Neutralizing Antibodies Inhibit the Binding of Basic Fibroblast Growth Factor to Its Receptor but Not to Heparin*

Polyclonal antibodies were prepared against recom- binant basic fibroblast growth factor (bFGF) that reacted only with bFGF but not acidic FGF. These anti- bodies were able to inhibit various biological activities of bFGF such as the ability of bFGF to stimulate DNA synthesis in 3T3 cells, proliferation and migration of bovine capillary endothelial cells (BCEC), and neurite extension in pheochromocytoma (PC12) cells. The anti-bFGF antibodies also inhibited the mitogenic activity of subendothelial cell extracellular matrix for BCEC, demonstrating that the growth factor compo-nent in extracellular matrix required for supporting BCEC proliferation was bFGF. Anti-bFGF antibodies inhibited the cross-linking of bFGF to its high affinity receptor on BCEC cells. However, these antibodies did not inhibit the binding of bFGF to heparin-Sepharose or to the low affinity receptors of BCEC which have been demonstrated to be heparin-like molecules. These results suggest that bFGF has distinct domains for binding to high affinity cellular receptors and for binding to heparin.

Polyclonal antibodies were prepared against recombinant basic fibroblast growth factor (bFGF) that reacted only with bFGF but not acidic FGF. These antibodies were able to inhibit various biological activities of bFGF such as the ability of bFGF to stimulate DNA synthesis in 3T3 cells, proliferation and migration of bovine capillary endothelial cells (BCEC), and neurite extension in pheochromocytoma (PC12) cells. The anti-bFGF antibodies also inhibited the mitogenic activity of subendothelial cell extracellular matrix for BCEC, demonstrating that the growth factor component in extracellular matrix required for supporting BCEC proliferation was bFGF. Anti-bFGF antibodies inhibited the cross-linking of bFGF to its high affinity receptor on BCEC cells. However, these antibodies did not inhibit the binding of bFGF to heparin-Sepharose or to the low affinity receptors of BCEC which have been demonstrated to be heparin-like molecules. These results suggest that bFGF has distinct domains for binding to high affinity cellular receptors and for binding to heparin. bFGF' has numerous biological activities that include the ability to stimulate cell migration, cell proliferation, and cell differentiation (1,2). The biological effects of bFGF are presumably mediated by high affinity cell surface receptors that have been identified, for example, on 3T3 (3), endothelial (4), baby hamster kidney (5), and PC12 cells (6). The biological activity of bFGF might also be mediated by its interaction with heparin or heparin-like molecules. bFGF is a heparinbinding protein that adheres tightly to columns of heparin-Sepharose (7,8) and is found to be associated with heparinlike molecules in subendothelial cell ECM (9,10) and in basement membrane (11). Heparin has been shown to potentiate the activity of aFGF (12) and to protect bFGF from denaturation (12). Endothelial cell-derived heparan sulfate * 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.
$ Supported by National Cancer Institute Grant CA37395 from the National Institutes of Health.
ll Supported by National Cancer Institute Grant CA 37392 from the National Institutes of Health. To whom correspondence should be addressed Children's Hospital, 300 Longwood Ave., Boston, MA 02115. Tel.: 617-735-7503.
The abbreviations used are: bFGF, basic fibroblast growth factor; aFGF, acidic fibroblast growth factor; BCEC, bovine capillary endothelial cells; PC12, pheochromocytoma; ECM, extracellular matrix; EC, endothelial cell(s); PBS, phosphate-buffered saline; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; DMEM, Dulbecco's modified Eagle's medium; HEPES, 4-(2-hydroxyethyl)-lpiperazineethanesulfonic acid. protects bFGF from proteolytic degradation (13). Moscatelli (4,14) has recently demonstrated the existence of low affinity bFGF-binding sites on numerous cell types including capillary endothelial and baby hamster kidney cells which appear to be cell-associated heparin-like molecules. These observations suggest that bFGF binding to heparin-like molecules on the cell surface might be important in regulating cellular responses.
