Activation of the FGF2/FGFR1 Autocrine Loop for Cell Proliferation and Survival in Uveal Melanoma Cells

Abstract

purpose. Constitutive activation of ERK1/2 controls proliferation of uveal melanoma cells. Because an autocrine fibroblast growth factor (FGF) activation loop controls ERK1/2 activation in many cancers, this study was conducted to examine the role of the FGF/FGF receptor autocrine loop in the ERK1/2-dependent proliferation and survival of uveal melanoma cells.

methods. Primary tumors and cell lines (OCM-1, MKT-BR, SP6.5, Mel270 and 92.1) were used to define the role of the FGF/FGFR system in human uveal melanoma. Cell proliferation was assessed by MTT-staining, and apoptosis was quantified by flow cytometry. Specific pharmacologic inhibitors of ERK1/2 and FGFR1, an anti-FGF2 neutralizing antibody and an antisense oligonucleotide directed against FGF2 were used to analyze signaling in the FGF/FGFR autocrine loop.

results. FGF1, FGF2, and their FGFR1 receptor were strongly expressed in the primary uveal melanomas. All five uveal melanoma cell lines expressed and secreted FGF2. They also expressed FGFR1. Cell proliferation was strongly reduced by the antisense oligonucleotide-mediated depletion of endogenous FGF2, immunoneutralization of secreted FGF2, and pharmacologic inhibition of FGFR1. The FGF2/FGFR1-mediated signaling pathway was identified by showing that inhibition of either FGF2 or FGFR1 reduced ERK1/2 activation, cell proliferation, and survival.

conclusions. The FGF/FGFR/ERK signaling pathway may be a target for therapeutic strategies against uveal melanoma.

Uveal melanoma is the most common primary ocular neoplasm in adults, with an estimated annual frequency of six to eight cases per million people in developed countries. ^{1 2 3} Its cause, however, is largely unknown, for no risk factors have yet been identified. Similarly, the cellular events leading to malignant transformation of uveal melanocytes also remain unclear. Cutaneous and uveal melanomas share similar histologic features. However, mutations in genes encoding major signaling pathway proteins that control cell proliferation, including those affected in cutaneous melanoma, are infrequent in uveal melanoma: no oncogene or tumor suppressor has yet been convincingly linked to uveal melanoma. ^{4 5} No evidence currently supports an association between Ras mutations and this cancer. ^{6 7} Because the activating mutation V600E in B-RAF (^W600EB-Raf) has been detected in only a few ocular melanomas, ^{8 9 10 11 12 13 14} therefore, molecular mechanisms involved in the activation of cell proliferation in uveal and cutaneous melanoma are absolutely different, and it appears likely that other molecular mechanisms are involved in autonomous proliferation of uveal melanoma cells. 

One such mechanism is the constitutive activation of extracellular signal–regulated kinases 1 and 2 (ERK1/2), which has been observed in both primary uveal melanoma tumors and uveal melanoma cell lines without either RAS or B-RAF mutations. ^{15 16 17} Inhibition of MAPK/ERK kinases 1 and 2 (MEK1/2)—the direct upstream activators of ERK1/2—and small interfering (si)RNA-mediated specific depletion of the wild-type form of B-Raf (^WTB-Raf)—the direct upstream activator of MEK1/2—both induce a substantial decrease in uveal melanoma cell line proliferation and thus identify the ^WTB-Raf/MEK/ERK signaling pathway as key to the control of proliferation of these cells. ^{18 19} Unlike uveal melanoma cells, normal uveal melanocytes (NUMs) do not express activated ERK1/2, and MEK1/2 inhibition does not affect their growth. ¹⁸ We recently showed that the stem cell factor (SCF)/c-Kit/ERK1/2 autocrine loop is activated and contributes to cell proliferation and transformation in ^WTB-Raf uveal melanoma cells. ¹⁶ Differences in c-Kit levels and responses to imatinib mesylate, a c-Kit inhibitor, in these cell lines suggest, however, that other growth factors play a role in melanoma cell proliferation. ¹⁶  

Inappropriate expression of fibroblast growth factor (FGF)-2 and its receptors causes aberrant cell proliferation in various cancers and many human tumor cell lines. ²⁰ FGF2 belongs to a family of at least 25 growth factors and oncogenes. ²¹ Different protein isoforms of FGF2 result from alternative translational initiation, giving rise to 21- to 24-kDa forms (collectively referred to as high-molecular-weight [HMW] isoforms) with limited tissue distribution and to the ubiquitously expressed 18-kDa form. ²¹ These isoforms all originate from alternative translation of a single mRNA. Tumor cells overexpress HMW FGF2 isoforms or have an elevated ratio of HMW isoforms to the 18-kDa isoform, compared with normal cells. The 18-kDa FGF2 is the predominant isoform released by cells. HMW isoforms, on the other hand, remain intracellular and appear to elicit different biological functions—including migration, proliferation, and transformation—than the 18-kDa isoform does. These functions are both dose- and cell-type-dependent. ^{20 21} The biological activity of the 18-kDa FGF2 requires the presence of both heparan sulfate proteoglycans (HSPGs) and FGF tyrosine kinase receptors (FGFRs) to transduce signals for cell proliferation. ^{21 22 23 24} FGFRs are encoded by four distinct genes (FGFR1–4), and the various associations between these growth factors and their receptors regulate the specificity of FGF-induced downstream signaling and biological activities. ²⁵ The regulatory mechanisms governing the expression of FGFs and FGFRs in normal and tumor cells are not yet well understood. FGFR1 and FGFR2 bind FGF2 with the greatest affinity, but the level of redundancy in receptor utilization within the FGF family is high. ²³ Most uveal melanomas express large amounts of FGF2, and uveal melanoma cultures secrete large amounts of it. ^{26 27 28 29} These findings suggest that an autocrine FGF2 loop may be involved in uveal melanoma tumorigenesis. In contrast, NUMs are quiescent in vivo, and exogenous FGF2 must be added to trigger their proliferation in vitro. ^{30 31} FGFR expression and FGF secretion may lead to autocrine/paracrine stimulation that sets up a constitutive activation loop of the FGF-intracellular signaling pathway and results in the expression of cell cycle machinery proteins. ³² To learn more about the role of FGF2 and its intracellular signaling in the control of uveal melanoma cell proliferation, we used pharmacologic, immunologic, and antisense oligonucleotide strategies that targeted either FGF2 or FGFR to investigate the roles of exogenous and endogenous FGF2 in regulating the proliferation of normal and malignant uveal melanocytes. We also investigated the role of the expression of ^W600EB-Raf in the expression of FGF2 and FGFRs in uveal melanoma cells and in the FGF2/FGFR autocrine loop that controls their proliferation and survival. Finally, we investigated the role of FGF2/FGFR in the control of ERK2 in uveal melanoma cells expressing either ^WTB-Raf or ^V600EB-Raf. ^{10 16}  

