0.3 µg/ml) were implanted subcutaneously on the dorsal flank of female mice (MAG; Novartis). The mice were treated with compounds (p.o. once daily) or vehicle (5% dimethyl sulfoxide, 1% Tween 80 in water) starting 1 day before implantation of the chamber and continuing for 5 days thereafter. At the end of the treatment period, the mice were killed, and the chambers were removed. The vascularized tissue growing around the chamber was removed carefully and weighed, and the blood content was assessed by measuring hemoglobin levels (Drabkins method; Sigma). The percentage inhibition of the angiogenic response (increase in tissue weight or total blood) was calculated as follows for individual animals receiving the drug treatment and chambers containing growth factor: (A  B)/(C  D) × 100, where A is the weight (or blood volume) of the tissue from a drug-treated mouse with a chamber containing growth factor; B is the mean weight (or blood volume) of the tissue from the group of drug-treated mice with chambers not containing growth factor; C is the mean weight (or blood volume) of the tissue from the group of vehicle-treated mice with chambers containing growth factor; and D is the mean background weight (or blood volume) of the tissue from the group of vehicle-treated mice with chambers not containing growth factor. EC₅₀ values were estimated from the dose response curves (% inhibition versus dose) plotted using a sigmoidal curve fitting program (Origin 6.0; Microcal Software, Inc., Northampton, NA). Each experiment was performed with six animals per dose group and each dose was tested in at least two independent experiments.

In Vitro Angiogenesis Assay. The in vitro angiogenesis assay was performed as described (Montesano and Orci, 1985). BME cells were seeded onto 500-µl three-dimensional rat type I collagen gels in 16-mm tissue culture wells (Nunc), at 0.5 to 1.0 × 10⁵ cells/well in 500 µl of MEM + 5% DCS. Upon reaching confluence (3 days), DCS was reduced to 2%, and the cells were treated with bFGF; VEGF; VEGF-C; neutralizing polyclonal anti-VEGF antibodies (R&D Systems, Minneapolis, MN); anti-trinitrophenol antibody (Pharmingen, San Diego, CA); preimmune rabbit -globulins; soluble VEGFR-1-IgG, VEGFR-3-IgG, or CD4-IgG; anti-VEGFR-2 antibodies; or VEGF-R2 kinase inhibitors. Medium, cytokines, and antagonists were renewed after 2 days, and after a further 2 days cultures were photographed under phase contrast microscopy using a Nikon Diaphot TMD inverted photomicroscope (Nikon, Tokyo, Japan). Quantitation was performed as described (Pepper et al., 1992a). Results are shown as the mean additive sprout length ± S.E.M. (in micrometers) for at least three experiments per condition. Mean values were compared using the Student's unpaired t test, and a significant p value was taken as <0.05. ED₅₀ values were estimated from the dose response curves plotted using a sigmoidal curve fitting program (Origin 6.0).

RNA Purification and RNase Protection Assay. Total cellular RNA was purified using Trizol reagent (Invitrogen). RNase protection assays were as described (Pepper et al., 1993a). [-³²P]dUTP cRNA probes were generated from partial bovine VEGF₁₆₄ (this paper), bovine VEGF-B, bovine VEGF-C, and bovine VEGF-D (M. S. Pepper and S. J. Mandriota, manuscript in preparation) and bovine acidic ribosomal phosphoprotein P0 cDNAs, the latter of which was used as an internal control (Pepper and Mandriota, 1998). Autoradiograms were scanned with a Laser ScanJet IIex Instrument (Hewlett Packard, Palo Alto, CA) and bands were quantitated using ImageQuant 3.3 (Molecular Dynamics, Sunnyvale, CA).