The ability of bFGF to bind to cell surface receptors as well as to heparin suggests the presence of multi-functional binding domains in bFGF. However, it is not clear whether the receptor-binding and heparin-binding domains are distinct or the same. This question is of importance in trying to ascertain the relative roles of cellular receptors and heparin in mediating the various biological activities of bFGF. Synthetic peptides corresponding to different parts of the bFGF molecule have been used to analyze bFGF domains (15,16). Some synthetic peptides, for example bFGF-(24-68) and bFGF-(106-115), were found to bind [3H]heparin and to inhibit lZ5I-bFGF binding to cell surface receptors. However, other synthetic peptides such as bFGF-(121-146) that did bind heparin did not inhibit '251-bFGF binding to surface receptors. The authors (15,16) concluded that there are multiple heparinbinding sites on bFGF, some of which are receptor-binding and some which are not.
We have used another approach to analyze heparin-binding and receptor-binding sites. Highly specific polyclonal antibodies have been developed that neutralize the ability of bFGF to stimulate cellular responses such as cell migration, cell proliferation, and neurite outgrowth. These antibodies will also block the ability of subendothelial cell ECM to support endothelial cell (EC) proliferation. In this report, we demonstrate that these antibodies inhibit the cross-linking of bFGF to high affinity EC receptors. However, these antibodies do not inhibit the ability of bFGF to bind to immobilized heparin or to heparin-like low affinity EC receptors. Therefore, it is possible to discriminate between the receptor-binding and heparin-binding domains of bFGF.

MATERIALS AND METHODS
Production of Polyclonal Antibodies-To produce anti-bFGF antibodies, male New Zealand White rabbits were given injections at multidorsal sites of human recombinant bFGF (140 pg/rabbit) which was emulsified in complete Freund's adjuvant (Difco). The rabbits were boosted after 2 months with subcutaneous injections of bFGF (140 pglrabbit) emulsified in incomplete Freund's adjuvant. To produce anti-aFGF polyclonal antibodies, a peptide representing bovine aFGF (59-90) (17) was conjugated to keyhole limpet hemocyanin and the conjugates were injected into rabbits (18). The rabbits were bled from the ear. Polyclonal antibodies used in this study were collected 7 days after the second booster injection. Non-immunized serum collected from the same rabbit prior to the first injection was used as control serum. The titer of the serum was determined using an enzyme-linked immunosorbent assay in which a 96-well plate (Fal-con) was coated with 50 ng of bFGF or synthetic aFGF peptide/50 p1 phosphate buffered saline (PBS) in each well at 4 "C overnight as described previously (18). Anti-bFGF IgG and normal rabbit sera IgG were purified by protein A-Sepharose chromatography (19). 2 ml of anti-bFGF serum or normal serum was applied to a protein A-Sepharose column (Pharmacia LKB Biotechnology Inc.; 2 ml bed volume) which had been equilibrated with 0.1 M Tris-HC1, pH 7.4. After washing the column with 0.1 M sodium acetate buffer, pH 6.0, bound IgG was eluted with 0.1 M sodium acetate buffer, pH 3.5. The peak protein fractions were collected and dialyzed against PBS. Protein concentration was measured by absorbance at 280 nm (A = 1.4 = 1 mg/ml IgG), and the IgG concentration was adjusted to obtain a stock solution of 4 mg/ml.
Iodination of bFGF-Human recombinant bFGF was iodinated by a modification of the method of Bolton and Hunter (20). A reaction mixture containing 6.7 pg (0.37 nmol) of bFGF, 0.61 nmol of monoiodinated 1z51-Bolton-Hunter reagent (2200 Ci/mmol, when first assayed; DuPont-New England Nuclear), and 0.1 M sodium phosphate, pH 8.3, in a total volume of 16 pl, was incubated on ice for 2.5 h. The unreacted Bolton-Hunter reagent was then quenched by adding 300 p1 of 0.2 M glycine and incubating for an additional 30 min on ice. 20 p1 of 1.5% gelatin was added to the vial, and the "'I-bFGF was purified by gel filtration on a column of Sephadex G-25 (LKB Biotechnology Inc., PD-10, 9.1 ml bed volume) equilibrated with 50 mM Tris-HC1,0.3 M NaCl, 0.1% gelatin, and 1 mM dithiothreitol, pH 7.5. The biological activity of "'I-bFGF was retained as determined by its ability to bind to heparin-Sepharose and to stimulate DNA synthesis in Balb/c/3T3 cells and in BCEC. The specific radioactivity of lZ5I-bFGF was 85 nCi/ng, when first assayed. Aliquots of '"I-bFGF were stored at -20 "C in the buffer used for gel filtration.