Materials and Methods

Cell Cultures

NUMs were isolated, as previously described ^{30 31} from human enucleated eyes (generously provided by Philippe Gain, Faculté de Médecine, Saint-Etienne, France). NUMs were cultured in F12 Ham’s medium supplemented with 20% fetal calf serum (FCS), 2.5 μg/mL amphotericin B, 2 mM l-glutamine (Invitrogen, Abingdon, UK), 10 ng/mL cholera toxin (CT; Calbiochem, Meudon, France), 0.1 mM isobutyl methylxanthine (IBMX; Sigma-Aldrich, Saint-Quentin Fallavier, France), and 10 ng/mL FGF2 (18-kDa FGF2 isoform, obtained from Hervé Prats, INSERM (Institut National de la Santé et de la Recherche Médicale) U397, Toulouse, France). All five melanoma cell lines—92.1 and Mel270 (kindly provided by Martine Jager, University of Leiden, Netherlands) and OCM-1, MKT-BR, and SP6.5 (kindly provided by François Malecaze, Ophthalmology Department, CHU (Center Hospitalier Université) Toulouse, France)—were grown in RPMI 1640 medium supplemented with 5% FCS, 2.5 μg/mL amphotericin B, 50 μg/mL gentamicin, and 2 mM l-glutamine (Invitrogen), as previously described. ^{33 34} NUMs and uveal melanoma cells were cultured at 37°C in a humidified air/CO₂ (19:1) atmosphere. 92.1 and Mel270 are uveal melanoma cell lines that express ^WTB-Raf, whereas OCM-1, MKT-BR and SP6.5 express ^V600EB-Raf (Table 1) . ¹⁹  

Cell Treatment and Proliferation Assay

The molecular aspects of cell proliferation were investigated by treating cells with specific pharmacologic inhibitors of signaling pathways or immunoneutralizing antibodies. When necessary, stock solutions of pharmacologic inhibitors were made up in dimethyl sulfoxide (DMSO) such that the final concentration of DMSO in the culture media did not exceed 0.1%, a concentration shown to have no effect on the proliferation of either the NUMs or the uveal melanoma cells. The cells were stimulated with FGF1, FGF5, VEGF (R&D Systems, Abingdon, UK) and PDGF-BB (kindly provided by Marijke Bryckaert, Hôpital Lariboisiere, INSERM U689, Paris, France) in serum-free culture medium for the uveal melanoma cell lines and in FGF2-free culture medium for the uveal melanocytes. Cells were used to seed 24-well plates in triplicate at a density of 35,000 cells/well for NUMs and 15,000 cells/well for the melanoma cells. We assessed the functionality of FGFR by adding FGF2 chemically coupled to saporin (CCF₂S) or saporin alone at the indicated concentrations 12 hours before cell stimulation and on day 0 of the culture period. In some experiments, we added protamine sulfate (Sigma-Aldrich) and suramin (a kind gift from Bayer, Leverkusen, Germany)—known to inhibit the autocrine stimulation of cell proliferation by FGFs—to cell cultures at stimulation and on day 3 of the culture period. We investigated the role of FGFR1 in cell proliferation by adding the FGFR1 tyrosine kinase inhibitor SU5402 (Calbiochem) at the indicated concentration 12 hours before cell stimulation and on day 0 of the culture period. The autocrine role of FGF2 in cell proliferation was investigated by adding an anti-FGF2 neutralizing monoclonal antibody (clone bFM-1; Upstate Biotechnology, Souffelweyersheim, France) at the indicated concentrations 12 hours before cell stimulation and on day 0 of the culture period. We quantified FGF2 secretion in uveal melanoma cell cultures by collecting conditioned medium after 72 hours of culture in serum-free medium containing 0.5% bovine serum albumin and then determining the amount of FGF2 (18-kDa FGF2 isoform) with ELISA, performed according to the kit manufacturer’s instructions (R&D Systems). 

Cell proliferation was assessed by determining the number of viable cells every 3 days by the MTT (3-(4,5-dimethyl thiazol-2-yl)-2,5-diphenyl tetrazolium bromide) colorimetric method. The percentage of growth inhibition was determined relative to DMSO-treated control cells. 

Cell Apoptosis

Cell cycle analyses were performed by propidium iodide (PI) staining after 24 hours in culture, to determine DNA content. The cells were washed in PBS and fixed by incubation in ice-cold 70% ethanol for 2 hours at 4°C. They were rehydrated in cold PBS, treated with 1 mg/mL RNase A and stained by incubation with 50 mg/mL PI for 15 minutes at 4°C. The stained cells were analyzed by flow cytometry (Epics ALTRA, Beckman, France). Apoptotic cells were detected in the sub-G₁ peak. 

Western Blot Analysis

NUMs and uveal melanoma cells were washed twice in PBS, lysed in ice-cold lysis buffer (20 mM Tris-HCl [pH 7.5], 100 mM NaCl, 1 mM EDTA, 0.5% NP-40, 0.5% deoxycholate, 50 mM β-glycerophosphate, 0.2 mM sodium orthovanadate, 50 mM sodium fluoride, 1 μg/mL leupeptin, 1 μg/mL pepstatin, 1 μg/mL aprotinin, 1 mM PMSF, and 2 μg/mL antipain) and centrifuged at 4°C for 10 minutes at 10,000g. Protein concentrations were determined with the Bradford method. Cell lysates were mixed with 3× Laemmli buffer and heated for 5 minutes at 95°C. They were then resolved by SDS-PAGE (10% or 15% polyacrylamide gels) and transferred to PVDF membrane (Immobilon; Millipore France, Molsheim, France) by electroblotting. Membranes were probed with polyclonal antibodies directed against FGF2 (dilution 1:100, Oncogene Research Products/Calbiochem) or FGFR1 (dilution 1:1000, Cell Signaling Technology, Beverly, MA). Polyclonal antibodies directed against phospho-FGFR1 (Y653/Y654) and phospho-ERK1/2 (T202/Y204) (1:1000, Cell Signaling Technology) were used to analyze the activation of these kinases during cell proliferation. The membranes were probed with a rat monoclonal antibody directed against α-tubulin (1:1000; AbD/Serotec, Oxford, UK) to control for equal loading. These primary antibodies were tagged with specific secondary horseradish peroxidase–conjugated antibodies. Antibody complexes were detected by enhanced chemiluminescence (ECL; GE Healthcare, Clermont-Ferrand, France), with the membrane placed against film (BioMax Light-1; Eastman Kodak, Chalon-sur-Saône, France). 

