Human melanoma cells express two prominent angiogenic factors, e.g., vascular endothelial growth factor (VEGF) and basic fibroblast growth factor (bFGF/fibroblast growth factor-2). In this study, we report on the relative contribution of these two factors to in vitro and in vivo growth of a tumorigenic melanoma cell line (WM164) and nontumorigenic, immortalized melanocytes (FM516SV). Overexpression of either cytokine significantly boosted tumorigenicity of WM164 cells in immunodeficient SCID mice. Attempting to overexpress bFGF antisense sequences produced no viable clones confirming earlier reports that autocrine bFGF is obligatory to melanoma cell survival and growth. By contrast, down-regulation of endogenous VEGF production did not affect growth of WM164 cells in vitro. In vivo expansion of WM164 cells expressing VEGF antisense was delayed but not abrogated. Forced expression of either bFGF or VEGF in immortalized but nontumorigenic melanocytes did not induce sustained tumor growth in vivo highlighting that neither of the two factors is sufficient for induction of tumorigenicity in this model system. Overexpression of either cytokine in WM164 cells led to the development of atypical large vessels but not to an increase in microvessel density. Taken together our results confirm an essential autocrine role of bFGF in human melanoma and indicate a beneficial but nonessential role of VEGF in the tumorigenic phenotype of human melanoma cells.

Human melanoma cells produce a plethora of angiogenic cytokines including VEGF3-A (1), bFGF/FGF2 (2, 3) placenta-derived growth factor; (4), platelet-derived growth factor (5), transforming growth factor α, and transforming growth factor β (6). In addition, melanoma cells produce extracellular matrix components and metalloproteases, which are relevant to angiogenesis, including tenascin (7), components of the plasminogen activator system (8, 9, 10), gelatinases (11), and CD44 (12, 13).

The multitude of angiogenic factors produced by this tumor type raises the question of whether any of them is rate-limiting in supporting melanoma-derived vessel formation and/or tumorigenicity in vivo. In this study, we focused on the relative contribution of two potent melanoma-derived angiogenic cytokines, VEGF and bFGF, to these processes. Selection of these two factors was guided by their documented importance to the angiogenic process in other experimental systems and the observation that both are frequently overexpressed in transformed as compared with normal melanocytes. Specifically, VEGF is perceived to be one of the most important mediators of tumor-associated angiogenesis (reviewed in Ref. 14). Initially named vascular permeability factor, it was renamed VEGF to highlight its mitogenic effects on endothelial cells. Recent work demonstrated that cultured human melanoma cells frequently and constitutively overexpress VEGF when compared with normal neonatal melanocytes (1, 15, 16, 17). Like VEGF, bFGF was initially identified as a growth factor for endothelial cells (18). In endothelial cells, bFGF has been shown to induce VEGF expression (19), and bFGF and VEGF synergize in inducing endothelial cell growth (20, 21, 22, 23). bFGF also exerts essential autocrine functions in supporting melanoma cell growth in vitro and in vivo(2, 3, 24, 25). Melanoma cells in vitro and in situ frequently coexpress bFGF and VEGF (17).

These findings prompted us to investigate the relative importance of tumor-derived VEGF and bFGF to the tumorigenic/angiogenic phenotype in a human melanoma cell line (WM164) that coexpresses both growth factors. In addition, we investigated whether forced expression of either VEGF or bFGF could induce tumorigenicity in an immortalized but nontumorigenic human melanocyte line.

Origin and Culture of Human Cells.

Normal human melanocytes were obtained from neonatal foreskins. WM164 melanoma cells originated from a metastatic lesion and were established in the laboratory of Dr. M. Herlyn at the Wistar Institute, Philadelphia, PA. FM516 melanocytes expressing the SV40 T gene (FM516SV) were described previously (26). Both cell lines were free of contaminating normal cells such as endothelial cells, immune/inflammatory cells, or keratinocytes.

Vector Construction and Transfection of Melanocytic Cells.

Full-length VEGF cDNA encoding VEGF165 was amplified by reverse transcription-PCR from a melanoma-derived RNA preparation (WM164) using VEGF sense and antisense primers described earlier (1). The VEGF primers correspond to sequences in the untranslated 5′ and 3′ regions of the VEGF gene resulting in amplification of four splice variants of 516, 648, 720, and 771 bp. The PCR product corresponding to VEGF165 was recovered from agarose gels and cloned into the pCRII vector using a TA cloning kit (Invitrogen BV, Holland). After verification of VEGF sequences by direct sequencing, the purified VEGF 165 cDNA fragment was subcloned in sense and antisense direction into the polylinker region of pCB6+ (27) containing the cytomegalovirus promoter/enhancer to drive expression of the transgene and a neomycin selection cassette driven by the SV40 promoter. bFGF cDNA encoding the Mr 17,800 form of bFGF (kindly provided by A. Sommer, Synergen Inc. Boulder, CO) was first subcloned into the pCRII vector and then subcloned in sense and antisense direction into the polylinker region of pCB6+. WM164 melanoma cells and FM516V melanocytes were transfected as described previously (28) either with the empty pCB6+ vector for control purposes or the bFGF and VEGF sense and antisense vector constructs. The cells were passaged 48 h after transfection and maintained in the presence of 800 μg/ml G418 for selection. Individual colonies as well as pooled colonies were expanded and tested for transgene expression as described below.

RNA Isolation and Northern Blot Analysis.

Total cellular RNA was prepared from cultured cells or melanocytic tumors recovered from inoculated mice using the Rneasy RNA kit (Quiagen, Hilden, Germany). Total cellular RNA (6 μg/lane) was separated on formaldehyde-containing agarose gels and transferred to nylon membranes. VEGF and bFGF RNA sense and antisense probes were prepared by in vitro transcription from the VEGFpCRII or bFGFpCRII vector by use of the SP6 and SP7 DNA polymerases, respectively. Probes were labeled with digoxigenin using a commercially available kit (Boehringer Mannheim, Mannheim, Germany) and used for hybridization analyses according to instructions provided by the manufacturer.

Preparation of CM.

Melamoma cells and FM516SV cells were maintained in W489 base medium consisting of four parts MCDB153 and one part L15 supplemented with 2% FCS and insulin (5 μg/ml). To prepare CM, cells were plated at 3 × 105/well in six-well 35-mm culture plates in 3 ml of W489 supplemented with 0.2% bovine serum albumine, fatty acid free. After 3 days supernatants were collected and centrifuged at 500 × g to remove cellular debris. Cell numbers were assessed after collection of CM using a Coulter Counter.

Determination of in Vitro VEGF, bFGF, and TSP-1 Protein Production.

Secretion of VEGF into tissue culture supernatants was measured after preparation of CM using a VEGF ELISA (R&D Systems, Rüsselsheim, Germany) according to the manufacturer’s instructions. bFGF production was measured in cell extracts using a bFGF ELISA (R&D Systems) according to the manufacturer’s instructions. For bFGF Western blot analysis, cells were grown in complete medium for 72 h and then lysed in radioimmunoprecipitation assay buffer (PBS, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, Aprotinin 30 μl/ml, sodium orthovanadate 10 μ/ml, 10 mg/ml phenylmethansulfonylfluoride). Samples in Laemmli loading buffer [1 ml of glycerol, 0.5 ml of β-mercaptoethanol, 3 ml of 10% SDS, 1.25 ml of 1 m Tris-HCl (pH 6.7), and 1–2 mg of bromophenol blue] were separated on a 15% SDS-polyacrylamide gel and transferred to a nitrocellulose membrane. Membranes were probed with an anti-bFGF antibody (Santa Cruz Biotechnology, Heidelberg, Germany). After incubation with a secondary goat antirabbit antibody linked to phosphatase (Sigma Chemical Co., Taufkirchen, Germany) blots were developed using the nitroblue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate chromogen system according to manufacturer’s instructions (Promega, Mannheim, Germany). For TSP-1 Western blot analysis cells were grown and processed as described above. Samples were separated on a 7.5% SDS-polyacrylamide gel and transferred to a nitrocellulose membrane. Membranes were probed with an anti-TSP-1 mAb (Calbiochem-Nova Biochem GmbH, Bad Soden, Germany). After incubation with a secondary antimouse antibody blots were developed using the nitroblue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate system as described above.