Cross-linking of '"I-bFGF to BCEC Receptors-BCEC were cultured on gelatinized 100-mm dishes as described previously (21). Cross-linking was performed with subconfluent cell layers (about 1.5 X lo6 cells/lOO-mm dish). Before addition of '"1-bFGF, cells were washed with PBS and preincubated at 4 "C for 10 min in 2 ml of binding buffer (PBS, 0.2% gelatin, pH 7.4) using gentle rotation on an orbital shaking platform. Subsequently, the cells were incubated for 3.5 h in 2 ml of fresh binding buffer containing '"I-hFGF (10 ng/ ml), washed rapidly with cold PBS, and incubated on an orbital shaker at room temperature for 30 min in 2 ml of PBS containing 0.15 mM disuccinimidyl suberate. The disuccinimidyl suberate was prepared immediately prior to addition as a 20 mM stock solution in dimethyl sulfoxide. The excess disuccinimidyl suberate was quenched by addition of 200 pl of 0.2 M glycine, 10 mM Tris-HC1, and 2 mM EDTA, pH 7.5. The solution was decanted, and 0.7 ml of PBS containing 1 mM EDTA and 1 mM phenylmethylsulfonyl fluoride was added to each dish. Cells were scraped from the dish with a Teflon cell scraper (Costar) and transferred to a 1.5-ml microcentrifuge tube. Each dish was washed with an additional 0.7 ml of the same buffer which was then added to the microcentrifuge tube. The samples were centrifuged at 14,000 X g for 1 min, and the cell pellets were solubilized by vortexing in 50 p1 of 10 mM Tris-HC1, 0.5% Nonidet P-40,O.l mM EDTA, and 1 mM phenylmethylsulfonyl fluoride, pH 7.0. Insoluble debris were removed by centrifugation at 14,000 X g for 3 min. Concentrated SDS-PAGE sample buffer (17 pl) was added to each supernatant, and the samples were analyzed by SDS-PAGE on 8% acrylamide gels followed by autoradiography. For neutralization studies 400 pg/ml anti-bFGF was added to the "'I-bFGF at the beginning of the 3.5-h incubation.
Binding of lZ5I-bFGF to Low Affinity Binding Sites-Analysis of low affinity binding sites in BCEC was carried out by a method similar to that of Moscatelli (4). Low affinity binding sites are defined as those sites from which bFGF can be removed by treatment with 2 M NaC1, pH 7.5 (4). On the other hand, high affinity binding sites are defined as those sites from which bFGF can be removed by extraction with acidic buffer, pH 4 (4). Confluent cultures of BCEC in 6.4-mm wells were washed with PBS and incubated for 10 min at 4 "C in binding buffer (PBS, 0.2% gelatin, pH 7.4). The cells were then incubated for 2 h at 4 "C in 100 p1 of binding buffer containing 10 ng/ml lZ5I-bFGF, with or without various concentrations of heparin, anti-bFGF IgG, or normal IgG. At the end of the incubation, low affinity binding of Iz5I-bFGF to cells was analyzed by washing the cells 3 times with PBS and once with 150 p1 of 2 M NaCl and 20 mM HEPES, pH 7.5. The radioactivity in 2 M NaCl extract was determined in a y scintillation counter (Beckman model 5500). All experimental measurements were run in triplicate.
Heparin-Sephurose Chromatography-Heparin-Sepharose chromatography was performed as described previously (22). "'I-bFGF (10 ng/ml) in PBS and 0.2% gelatin was incubated either with anti-bFGF IgG (400 pg/ml), normal IgG (400 pg/ml), or without addition at room temperature for 1.5 h. The reaction mixture (total counts; 6.3 X lo5 cpm/column) was applied to a small heparin-Sepharose column (Pharmacia LKB Biotechnology Inc.; 0.5 ml bed volume) which had been equilibrated with 0.1 M NaC1, 10 mM Tris-HC1, pH 7.5. After application of the sample, the column was washed with 5 ml of 0.6 M NaCl, 10 mM Tris-HC1, pH 7.5, followed by 5 ml of 2.5 M NaCl, 10 mM Tris-HC1, pH 7.5. Fractions of 1 ml were collected, and the radioactivity in each fraction was measured in a y scintillation counter.