Oligonucleotides and Oligonucleotide Treatment of the Cells

Phosphorothioate sense and antisense oligonucleotides (ODNs) directed against FGF2 were designed according to the published sequence of the human FGF2 gene. ³⁵ The first 15-mer antisense ODN (5′-GGC-TGC-CAT-GGT-CCC-3′, referred to as AS FGF2) was directed against the translation start site (AUG codon) and surrounding nucleotides. AS FGF2 inhibits the synthesis of all endogenous FGF2 isoforms. The corresponding sense ODN was 5′-GGG-ACC-ATG-GCA-GCC-3′ (referred to as S FGF2). We used lipofectin (Invitrogen), a cationic lipid, to deliver the ODNs, because this method results in high levels of ODN uptake and stability in the intracellular compartment without affecting the final nuclear location of the ODN after endocytosis and release from the endocytic compartment. All ODNs were synthesized commercially (Life Technology-Invitrogen) and purified by HPLC. Lipofectin-ODN complexes were produced according to the manufacturer’s instructions. All results were obtained with 2.5 μM ODNs and 7 μg/mL lipofectin. The cells were cultured in medium supplemented with heat-inactivated calf serum for 24 hours and were then transfected for 4 hours in a serum-free medium containing ODNs/lipofectin. The cells were washed twice and fresh complete culture medium containing the appropriate concentration of ODNs was added. The cells were cultured for 3 days. 

RT-PCR Analysis

Total RNA was extracted from uveal melanoma cells with a single-step method (Trizol; Invitrogen) and treated with amplification grade DNase I at room temperature. We extracted mRNA directly from the NUMs, using an mRNA isolation kit (μMACS; Miltenyi Biotec SAS, Paris, France). We subjected 1 μg of total RNA (for uveal melanoma cells) and 100 ng of mRNA (for NUMs) to reverse transcription for 1 hour at 42°C, using oligo-dT primers and Moloney murine leukemia virus reverse transcriptase. We then analyzed expression of the FGF1, FGF2, and FGFR1–4 genes by amplifying the cDNA obtained by semiquantitative PCR for 27 to 35 cycles. β-Actin levels, determined after 22 cycles of amplification, were used to control for loading and constant expression. The primers used in this study are summarized in Table 2 . 

Results

Expression of FGF1, FGF2, and FGFR1–4 by Primary Uveal Melanomas

Because we thought that FGF2 might be involved in uveal melanoma growth, we began by using RT-PCR to analyze the expression of FGF1, FGF2, and their receptors FGFR1–4 in a first set of nine primary uveal melanomas (patients 1–9). Seven of them expressed FGF2 (Fig. 1A) . FGF1 was detected in both FGF2-negative tumors and in five of the seven tumors that expressed FGF2 (Fig. 1A) . All but two of the tumors expressing FGF2 also expressed FGFR1. Of the seven FGFR1-positive tumors, six coexpressed FGFR2, five FGFR3, and four FGFR4. Three primary uveal melanomas were positive for all four FGFRs, as well as for FGF1 and -2. These findings demonstrate that FGF and FGFR are combined in various ways in primary uveal melanomas and thus suggest the possible involvement of FGF/FGFR in uveal melanoma. Then, we analyzed the expression of FGF2 protein in a second set of nine other primary uveal melanomas (patients A–I; Fig. 1B ). Western blot showed the expression of FGF2 in these nine melanomas at the protein level. Note that seven of nine primary tumors expressed both the 18-kDa and the HMW (22–24-kDa) FGF2 isoforms, whereas the other two expressed only the 22-kDa FGF2 isoform. 

Exogenous FGF2 as a Mitogenic Factor for Normal Uveal Melanocytes and Uveal Melanoma Cells

Although Hu et al. ^{30 31} demonstrated that NUMs require the 18-kDa FGF2 isoform to proliferate in vitro, the effects of this growth factor (and other mitogenic factors) on proliferation of uveal melanoma cells remain unknown. We used primary cultures of NUM and five primary uveal melanoma tumor-derived cell lines to characterize the effects of FGF1, -2, and -5; VEGF₁₆₅; PDGF-BB; and EGF on cell proliferation. We first confirmed that the addition of FGF2 induced NUM proliferation. Rising concentrations of FGF2 (0.01–50 ng/mL) increased NUM proliferation dose dependently in cell cultures containing serum, IBMX, and cholera toxin (Fig. 2A) . The mitogenic activity of FGF2 was at is maximum at a concentration of 20 ng/mL FGF2, and the EC₅₀ of FGF2 was 3 ng/mL after 3 days of culture (Fig. 2B) . By contrast, replacing FGF2 with FGF1 or FGF5 (20 ng/mL each) resulted in very slight stimulation (Fig. 2C) . Heparin (10 μg/mL), which is required for FGFs to bind to their tyrosine kinase receptors and for full FGF-induced cell proliferation in some cell types, ^{22 23 24} did not increase the mitogenic activity of FGF2, -1, or -5 in NUMs (Fig. 2C) . The failure of VEGF₁₆₅, PDGF-BB, and EGF (20 ng/mL each), either alone or in combination, to induce NUM proliferation after 3 days of culture (Fig. 2Cand data not shown), highlights the specific mitogenic effect of FGF2 on NUMs. 

FGF2 is also required for the proliferation of normal cutaneous melanocytes, and downregulation of FGFR1 inhibits their FGF2-induced proliferation. ³⁸ When we treated NUMs with SU5402, a specific inhibitor of the tyrosine kinase activity of FGFR1, ³⁹ FGFR1 activation (Fig. 2D) , and FGF2-induced cell proliferation both decreased substantially in NUM cultures (Fig. 2E) . These findings demonstrate the key role of FGFR1 activation in mediating FGF2-induced cell proliferation in uveal melanocytes. 

Next, we investigated the mitogenic effects of FGF1, FGF2, FGF5, VEGF₁₆₅, PDGF-BB, and EGF on cell proliferation in uveal melanoma cell lines. We used five uveal melanoma cell lines, which express different combinations of B-Raf genotypes and SCF/c-Kit autocrine loop to investigate the effects of activating mutation in B-Raf (Mel270 and 92.1 are ^WTB-Raf melanoma cell lines, and OCM-1, MKT-BR, and SP6.5 are mutant ^V600EB-Raf melanoma cell lines) and activated SCF/c-Kit autocrine loop (three SCF/c-Kit melanoma cell lines: SP6.5, Mel270, and 92.1) on the role of FGF/FGFR in uveal melanoma (Table 1) . ^{10 16 19} As previously observed in NUMs, we found little or no effect of FGF1 and -5, VEGF₁₆₅, PDGF-BB, or EGF on melanoma cell proliferation, even at the high concentration of 100 ng/mL (Fig. 3Aand data not shown). All five uveal melanoma cell lines, however, were responsive to exogenous FGF2 (20 ng/mL), although not as responsive as NUM cells. Treatment with FGF2 increased uveal melanoma cell proliferation by only 8% to 19%, compared with untreated melanoma cells (Fig. 3A) . 