Melanoma Growth Assays.

To determine in vitro growth, melanoma cells were seeded at 0.5 × 105/well (24-well plates) in W489 base medium (total volume 0.5 ml) supplemented with 0.2% BSA. Cell growth was assayed after 3 and 7 or 10 days by counting the number of viable cells using a Coulter counter.

Tumorigenicity Experiments in SCID Mice.

Tumorigenicity of cells in which bFGF or VEGF production was altered by expression of sense or antisense mRNA was determined in SCID mice. CB.17 SCID mice (4–6 weeks of age) were inoculated s.c. with transfectants of WM164 tumor cells (5 × 106 cells) or transfectants of FM516SV cells (1 × 107 cells). Each experimental group consisted of five mice each. Mock-transfected WM164 and FM516SV cells were used as controls. Tumor sizes were measured with calipers at regular time intervals up to 3 months after inoculation. Tumors were removed and snap-frozen in liquid nitrogen at different time points as specified.

Immunohistochemistry.

Immunohistochemistry was performed in addition to conventional staining (H&E, elastica van Gieson, and periodic acid Schiff) using paraffin-embedded tumor tissue. Sections (4 μm) were deparaffinized and rehydrated using xylene and decreasing ethanol concentrations. The slides were then incubated for 20 min in 1.8% hydrogen peroxide, washed in PBS, and blocked with normal serum (Zymed Laboratories Inc., San Francisco, CA). Primary antibodies used were: anti-CD31 (Clone JC/70A; Dako, Hamburg, Germany) 1:20 dilution; anti-VEGF (Santa Cruz Biotechnology, Heidelberg, Germany) 1:100 dilution; anti-bFGF (Santa Cruz Biotechnology) 1:15 dilution; and anti-Melan A (Novocastra, Newcastle, United Kingdom) 1:20 dilution. Visualization of antibody binding using a biotinylated secondary antibody and the avidin-biotin complex method was according to the manufacturer’s instructions (Zymed Laboratories Inc.). After immunohistochemistry the sections were examined at low magnification (×100) to identify areas with high vessel density. In these areas the number of microvessels in 10 high power fields (×400) was counted.

In Vitro Angiogenesis Assay.

To assess in vitro angiogenesis we used an in vitro angiogenesis assay (Chemicon International, Temecula, CA) as described by the manufacturer. WM164M, WM164Vs, WM164Vas2, WM164Fs, and HUVECs (CellSystems Biotechnologie Vertrieb GmbH, St. Katharinen, Germany) were seeded at 7.5 × 103 cells/96 well on ECMatrix (Chemicon International), prepared from Engelbreth Holm-Swarm mouse tumor, and grown in 200 μl of endothelial growth medium (Chemicon International) in a CO2 incubator. Tube formation was monitored at different time points from 6 to 72 h. Photographs were taken after 36 and 72 h at low magnification (×100) without additional staining using an inverted light microscope.

Characterization of Cell Lines Transfected with VEGF and bFGF Constructs.

We chose two cell lines representing different stages of melanoma development to evaluate cytokine effects on tumorigenicity and angiogenesis. WM164 cells were derived from a metastatic lesion, proliferate in serum-free medium, grow anchorage-independently in semisolid medium, and are tumorigenic when inoculated into immunodeficient mice (8). FM516SV cells were derived from a normal neonatal melanocyte strain by transfection with a plasmid encoding the SV40T antigen. These cells are postcrisis and immortalized but do not form tumors when inoculated into nude or SCID mice (26). On the basis of these criteria, we consider them to represent an intermediate stage of melanoma development. Both cell lines were stably transfected with plasmids encoding either full-length bFGF or VEGF165 cDNA. In addition, WM164 cells were transfected with bFGF and VEGF sequences in antisense orientation. Expression of transgenes was confirmed by Northern blot analyses using appropriate sense and antisense riboprobes as shown in Fig. 1, A and B. Both VEGF sense-transfected WM164 and FM516SV cells secreted ∼100-fold higher VEGF levels when compared with mock-transfected cells (Fig. 2,A). Conversely, WM164 cells transfected with antisense VEGF secreted significantly less VEGF (2–4% of mock-transfected controls) depending on the clone (Vas2 versus Vas4) investigated. Similarly, transfection of bFGF sense cDNA into WM164 and FM516SV cells led to robust expression both on the mRNA (Fig. 1,B) and the protein levels (Fig. 3). By contrast, forced expression of bFGF antisense sequences produced no viable clones consistent with the critical role of bFGF in melanoma cell growth described previously (2, 3).

Earlier studies have demonstrated that bFGF can induce VEGF production in glioma cells (29). Thus, we determined whether WM164 or FM516SV cells overexpressing bFGF secreted VEGF at higher levels than mock-transfected cells. As shown in Fig. 2,B, forced expression of bFGF did not increase VEGF secretion in these cells. For WM164 cells overexpression of bFGF even led to a slight decrease of VEGF production. Similarly, overexpression of VEGF did not affect bFGF expression in WM164 and FM516SV cells as determined by Western blot (Fig. 3). Because constitutive bFGF protein levels in melanoma cells were barely detectable by Western blot analyses (Fig. 3) we also used a bFGF ELISA to test for potential changes in bFGF production that could not be detected by Western blotting. The bFGF ELISA was performed on cell extracts because of the fact that bFGF does not have a signal sequence for secretion. As shown in Fig. 2 C, bFGF protein levels in VEGF sense or antisense mRNA-transfected cells were comparable with those of mock-transfected cells.

To additionally investigate functional aspects of melanoma-derived VEGF, we selected clone WM164-Vas2 based on the dramatic reduction (>98%) of endogenous VEGF production observed for this clone. In the case of VEGF- and bFGF-overexpressing cells we used pools of transfected cells for additional analysis.

In Vitro Growth and Phenotype of Transfectants.

Both bFGF (2, 3) and VEGF (30, 31) have been reported previously to support melanoma growth in an autocrine manner. To account for possible autocrine effects of these factors we first determined growth characteristics of all of the transfected cells in serum-free medium (Fig. 4). No significant differences in cell growth were observed between mock-transfected and bFGF- or VEGF-overexpressing cells. However, it should be noted that bFGF-overexpressing FM516SV cells accumulated slightly but consistently higher cell numbers than controls during the observation period of 10 days. Similarly, down-regulation of VEGF expression in WM164-Vas2 cells had no discernible effect on in vitro cell growth. (Fig. 4; Ref. 1) These results indicate that endogenous VEGF has no obvious autocrine role as it relates to melanoma proliferation in vitro. None of the transfected cells demonstrated changes in cell morphology, shape, or attachment as observed by light microscopy (results not shown).

In Vivo Tumorigenicity of Transfectants.