Cell Proliferation-The proliferation of BCEC was measured in gelatinized 24-well plates (16-mm wells, Costar) (21,22). BCEC (10' cells/well) were plated in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% calf serum (CS). After attachment of the cells, the media was changed to fresh DMEM/10% CS and bFGF (2 ng/ml) was added. In antibody neutralization studies anti-bFGF IgG (400 pg/ml) or normal rabbit I g G (400 pg/ml) was added at the same time as bFGF. 3 days later, the BCEC were trypsintreated and counted in a cell counter (Coulter Electronics). DNA synthesis was measured in Balb/c/3T3 cells (22,23). Confluent monolayers of 3T3 cells were prepared in 96-well plates (Costar), and after addition of bFGF or aFGF the incorporation of [3H]thymidine into 3T3 DNA was measured in a 36-48-h period. 1 unit of activity was defined as the amount of growth factor needed to stimulate halfmaximal synthesis of 3T3 DNA. For antibody neutralization studies, anti-bFGF IgG (400 pg/ml) or normal rabbit IgG (400 pg/ml) were added just prior to the FGF samples.
Chemotaxis-Cell chemotaxis was measured in a 48-microwell chamber apparatus (Neuroprobe) (24). Test samples of bFGF (0.01-1000 ng/ml) in DMEM containing 1% calf serum, in a total volume of 27 pl, were loaded into each bottom well. Anti-bFGF IgG (400 pg/ ml) was added in neutralization studies. BCEC were prepared by mild trypsin treatment, washed with DMEM/10% CS, collected by centrifugation, and resuspended in 1% calf serum/DMEM. Cell suspensions (7500 cells/50 pl) were loaded into each top chamber of the microwell plate. The apparatus was incubated at 37 "C with 10% CO2 for 4 h. The top (non-penetrating) layer of cells was removed from the membrane (8-pm pores, polyvinylpyrrolidone-free; Nucleopore) using rubber wiper blades and cotton swabs. The migrating cells were fixed by immersing the membrane for 45 min in 10% buffered formalin phosphate (Fisher Scientific) and stained with 0.5% Nuclear Fast Red solution overnight. The membrane was washed with distilled water and mounted on a slide. Cells in the entire area of each well (8-mm2) were counted at magnification X 40.
Neurite Outgrowth"PC12 cells, kindly provided by Dr. S. Rogelj (Whitehead Institute, Cambridge, MA), were seeded (5 X lo4 cells/ dish, 30-mm dish) in DMEM/10% CS, 50 units/ml penicillin, and 25 pg/ml streptomycin. The following day the culture medium was replaced with fresh medium that contained 25 ng/ml bFGF in the absence or in the presence of anti-bFGF antiserum (10 pl of antiserum/l.5 ml medium). Photographs were taken 5 days later with a Nikon camera; magnification was X 100. Extracellular Matrix-6-well cluster dishes (35 mm/well) containing cell-free preparations of subendothelial ECM were generously provided by Dr. Israel Vlodavsky (Jerusalem, Israel) and had been prepared as described previously (9).
bFGF and aFGF-Human recombinant bFGF (25) was a generous gift of the Takeda Chemical Company (Osaka, Japan). It had a specific activity of about 1-5 units/ng in the Balb/c/3T3 DNA synthesis assay described above. SK hepatoma cell-derived bFGF was prepared as described previously (26). Bovine brain bFGF and aFGF (27) were kindly provided by Dr. Patricia D'Amore (Children's Hospital, Boston, MA).

RESULTS
Specificity of Antibody of hFGF"Polyclona1 antiserum directed against human recombinant bFGF was analyzed by SDS-PAGE for the ahility to react with bFGF and aFGF (Fig.  1). Anti-bFGF antibodies reacted with bFGF (lune 4 ) but not aFGF ( h e *?). On the other hand, anti-aFGF antibodies reacted with aFGF (lane I ) but not bFGF (lune 2). Thus, the anti-bFGF antibodies used in the studies to be described below are highly specific in that they cross-react with bFGF, but not aFGF, despite the 55% homology (17) that exists between t,hese two proteins.