Reverse Transcription-Polymerase Chain Reaction (RT-PCR). Two micrograms of total cellular RNA were combined with random exanucleotides (Boehringer Mannheim, Mannheim, Germany) and Moloney Murine Leukemia Virus reverse transcriptase (Promega, Madison, WI) according to the manufacturer's instructions. VEGF cDNAs were amplified from 1/20 of the volume of the RT product using PrimeZyme DNA polymerase (Biometra, Göttingen, Germany) and a Biometra TRIO-Thermoblock instrument. The primer sequences used were: for VEGF, forward: 5'-CCGGAATTCCAGGAGTACCCAGATGAG; reverse: 5'-CGCGGATCCGGCTCACCGCCTCGGCTTGTC, located in the NH₂-terminal region or in the translational-end region, respectively, of the bovine VEGF cDNA, the former upstream and the latter downstream of the alternative splicing site (Leung et al., 1989). In a parallel PCR reaction, 1/20 of the volume of the same RT products was amplified by a pair of primers (forward: 5'-CCGGAATTCATGCGTTACGTTGCCTCCTAC; reverse: 5'-CGCGGATCCGGCCAGCTTGCCGATACCCTG) located in the coding region of the bovine acidic ribosomal protein P2 (GenBank accession number U17836). Cycles employed for PCR were as follows: for VEGF: 95°C, 3 min (one time); 95°C, 1 min; 62°C, 1 min; 72°C, 1 min (27 times); 72°C, 10 min (one time); for P2: 95°C, 3 min (one time); 95°C, 45 s; 64°C, 45 s; 72°C, 45 s (27 times); 72°C, 10 min (one time). PCR products were electrophoresed in 6% (VEGF) or 8% (P2) polyacrylamide gels and stained with ethidium bromide. For all the semiquantitative RT-PCR reactions, preliminary experiments were performed to determine the number of PCR cycles at which saturation occurred, and the experiments mentioned were performed with a number of cycles that precedes saturation. A ~400-bp RT-PCR product from bovine brain, amplified with the VEGF primer pair described above, and corresponding in size to the expected M_r of the 164-amino acid VEGF isoform (411 bp), was purified, digested with EcoRI and BamHI (which are restriction sites included in the 5' ends of the primers), and subcloned into the corresponding sites of pBluescriptKS() (Stratagene, La Jolla, CA). The insert was sequenced partially on both strands by the chain termination method (Sanger et al., 1977). Analysis of a total of 323 bases revealed 99% identity with the bovine VEGF₁₆₄ cDNA (Leung et al., 1989).

Immunoprecipitation and Western Blotting. Immunoprecipitation and detection of VEGFR-2 phosphorylation were performed as previously described (Pepper and Mandriota, 1998). BME and BAE cell VEGFR-2 was immunoprecipitated from cell lysates with a polyclonal antibody (sc-504; Santa Cruz Biochemicals, Santa Cruz, CA) recognizing amino acids 1158 to 1345 in the mouse VEGFR-2 carboxy terminus. Immunoprecipitates were boiled for 3 min in the presence of 100 mM DTT, run in a 5% SDS polyacrylamide gel, and transferred to a PVDF membrane (Bio-Rad, Hercules, CA). Western blotting analysis was performed with an anti-phosphotyrosine antibody (catalog number 05-321; Upstate Biotechnology, Lake Placid, NY) and a polyclonal anti-VEGFR-2 antibody (sc-315; Santa Cruz Biochemicals) recognizing the mouse carboxyl-terminal amino acids 1348 to 1367. For detection of VEGF and VEGF-C, confluent BME and BAE cell monolayers in 100-mm tissue culture dishes were incubated with bFGF (10 ng/ml) or cytokine-free medium for 8 and 15 h. The cells were lysed in 50 mM Tris-HCl, pH 8.0, 0.5% SDS, 5 mM EDTA, 0.5% NP40, 1 mM phenylmethylsulfonyl fluoride, 200 Kallikrein Inhibitory Units aprotinin, and 1 µg/ml pepstatin A. Cell extracts (100 µg) were electrophoresed in 12% SDS polyacrylamide gels under nonreducing conditions and then transferred to a PVDF membrane for 1 h at 100 V. The membrane was incubated overnight at 4°C in 1% skim milk and 1% BSA in 10 mM Tris-HCl, pH 7.4, 0.15 M NaCl, and 0.1% Tween 20. VEGF or VEGF-C were detected by incubating the membrane with a rabbit polyclonal anti-human VEGF (1 µg/ml; sc-507; Santa Cruz) or anti-human VEGF-C (1 µg/ml; sc-9047; Santa Cruz) antibody for 1 h at room temperature. After incubation with horseradish peroxidase-conjugated goat anti-mouse IgG (diluted 1:10,000; sc-2004; Santa Cruz) for 1 h at room temperature, immune complexes were detected with the enhanced chemiluminescence ECL detection system (Amersham Pharmacia Biotech). Membranes were exposed to audiographic films (Kodak, Rochester, NY). Recombinant human VEGF or normal rabbit serum were used as a control.