Next, we determined tumorigenic properties in athymic mice of VEGF and bFGF overexpressing WM164 and FM516SV cells and of WM164 cells in which VEGF production was ablated. As shown in Fig. 5,A, forced expression of bFGF significantly accelerated tumor development of WM164 cells inoculated into SCID mice. Overexpression of VEGF had an even more dramatic effect on tumorigenicity of WM164 cells leading to rapid development of large tumors at the inoculation site (Fig. 5,B). At autopsy, the internal organs of tumor-bearing mice were scanned by inspection and palpation for the presence of tumor nodules; no evidence for macroscopic metastases was found in any of the transfectants investigated. Expression of the VEGF antisense transgene in WM164-Vas2 cells resulted in significant delay in tumor formation relative to mock-transfected cells (Fig. 5 B). For ∼4 weeks after inoculation no tumor masses were palpable. However, ∼50 days after inoculation palpable tumor masses appeared in all of the mice inoculated and subsequently expanded at rates comparable with those observed for mock-transfected WM164 cells.

Overexpression of VEGF in FM516SV was not associated with tumor formation during the observation time of 3 months. By contrast, FM516SV cells overexpressing bFGF revealed transient and reproducible tumorigenicity within 4–10 days after inoculation (Fig. 6). However, after 8 days, tumor volumes decreased progressively, and residual small palpable masses disappeared gradually over a time period of 2 weeks. Additional observation did not reveal any recurrence of tumors within 3 months after inoculation. Results described below on FM516SV tumors relate to tumors excised at day 7, i.e., at the peak of tumor formation.

Transgene Expression in Vivo.

Expression of transgenes in tumors was monitored by in situ immunohistochemical analyses using antibodies to bFGF and VEGF, respectively. Results on staining for CD31 are presented in the “tumor vascularity in vivo” section. Immunohistochemistry for VEGF demonstrated moderate staining of mock-transfected WM164 cells for VEGF consistent with the expression pattern observed in vitro (Fig. 7,a). bFGF staining in tumors from WM164 mock-transfected cells was barely detectable (Fig. 7,b), which is consistent with low bFGF expression as determined by Western blot and ELISA analysis (Figs. 2,c and 3). As expected, in tumors formed by VEGF-overexpressing clones, the cells reacted strongly with a VEGF specific antibody (Fig. 7,d). bFGF expression in VEGF-overexpressing tumors was not detectably different from that observed in mock-transfected cells (Fig. 7, b and e). Importantly, WM164-Vas2 tumors did not react with VEGF antibodies demonstrating that these tumors had not regained expression of VEGF (Fig. 7,g). Furthermore, bFGF expression in 164-Vas2 tumors was comparable with controls in SCID mice indicating that delayed tumor growth was not contributed to up-regulated bFGF expression (Fig. 7, b and h). VEGF expression in bFGF-overexpressing WM164-Fs tumors was comparable with that observed in mock-transfected cells (Fig. 7, a and j). Thus, no evidence of in vivo cross-induction between melanoma-derived VEGF and bFGF was found. As expected, tumors formed by bFGF-overexpressing cells (WM164-Fs) revealed intense bFGF immunoreactivity (Fig. 7 k).

Expression of VEGF sense and antisense transgenes in tumor tissues was additionally confirmed by Northern blot analysis using strand-specific probes (Fig. 8). We observed robust expression of VEGF antisense sequences in WM164-Vas2 tumors consistent with the immunostaining results shown in Fig. 7 g.

Tumors developing after inoculation of FM516SV bFGF-overexpressing cells were removed on day 7, because spontaneous regression occurred beyond this time point as shown in Fig. 6. The majority of cells in these tumors stained strongly with an antibody to Melan A (Fig. 9,b), a marker for human melanocytic cells. They also expressed high levels of bFGF (Fig. 9,d) and moderate levels of VEGF comparable with the in vitro data (Fig. 9 c).

Tumor Vascularity in Vivo.

To assess the effects of VEGF and bFGF overexpression on tumor-associated neovascularization we stained intratumoral vessels with CD31/PECAM-1-specific antibodies. Representative results of this analysis are shown in Figs. 7 and 9, and quantitative results of the microvessel density analysis are given in Fig. 1. VEGF overexpression in WM164 was associated with the formation of large vessels lined with CD31+ cells, which dominated the appearance of the tumors (Fig. 7,f). To a lesser extent large lacunar vessels were also observed in tumors obtained by inoculation with bFGF overexpressing WM164 cells (Fig. 7,l). These lacunar vessels caused a net increase in vascular surface area in these tumors when compared with mock-transfected controls similar to results recently described by Grunstein et al.(32) for VEGF-transfected fibrosarcoma cells. By contrast, tumors formed by either WM164 mock control or WM164-Vas cells did not contain these large vessels (Fig. 7, c and i). When microvessel density was determined according to Weidner (33) we observed no relationship between VEGF or bFGF expression and vessel density in WM164 cells. Notably, even WM164-Vas2 cells formed tumors with microvessel densities comparable with controls although they produced negligible amounts of VEGF. Furthermore, WM164-Vs cells formed tumors with the lowest microvessel density. The highest microvessel count was observed in the spontaneously regressing tumors formed by FM516SV cells overexpressing bFGF.

Expression of TSP-1.

The angiogenic phenotype of a given cell is determined by the balance between angiogenesis-inducing and -inhibiting factors. Recent studies have shown that TSP-1 is a strong inhibitor of angiogenesis and potent suppressor of malignant growth (34, 35). Therefore, we determined whether WM164 and FM516SV cells displayed marked differences in TSP-1 protein expression. As shown in Fig. 11, WM164 cells produced no detectable TSP-1, and forced expression of either sense VEGF mRNA or sense bFGF mRNA did not affect TSP-1 production in these cells. By contrast, FM516SV cells expressed TSP-1. Similarly, FM516SV-Fs cells, which gave rise to spontaneously regressing tumors, expressed TSP-1 (Fig. 11).

In Vitro Angiogenesis Assay.

To test whether VEGF antisense mRNA-transfected WM164 melanoma cells were capable of vascular mimicry and thereby promoted tumor growth even in the absence of endogenous VEGF production, we tested the WM164 transfectants using an in vitro angiogenesis assay. HUVECs served as positive control. As shown in Fig. 12, HUVEC endothelial cells rapidly aligned and formed well-delineated tube-like structures (Fig. 12, g and h). Of the WM164 transfectants tested, both WM164-Fs (Fig. 12, e and f) and WM164-Vs (Fig. 12, i and j) and, to a lesser extent, WM164-M cells (Fig. 12, a and b), showed a tendency to form tube-like structures albeit not as delineated as HUVEC cells. Especially after 72 h the majority of melanoma cells remain poorly organized in irregular-shaped cell clusters. By contrast, the VEGF AS mRNA-transfected WM164-Vas2 cells formed loose cell aggregates with no evidence of tube formation indicative for an endothelial-like phenotype or vascular mimicry (Fig. 12, c and d).

This study demonstrates that both VEGF and bFGF, when overexpressed by melanoma cells, support melanoma cell growth in vivo. These results are consistent with previous studies in which associations between VEGF (36) and/or bFGF (37) production and biological aggressiveness of human melanomas have been described. However, our results also indicate that of these two melanoma-derived factors only bFGF is critically important for tumor formation in vivo, whereas VEGF is dispensable. These conclusions are based on the findings that WM164-Vas cells formed rapidly growing tumors after a lag period of several weeks as compared with controls (1), and we could not establish a WM164 variant in which bFGF protein production was successfully and consistently ablated (2). The critical importance of autocrine bFGF production by melanoma cells has been demonstrated earlier (2, 3, 25, 38). By contrast, it was surprising that VEGF production was not essential to either growth of WM164 cells in vitro through an autocrine mechanism or WM164 tumor formation through paracrine mechanisms in vivo. Although suppression of VEGF production significantly delayed tumorigenicity of WM164 cells in mice, the tumors that eventually formed expanded at rates that were comparable with those achieved by control mock-transfected cells. These tumors continued to express VEGF antisense transcripts and produced no VEGF protein attesting to the fact that they did not rely on VEGF production for tumor formation. Our results extend earlier studies, which reported stunted tumor development in melanoma cells with down-regulated VEGF expression (39, 40). As in the present study, this earlier report also demonstrated similar in vitro growth rates for mock-transfected control and for VEGF sense and antisense-transfected human melanoma cells. However, some differences between these earlier studies and the results presented here deserve attention.