We wondered whether this limited effect might be overcome by higher FGF2 concentrations. Cells were treated with FGF2 at a concentration of 100 ng/mL, alone or in combination with heparin at 10 μg/mL, a concentration known to increase FGF mitogenic activity in some types of cells. ²³ Proliferation was not enhanced by either treatment (Fig. 3) . This result strongly suggests that the weak stimulatory effect of FGF2 on proliferation of melanoma cells compared with NUMs is not due to its low concentration. Of note, responsiveness to FGF2 was equally low in ^WTB-Raf and ^V600EB-Raf cells (Fig. 3) . The presence of the activating V600E mutation of B-Raf apparently did not influence cell response to FGF2 stimulation. 

Expression of a Functional FGF2/FGFR System in Uveal Melanoma Cells

A possible explanation for the reduced sensitivity to FGF2 of melanoma cells compared with NUMs is that FGFR expression is altered in melanoma cells. This hypothesis was tested by RT-PCR analysis of the expression of all four FGFR isoforms in NUMs and in uveal melanoma cell lines. We used the ARPE-19 cell line as a positive control, because it expresses all four FGFRs. ⁴⁰ Transcripts were detected for every FGFR isoform except FGFR2 in the NUMs and in all five melanoma cell lines (Fig. 4A) . Less FGFR3 was expressed in the melanoma cells than in the NUMs, however (Fig. 4A) . 

We then used FGF2-saporin (CCF₂S) to test the functionality of the FGF/FGFR system—that is, the capacity of cell surface FGFR to bind and internalize FGF2. CCF₂S, a chemical conjugate of FGF2, is both a cytotoxin and a mitotoxin and is taken up by cells after binding to functional FGFRs, whereas exogenous saporin alone cannot enter the cells and thus has no effect on them. ^{41 42} CCF₂S treatment of ^V600EB-Raf uveal melanoma cells reduced cell proliferation in a concentration-dependent manner (30%–54% inhibition with 5 ng/mL CCF₂S and 74%–86% with 20 ng/mL CCF₂S), whereas treatment with 50 ng/mL saporin alone had no cytotoxic effect over a 3-day culture period (Fig. 4B) . The effects of CCF₂S on ^WTB-Raf uveal melanoma cells were similar: 27%–63% inhibition with 5 ng/mL CCF₂S and 77%–95% with 20 ng/mL CCF₂S; Fig. 4B ). CCF₂S treatment of SP6.5, Mel270 and 92.1 (which had the lowest levels of FGFR3 mRNA) induced cytotoxic effects similar to those observed with the other two melanoma cell lines and thus suggests that low levels of FGFR3 expression do not affect FGF2 binding to FGFRs or its internalization (Fig. 4B) . Similar results were obtained with the FGF2-saporin (FPF₂S) fusion protein (data not shown). All five melanoma cell lines expressed functional FGF2 receptors. It thus appears that the presence of the V600E mutation of B-Raf influenced neither the expression of FGF/FGFR nor the functionality of FGFR. 

Autocrine Growth Factor Activation Loop in the Proliferation and Transformation of Uveal Melanoma Cells

We then investigated whether autocrine stimulation of melanoma cells by secreted endogenous growth factors makes the cells insensitive to further stimulation with exogenous FGF2 and is thus responsible for the weakness of its mitogenic effect. Uveal melanoma cell lines were treated with 10 μg/mL protamine sulfate, a basic molecule that inhibits paracrine mitogenic stimulation of several types of neoplastic cells at this concentration by preventing growth factors from binding to their cell surface receptors. ⁴³ Protamine sulfate reduced melanoma cell proliferation by approximately half, regardless of the mutational status of B-Raf (Fig. 5A) . Results were similar for ^V600EB-Raf and ^WTB-Raf uveal melanoma cells treated with suramin (50 μg/mL; Fig. 5A ), an antitumor agent that also blocks the paracrine mitogenic effects of many growth factors, including FGF2. Our data therefore suggest that uveal melanoma cells secrete growth factors and that these stimulate cell proliferation through an autocrine activation loop. We also tested whether suramin affects the ability of cells to proliferate under anchorage-independent conditions and to form colonies in soft agar, a characteristic associated with cell transformation. Cell treatment with 50 μg/mL suramin decreased colony formation independent of the presence of the activating V600E mutation in B-Raf (inhibition by 54%–85% in ^V600EB-Raf Raf melanoma cells and by 64 to 78% in ^WTB-Raf melanoma cells; Fig. 5B ). Taken together, these data indicate that stimulation of uveal melanoma cells by endogenous growth factor is involved in both proliferation and transformation. 

Involvement of the FGF2/FGFR1 Autocrine Loop in Uveal Melanoma Cell Proliferation, Transformation, and Survival

To determine whether FGF2 participates in this autocrine activation loop, we first investigated FGF2 expression in uveal melanoma cell lines. RT-PCR analysis showed that all five uveal melanoma cell lines expressed FGF2 (Fig. 6A) . Western blot and ELISA further showed that all the uveal melanoma cell lines, as well as NUM, produced (Fig. 6B)and secreted (Table 1)FGF2. All five melanoma cell lines, but not NUMs, produced the HMW FGF2 isoforms (220 and 24 kDa; Fig. 6B ). OCM-1, MKT-BR, and Mel270 secreted the largest amounts of 18-kDa FGF2 (Table 1) . Note that the lack of any relation between the quantity of secreted FGF2, expression of mutant ^V600EB-Raf, and expression of the SCF/c-Kit autocrine loop suggests that these events are independent. 