Zymography and Reverse Zymography. BME cells were grown to confluence in 35-mm dishes in MEM + 5% DCS. At confluence, medium was removed and the cells were washed twice with phosphate-buffered saline and then incubated with serum-free medium containing 200 Kallikrein Inhibitory Units/ml Trasylol (Bayer-Pharma AG, Zurich, Switzerland), cytokines and inhibitors for 15 h. Supernatants and cells lysates were harvested and analyzed by zymography and reverse zymography as previously described (Pepper et al., 1990).

Migration Assay. Migration assays was performed in a 48-well micro-chemotaxis chamber (AP48; NeuroProbe, Gaithersburg, MD). Eight-micrometer polyvinylpyrrolidine-free polycarbonate filters (NeuroProbe) were coated with collagen type I (100 µg/ml) in 0.2% acetic acid for 30 min at room temperature and air dried. The filter was placed over the bottom chamber containing VEGF (100 ng/ml) or bFGF (10 ng/ml) in DMEM + 0.1% BSA-serum-free medium. BAE cells were suspended in the same medium at 50,000 cells in 50 µl and were added to the upper chamber. For assessment of the inhibitory activity of VEGFR-2 kinase inhibitors, BAE cells were preincubated in DMEM before addition to the upper chamber. The chamber was incubated for 24 h at 37°C. Nonmigrated cells on the upper surface of the filter were removed by scraping, and migrated cells on the lower surface were fixed for 1 min in ethanol and stained for 1 min in 4% Toludine blue. Filters were scanned, and spots were quantitated using Scion Image software. Results are shown as the number of migrating cells ± S.E.M. (as a percentage of controls) for at least three wells per condition.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Characterization of the VEGF Receptor Tyrosine Kinase Inhibitors. All compounds are potent inhibitors of the human VEGFR-2 tyrosine kinase with similar or weaker activity against its mouse counterpart (Table 1). These compounds are also moderately active against VEGFR-1 and VEGFR-3, but have little or no activity against PDGFR- tyrosine kinases, and are all inactive against FGFR-1, FGFR-3, FGFR-4, PDGFR-, and Tie-2 tyrosine kinases at concentrations of up to 10 µM (Table 1).

View larger version (47K): [in this window] [in a new window]   Fig. 1.   Inhibition of VEGFR-2 tyrosine phosphorylation by VEGFR-2 tyrosine kinase inhibitors. BAE cells were incubated with 10 µM PTK787 in the presence or absence of 50 ng/ml VEGF or 10 ng/ml bFGF for 1 h at 4°C and then for 8 min at 37°C. VEGFR-2 was immunoprecipitated from cell lysates. Western blot analysis was performed on immunoprecipitates with an anti-phosphotyrosine antibody (upper panel) and a polyclonal anti-VEGFR-2 antibody (lower panel). Three experiments were performed and similar results were obtained.

VEGFR Tyrosine Kinase Inhibitors Inhibit Angiogenesis in Vivo. The VEGFR-2 tyrosine kinase inhibitors (Table 1) were tested in an in vivo model of angiogenesis. In this model, VEGF and bFGF induce growth of vascularized tissue around a subcutaneous implant. The angiogenic response is quantitated by measuring the weight and blood content of this tissue. All compounds given in daily oral doses for 6 days blocked VEGF-induced angiogenesis in a dose-dependent manner (Table 2; Fig. 2). All of the compounds also inhibited the response to bFGF to some extent, but the dose-response curve was not linear for all compounds (Fig. 2).

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Inhibition of VEGF- and bFGF-induced angiogenesis in vivo by VEGFR-2 tyrosine kinase inhibitors. Porous Teflon chambers (volume, 0.5 ml) filled with 0.8% w/v agar containing 20 U/ml heparin with or without 2 µg/ml VEGF or 0.3 µg/ml bFGF were implanted subcutaneously on the dorsal flank of female mice. The mice were treated with VEGFR-2 tyrosine kinase inhibitors at the indicated dose (p.o. once daily) or vehicle. Five days after implantation, the mice were killed, and the chambers were removed. The vascularized tissue growing around the chamber was removed carefully and weighed, and the blood content assessed by measuring hemoglobin. Results are shown as percent inhibition of increase in blood content around the chamber. Each experiment was performed with six animals per dose group, and each dose was tested in at least two independent experiments.