Oku et al.(40) used SK-Mel-2 human melanoma cells, which produce only very low levels of VEGF under normoxic or hypoxic conditions, whereas the WM164 cells used in the present study secrete VEGF constitutively at high levels (>1000 pg/ml). Thus, down-regulation of VEGF expression in WM164 was expected to have a more dramatic effect on tumorigenicity than down-regulation of VEGF production in SK-Mel-2 cells. The complete absence of WM164-Vas2 tumor masses during the first 3–4 weeks after inoculation is consistent with this prediction. However, the growth rate of tumors that eventually developed was rapid and indistinguishable from controls. WM164-Vas2 tumor development occurred despite continued expression of antisense VEGF sequences in WM164-Vas2 tumors associated with absent VEGF protein production. Thus, the fast growth rate of these tumors is likely to be attributable to the development of VEGF-independent mechanisms for sustained tumor growth and angiogenesis. The presence of at least two independent angiogenic pathways in melanoma cells mediated by either VEGF receptors or members of the Tie family has been documented very recently to contribute to melanoma-associated angiogenesis (31). Forced expression of soluble forms of either VEGFR-1 or Tie2, which acted to neutralize ambient VEGF or angiopoietin-1, respectively, were shown to inhibit in vivo expansion of A375 melanoma cells. Importantly, as in the present study, melanoma formation was not abolished. Instead, poorly vascularized, relatively small and slow growing tumor masses developed. By contrast, we describe here the in vivo emergence of fast-growing WM164 variant cells, which expanded in vivo despite sustained suppression of VEGF production by antisense sequences. Emergence of these variants was not an isolated phenomenon but occurred with similar kinetics in all of the mice inoculated with WM164-Vas2 melanoma cells. In conclusion, the delayed but rapid expansion of WM164-Vas cells in vivo must have been attributable to alternative angiogenic mechanisms that substituted for tumor-derived VEGF. Taken together, our results highlight that, at least in WM164 melanoma cells, VEGF production is beneficial but not required for tumorigenicity in vivo.

This is in contrast to bFGF, which was absolutely required for cell growth even in vitro. We were unable to establish clones that stably expressed bFGF antisense sequences consistent with an essential autocrine role of bFGF in melanoma development as described earlier (2, 3, 25). By contrast, WM164 cells overexpressing bFGF were established easily. Interestingly, these cells had population-doubling rates that were indistinguishable from mock-transfected or parental WM164 cells. This result indicates that the level of endogenous bFGF produced by WM164 cells is sufficient for maximal autocrine growth stimulation, at least under the in vitro conditions used here. However, bFGF-transfected WM164 cells formed tumors at a faster rate than mock-transfected cells. This result suggests significant paracrine effects of overexpressed and, implicitly, of endogenous bFGF in WM164 cells.

Forced expression of either bFGF or VEGF did not induce sustained tumor formation of FM516SV cells in mice. Interestingly, forced expression of bFGF in FM516SV cells caused transient tumor formation associated with some degree of neovascularization. A recent report described that forced expression of bFGF in normal neonatal melanocytes induced a transformed phenotype as determined by growth factor-independent proliferation in vitro and anchorage-independence in soft agar (41). In addition, these cells formed nests of proliferating melanocytes when injected into the dermis of human skin grafts on SCID mice. Our results extend this earlier report insofar as we observed macroscopic albeit transitory tumor formation of bFGF-transduced FM516SV cells. The FM516SV bFGF tumors spontaneously and reproducibly regressed indicating that stable bFGF overexpression alone is not sufficient to support tumorigenicity of FM516SV cells defined as continuous and inexorable expansion of transformed cells in vivo. Unfortunately, no data are available for comparison relating to the long-term persistence and growth of bFGF-overexpressing melanocytes described by Nesbit et al.(41). At present, it is unclear why overexpression of bFGF or VEGF does not induce tumorigenicity in FM516SV cells. It appears possible that factors other than either bFGF or VEGF are also required for melanoma growth in vivo to occur. Conversely, the production of antiangiogenic factors by normal cell types may limit the angiogenic propensity of these cells. In support of this notion we describe here that FM516SV melanocytes produce TSP-1, whereas WM164 do not.

A similar expression pattern of pro- and antiangiogenic factors has been described in bladder cancer and normal urothelial cells. (42). Campbell et al.(42) demonstrated that normal urothelial cells, which failed to induce endothelial growth in a rat corneal assay, expressed VEGF at the same level as bladder carcinoma cells but retained TSP-1 production, which was reduced by >94% in angiogenic bladder carcinoma cells. Additional evidence for the significance of the differential expression of TSP-1 in FM516SV cells as opposed to WM164 cells is provided by the recent finding that TSP-1 inhibits angiogenesis in vivo by inducing apoptosis of activated endothelial cells (43). Induction of apoptosis in activated endothelial cells is one possible explanation for the spontaneous regression of tumors formed after s.c. inoculation FM516SV-Fs cells in SCID mice. Taken together our data indicate that the loss of TSP-1 might be a key event in the switch from an antiangiogenic to an angiogenic phenotype and even more critical for melanoma formation than the capacity to secrete VEGF. Future studies will determine whether forced expression of TSP-1 in malignant melanomas neutralizes the angiogenic/tumorigenic potential of these cells as well, as has been shown for cutaneous squamous cell carcinomas (35).

Of note, both VEGF- and bFGF-overexpressing WM164 cells formed rapidly expanding tumors that contained numerous larger vessels lined with CD31-positive endothelial cells similar to those described by Grunstein et al.(32) for VEGF-transfected fibrosarcoma cells. These lacunar endothelia were induced by transgene expression, because they were absent from tumors formed by either WM164 mock control or WM164-Vas2 cells. However, when microvessel density was determined according to Weidner (33) we observed no relationship between VEGF or bFGF expression and vessel density in WM164 cells. Notably, even WM164-Vas2 cells, which express no VEGF, formed tumors with microvessel densities comparable with controls. Furthermore, WM164-Vs cells formed tumors with the lowest microvessel density, and the highest microvessel count was observed in the spontaneously regressing tumors formed by FM516SV cells overexpressing bFGF. These results indicate induction of unusual vessels characterized by large lumina but not microvessels by bFGF or VEGF overexpression in the two melanoma cell lines tested. This result raised the issue of whether bFGF or VEGF when overexpressed elicited the formation of atypical vessels. Specifically, Maniotis et al.(44) described recently that melanoma cells are capable of “vascular mimicry,” i.e., the formation of vascular structures without endothelial lining. We addressed this possibility by testing WM164 transfectants in an in vitro angiogenic assay. The results presented here are consistent with the notion that melanoma cells may participate in vascular mimicry as described recently (44). Overexpression of bFGF or VEGF appears to induce tube formation under the assay conditions used, additionally raising the question of whether these two cytokines participate in vascular mimicry by melanoma cells. This issue is presently under investigation.

Taken together, our results highlight that of the two melanoma-derived growth factors investigated only bFGF is critically important to tumorigenicity of WM164 cells in vivo. In contrast to bFGF, VEGF serves no essential autocrine or paracrine functions although it accelerates tumor expansion in vivo dramatically when overexpressed. Furthermore, our results demonstrate the potential importance of endogenous angiogenesis inhibitors like TSP-1 because the expression of VEGF and bFGF could only sustain or accelerate tumor growth in the absence of TSP-1.