Inhihition of bFGF-induced Cell Proliferation by Anti-bFGF Antibody-The ahility of the antibody preparation that crossreacted with bFGF on a Western blot (Fig. 1) to neutralize hFGF-induced DNA synthesis in 3T3 cells and to neutralize RCEC proliferation in a dose-dependent manner was tested (Fig. 2). An IgG fraction of anti-bFGF at 400 pg/ml neutralized the mitogenic activity for 3T3 cells of human recombinant bFGF hy 100% (Fig. 2 A ) and of human hepatoma cellderived bFGF hy 80% (Fig. 2R). The antibody preparation did not, inhibit aFGF at all (Fig. 2C), nor platelet-derived growth factor (not shown), nor any other calf serum-derived

FIG. 1. Specificity of antibodies for bFGF and aFGF: Western blot analysis. Rovine hrain aFGF (lanes 1 and 3) and bovine hrain hFGF (Ianes 2 and 4 ) were electrophoresed on SDS-PAGE and electrophoretically transferred to nitrocellulose. The nitrocellulose sheets were incuhated with either anti-aFGF antiserum prepared against a synthetic peptide corresponding to amino acids 59-90 of aFGF (/ones I and 2) or anti-hFGF antiserum prepared against human recomhinant hFGF (/ones 3 and 4 ) . aFGF and bFGF were visualized as descrihcd under "Materials and Methods."
mitogen for 3T3 cells (not shown). The slightly decreased ability of the antibodies prepared against recombinant bFGF to inhibit hepatoma-derived bFGF probably reflects the structural differences in the two bFGF preparations, the recombinant bFGF being the truncated bFGF-(1-146) (17) and the hepatoma-derived bFGF being the intact bFGF-(1-154) (26).
The antibodies also neutralized the ability of bFGF to stimulate the proliferation of BCEC (Fig. 20).
Inhibition of ECM-induced Cell Proliferation by Anti-bFGF Antibody-Cultured bovine corneal and aortic EC deposit an ECM which replaces the requirement of exogenous FGF for supporting the proliferation of these cells at clonal densities (9). Since subendothelial cell ECM contains bFGF, it was suggested, but not demonstrated, that this growth factor may be one of the ECM components needed to support EC proliferation (9). In the present study, plates containing cell-free preparations of corneal subendothelial cell ECM were incubated with polyclonal antibodies directed against bFGF and BCEC proliferation on these plates was measured (Fig. 3). There was a 14-fold increase of BCEC when grown on subendothelial cell ECM in 5 days (Fig. 3A, lune I). The presence of anti-bFGF IgG decreased ECM-stimulated BCEC proliferation by over 75% (Fig. 3A, lane 3 ) . Normal rabbit IgG showed very little inhibition (Fig. 3A, lune 2). It was concluded that ECM-derived bFGF was required to support BCEC proliferation.
In a parallel study (Fig. 3B), neutralizing anti-bFGF antibody (lune 3 ) , but not normal rabbit antibody (lune 2), inhibited the proliferation of BCEC stimulated by exogenous bFGF to the same extent that they inhibited proliferation of BCEC stimulated by ECM in the absence of exogenous bFGF (Fig.  3A). The antibodies inhibited bFGF-stimulated but not basal BCEC proliferation. The basal proliferation of BCEC in 10% CS and in the absence of exogenous bFGF is about 4-fold in 5 days, i.e. from 1000 cells plated to 4000 cells/dish. The inability of anti-bFGF antibodies to neutralize BCEC basal proliferation is evidence that the antibodies are inhibiting bFGF action rather than being toxic to BCEC.
Inhibition of bFGF-induced Chemotaxis by Anti-bFGF Antibodies-Increasing concentrations of bFGF were tested for their ability to stimulate BCEC migration in a multiple-well Boyden chamber (Fig. 4). bFGF-induced chemotaxis of BCEC followed a bell-shaped curve with maximal response at 0.1-1.0 ng/ml. Anti-bFGF antibodies inhibited the chemotaxis stimulated by 1 ng/ml bFGF (Fig. 4, bur graph) and by lower amounts of bFGF (not shown), but did not inhibit basal migration in 10% CS.
Inhibition of bFGF-induced Neurite Outgrowth by Anti-  Some wells containing 1 ng/ml bFGF also received 400 pg/ml anti-bFGF IgG (bar graph). The amount of anti-bFGF antibody used was not effective against larger concentrations of bFGF, such as 10 ng/ml, which represented a 1000fold excess of bFGF needed to stimulate chemotaxis.