VEGFR Tyrosine Kinase Inhibitors Inhibit Angiogenesis in Vitro. We have shown previously that VEGF family members and bFGF induce BME and BAE cells to invade three-dimensional collagen gels as capillary-like, tubular structures (in vitro angiogenesis). More precisely, VEGF and bFGF induced invasion of BME cells (which express VEGFR-1 and -2 and FGFR-1), whereas VEGF, VEGF-C, and bFGF induced invasion of BAE cells (which express VEGFR-1, -2, and -3 and FGFR-1) (Pepper et al., 1992a, 1993b, 1998; Mandriota et al., 1996; Pepper and Mandriota, 1998). In addition, bFGF and VEGF induced a synergistic response in BME and BAE cells, VEGF-C potentiated the effect of bFGF in BME cells, and VEGF and VEGF-C had a synergistic effect on BAE cells (Pepper et al., 1992a, 1995, 1998). Using a variety of VEGF mutants that specifically bind VEGFR-1, -2, or -3 and neutralizing anti-VEGFR-2 antibodies, we have observed that VEGF- and VEGF-C (NC) in vitro angiogenic activities are mediated by VEGFR-2 (J.-C. Tille and M. S. Pepper, manuscript in preparation). We therefore investigated the effect of the VEGFR-2 tyrosine kinase inhibitors listed in Table 1 on VEGF-, VEGF-C-, and bFGF- induced BME or BAE cell collagen gel invasion.

View larger version (224K): [in this window] [in a new window]   Fig. 3.   Inhibition of VEGF-induced in vitro angiogenesis by PTK787. Confluent monolayers of BME cells on three-dimensional collagen gels were treated for 4 days with 30 ng/ml VEGF (c) or 10 ng/ml bFGF (e) alone or in combination with 10 µM PTK787 (d and f). The resulting capillary-like tubular structures were viewed by phase-contrast microscopy. In b, in which the plane of focus is at the level of the surface of the collagen gel, cells treated with 10 µM PTK787 show no signs of cytotoxicity (similar results were obtain with the other inhibitors). No invasion occurred in untreated cultures (a). Bar, 75 µm.

View larger version (26K): [in this window] [in a new window]   Fig. 4.   Inhibition of in vitro angiogenesis by VEGFR-2 inhibitors. Confluent monolayers of BME cells on three-dimensional collagen gels were treated with 10 ng/ml bFGF, 30 ng/ml VEGF, or both cytokines in combination, whereas BAE cells were treated with 3 ng/ml bFGF, 30 ng/ml VEGF, or 30 ng/ml VEGF-C. VEGFR-2 inhibitors were added at the concentrations indicated. Invasion was measured after 4 days and is expressed as percent inhibition of sprouting induced by cytokines alone. Results are shown as the means ± S.E.M. from at least three experiments per condition.

Effect of bFGF on Endogenous VEGF and VEGF-C mRNA and Protein Expression in Endothelial Cells. Based on the observation that bFGF increases VEGFR-2 expression at the levels of mRNA and total protein in BME cells (Pepper and Mandriota, 1998), we hypothesized 1) that BME and BAE cells may express VEGF and/or VEGF-C; 2) that bFGF increases VEGF and VEGF-C in these cells; and 3) that the concomitant expression of VEGF, VEGF-C, and VEGFR-2 promotes an autocrine loop of VEGFR-2 activation in endothelial cells.

View larger version (61K): [in this window] [in a new window]   Fig. 5.   Expression of VEGF family members by BME and BAE cells. VEGF, VEGF-B, VEGF-C, and VEGF-D mRNA expression was studied by RNase protection analysis in confluent monolayers of BME and BAE cells. Purified 32P-labeled bovine cRNA probes were hybridized to hybridization mix (probe + h.m.), yeast tRNA (tRNA), or total RNA from BME or BAE cells. Positive control: bovine lung for VEGF-A, VEGF-C, and VEGF-D; bovine heart for VEGF-B. Parallel cultures were incubated for 4 h in the presence of 10 ng/ml bFGF.

View larger version (34K): [in this window] [in a new window]   Fig. 6.   Regulation of VEGF and VEGF-C mRNA by bFGF in BME and BAE cells. A, 2 µg of total RNA from adult bovine kidney or from BME and BAE cells incubated for the times indicated in the presence of 10 ng/ml bFGF (F) or cytokine-free medium (C) were combined with random exanucleotides in the presence of reverse transcriptase (RT). Where indicated, RT was omitted (RT). Samples or an equivalent volume of H2O were subjected to polymerase chain reaction (upper panel). In a parallel PCR reaction consisting an equal volume of the same RT product was amplified with a pair of primers of the bovine acidic ribosomal protein P2. PCR products were electrophoresed polyacrylamide gels and stained with ethidium bromide (lower panel). B and C, confluent monolayers of BME cells were incubated for 4 or 15 h in the presence of 10 ng/ml bFGF (F) or cytokine-free medium (C). At the end of the incubation VEGF-B (B) and VEGF-C (C) mRNA expression was assessed by RNase protection analysis. Purified 32P-labeled bovine cRNA probes (pr.) were hybridized to hybridization mix (pr. + h.m.), yeast tRNA (tRNA), or total RNA from BME cells. Bovine acidic ribosomal protein P0 was used as an internal control.