We next treated cell cultures with a monoclonal anti-FGF2 immunoneutralizing antibody to inhibit activation of any FGF2 autocrine loop in the uveal melanoma cells. Neutralization of FGF2 in the culture medium by an anti-FGF2 immunoneutralizing antibody reduced cell proliferation by similar percentages in ^V600EB-Raf (inhibition of 59%–88%) and ^WTB-Raf (inhibition of 87%–88%) melanoma cells (Fig. 7A)and thus suggests that FGF2 secreted by uveal melanoma cells plays a key role in their autonomous proliferation. Immunoneutralization of FGF2 also reduced the number of cell colonies in soft agar (reduction of 13%–28% in ^V600EB-Raf and of 19%–24% in ^WTB-Raf melanoma cells), but to a lesser extent than it reduced cell proliferation (Fig. 7B) . Flow cytometry analysis also showed that uveal melanoma cells underwent apoptosis (sub-G₁ DNA apoptotic peak) after a 3-day period of treatment with the anti-FGF2 immunoneutralizing antibody (Fig. 7C) . Neutralization of FGF2 in the culture medium induced apoptosis in ^V600EB-Raf (increase of 27%–29%) and ^WTB-Raf (increase of 13%–45%) melanoma cells. Control experiments with nonimmune antibody at the same concentration did not affect either cell proliferation or cell transformation (data not shown). 

All five uveal melanoma cell lines produced FGFR1 (Fig. 6B) . To evaluate the role of this receptor in activation of the FGF2 autocrine loop, we used SU5402 to inhibit FGFR1 activation. This treatment reduced the rate of cell proliferation by 67% to 76% in ^V600EB-Raf and by 47% to 55% in ^WTB-Raf melanoma cells (Fig. 7A) . Inhibition of FGFR1 also reduced the number of uveal melanoma cell colonies in soft agar to a similar extent in ^V600EB-Raf (inhibition of 27%–49%) and ^WTB-Raf (inhibition of 35%–41%) melanoma cells (Fig. 7B) . Flow cytometry analysis showed that inhibition of FGFR1 also increased apoptosis of uveal melanoma cells after 3 days of treatment (Fig. 7C) . This treatment increased apoptosis by 34%–37% in ^V600EB-Raf and by 21%–56% in ^WTB-Raf melanoma cells (Fig. 7C) . 

Taken together, these data demonstrate that the FGF2/FGFR1 autocrine activation loop plays a key role in controlling uveal melanoma cell proliferation and survival. 

The ERK1/2 Intracellular Signaling Pathway in the FGF2/FGFR1 Autocrine Activation Loop in Uveal Melanoma Cells

The intracellular HMW FGF2 isoforms induce cell proliferation and transformation by molecular mechanisms distinct from, but complementary to, those of the 18-kDa FGF2-secreted isoform. ^{21 44} To verify that the inhibition of cell proliferation that we observed was due to depletion of all FGF2 isoforms, we inhibited the synthesis of all endogenous FGF2 isoforms by applying an AS ODN strategy. We confirmed that FGF2 AS ODN-transfected OCM-1 expressed significantly smaller amounts of the various FGF2 isoforms (Figs. 8A) . The reduction of FGF2 expression mediated by FGF2 AS ODN decreased cell proliferation by 55%, whereas treatment with the FGF2 sense (S) ODN had no marked effect on FGF2 levels and did not decrease cell proliferation below the level of lipofectin-treated control cells (Figs. 8A 8B) . The absence of any difference in the inhibition by FGF2 AS ODN of proliferation of OCM-1 ^V600EB-Raf melanoma cells and of Mel270 ^WTB-Raf melanoma cells (Fig. 8B)confirms the results of the cells treated with anti-FGF2 immunoneutralizing antibody. The FGF2 AS ODN-mediated reduction of FGF2 expression decreased the formation of uveal melanoma cell line colonies by 37% to 51%, whereas treatment with the FGF2 S ODN had no marked effect on cell transformation, compared with lipofectin-treated control uveal melanoma cells (data not shown). 

ERK1/2 is constitutively activated in primary tumors, and this activation plays a key role in ^V600EB-Raf melanoma cells, due to the constitutive activation of the B-Raf mutant, and in ^WTB-Raf melanoma cells. ^{10 15 16 17 19} We thus speculated that the FGF2/FGFR1 autocrine activation loop is responsible for ERK1/2 activation in ^WTB-Raf melanoma cells. To test this theory, we used Western blot analysis to assess the effect of FGF2 depletion on ERK1/2 phosphorylation status. We confirmed the constitutive activation of ERK1/2 in control FGF2 S ODN-treated uveal melanoma cells (Fig. 8A) . The FGF2 AS ODN-mediated depletion of endogenous FGF2 significantly reduced the levels of ERK1/2 phosphorylation in both OCM-1 ^V600EB-Raf and Mel270 ^WTB-Raf melanoma cells (Figs. 8A 8C) , thereby demonstrating that FGF2 activates ERK1/2 in uveal melanoma cells, regardless of the mutational status of B-Raf. Finally, to verify that FGFR1 activation mediated the FGF2-induced activation of ERK1/2, we inhibited FGFR1 with SU5402. This reduced ERK1/2 activation in both ^V600EB-Raf and ^WTB-Raf melanoma cells (Fig. 8D)and thereby confirmed that FGFR1 transduces proliferation signaling by activating ERK1/2. Taken together, these data demonstrate that activation of FGFR1, mediated by secreted FGF2, induces ERK1/2 activation and plays a key role in the proliferation of uveal melanoma cells. 

Discussion

The B-Raf Mutation and Growth Factor Autocrine Activation Loop Have Cooperative Rather Than Mutually Exclusive Effects

Activating mutations in B-RAF and RAS are associated with approximately 60% and 20% of cutaneous melanomas, respectively. ^{45 46 47} These mutations are largely responsible for the constitutive activation of ERK1/2 and for the subsequent acquisition of growth signal autonomy in human cutaneous melanoma. ^{47 48 49} That this is absolutely not the case for uveal melanoma, despite the similarity of the histologic features of cutaneous and uveal melanomas, ^{6 7 8 9 10 11 12 13 14} suggests that other molecular mechanisms activate the Ras/Raf/MEK/ERK signaling pathway in the latter. Autocrine loops for growth factors may well be responsible for the acquisition of autonomous proliferation in these tumors. Insulin-like growth factor-1 receptor (IGF-1R) is expressed in uveal melanoma. ⁵⁰ Inhibition of IGF-1R kills uveal melanoma cell lines and inhibits melanoma growth in an animal model of uveal melanoma. This finding suggests that an IGF/IGF-1R autocrine loop induces the proliferation of uveal melanoma cells. ^{51 52} Moreover, activation of the hepatocyte growth factor (HGF)/c-Met axis stimulates the migration of uveal melanoma cells and thus suggests that an HGF-dependent signaling loop plays a role in the migration and invasiveness of uveal melanoma. ⁵³ We recently demonstrated that the SCF/c-Kit/ERK1/2 autocrine loop is activated and contributes to cell proliferation and transformation in uveal melanoma cells that express either ^WTB-Raf or ^V600EB-Raf and thus showed that the B-Raf mutation and SCF/c-Kit autocrine activation loop have cooperative rather than mutually exclusive effects on uveal melanoma cells. ¹⁶  