Effects of Anti-bFGF Antibody on bFGF Cross-linking to Receptors-Cell migration and proliferation are thought to be mediated by growth factor interaction with cell surface receptors. Cross-linking experiments have previously identified a receptor for bFGF that appears as a doublet on SDS-PAGE with molecular sizes that range from 115,000 to 140,000 daltons (28). Biologically active T -b F G F was cross-linked to BCEC cell surface receptors with disuccinimidyl suberate, and the radioactive species were analyzed by SDS-PAGE and autoradiography (Fig. 6). The resulting radioactive crosslinked species was a doublet with apparent molecular sizes of 130,000 and 150,000 daltons (Fig. 6, lane I ) . This is in good agreement with the molecular mass characteristics of bFGF  6. Effects of anti-bFGF antibody on "'I-bFGF crosslinking to its receptor. '"I-bFGF (10 ng/ml) was added to cultures of BCEC (100-mm dish) either in the absence (lane I) or presence (lane 2 ) of 400 pg/ml anti-bFGF IgG. BCEC were incubated for 3.5 h at 4 "C, and "'1-bFGF was cross-linked to its receptor by incubation with disuccinimidyl suberate for 30 min at room temperature. The BCEC were washed and lysed, and the lysates were analyzed by SDS-PAGE and autoradiography. receptors identified in other cells (5,28). Cross-linking of "'I-bFGF to the high affinity receptor was abolished by the anti-bFGF antibodies (Fig. 6, lane 2). T h e l2'1-bFGF receptor crosslinking was also abolished by a 100-fold excess of unlabeled bFGF (not shown) but was not affected by the control preimmune IgG (not shown). In lane 2 the cross-linked bands with molecular mass of about 85-90 kDa are probably T -b F G F linked to IgG heavy chains. The large amount of radioactivity running at the front line in both lanes is non-cross-linked "'I-bFGF.

116-
Effects of Anti-bFGF Antibody on bFGF Binding to Heparin-bFGF binds tightly to heparin as well as to its cell surface receptors. Samples of biologically active Y -b F G F incubated with anti-bFGF IgG, normal rabbit IgG, or not incubated with either were applied to columns of heparin-Sepharose and eluted batchwise with 0.6 M NaCl followed by 2.0 M NaCl, the concentration of NaCl required to elute bFGF from the column (Fig. 7). Anti-bFGF antibodies did not inhibit the binding of FGF to heparin. These results suggest that the heparinbinding domain which is not blocked by anti-bFGF antibodies and the cell receptor-binding domain which is blocked by these same antibodies must represent different parts of the bFGF molecule.
It has been suggested that the low affinity binding sites for bFGF are heparin-like molecules (4,14). The observation that anti-bFGF antibodies do not block the ability of bFGF to bind to immobilized heparin suggests that these antibodies will not block the binding of bFGF to the low affinity binding sites either. Binding of lZ5I-bFGF to low affinity binding sites was inhibited by heparin in a dose-dependent manner, confirming previous reports (4,14), but not by anti-bFGF antibodies (Table I)  Samples were applied to columns of heparin-Sepharose (0.5 ml bed volume) which were washed first with 0.6 M NaCl (5 ml) and then 2.5 M NaCl (5 ml). Fractions of 1 ml were collected and counted in a y counter. The recovery of l2'I-bFGF from each column in the 2.5 M NaCl eluate was about 60%.

TABLE I
Effect of heparin and anti-bFGF antibody on the binding of "'Z-FGF to low affinity binding sites in BCEC T -b F G F (10 ng/ml) alone or with heparin, or with anti-bFGF IgG, or with normal IgG was added to confluent cultures of BCEC (4 X lo4 cells, 96-well plate). After 2 h incubation at 4 "C, the unbound label was removed and '"I-bFGF bound to low affinity sites was released by 2 M NaCl at neutral pH and quantified (4). proliferation, migration, neurite extension, and receptor cross-linking.

DISCUSSION
bFGF binds strongly to two different types of macromolecules: cell surface receptors which are proteins, and heparin and heparan sulfate which are glycosaminoglycans. Polyclonal antibodies, capable of neutralizing a number of the biological activities of bFGF such as the ability to stimulate BCEC migration, BCEC proliferation and neurite outgrowth in PC12 cells, inhibit the cross-linking of bFGF to its cell surface receptors but do not prevent binding of bFGF to immobilized heparin or to heparin-like molecules on the surface of BCEC. These results suggest that the cell surface receptor-binding domains and heparin-binding domains of bFGF are not the same. In previous studies using synthetic peptides representing regions of bFGF, it was observed that some heparinbinding bFGF peptides could also bind cell surface receptors while others could not (15,16). These authors concluded that heparin-binding and receptor-binding domains overlapped in certain regions of bFGF but not in others.