Fig. 1.

A, expression of VEGF sense or antisense mRNA in WM164 or FM516SV cells. After stable transfection total cellular RNA was probed with VEGF sense or antisense riboprobes for vector-derived VEGF mRNA expression. Using a VEGF antisense riboprobe a single band of the predicted size (650 bp) for vector-derived VEGF sense RNA was detected only WM164-Vs and FM516SV-Vs cells (top panel). In the WM164-Vas cells VEGF antisense mRNA expression was confirmed using a VEGF sense riboprobe, which detected a single band of 650 bp (bottom panel). B, expression of vector-derived bFGF mRNA after transfection of WM164 and FM516SV. After stable transfection total cellular RNA was probed with a bFGF antisense riboprobe for vector-derived bFGF mRNA expression. A single band of the predicted size (480 bp) for vector-derived bFGF sense RNA was detected only WM164-Fs and FM516SV-Fs cells.

Fig. 1.

A, expression of VEGF sense or antisense mRNA in WM164 or FM516SV cells. After stable transfection total cellular RNA was probed with VEGF sense or antisense riboprobes for vector-derived VEGF mRNA expression. Using a VEGF antisense riboprobe a single band of the predicted size (650 bp) for vector-derived VEGF sense RNA was detected only WM164-Vs and FM516SV-Vs cells (top panel). In the WM164-Vas cells VEGF antisense mRNA expression was confirmed using a VEGF sense riboprobe, which detected a single band of 650 bp (bottom panel). B, expression of vector-derived bFGF mRNA after transfection of WM164 and FM516SV. After stable transfection total cellular RNA was probed with a bFGF antisense riboprobe for vector-derived bFGF mRNA expression. A single band of the predicted size (480 bp) for vector-derived bFGF sense RNA was detected only WM164-Fs and FM516SV-Fs cells.

Close modal
Fig. 2.

A, effect of VEGF sense or antisense mRNA expression on in vitro VEGF protein secretion. FM516SV and WM164 cells transfected with the empty vector (FM516SV-M and WM164-M), pool cultures stably expressing sense VEGF mRNA (FM516SV-Vs and WM164-Vs), and clones stably expressing antisense VEGF mRNA (WM164-Vas2 and -Vas4); bars, ± SE. B, effect of bFGF sense mRNA expression on in vitro VEGF protein secretion. FM516SV and WM164 cells transfected with the empty vector (FM516SV-M and WM164-M) and pool cultures stably expressing sense bFGF mRNA (FM516SV-Fs and WM164-Fs). For A and B cells were grown in W489 medium supplemented with 0.2% BSA at 3 × 105/well in six-well 35-mm culture plates. After 3 days the medium was collected, and cell numbers were assessed. VEGF concentrations were measured by ELISA; bars, ± SE. C, effect of VEGF sense or antisense mRNA and bFGF sense mRNA expression on in vitro bFGF protein production. Result shown are for FM516SV and WM164 cells mock-transfected with the empty vector (FM516SV-M and WM164-M), pool cultures stably expressing VEGF sense mRNA (FM516SV-Vs and WM164-Vs) and bFGF sense mRNA (FM516-Fs and WM164-Fs), and clones stably expressing antisense VEGF mRNA (WM164-Vas2). Cells were grown in W489 medium supplemented with 0.2% BSA at 3 × 105/well in six-well 35-mm culture plates. After 3 days the cells were harvested. bFGF concentrations were measured by ELISA using cell extracts, bars, ± SE.

Fig. 2.

A, effect of VEGF sense or antisense mRNA expression on in vitro VEGF protein secretion. FM516SV and WM164 cells transfected with the empty vector (FM516SV-M and WM164-M), pool cultures stably expressing sense VEGF mRNA (FM516SV-Vs and WM164-Vs), and clones stably expressing antisense VEGF mRNA (WM164-Vas2 and -Vas4); bars, ± SE. B, effect of bFGF sense mRNA expression on in vitro VEGF protein secretion. FM516SV and WM164 cells transfected with the empty vector (FM516SV-M and WM164-M) and pool cultures stably expressing sense bFGF mRNA (FM516SV-Fs and WM164-Fs). For A and B cells were grown in W489 medium supplemented with 0.2% BSA at 3 × 105/well in six-well 35-mm culture plates. After 3 days the medium was collected, and cell numbers were assessed. VEGF concentrations were measured by ELISA; bars, ± SE. C, effect of VEGF sense or antisense mRNA and bFGF sense mRNA expression on in vitro bFGF protein production. Result shown are for FM516SV and WM164 cells mock-transfected with the empty vector (FM516SV-M and WM164-M), pool cultures stably expressing VEGF sense mRNA (FM516SV-Vs and WM164-Vs) and bFGF sense mRNA (FM516-Fs and WM164-Fs), and clones stably expressing antisense VEGF mRNA (WM164-Vas2). Cells were grown in W489 medium supplemented with 0.2% BSA at 3 × 105/well in six-well 35-mm culture plates. After 3 days the cells were harvested. bFGF concentrations were measured by ELISA using cell extracts, bars, ± SE.

Close modal
Fig. 3.

Effect of bFGF sense mRNA expression and effect of VEGF sense or antisense mRNA expression on in vitro bFGF protein production. FM516SV and WM164 cells transfected with the empty vector (FM516SV-M and WM164-M), pool cultures stably expressing sense bFGF mRNA (FM516SV-Fs and WM164-Fs), pool cultures stably expressing sense VEGF mRNA (FM516SV-Vs and WM164-Vs), and clones stably expressing antisense VEGF mRNA (WM164-Vas2) were grown as described in Fig. 2. After 3 days cell extracts were electrophoresed on a 15% SDS gel, electroblotted to a nitrocellulose membrane and probed with an anti-bFGF Ab as described in “Materials and Methods.”

Fig. 3.

Effect of bFGF sense mRNA expression and effect of VEGF sense or antisense mRNA expression on in vitro bFGF protein production. FM516SV and WM164 cells transfected with the empty vector (FM516SV-M and WM164-M), pool cultures stably expressing sense bFGF mRNA (FM516SV-Fs and WM164-Fs), pool cultures stably expressing sense VEGF mRNA (FM516SV-Vs and WM164-Vs), and clones stably expressing antisense VEGF mRNA (WM164-Vas2) were grown as described in Fig. 2. After 3 days cell extracts were electrophoresed on a 15% SDS gel, electroblotted to a nitrocellulose membrane and probed with an anti-bFGF Ab as described in “Materials and Methods.”

Close modal
Fig. 4.

Effect of VEGF sense or antisense mRNA expression or bFGF sense mRNA expression on in vitro growth of WM164 or FM516SV cells. A, WM164 cells transfected with the empty vector (WM164-M), stably expressing VEGF sense (WM164-Versus), or antisense mRNA (WM164-Vas2), bars, ± SE. B, WM164 cells transfected with the empty vector (WM164-M) or stably expressing bFGF sense mRNA (WM164-Fs), bars, ± SE. C, FM516SV cells transfected with the empty vector (FM516SV-M) or stably expressing VEGF sense mRNA (FM516SV-Vs), or D, stably expressing bFGF sense mRNA (FM516SV-Fs), bars, ± SE. All cells were seeded at 0.5 × 105/well (six-well plates) in W489 base medium (total volume of 2 ml) supplemented with 0.2% BSA. Cell growth was assayed after 3 and 7 or 10 days by counting the number of viable cells using a Coulter counter.

Fig. 4.

