Potentiation of Endothelial Cell Proliferation by Fibrin(ogen)-bound Fibroblast Growth Factor-2*

Endothelial cell growth is stimulated by fibroblast growth factor-2 (FGF-2), and both adhesion and proliferation are modulated by interactions with fibrinogen and fibrin. Previous evidence indicates that FGF-2 binds specifically and with high affinity to fibrinogen and fibrin, suggesting that their effects on endothelial cells may be coordinated. In this study, we have, therefore, investigated the ability of FGF-2 bound to fibrinogen and fibrin to stimulate proliferation of endothelial cells. Human umbilical vein endothelial cells were cultured in the presence of FGF-2 with or without fibrinogen, and proliferation was assessed by microscopic examination of cultures, incorporation of [3H]thymidine and by cell counting. Cells cultured in the presence of both FGF-2 and fibrinogen proliferated more rapidly than those with FGF-2 alone and exhibited a decreased population doubling time. At concentrations of FGF-2 up to 150 ng/ml, there was greater endothelial cell proliferation in the presence of fibrinogen than in its absence with the most pronounced effect below 1 ng/ml. The maximum effect of fibrinogen was observed at a molar ratio of fibrinogen to FGF-2 of 2:1, corresponding to the maximum molar binding ratio. Endothelial cells proliferated when plated on fibrin or surface-immobilized fibrinogen with FGF-2, indicating that FGF-2 bound to surface-associated fibrin(ogen) retained activity. We conclude that fibrinogen- or fibrin-bound FGF-2 is able to support endothelial cell proliferation and that fibrinogen potentiates the proliferative capacity of FGF-2.

Endothelial cells normally have a low rate of proliferation in the adult with a life span of 100 -10,000 days (1), but the endothelium retains its capacity for proliferation, which occurs physiologically in the corpus luteum and uterus and also during wound healing. Endothelial cell proliferation, differentiation and migration are also needed for angiogenesis, an important process in many pathologic conditions including tumor growth, diabetic retinopathy, inflammation, and ischemic thrombotic diseases. Polypeptide growth factors play an important regulatory role in angiogenesis, and several stimulatory and inhibitory molecules have been identified (2,3) including fibroblast growth factor-2 (FGF-2, 1 basic fibroblast growth fac-tor), an 18-kDa polypeptide of the FGF family, which exerts a variety of effects on many cells and organ systems (4,5).
Endothelial cells in culture require an FGF to support proliferation and to prevent apoptosis (6). In addition, FGFs promote endothelial cell migration (7,8) and increase synthesis of several proteins that are important in degradation of the extracellular matrix during cell migration or angiogenesis including collagenase (8,9), urokinase plasminogen activator, urokinase plasminogen activator receptor (8 -11), and plasminogen activator inhibitor-1 (12,13). FGFs also regulate endothelial cell adhesion by modulating expression of integrin receptors (14), and they can increase expression of vascular endothelial growth factor, another angiogenic peptide (15). Although FGF-1 and FGF-2 lack signal peptides, they are both active in the pericellular environment and bind with high affinity to specific receptor tyrosine kinases (16). FGFs are released from vascular cells following injury, and FGF-2 mRNA is up-regulated in atherosclerotic arteries (17) and following vessel injury (18).
The hemostatic system is activated at sites of tissue injury, and this results in local formation of fibrin, which also plays a role in regulating the endothelial cell responses required for healing and angiogenesis. Endothelial cells adhere, spread and proliferate on both fibrinogen and fibrin in vitro through binding to integrins ␣ v ␤ 3 and ␣ 5 ␤ 1 (19,20), and fibrinogen supports leukocyte tethering to endothelial cells at sites of inflammation through ICAM-1 (21). Although endothelial cells are physiologically stable, when exposed normally to a high concentration of fibrinogen in vivo, they interact with fibrin uncommonly, and marked phenotypic change results. Specifically, fibrin exposure in vitro results in loss of monolayer organization, cell retraction, and migration (22,23). Fibrin also stimulates synthesis and secretion of tissue plasminogen activator and prostacyclin (24,25), induces interleukin-8 expression (26), suppresses plasminogen activator inhibitor-1 release (13,27), and causes rapid mobilization of high molecular weight von Willebrand factor from Weibel-Palade bodies (28). These responses require the thrombin-induced cleavage of fibrinopeptide B from the fibrinogen B␤ chain, which results in exposure of a reactive site at the new amino terminus of the fibrin ␤ chain (29). An endothelial cell receptor interacting with a site within the 15-42 region of the ␤ chain has been identified (30) and recently shown to be VE-cadherin (31).
We have recently demonstrated specific high-affinity binding of FGF-2 to fibrinogen and fibrin (32). This suggests a mechanism by which fibrin can localize FGF-2 at sites of inflammation or tissue injury to coordinate endothelial cell responses. Previous studies have shown that FGF-2 specifically binds to extracellular matrix and is released by heparinase and by heparin, and the proliferative activity of FGF-2 is enhanced by binding to heparin (33)(34)(35). In this study we have investigated the ability of fibrinogen-and fibrin-associated FGF-2 to stimulate proliferation of endothelial cells in vitro. The results indicate that FGF-2 associated with surface-immobilized fibrinogen or fibrin retains its mitogenic activity, and that the endothelial cell proliferative response to FGF-2 is potentiated by fibrinogen.