One possible advantage in using antibodies as probes, as compared to using synthetic peptides, is that the antibodies might be more sensitive and accurate in recognizing multiple activities dependent on the three-dimensional conformation of bFGF. Because the synthetic peptides probably do not have the conformation that these regions have in native bFGF, they may be less specific in binding, inhibition, and mitogenic assays. For example, an amino acid that is available in a synthetic peptide to bind to heparin might be inaccessible in the folded native molecule.
Neutralizing antibodies block bFGF stimulation of BCEC cell migration, BCEC proliferation, and PC12 neurite extension, as well as cross-linking of bFGF to 130,000-150,000dalton BCEC receptors. However, they do not inhibit bFGF binding to heparin. These results suggest strongly that it is the high affinity 130,000-150,000-dalton receptors and not heparin which mediate these biological activities of bFGF. There may be other biological activities of bFGF that are mediated by binding to heparin, although it is not clear at present what these biological activities may be. One biological role of heparin might be to bind bFGF to the cell surface with low affinity. Moscatelli (4) has provided evidence that cells such as baby hamster kidney and BCEC have low affinity receptors from which binding of lz5I-bFGF is displaced by heparin (4). These low affinity binding sites appear to be heparin-like, number about 600,000-1,000,000/cell, and have a Kd of about 2,000 PM as compared to the high affinity binding sites which number 8,000-82,000/cell and have a Kd of about 20 PM. We also find that heparin inhibits binding of bFGF to low affinity sites on BCEC. It has been suggested that heparin-like molecules on the cell surface and in the extracellular matrix act as a reservoir of bFGF. "Stored" bFGF might be transferred from heparin-binding sites of relatively low affinity to cell surface receptor binding sites of relatively high affinity (2,9,14,22). Thus, the heparin-binding domain of bFGF might act as a functional entity which allows it to bind to heparin-like molecules, but relatively weakly. Binding of bFGF to heparin-like molecules may be a preliminary step in the transfer of bFGF to high affinity receptors at another locus on the cell surface resulting in cell migration, proliferation, or differentiation.
The use of neutralizing antibodies directed against bFGF has given us several insights into the regulation of EC growth. For example, EC plated on subendothelial cell ECM no longer have an exogenous requirement for bFGF (9). The presence of bFGF in subendothelial ECM suggested, but did not prove, that this growth factor might play a role in ECM-mediated EC proliferation (9, 10). As demonstrated in this report, the ability of anti-bFGF antibodies to inhibit ECM-stimulated BCEC cell growth suggests strongly that bFGF is the active mitogen in ECM required for supporting the proliferation of EC.
Another observation is that anti-bFGF antibodies inhibit exogenous bFGF-stimulated EC migration and proliferation but not basal EC migration and proliferation. Lack of inhibition of basal growth by anti-bFGF antibodies is consistent with previous reports (29). The inability of anti-bFGF antibodies to inhibit EC basal growth suggests two possibilities. Since EC synthesize relatively large amounts of bFGF which remain cell-associated (9,30,31), it may be that basal growth of EC is due to cell-associated bFGF which is not accessible to antibodies. Alternatively, basal EC growth may be due to growth factors that are not bFGF, for example, serum-derived high density lipoprotein (32).
bFGF is an ECM-associated protein that has at least two functional domains, a heparin-binding domain and a cell surface receptor-binding domain. Interestingly, bFGF has structural and biological features similar to the ECM proteins fibronectin and laminin. Both fibronectin and laminin have heparin-binding and cell-binding domains (33, 34) and both are adhesion molecules (33, 34), as is bFGF (15, 16). Fibronectin and laminin also have some of the biological properties normally associated with bFGF. For example, fibronectin promotes cell migration and proliferation (33,34) while laminin stimulates neurite outgrowth (35). Our antibody studies suggest that bFGF might be structured like laminin and fibronectin to have different domains which are characteristic of ECM structural proteins in general. These are domains that interact with heparan sulfate proteoglycan and cell surface receptors as well as structural elements found on other ECM proteins.
While neutralizing antibody studies have helped to establish the presence of distinct functional domains in bFGF, they have not identified where these domains are. Studies which, for example, include the use of antibodies directed against specific sites in bFGF and techniques such as in vitro mutagenesis may be helpful in mapping these domains.