Effect of VEGF sense or antisense mRNA expression or bFGF sense mRNA expression on in vitro growth of WM164 or FM516SV cells. A, WM164 cells transfected with the empty vector (WM164-M), stably expressing VEGF sense (WM164-Versus), or antisense mRNA (WM164-Vas2), bars, ± SE. B, WM164 cells transfected with the empty vector (WM164-M) or stably expressing bFGF sense mRNA (WM164-Fs), bars, ± SE. C, FM516SV cells transfected with the empty vector (FM516SV-M) or stably expressing VEGF sense mRNA (FM516SV-Vs), or D, stably expressing bFGF sense mRNA (FM516SV-Fs), bars, ± SE. All cells were seeded at 0.5 × 105/well (six-well plates) in W489 base medium (total volume of 2 ml) supplemented with 0.2% BSA. Cell growth was assayed after 3 and 7 or 10 days by counting the number of viable cells using a Coulter counter.

Close modal
Fig. 5.

Effect of VEGF sense or antisense mRNA expression or bFGF sense mRNA expression on in vivo growth of WM164 cells when implanted s.c. into SCID mice. A, WM164 cells transfected with the empty vector (WM164-M) or stably expressing bFGF sense mRNA (WM164-Fs), bars, ± SE. B, WM164 cells transfected with the empty vector (WM164-M) or stably expressing VEGF sense (WM164-Vs) or antisense mRNA (WM164-Vas2), bars, ± SE. Cells (5 ×106)/per animal were implanted s.c. into SCID mice (n = 5). In each case the experiment was repeated twice with similar results.

Fig. 5.

Effect of VEGF sense or antisense mRNA expression or bFGF sense mRNA expression on in vivo growth of WM164 cells when implanted s.c. into SCID mice. A, WM164 cells transfected with the empty vector (WM164-M) or stably expressing bFGF sense mRNA (WM164-Fs), bars, ± SE. B, WM164 cells transfected with the empty vector (WM164-M) or stably expressing VEGF sense (WM164-Vs) or antisense mRNA (WM164-Vas2), bars, ± SE. Cells (5 ×106)/per animal were implanted s.c. into SCID mice (n = 5). In each case the experiment was repeated twice with similar results.

Close modal
Fig. 6.

Effect of bFGF sense mRNA expression on in vivo growth of FM516SV cells when implanted s.c. into SCID mice. FM516SV cells transfected with the empty vector (FM516SV-M) or stably expressing bFGF sense mRNA (FM516SV-Fs). Cells (5 × 106)/per animal were implanted s.c. into SCID mice (n = 5). In each case the experiment was repeated twice with similar results. Bars, ± SE. *, animals of which their tumors regressed spontaneously were killed on day 7 for additional analysis.

Fig. 6.

Effect of bFGF sense mRNA expression on in vivo growth of FM516SV cells when implanted s.c. into SCID mice. FM516SV cells transfected with the empty vector (FM516SV-M) or stably expressing bFGF sense mRNA (FM516SV-Fs). Cells (5 × 106)/per animal were implanted s.c. into SCID mice (n = 5). In each case the experiment was repeated twice with similar results. Bars, ± SE. *, animals of which their tumors regressed spontaneously were killed on day 7 for additional analysis.

Close modal
Fig. 7.

Immunohistochemistry of s.c. tumors formed by WM164 cells after transfection with the empty vector (WM164-M, a–c), stable VEGF sense mRNA expression (WM164-Vs, d–f), stable VEGF antisense mRNA expression (WM164-Vas2, g–i), or stable bFGF sense mRNA expression (WM164-Fs, j–l). a, d, g, and j, staining with an anti-VEGF mAb; b, e, h, and k, staining with an anti-bFGF Ab; c, f, i, and l, staining with an anti-CD31 mAb. a–d, f–j, and l, ×40; k, ×100; e, ×200.

Fig. 7.

Immunohistochemistry of s.c. tumors formed by WM164 cells after transfection with the empty vector (WM164-M, a–c), stable VEGF sense mRNA expression (WM164-Vs, d–f), stable VEGF antisense mRNA expression (WM164-Vas2, g–i), or stable bFGF sense mRNA expression (WM164-Fs, j–l). a, d, g, and j, staining with an anti-VEGF mAb; b, e, h, and k, staining with an anti-bFGF Ab; c, f, i, and l, staining with an anti-CD31 mAb. a–d, f–j, and l, ×40; k, ×100; e, ×200.

Close modal
Fig. 8.

In vivo expression of VEGF sense or antisense mRNA. After recovery of s.c.-formed tumors total cellular RNA was probed with VEGF sense or antisense riboprobes for sustained vector derived VEGF mRNA expression. Using a VEGF antisense riboprobe a single band of the predicted size (650 bp) for VEGF sense RNA was detected only WM164-Vs tumors (top panel). In the WM164-Vas2 tumors that formed after 50 days, sustained VEGF antisense mRNA expression was confirmed using a VEGF sense riboprobe, which detected a single band of 650 bp. (bottom panel).

Fig. 8.

In vivo expression of VEGF sense or antisense mRNA. After recovery of s.c.-formed tumors total cellular RNA was probed with VEGF sense or antisense riboprobes for sustained vector derived VEGF mRNA expression. Using a VEGF antisense riboprobe a single band of the predicted size (650 bp) for VEGF sense RNA was detected only WM164-Vs tumors (top panel). In the WM164-Vas2 tumors that formed after 50 days, sustained VEGF antisense mRNA expression was confirmed using a VEGF sense riboprobe, which detected a single band of 650 bp. (bottom panel).

Close modal
Fig. 9.

Immunohistochemistry of s.c. tumors formed by FM516SV cells after transfection with the bFGF sense mRNA expression vector (FM516SV-Fs). a, H&E staining; b, staining with an antibody to Melan-A (arrows), a melanocytic marker; c, staining with an anti-VEGF mAb; d, staining with an anti-bFGF Ab; e, staining with an anti-CD31 mAb. a, c, and d, ××40; b and d, ×200)

Fig. 9.

Immunohistochemistry of s.c. tumors formed by FM516SV cells after transfection with the bFGF sense mRNA expression vector (FM516SV-Fs). a, H&E staining; b, staining with an antibody to Melan-A (arrows), a melanocytic marker; c, staining with an anti-VEGF mAb; d, staining with an anti-bFGF Ab; e, staining with an anti-CD31 mAb. a, c, and d, ××40; b and d, ×200)

Close modal
Fig. F10-1552.

Vascular density in the different tumors generated after s.c inoculation of WM164 or FM516SV cells transfected with the empty vector (WM164-M), VEGF sense vector (WM164-Vs), VEGF antisense vector (WM164-Vas2), or bFGF sense vector (WM164-Fs and FM516SV-Fs); bars, ± SE. Microvessel density was determined as described in “Materials and Methods.”

Fig. F10-1552.

Vascular density in the different tumors generated after s.c inoculation of WM164 or FM516SV cells transfected with the empty vector (WM164-M), VEGF sense vector (WM164-Vs), VEGF antisense vector (WM164-Vas2), or bFGF sense vector (WM164-Fs and FM516SV-Fs); bars, ± SE. Microvessel density was determined as described in “Materials and Methods.”

Close modal
Fig. 11.

TSP-1 expression in WM164 and FM516SV cells. FM516SV and WM164 cells transfected with the empty vector (WM164-M and FM516SV-M) and pool cultures stably expressing sense bFGF mRNA (WM164-Fs and FM516SV-Fs) or sense VEGF mRNA (WM164-Vs) were grown as described in Fig. 2. After 3 days cell extracts where electrophoresed on a 7.5% SDS polyacrylamide gel, electroblotted to a nitrocellulose membrane, and probed with an anti-TSP-1 mAb as described in “Materials and Methods.”