Imatinib mesylate treatment of uveal melanoma cell lines inhibits cell proliferation. Unfortunately, this drug was recently shown to be ineffective in the treatment of metastatic uveal melanoma. ^{54 55 56} Differences in c-Kit levels and responses to it in uveal melanoma cell lines suggest that other growth factors and chemokines play a role in melanoma cell proliferation. ^{16 50} The large quantity of FGF2 secreted by most primary uveal melanomas and uveal melanoma cell cultures ^{26 27 28} suggests that an FGF2/FGFR autocrine activation loop exists in these cells. Because our intent was precisely to investigate the possible existence and role of such a loop, we assessed the specific levels of expression of the various FGFRs. We found that FGF2 and FGFR1 are expressed in substantial quantities in primary uveal melanomas and that all five uveal melanoma cell lines studied secreted FGF2. Using complementary strategies, including AS ODN-mediated FGF2 depletion, FGF2 immunoneutralization and pharmacologic inhibition of the tyrosine kinase activity of FGFR1, we demonstrated that the FGF2/FGFR1 autocrine activation loop is involved in autonomous cell proliferation in uveal melanoma cells. 

A recent study found a higher prevalence of activated mutant ^V600EB-Raf in uveal melanoma than previously reported. ⁵⁷ This finding indicates the importance of ours that the ^V600EB-Raf mutation in uveal melanoma cells does not prevent activation of this autocrine loop: Expression of both FGF2 and FGFR1 is independent of this mutation. We cannot, however, rule out the possibility of a cooperative effect between the gain-of-function B-Raf mutation and the FGF2/FGFR1 autocrine activation loop in the acquisition of growth signal autonomy in uveal melanoma cells. Consistent with this hypothesis, we observed that AS ODN-mediated FGF2 depletion and inhibition of the tyrosine kinase activity of FGFR1 decreased ERK1/2 phosphorylation levels by only approximately 50% in ^V600EB-Raf uveal melanoma cells. However, inhibition of the tyrosine kinase activity of FGFR1 also cut ERK1/2 phosphorylation levels approximately in half in ^WTB-Raf uveal melanoma cells, suggesting that other complementary mechanisms are involved in ERK1/2 activation in these cells. 

NUMs Secrete FGF2 but Need Exogenous FGF2 for Proliferation

FGF2 expression and secretion were also observed in NUMs, which expressed FGFR1, -3, and -4 on their surfaces. An FGF2/FGFR autocrine activation loop may therefore be possible in these cells. However, inducing proliferation of these cells required the exogenous 18-kDa FGF2 isoform. This finding is consistent with reports that normal cutaneous melanocytes expressing the 18-kDa FGF2 isoform nonetheless require exogenous FGF2 to grow. ⁵⁸ This phenomenon has not yet been clearly elucidated. Our finding that NUMs expressed only the secreted 18-kDa FGF2 isoform while uveal melanoma cell lines and primary tumors of uveal melanoma expressed both the 18-kDa isoform and the HMW FGF2 isoforms suggests that the latter may be important in the induction of cell proliferation. Transfection of cells with HMW FGF2 isoforms from pancreatic acinar cancer cells that express FGFRs but not FGF2 enables the transfected cells to grow in serum-free medium and thereby confirms the key role of the intracellular HMW FGF2 isoforms. ⁵⁹ The expression of HMW FGF2 isoforms may therefore increase the responsiveness of uveal melanoma cells to the 18-kDa form, as it appears to do in pancreatic cancer cells. The specific effects of the HMW isoforms depend, however, on cell type and context. ⁴⁴ Expression of FGF2 with a construct containing all FGF2 isoforms induced proliferation, migration, and cluster formation in a model of skin reconstructs, thereby suggesting that endogenous FGF2 isoforms participate in mechanisms involved in the progression from cutaneous melanocytes to melanoma. ⁶⁰ We did not, however, observe that NUM stimulation by the exogenous 18-kDa FGF2 isoform induced NUM migration or clusters of NUM in culture (Mascarelli F, unpublished findings, 2008), although it did induce their proliferation. 

The capacity of the 18-kDa FGF2 isoform to bind to the cell surface regulates FGF2 biological activity. ^{22 23 24} HSPGs control both the stability of FGF2 and its binding to FGFRs. Downregulation of perlican, an HSPG, reduces FGF2 binding to cell surfaces and thus FGF2 mitogenic activity in cutaneous melanoma cells. ⁶¹ By analogy, degradation of HSPGs abolishes ERK1/2 activation and proliferation in response to the 18-kDa FGF2 isoform in metastatic cutaneous melanoma. ⁶² Therefore, differences between NUM and uveal melanoma cells in the cell surface HSPGs may also explain why NUMs require the addition of exogenous 18-kDa FGF2 to proliferate. 

Moreover, FGFR1 was present in two bands in the three uveal melanoma cell lines, SP6.5, Mel270, and 92.1, but only in the HMW band for the other two melanoma cell lines OCM-1 and MKT-BR. In NUM, it was present only in the lower molecular weight. These differences may be due to posttranscriptional and posttranslational modifications of FGFR1 and lead to differences in the affinity for FGF2 for FGFR1 and thus in FGF2 mitogenic activity. 

Is the ERK1/2 Activation That Is Mediated by the FGF2/FGFR1 Autocrine Activation Loop a Key Signaling Pathway Common to Uveal and Cutaneous Melanomas?

Our data on the gain of function of FGF2 autocrine loops in uveal melanoma cells seem to conflict with data showing that the V600E mutation in B-Raf is the major oncogenic event in the transformation of cutaneous melanocytes. ^{48 49} It was recently shown, however, that FGF2 is not essential for cutaneous melanoma formation in a transgenic mouse model. ⁶³ Transgenic mice expressing dominant-active N-Ras, which are deficient for FGF2, do not differ from FGF2 wild-type mice in their development of cutaneous metastasizing melanoma. This observation suggests that FGF2 does not play a role in activated Ras-expressing cutaneous melanoma. Because FGF2 activates the Ras/ERK signaling pathway in numerous cells, including cutaneous melanocytes, retinal pigmented epithelial cells, and choroidal vascular endothelial cells, ^{46 63 64 65 66} the question is whether there is a potential overlap between FGF2 signaling and activated N-Ras signaling, essentially through the activation of ERK1/2. Because of the redundancy in FGF utilization, other FGF family members may compensate for the lack of FGF2 during tumorigenesis, which would explain why FGF2-null mice are viable, cannot be distinguished from their wild-type littermates by gross examination, survive to adulthood, and are fertile. ⁶⁷  