View larger version (24K): [in this window] [in a new window]   Fig. 7.   bFGF does not alter VEGF and VEGF-C protein levels in BME and BAE cells. A, Western blot analysis of VEGF in confluent BME cell monolayers incubated with 10 ng/ml bFGF or cytokine-free medium for 8 h. At the end of the incubation, cells were processed as described under Materials and Methods. Upper panel, probed with anti-VEGF polyclonal antibody; lower panel, antibody preincubated with rhVEGF165 protein. B, Western blot analysis of VEGF-C in confluent BME and BAE monolayers incubated with 10 ng/ml bFGF or cytokine-free medium for 8 h. At the end of the incubation, cells were processed as described under Materials and Methods. Upper panel, probed with anti-VEGF-C polyclonal antibody; lower panel, incubated with normal rabbit serum. This experiment was repeated three times with similar results.

VEGF and VEGF-C Antagonists Inhibit bFGF-Induced Angiogenesis in Vitro. To assess the possibility that the in vitro angiogenic effect of bFGF may be dependent in part on an autocrine VEGF and/or VEGF-C loop, a chimeric molecule consisting of the extracellular domain of human VEGFR-1 fused to a fragment of the heavy chain of human IgG1 (VEGFR-1-IgG), which is a potent inhibitor of VEGF activity (Aiello et al., 1995; Shima et al., 1995), was tested for its effect on bFGF-induced BME cell collagen gel invasion. VEGFR-1-IgG, but not a CD4-human IgG1 fusion protein, inhibited bFGF-induced in vitro angiogenesis in a dose-dependent manner, with a maximal inhibition of 51% at 10 µg/ml (Figs. 8 and 9A). Similarly, up to 48% inhibition was observed with a neutralizing anti-human VEGF polyclonal antibody, whereas preimmune rabbit -globulins had no effect (Figs. 8 and 9B). No toxicity was observed in VEGFR-1-IgG- or anti-VEGF antibody-treated cells (Fig. 8). VEGF-induced BME cell invasion was inhibited almost completely by both VEGFR-1-IgG and the anti-human VEGF antibody (Fig. 9, A and B). To assess the role of endogenous VEGF-C in bFGF-induced in vitro angiogenesis, a similar chimeric molecule (VEGFR-3-IgG) (Lee et al., 1996) was tested on bFGF-induced BAE cell invasion. VEGFR-3-IgG inhibited bFGF-induced in vitro angiogenesis in a dose-dependent manner, with a maximal inhibition of 50% at 10 µg/ml, whereas control IgGs had no effect (Fig. 9C). No toxicity was observed in VEGFR-3-IgG-treated cells (data not shown). VEGFR-1-IgG did not alter significantly VEGF-C-induced invasion, and or VEGFR-3-IgG had no effect on invasion induced by VEGF, thereby demonstrating that VEGF-C- induced angiogenesis is not dependent on endogenous VEGF, and vice versa (data not shown).

View larger version (229K): [in this window] [in a new window]   Fig. 8.   Partial inhibition of bFGF-induced in vitro angiogenesis by Flt-IgG and neutralizing anti-VEGF antibodies. Confluent monolayers of BME cells on three-dimensional collagen gels were treated for 4 days with 10 ng/ml bFGF (a) or in combination with 10 µg/ml VEGFR-1-IgG (c and d), 30 µg/ml neutralizing anti-human VEGF antibodies (f), 10 µg/ml CD4-IgG (b), or 30 µg/ml preimmune -globulins (e). The resulting capillary-like tubular structures were viewed by phase-contrast microscopy. In c, the plane of focus is at the level of the surface of the collagen gel; cells treated with 10 µg/ml VEGFR-1-IgG show no sign of cytotoxicity. d, the same field photographed beneath the surface monolayer. Arrows indicate refringent lumina within endothelial cell cords that have invaded the collagen gel. No invasion occurred in untreated cultures (data not shown). Bar, 100 µm.