Fig. 11.

TSP-1 expression in WM164 and FM516SV cells. FM516SV and WM164 cells transfected with the empty vector (WM164-M and FM516SV-M) and pool cultures stably expressing sense bFGF mRNA (WM164-Fs and FM516SV-Fs) or sense VEGF mRNA (WM164-Vs) were grown as described in Fig. 2. After 3 days cell extracts where electrophoresed on a 7.5% SDS polyacrylamide gel, electroblotted to a nitrocellulose membrane, and probed with an anti-TSP-1 mAb as described in “Materials and Methods.”

Close modal
Fig. 12.

In vitro angiogenesis assay. WM164 cells transfected with the empty vector (WM164-M; a and b), transfected with VEGF sense mRNA (WM164-Vs; i and j), transfected with VEGF antisense mRNA (WM164-Vas2; c and d), or with bFGF sense mRNA (WM164-Fs; e and f) were tested for in vitro tube formation as described in “Material and Methods.” Normal HUVECs showing alignment and tube-like structures served as control (g and h). Pictures were taken after 36 and 72 h.

Fig. 12.

In vitro angiogenesis assay. WM164 cells transfected with the empty vector (WM164-M; a and b), transfected with VEGF sense mRNA (WM164-Vs; i and j), transfected with VEGF antisense mRNA (WM164-Vas2; c and d), or with bFGF sense mRNA (WM164-Fs; e and f) were tested for in vitro tube formation as described in “Material and Methods.” Normal HUVECs showing alignment and tube-like structures served as control (g and h). Pictures were taken after 36 and 72 h.

Close modal

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

Supported by Deutsche Forschungsgemeinschaft Grant Gr 984/3-1 and FORUM University Bochum Grant (to U. G.); and by grants from the NIH and the W. W. Smith Foundation (to U. R.).

3

The abbreviations used are: VEGF, vascular endothelial growth factor; bFGF, basic fibroblast growth factor; CM, conditioned medium; mAb, monoclonal antibody; TSP-1, thrombospondin-1; HUVEC, human umbilical vein endothelial cell.