Moreover, high levels of FGF2 and FGFR1 mRNA have been detected in cutaneous melanomas. ^{20 68 69} In contrast, the failure to detect FGF2 in normal cutaneous melanocytes suggests that the FGF2 activation autocrine loop may be one of the molecular mechanisms involved in the growth of human cutaneous melanoma in the absence of activating RAS and B-RAF mutations. Long before the V600E activating mutation in B-Raf was identified, proliferation in cutaneous melanoma cell lines was found to be inhibited by treatment with AS ODNs that target FGF2 and FGFR1 mRNA. ^{31 70} AS ODN-mediated depletion of FGF2 reduced cell proliferation in the WM75, WM983-A, and 983-B cutaneous melanoma cell lines, ⁷⁰ later shown to express ^V600EB-Raf. ⁴⁵ Retrospectively, these data demonstrate that expression of ^V600EB-Raf and that of FGF2/FGFR1 are not mutually exclusive in cutaneous melanoma cells and confirm our observations of growth factor autocrine activation loops, irrespective of whether the activating V600E mutation in B-Raf is present. Our hypothesis is also supported by a recent study showing that ERK1/2 activation is reduced by an AS FGF2-expressing adenovirus in the 1205Lu cutaneous melanoma cell line that expresses ^V600EB-Raf. ⁴⁵ It has also been shown that exogenous FGF2 may rescue cutaneous melanoma cells from the apoptosis induced by siRNA-mediated ^V600EB-Raf knockdown. ⁷¹ This implies that B-Raf/ERK1/2 mediates cell survival and that exogenous FGF2 may rescue cells from apoptosis by activating ERK1/2. The antiapoptotic role of exogenous FGF2 in cutaneous melanoma cells is consistent with our data, which show that pharmacologic blockade of the tyrosine kinase activity of FGFR1 induces massive apoptosis in uveal melanoma cells. Overall, these data are consistent with a cooperative effect between B-Raf signaling and the FGF2/FGFR1 autocrine activation loop in uveal and cutaneous melanoma cells. 

We found that suramin and conjugated FGF2-saporin reduced uveal melanoma cell proliferation rapidly. This combination inhibits the proliferation mediated by the FGF autocrine activation loop in a murine model of cutaneous melanoma. ⁷² A combination therapy including FGF-saporin together with nonspecific compounds, such as suramin or specific inhibitors of proliferation/survival signaling pathways, may therefore be effective in the treatment of uveal melanoma. 

In conclusion, in this study that stimulation of FGFR1 played a role in the proliferation and survival of uveal melanoma cells. Constitutive ERK1/2 activation played an important role in the FGF2/FGFR1 autocrine loop that leads to the acquisition of autonomous growth of uveal melanoma cells. The FGF/FGFR/ERK signaling pathway may therefore have important implications for the design of specific therapeutic strategies to control the growth and progression of uveal melanoma. 

⁴  Present affiliation: Department of Pathology, University of Michigan, Ann Arbor, Michigan.

⁵  Contributed equally to the work and therefore should be considered equivalent authors.

 Supported by a grant from the Ministère Français des Affaires Etrangères (NB).

 Submitted for publication June 4, 2008; revised September 12 and October 14, 2008; accepted January 19, 2009.

 Disclosure: G. Lefèvre, None; N. Babchia, None; A. Calipel, None; F. Mouriaux, None; A.-M. Faussat, None; S. Mrzyk, None; F. Mascarelli, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

FGF2 and FGFR Expression and FGF2 Secretion in Normal Melanocytes and Uveal Melanoma Cell Lines That Express ^WTB-Raf and ^V600EB-Raf

Cells	SCF/c-Kit Axis	FGF2/FGFR-1 Axis					B-Raf Mutation
				FGF2	FGFR	FGF2 Secretion (pg/mL)
NUM	−	+	+	49.4	−
OCM-1	−	+	+	216.3	V600E
MKT-BR	−	+	+	114.6	V600E
SP6.5	+	+	+	6.8	V600E
Mel270	+	+	+	143.6	−
92.1	+	+	+	34.9	−

 RT-PCR amplification and genomic DNA sequencing were used to analyze the expression of FGF2 and the various FGFR isoforms in NUM and uveal melanoma cell lines. FGF2 secretion into the culture media of NUM and uveal melanoma cells was analyzed by ELISA. FGF2 secretion into the cell-conditioned culture medium was quantified after 72 hours of culture and is expressed in picograms/milliliter of culture medium. +, presence; −, absence.

Table 2.

 

Primer Sequences Used to Amplify the Transcripts

Transcript	Primer	Sequence	PCR Fragment (bp)	Annealing (°C)	Reference
FGF2	Forward	5′-GCTCTTAGCAGACATTGGAAG-3′	375	55	Sakai et al. 36
	Reverse	5′-GTGTGTGCTAACCTTACCT-3′
FGFR1	Forward	5′-AACCTGCCTTATGTCCAGATCT-3′	215	56	Tartaglia et al. 37
	Reverse	5′-AGGGGCGAGGTCATCACTGC-3′
FGFR2	Forward	5′-GGCTGCCCTACCTCAAGGTTC-3′	210	60
	Reverse	5′-AGTCTGGGGAAGCTGTAATCTC-3′
FGFR3	Forward	5′-GCACACCCTACGTTACCGTG-3′	208	60
	Reverse	5′-GCCTCGTCAGCCTCCACCAG-3′
FGFR4	Forward	5′-GTTTCCCCTATGTGCAAGTCC-3′	269	60
	Reverse	5′-GCGCTGCTGCGGTCCATGT-3′
β-Artin	Forward	5′-AGGAGAAGCTGTGCTACGTC-3′	470	56	Lefèvre et al. 16
	Reverse	5′-AGGGGCCGGACTCGTCATAC-3′

Figure 1.

 

Expression of FGF1 and -2 and FGFR1, -2, -3, and -4 in primary tumors of uveal melanomas. (A) The expression of FGF1 and -2 and of the four isoforms of FGFR in a set of nine primary human uveal melanomas was analyzed by RT-PCR. (B) The expression of the FGF2 isoforms was analyzed by Western blot in another set of nine primary human uveal melanomas. Therefore, the Western blot analysis is not a direct validation of the RT-PCR results.

Figure 2.