View larger version (32K): [in this window] [in a new window]   Fig. 9.   Partial inhibition of bFGF-induced in vitro angiogenesis by Flt-IgG, neutralizing anti-VEGF antibodies, and Flt4-IgG. Confluent monolayers of BME cells on three-dimensional collagen gels were treated with 10 ng/ml bFGF alone or in combination with antagonists for 4 days, and invasion, expressed as total additive sprout length (in micrometers), was quantitated as described under Materials and Methods. A, VEGFR-1-IgG (VR1-IgG) was added to VEGF- or bFGF-treated BME cells. VEGFR-1-IgG (1 µg/ml) inhibited VEGF-induced invasion by 96% and when added at 10 µg/ml, inhibited bFGF-induced invasion by 51%; CD4-IgG at 10 µg/ml had no effect on invasion. B, neutralizing anti-VEGF antibodies were added to VEGF- or bFGF-treated BME cells. The antibodies (30 µg/ml) inhibited VEGF-induced invasion by 98% and bFGF-induced invasion by 48%; preimmune -globulins (NRG) had no effect on invasion at the same concentration. C, when added to 30 ng/ml VEGF-C-treated BAE cells,10 µg/ml VEGFR-3-IgG (VR3-IgG) inhibited invasion by 98%; VEGFR-3-IgG was added to bFGF-treated BME cells at the concentrations indicated, and inhibited bFGF-induced invasion by 50% at 10 µg/ml; control IgGs (anti-TNP) had no effect on invasion at the same concentration. Results are shown as mean total additive sprout length in micrometers ± S.E.M. from at least three separate experiments per condition (three wells per experiment). *, p < 0.05; **, p < 0.01; ***, p < 0.001 (Student's unpaired t test) when compared with invasion induced by bFGF or VEGF alone.

View larger version (28K): [in this window] [in a new window]   Fig. 10.   Partial inhibition of bFGF-induced in vitro angiogenesis by blocking anti-VEGFR-2 antibodies. Confluent monolayers of BME and BAE cells on three-dimensional collagen gels were treated with VEGF (BME at 100 ng/ml and BAE at 50 ng/ml) or bFGF (BME at 10 ng/ml BAE at 3 ng/ml) alone or in combination with neutralizing anti-VEGFR-2 antibodies for 4 days. A, anti-human VEGFR-2 antibodies p1C11 and 6.64 were added to VEGF-treated BME and BAE cells. p1C11 and 6.64 (20 µg/ml) inhibited VEGF-induced BME and BAE cell invasion by 99%. B, neutralizing anti-VEGFR-2 antibodies were added to bFGF-treated BME cells. p1C11 and 6.64 (20 µg/ml) inhibited bFGF-induced invasion by 48 and 49%, whereas the isotope control antibody (C225) at the same concentration had no effect on invasion. C, when added to bFGF-treated BAE cells, p1C11 and 6.64 (20 µg/ml) inhibited invasion by 23 and 37%, respectively; the isotype control antibody (C225) had no effect on invasion at the same concentration. Results are shown as mean total additive sprout length in micrometers ± S.E.M. from at least three separate experiments per condition (three wells per experiment). *, p < 0.05; **, p < 0.01; ***, p < 0.005, ****, p < 0.001 (Student's unpaired t test) when compared with invasion induced by bFGF alone.

The Role of VEGFR-2 Signaling in bFGF-Induced Endothelial Cell Activation. Angiogenesis is dependent on a triad of endothelial cell proliferation, migration, and extracellular proteolytic activity, all of which are increased by VEGF and bFGF (for review, see Pepper et al., 1996a, 1996b). To assess the effect of VEGFR-2 tyrosine kinase inhibitors on endothelial cell proliferation, low density cultures of VEGF- and bFGF-treated BME cells were treated with the inhibitors for 5 days. Cell number was increased by 53% in VEGF- and 280% in bFGF-stimulated cultures (Fig. 11A). In VEGF-treated cultures, addition of the VEGFR-2 inhibitor AAC789 reduced BME cell number to baseline levels from 1 µM (Fig. 11A). Likewise, bFGF-induced BME cell proliferation was reduced markedly by AAC789 from 1 to 10 µM, without however reaching basal levels (Fig. 11A). Similar results were obtained with the other VEGFR-2 tyrosine kinase inhibitors described in this paper (data not shown).