1
Graeven U., Fiedler W., Karpinski S., Ergun S., Kilic N., Rodeck U., Schmiegel W., Hossfeld D. K. Melanoma-associated expression of vascular endothelial growth factor and its receptors FLT-1 and KDR.
J. Cancer Res. Clin. Oncol.
,
125
:
621
-629,  
1999
.
2
Becker D., Meier C. B., Herlyn M. Proliferation of human malignant melanomas is inhibited by antisense oligodeoxynucleotides targeted against basic fibroblast growth factor.
EMBO J.
,
8
:
3685
-3691,  
1989
.
3
Becker D., Lee P. L., Rodeck U., Herlyn M. Inhibition of the fibroblast growth factor receptor 1 (FGFR-1) gene in human melanocytes and malignant melanomas leads to inhibition of proliferation and signs indicative of differentiation.
Oncogene
,
7
:
2303
-2313,  
1992
.
4
Graeven U., Rodeck U., Karpinski S., Andre N., Schmiegel W. Expression patterns of placenta growth factor-1 in human melanocytic cell lines.
J. Investig. Dermatol.
,
115
:
118
-123,  
2000
.
5
Westermark B., Johnsson A., Paulsson Y., Betsholtz C., Heldin C. H., Herlyn M., Rodeck U., Koprowski H. Human melanoma cell lines of primary and metastatic origin express the genes encoding the chains of platelet-derived growth factor (PDGF) and produce a PDGF-like growth factor.
Proc. Natl. Acad. Sci. USA
,
83
:
7197
-7200,  
1986
.
6
Rodeck U., Melber K., Kath R., Menssen H. D., Varello M., Atkinson B., Herlyn M. Constitutive expression of multiple growth factor genes by melanoma cells but not normal melanocytes.
J. Investig. Dermatol.
,
97
:
20
-26,  
1991
.
7
Herlyn M., Graeven U., Speicher D., Sela B. A., Bennicelli J. L., Kath R., Guerry D. Characterization of tenascin secreted by human melanoma cells.
Cancer Res.
,
51
:
4853
-4858,  
1991
.
8
Herlyn D., Iliopoulos D., Jensen P. J., Parmiter A., Baird J., Hotta H., Adachi K., Ross A. H., Jambrosic J., Koprowski H., et al In vitro properties of human melanoma cells metastatic in nude mice.
Cancer Res.
,
50
:
2296
-2302,  
1990
.
9
de Vries T. J., Kitson J. L., Silvers W. K., Mintz B. Expression of plasminogen activators and plasminogen activator inhibitors in cutaneous melanomas of transgenic melanoma-susceptible mice.
Cancer Res.
,
55
:
4681
-4687,  
1995
.
10
Nip J., Rabbani S. A., Shibata H. R., Brodt P. Coordinated expression of the vitronectin receptor and the urokinase-type plasminogen activator receptor in metastatic melanoma cells.
J. Clin. Investig.
,
95
:
2096
-2103,  
1995
.
11
Montgomery A. M., De Clerck Y. A., Langley K. E., Reisfeld R. A., Mueller B. M. Melanoma-mediated dissolution of extracellular matrix: contribution of urokinase-dependent and metalloproteinase-dependent proteolytic pathways.
Cancer Res.
,
53
:
693
-700,  
1993
.
12
Knutson J. R., Iida J., Fields G. B., McCarthy J. B. CD44/chondroitin sulfate proteoglycan and α2β1 integrin mediate human melanoma cell migration on type IV collagen and invasion of basement membranes.
Mol. Biol. Cell
,
7
:
383
-396,  
1996
.
13
Takahashi K., Eto H., Tanabe K. K. Involvement of CD44 in matrix metalloproteinase-2 regulation in human melanoma cells.
Int. J. Cancer
,
80
:
387
-395,  
1999
.
14
Neufeld G., Cohen T., Gengrinovitch S., Poltorak Z. Vascular endothelial growth factor (VEGF) and its receptors [Review].
FASEB J.
,
13
:
9
-22,  
1999
.
15
Kumar R., Kuniyasu H., Bucana C. D., Wilson M. R., Fidler I. J. Spatial and temporal expression of angiogenic molecules during tumor growth and progression.
Oncol. Res.
,
10
:
301
-311,  
1998
.
16
Bayer-Garner I. B., Hough A. J., Smoller B. R. Vascular endothelial growth factor expression in malignant melanoma: prognostic versus diagnostic usefulness.
Mod. Pathol.
,
12
:
770
-774,  
1999
.
17
Birck A., Kirkin A. F., Zeuthen J., Hou-Jensen K. Expression of basic fibroblast growth factor and vascular endothelial growth factor in primary and metastatic melanoma from the same patients.
Melanoma Res.
,
9
:
375
-381,  
1999
.
18
Bohlen P., Baird A., Esch F., Ling N., Gospodarowicz D. Isolation and partial molecular characterization of pituitary fibroblast growth factor.
Proc. Natl. Acad. Sci. USA
,
81
:
5364
-5368,  
1984
.
19
Seghezzi G., Patel S., Ren C. J., Gualandris A., Pintucci G., Robbins E. S., Shapiro R. L., Galloway A. C., Rifkin D. B., Mignatti P. Fibroblast growth factor-2 (FGF-2) induces vascular endothelial growth factor (VEGF) expression in the endothelial cells of forming capillaries: an autocrine mechanism contributing to angiogenesis.
J. Cell Biol.
,
141
:
1659
-1673,  
1998
.
20
Pepper M. S., Ferrara N., Orci L., Montesano R. Potent synergism between vascular endothelial growth factor and basic fibroblast growth factor in the induction of angiogenesis in vitro.
Biochem. Biophys. Res. Commun.
,
189
:
824
-831,  
1992
.
21
Asahara T., Bauters C., Zheng L. P., Takeshita S., Bunting S., Ferrara N., Symes J. F., Isner J. M. Synergistic effect of vascular endothelial growth factor and basic fibroblast growth factor on angiogenesis in vivo.
Circulation
,
92
:
II365
-II371,  
1995
.
22
Koolwijk P., van Erck M. G., de Vree W. J., Vermeer M. A., Weich H. A., Hanemaaijer R., van Hinsbergh V. W. Cooperative effect of TNFα, bFGF, and VEGF on the formation of tubular structures of human microvascular endothelial cells in a fibrin matrix. Role of urokinase activity.
J. Cell Biol.
,
132
:
1177
-1188,  
1996
.
23
Pepper M. S., Mandriota S. J., Jeltsch M., Kumar V., Alitalo K. Vascular endothelial growth factor (VEGF)-C synergizes with basic fibroblast growth factor and VEGF in the induction of angiogenesis in vitro and alters endothelial cell extracellular proteolytic activity.
J. Cell. Physiol.
,
177
:
439
-452,  
1998
.
24
Halaban R., Langdon R., Birchall N., Cuono C., Baird A., Scott G., Moellmann G., McGuire J. Basic fibroblast growth factor from human keratinocytes is a natural mitogen for melanocytes.
J. Cell Biol.
,
107
:
1611
-1619,  
1988
.
25
Dotto G. P., Moellmann G., Ghosh S., Edwards M., Halaban R. Transformation of murine melanocytes by basic fibroblast growth factor cDNA and oncogenes and selective suppression of the transformed phenotype in a reconstituted cutaneous environment.
J. Cell Biol.
,
109
:
3115
-3128,  
1989
.
26
Melber K., Zhu G., Diamond L. SV40-transfected human melanocyte sensitivity to growth inhibition by the phorbol ester 12-O-tetradecanoylphorbol-13-acetate.
Cancer Res.
,
49
:
3650
-3655,  
1989
.
27
Anderson S., Davis D., Dahlbäck H., Jörnvall H., Russel D. W. Cloning, structure and expression of the mitochondrial cytochrome P-450 sterol 26-hydroxylase, a bile acid biosynthetic enzyme.
J. Biol. Chem.
,
264
:
8222
-8229,  
1989
.
28
Kralli A., Ge R., Graeven U., Ricciardi R. P., Weinmann R. Negative regulation of the major histocompatibility complex class I enhancer in adenovirus type 12-transformed cells via a retinoic acid response element.
J. Virol.
,
66
:
6979
-6988,  
1992
.
29
Ryuto M., Ono M., Izumi H., Yoshida S., Weich H. A., Kohno K., Kuwano M. Induction of vascular endothelial growth factor by tumor necrosis factor α in human glioma cells. Possible roles of SP-1.
J. Biol. Chem.
,
271
:
28220
-28228,  
1996
.
30
Liu B., Earl H. M., Baban D., Shoaibi M., Fabra A., Kerr D. J., Seymour L. W. Melanoma cell lines express VEGF receptor KDR and respond to exogenously added VEGF.
Biochem. Biophys. Res. Commun.
,
217
:
721
-727,  
1995
.
31
Siemeister G., Schirner M., Weindel K., Reusch P., Menrad A., Marme D., Martiny-Baron G. Two independent mechanisms essential for tumor angiogenesis: inhibition of human melanoma xenograft growth by interfering with either the vascular endothelial growth factor receptor pathway or the Tie-2 pathway.
Cancer Res.
,
59
:
3185
-3191,  
1999
.
32
Grunstein J., Roberts W. G., Mathieu-Costello O., Hanahan D., Johnson R. S. Tumor-derived expression of vascular endothelial growth factor is a critical factor in tumor expansion and vascular function.
Cancer Res.
,
59
:
1592
-1598,  
1999
.
33
Weidner N. Current pathologic methods for measuring intratumoral microvessel density within breast carcinoma and other solid tumors.
Breast Cancer Res. Treat.
,
36
:
169
-180,  
1995
.
34
Weinstat S. D., Zabrenetzky V. S., vanHoutte K., Frazier W. A., Roberts D. D., Steeg P. S. Transfection of thrombospondin 1 complementary DNA into a human breast carcinoma cell line reduces primary tumor growth, metastatic potential, and angiogenesis.
Cancer Res.
,
54
:
6504
-6511,  
1994
.
35
Streit M., Velasco P., Brown L. F., Skobe M., Richard L., Riccardi L., Lawler J., Detmar M. Overexpression of thrombospondin-1 decreases angiogenesis and inhibits the growth of human cutaneous squamous cell carcinoma.
Am. J. Pathol.
,
155
:
441
-452,  
1999
.
36
Potgens A. J., Lubsen N. H., van Altena M. C., Schoenmakers J. G., Ruiter D. J., de Waal R. M. Vascular permeability factor expression influences tumor angiogenesis in human melanoma lines xenografted to nude mice.
Am. J. Pathol.
,
146
:
197
-209,  
1995
.
37
Albino A. P., Davis B. M., Nanus D. M. Induction of growth factor RNA expression in human malignant melanoma: markers of transformation.
Cancer Res.
,
51
:
4815
-4820,  
1991
.
38
Yayon A., Ma Y. S., Safran M., Klagsbrun M., Halaban R. Suppression of autocrine cell proliferation and tumorigenesis of human melanoma cells and fibroblast growth factor transformed fibroblasts by a kinase-deficient FGF receptor 1: evidence for the involvement of Src-family kinases.
Oncogene
,
14
:
2999
-3009,  
1997
.
39
Claffey K. P., Brown L. F., del Aguila L. F., Tognazzi K., Yeo K. T., Manseau E. J., Dvorak H. F. Expression of vascular permeability factor/vascular endothelial growth factor by melanoma cells increases tumor growth, angiogenesis, and experimental metastasis.
Cancer Res.
,
56
:
172
-181,  
1996
.
40
Oku T., Tjuvajev J. G., Miyagawa T., Sasajima T., Joshi A., Joshi R., Finn R., Claffey K. P., Blasberg R. G. Tumor growth modulation by sense and antisense vascular endothelial growth factor gene expression: effects on angiogenesis, vascular permeability, blood volume, blood flow, fluorodeoxyglucose uptake, and proliferation of human melanoma intracerebral xenografts.
Cancer Res.
,
58
:
4185
-4192,  
1998
.
41
Nesbit M., Nesbit H. K. E., Bennett J., Andl T., Hsu M. Y., Dejesus E., McBrian M., Gupta A. R., Eck S. L., Herlyn M. Basic fibroblast growth factor induces a transformed phenotype in normal human melanocytes.
Oncogene
,
18
:
6469
-6476,  
1999
.
42
Campbell S. C., Volpert O. V., Ivanovich M., Bouck N. P. Molecular mediators of angiogenesis in bladder cancer.
Cancer Res.
,
58
:
1298
-1304,  
1998
.
43
Jimenez B., Volpert O. V., Crawford S. E., Febbraio M., Silverstein R. L., Bouck N. P. Signals leading to apoptosis-dependent inhibition of neovascularization by thrombospondin-1.
Nat. Med.
,
6
:
41
-48,  
2000
.
44
Maniotis A. J., Folberg R., Hess A., Seftor E. A., Gardner L. M. G., Pe’er J., Trent J. M., Meltzer P. S., Hendrix M. J. C. Vascular channel formation by human melanoma cells in vivo and in vitro: vasculogenic mimicry.
Am. J. Pathol.
,
155
:
739
-752,  
1999
.