 

Effects of FGF2 stimulation and FGFR1 inhibition in NUMs. (A) Effect of various concentrations of FGF2 on the proliferation of NUMs over a 4-day culture period. (B) The maximum mitogenic effect of FGF2 (100%) and the concentration of FGF2 necessary to increase cell proliferation by 50% (EC₅₀) were determined after a 3-day culture period. (C) Comparative effects of cell stimulation with 20 ng/mL FGF1, FGF5, VEGF165, PDGF-BB, EGF, and FGF2, in the presence or absence of heparin (10 μg/mL). (D) Effects of the inhibitor of FGFR1 kinase activity SU5402 on FGFR1 activation. NUMs were treated with SU5402 (20 μM) for 1 hour and FGFR1 phosphorylation levels were analyzed by Western blot analysis. (E) Dose effects of SU5402 on the proliferation of NUM after 3 days of treatment. NUMs were cultured in FIC medium for 3 days, deprived of FGF2 for 12 hours, and stimulated with growth factors (A–C) or treated with the FGFR1 inhibitor for 1 hour (D) or 3 days (E). Cell proliferation was assessed with the MTT colorimetric assay. The results presented are representative of three independent experiments.

Figure 3.

 

Effects of different isoforms of FGF on cell proliferation in uveal melanoma cells. Cells were cultured in RPMI 1640 medium supplemented with 5% FCS for 3 days. Cell cultures were then incubated with RPMI 1640 medium supplemented with 0.5% FCS, in the presence or absence of growth factors at the indicated concentrations and with or without heparin (10 μg/mL) for 3 days. Cell proliferation was investigated with the MTT colorimetric method. (A) The effects of FGF2 (20 or 100 ng/mL) and FGF1, -2, and -5 (100 ng/mL each). (B) The effects of FGF1, -2, and -5 (100 ng/mL each) in combination with heparin (10 μg/mL). Percentage growth induction was calculated in comparison with control FGF2-untreated melanoma cells (100%). The data presented are representative of results in three independent experiments.

Figure 4.

 

Expression of FGFR mRNA and effects of FGF2-coupled saporin on cell proliferation in uveal melanoma cells. Cells were cultured with FIC medium supplemented with 20% FCS (NUM) or in RPMI 1640 medium supplemented with 5% FCS (melanoma cells) for 3 days. (A) The levels of FGFR1–4 mRNA were determined by RT-PCR. We used the ARPE-19 cell line as a positive control, as this cell line expresses all four FGFRs. (B) Uveal melanoma cell lines were treated with FGF2-saporin (CCF₂S) to test the functionality of the FGFRs. The dose-dependent inhibitory effects of CCF₂S on cell proliferation were investigated with the MTT colorimetric method, after a 3-day treatment period. The data presented are representative of three independent experiments.

Figure 5.

 

Proliferation and transformation of uveal melanoma cells involved a growth factor autocrine loop. (A) Effects of protamine sulfate or suramin on proliferation of the various uveal melanoma cell lines. Cells were cultured in RPMI 1640 medium supplemented with 5% FCS and then incubated with RPMI 1640 medium supplemented with 0.5% FCS in the presence or absence of suramin or protamine sulfate at the indicated concentration. The inhibitory effects of protamine sulfate (10 μg/mL) and suramin (50 μg/mL) on cell proliferation were investigated with the MTT colorimetric method after a 3-day treatment period. The percentage of growth inhibition was calculated compared with that in untreated control cells. The data are representative of results in three independent experiments. (B) The capacity of suramin (50 μg/mL) to affect cell transformation was tested under anchorage–independent conditions. Uveal melanoma cells were cultured for 3 days in complete growth medium. Melanoma cells were then resuspended in complete medium containing 0.3% agar and either suramin or vehicle and then plated on a layer of 0.7% agar in complete medium. The plates were incubated at 37°C for 3 weeks. Macroscopic colonies were counted by microscope on day 18. Similar results were obtained in two independent experiments.

Figure 6.

 

Expression of FGF2 mRNA and proteins and of FGFR1 protein in NUMs and uveal melanoma cells. (A) FGF2 mRNA expression was analyzed by RT-PCR in NUMs and five uveal melanoma cell lines. The ARPE-19 cell line was used as the positive control, as this cell line is known to express FGF2. (B) The levels of FGF2 and -1 proteins were determined by Western blot analysis. The data presented are representative of results in three independent experiments.

Figure 7.

 

Effects of FGF2 neutralization and FGFR1 inhibition on the proliferation and death of uveal melanoma cells. (A) The effects of FGF2 immunoneutralization and the inhibition of FGFR1 kinase activity were investigated by culturing uveal melanoma cells in the presence of 0.5% FCS and treating them with an anti-FGF2 immunoneutralizing antibody or the FGFR1 kinase inhibitor SU5402, at the indicated concentrations. Cell proliferation was investigated, with the MTT colorimetric method, after a 3-day treatment period. The percentage of growth inhibition was calculated compared with that of control cells. (B) The capacity of FGF2 neutralization and FGFR1 inhibition to affect cell transformation was tested under anchorage–independent conditions. Uveal melanoma cells were cultured for 3 days in complete growth medium. Melanoma cells were then resuspended in complete medium containing 0.3% agar and either with an anti-FGF2 immunoneutralizing antibody or SU5402, and then plated on a layer of 0.7% agar in complete medium. The plates were incubated at 37°C for 3 weeks. Macroscopic colonies were counted under a microscope on day 18. (C) The effects of FGF2 neutralization and FGFR1 inhibition on uveal melanoma cell death were investigated by fluorescence-activated cell sorting analysis after a 3-day period of treatment with an anti-FGF2 immunoneutralizing antibody or SU5402. Control experiments were performed with a nonimmune antibody at the same concentration in the FGF2 immunoneutralization experiments or with DMSO alone at the same concentration in the FGFR1 inhibition experiments. These treatments did not affect cell proliferation, cell transformation, or apoptosis (data not shown). The data presented are representative of results in three independent experiments.

Figure 8.

 

Effects of FGF2 depletion or FGFR1 inhibition on ERK1/2 activation and cell proliferation in uveal melanoma cells. (A–C) An AS ODN strategy was used to investigate the effects of FGF2 depletion. OCM-1 cells were plated in complete culture medium at 50% to 60% confluence and incubated for 24 hours. The cells were transiently transfected with lipofectin and AS FGF2 ODN (AS) or S FGF2 ODN (S). (A) FGF2 downregulation was assessed 3 days after transfection, by Western blot analysis. (B) Melanoma cell proliferation was investigated in the OCM-1 and Mel270 cell lines with the MTT colorimetric method, after 3 days of treatment. The effect of FGF2 downregulation (C) and of FGFR1 inhibition with 20 μM SU5402 (SU) (D) on ERK1/2 activation was analyzed by Western blot analysis with an antibody directed against phospho-ERK1/2 (T202/Y204) in the OCM-1 and Mel270 cell lines. The data presented are representative of results in three independent experiments.