View larger version (28K): [in this window] [in a new window]   Fig. 11.   Partial inhibition of bFGF-induced endothelial cell proliferation and PA activity but not migration by VEGFR-2 tyrosine kinase inhibitor. A, BME cells were induced to proliferate with 100 ng/ml VEGF or 3 ng/ml bFGF in the presence or absence of the indicated concentrations of tyrosine kinase inhibitors, and cell number was determined after 5 days. This experiment was repeated twice in duplicate wells. Results are shown as the mean cell number ± S.E.M. B, BME cells were incubated with indicated concentrations of PTK787 for 15 h, alone or in the presence of 30 ng/ml VEGF or 10 ng/ml bFGF. At the end of the incubation, cell extracts (CE) and culture supernatants (SN) were subjected to zymography. Two experiments were performed and similar results were obtained. C, BAE cells were induced to migrate with 100 ng/ml VEGF or 10 ng/ml bFGF with the indicated concentration of AAD777 for 24 h. Migrated cells were quantitated and results are shown as migrated cell ± S.E.M. (percentage of control) for at least of three wells per conditions. *, p < 0.001 (Student's unpaired t test) when compared with invasion induced by VEGF alone.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In agreement with previous studies (Bold et al., 2000a; Wood et al., 2000), we demonstrate in this report that VEGFR-2 tyrosine kinase inhibitors block angiogenesis induced by VEGF in vivo and in vitro. These inhibitors also blocked VEGF-induced endothelial cell proliferation, protease (uPA and tPA) activity, and migration. Surprisingly however, we found that these compounds also inhibited bFGF-induced angiogenesis and endothelial cell proliferation in the same models.

Because the VEGFR-2 tyrosine kinase inhibitors lack activity against FGFRs, we explored the possibility that they may act by interrupting an autocrine VEGF loop in cultured endothelial cells (Seghezzi et al., 1998). To this end, we used a well-characterized model of in vitro angiogenesis, in which quiescent endothelial cells can be stimulated to invade a three-dimensional collagen gel within which they form capillary-like tubes (Montesano and Orci, 1985). We observed that two different VEGF antagonists, namely VEGFR-1-IgG (Aiello et al., 1995; Shima et al., 1995) and a neutralizing polyclonal anti-VEGF antibody, but not CD4-IgG or preimmune IgG, inhibited bFGF-induced angiogenesis by approximately 50%, and that this occurred in the absence of overt cytotoxicity. VEGFR-3-IgG also inhibited bFGF-stimulated BME cell collagen gel invasion by approximately 50%. Furthermore, when blocking antibodies to VEGFR-2 were added to the system, bFGF-induced invasion was inhibited in a similar manner. These observations suggested that endogenous VEGF and VEGF-C, acting in an autocrine manner, are required for the in vitro angiogenic effect induced by exogenously added bFGF.

We next assessed whether our cells express endogenous VEGF and VEGF-C, and if so, whether expression could be modulated by exogenously added bFGF. We found that BME and BAE cells express VEGF, VEGF-B, and VEGF-C, but not VEGF-D. Exogenously added bFGF induced a modest increase in VEGF and VEGF-C mRNA at early (4 h) but not at late (15 h) time points. Despite this increase, levels of VEGF and VEGF-C protein were unaffected. Exogenously added bFGF did not increase VEGF-B mRNA nor did it induce expression of VEGF-D. Taken together with our observations on in vitro angiogenesis, these findings strongly suggest that collagen gel invasion induced by exogenously added bFGF is dependent on endogenous VEGF and VEGF-C. In the present study we also considered the possibility that bFGF may induce trans-phosphorylation of VEGFR-2. No transphosphorylation was observed. However, we have demonstrated previously that bFGF strongly increases VEGFR-2 mRNA and protein levels in BME and BAE cells (Pepper and Mandriota, 1998). This may imply that even though levels of endogenous VEGF and VEGF-C are unaltered, signaling via VEGFR-2 is increased in the presence of exogenous bFGF as a consequence of an increase in VEGFR-2.

In contrast to the approximately 50% inhibition of bFGF-induced invasion observed with VEGFR-1-IgG, neutralizing anti-VEGF antibodies, and VEGFR-s blocking antibodies, VEGFR-2 tyrosine kinase inhibitors decreased bFGF-induced angiogenesis by approximately 90%. This difference may reflect partial ligand-independent VEGFR-2 tyrosine kinase activity in response to bFGF, because bFGF strongly increases VEGFR-2 in our cells (Pepper and Mandriota, 1998). This possibility is not unprecedented, because many tyrosine kinase receptors are capable of ligand-independent dimerization and phosphorylation. Furthermore, the stability of the VEGFR-2 tyrosine kinase inhibitors is greater that VEGFR-1-IgG and neutralizing anti-VEGF antibodies, which implies that prolonged activity of the tyrosine kinase inhibitors can be expected once they have entered cells. Finally, the partial inhibition seen with Flt-IgG and neutralizing anti-VEGF antibodies may simply reflect invasion, which occurs as a consequence of the direct effect of bFGF on FGFR tyrosine kinases and which cannot be inhibited by these antagonists. We have no explanation for the observation that the IC₅₀ values were greater in BAE than in BME cells.

The finding that endogenous endothelial cell VEGF and VEGF-C account in part for the angiogenic effect of bFGF implies that bFGF-induced angiogenesis per se is the consequence of a "synergism" between these cytokines. In the same way, because cultured endothelial cells produce their own bFGF, angiogenesis induced by VEGF or VEGF-C could also be considered as a form of "synergism." In other words, although bFGF or VEGF is angiogenic in vitro, both require the presence of the other cytokine to mediate their effect. This hypothesis is supported by the observation that inhibitory antibodies to bovine bFGF completely prevent VEGF-induced in vitro angiogenesis (Mandriota and Pepper, 1997). Although it has been reported previously that bFGF stimulates VEGF production in BAE cells and bovine capillary endothelial cells (Seghezzi et al., 1998), in our experiments, bFGF did not increase VEGF or VEGF-C protein in BME or BAE cells. The reason for these differences are not known.

The finding that bFGF-induced invasion is dependent on endogenous VEGF and VEGF-C and that bFGF increases VEGFR-2, may explain, at least in part, why bFGF and VEGF synergize in the induction of angiogenesis in vitro, and why VEGF-C potentiates the effect of bFGF (Pepper et al., 1992a, 1998; Goto et al., 1993). Thus, when bFGF and VEGF or VEGF-C are tested in combination, the increase in VEGFR-2 expression and its concomitant occupancy by both endogenous and exogenous VEGF or VEGF-C may stimulate angiogenesis to a greater extent than when either cytokine is tested alone. This model also explains why at a constant dose of bFGF, synergism is dependent on the amount of exogenously added VEGF or VEGF-C, and vice versa (Pepper et al., 1992a, 1998). In the former case, the endogenous autocrine loop may sequester a constant number of VEGFR-2 molecules, thus allowing the remaining receptors, which are also constant in number, to be activated proportionately to the amount of exogenous VEGF or VEGF-C added; in the latter case, at a constant dose of VEGF or VEGF-C, increasing concentrations of bFGF would increase VEGFR-2 in a dose-dependent manner, thus increasing VEGFR-2-mediated signaling relative to the amount of bFGF added.

We have observed that the IC₅₀ values required to block VEGF-induced HUVE cell proliferation are lower than those required for inhibition of kinase activity in the cell-free system. It should be pointed out that the kinase assays are performed under very artificial conditions using only the kinase domain of the receptor and nonphysiological concentrations of ATP. The cellular assays represent more physiological conditions of the receptor and may explain the difference in potency between the isolated kinase assays and the cellular assays. Moreover, we and other investigators have observed that many of these compounds are taken up avidly into cells and are not readily washed out again (Mendel et al., 2000). Hence the intracellular concentration may well be higher than that in the medium. Finally, we cannot exclude the possibility that the differences in potency between the cell-based versus kinase assays reflect the fact that the inhibitors are acting via additional mechanisms in cell culture.

In summary, the findings that endothelial cells express VEGF and VEGF-C and that the activity of exogenously added bFGF is partially dependent on the activity of autocrine VEGF and VEGF-C may explain why VEGFR-2 antagonists have the capacity to inhibit partially bFGF-induced angiogenesis in vivo and in vitro. The potential in vivo inhibitory effect of VEGFR-2 kinase inhibitors on a paracrine VEGF and/or VEGF-C loop between fibroblasts/smooth muscle cells and endothelial cells was not explored in our studies. Nonetheless, we hypothesize that in vivo, bFGF-induced angiogenesis may be mediated, in part, either by autocrine or paracrine VEGF, the former being released by endothelial cells, and the latter by fibroblasts or smooth muscle cells. This implies that the therapeutic applications of VEGFR-2 tyrosine kinase inhibitors are not limited to settings in which angiogenesis is induced by VEGF but can also be widened to settings in which bFGF is the primary or costimulus.