EXPERIMENTAL PROCEDURES
Cell Culture-Primary endothelial cells were obtained from human umbilical veins as described previously (36), seeded on 0.2% w/v gelatincoated 25-cm 2 tissue culture flasks and cultured in McCoy's 5A medium (Flow Laboratories, McLean, VA) containing 20% fetal bovine serum (FBS), 50 g/ml endothelial cell growth supplement (ECGS) (Collaborative Research, Inc., Bedford, MA) and 100 g/ml heparin (Sigma) until they reached confluence, typically within 4 -5 days. The cells were passaged up to two times before use and then placed in suspension by trypsinization of monolayers. Cells were suspended by rinsing in Hanks' balanced salt solution followed by brief incubation with trypsin-EDTA (Life Technologies, Inc.). The cells were pelleted by centrifugation for 10 min at 500 ϫ g and resuspended in McCoy's 5A medium. This wash procedure was repeated twice before use in experimental protocols.
[ 3 H]Thymidine Incorporation-Cell proliferation using [ 3 H]thymidine incorporation was quantitated as described previously (37). Briefly, approximately 2 ϫ 10 4 endothelial cells suspended in McCoy's 5A medium supplemented with 20% FBS, 50 g/ml ECGS, and 100 g/ml heparin were plated in gelatin-coated 24-well, nontissue culturetreated plates (Becton Dickinson & Co., NJ) and allowed to adhere for 6 h. The medium was then removed, and the cells were washed twice with serum-free McCoy's 5A medium. Serum-free medium was then added containing 1% Nutridoma ® (Roche Molecular Biochemicals), 25 ng/ml human recombinant FGF-2 (R&D Systems, Inc., Minneapolis, MN) and 1 Ci/ml [ 3 H]thymidine (NEN Life Science Products) in the presence or absence of 10 g/ml fibrinogen. After incubation at 37°C for 24 h, nonadherent cells were removed by washing twice with ice-cold phosphate-buffered saline (PBS). To each well was then added 500 l of 10% ice-cold trichloroacetic acid, and precipitates were collected on a filter using a manifold. Filters were washed twice with ice-cold 5% trichloroacetic acid, followed by 95% ethanol, allowed to air dry, and then suspended in scintillation fluid. Acid precipitable counts per min (cpm) were quantitated using a scintillation counter.
Fibrinogen and Fibrin Preparation-Human fibrinogen was obtained from American Diagnostica (Greenwich, CT), and copurifying fibronectin was removed by gelatin-Sepharose chromatography (38). Residual fibronectin remaining was further depleted by immunoaffinity chromatography as described elsewhere (39). The fibronectin concentration was determined by enzyme-linked immunosorbent assay (American Diagnostica) and represented less than 0.02% of the total protein. Cell culture wells were coated by incubation for 1 h at 25°C with 0.4 ml of 10 g/ml fibrinogen in McCoy's 5A medium. Excess fibrinogen was aspirated, and the wells were washed twice with McCoy's 5A medium before the cells were plated. Fibrin-coated wells were prepared using 1 mg/ml fibrinogen in McCoy's 5A medium to which 1 unit/ml thrombin (Calbiochem-Novabiochem Corp.) was added, mixed, and rapidly pipetted into 24-well cell culture plates. The solution was aspirated after 45 s and before polymerization, leaving a thin coating of fibrin on the sur-face. Wells coated with fibrinogen or fibrin with FGF-2 were prepared in the same way except that 25 ng/ml FGF-2 was added to the fibrinogen solution and incubated for 20 min at 37°C before coating wells. Fibrincoated wells were treated with 1 g/ml D-phenylalanyl-L-prolyl-L-arginylchloromethyl ketone (Bachem, Torrance, CA), a synthetic specific thrombin inhibitor, for 30 min to inhibit any remaining thrombin, and this was followed by two washes with McCoy's 5A medium before plating the cells.

Measurement of Apoptotic Nuclei by Terminal Deoxynucleotidyl Transferase in Situ Labeling (TUNEL)-Endothelial cells cultured on
Thermanox ® coverslips (Marsh Biomedical Company, Rochester, NY) were fixed in 3.7% formaldehyde in PBS for 20 min, postfixed in ethanol-acetic acid (2:1) at Ϫ20°C for 5 min and rinsed three times in PBS. Coverslips were then stained using the TUNEL method (40) using the Apoptag ® kit (Oncor ® , Gaithersburg, MD). Coverslips were then incubated with terminal deoxynucleotidyl transferase and digoxigenin-dUTP and stained with fluorescein isothiocyanate-anti-digoxigenin antibody according to the manufacturer's instructions. They were then mounted (cells facing up) on glass microscope slides using Gel/Mount ® (Biømedia Corp., Foster City, CA). Propidium iodide counterstain (3 g/ml in PBS) was applied, and the slide was covered with a glass coverslip; the edges were sealed using rubber cement, and the slides were stored at Ϫ20°C. Cells were viewed using a Nikon Eclipse E-800 fluorescence microscope equipped with a dual wavelength filter cube. Normal nuclei exhibit orange fluorescence due to propidium iodide staining, whereas apoptotic nuclei, which have incorporated the digoxigenin labeled nucleotides, exhibit green fluorescence.
Determination of Population Doubling Time-Endothelial cells were grown in McCoy's 5A medium containing 20% FBS, 25 ng/ml FGF-2, and 1 unit/ml hirudin with or without 10 g/ml fibrinogen. Cells were passaged at a ratio of 1:3 and plated in the presence or absence of fibrinogen up to passage 10. Before each passage, photographs were taken, and the number of cells per field was counted to determine the fold increase. Population doubling time was calculated using the following formula: population doubling time ϭ days in culture/fold increase in cell number. The fold increase ϭ 2 n , where n ϭ number of doublings.
Statistical Analysis-Each experiment was performed at least three times, and either triplicate or quadruplicate wells were used in each experiment. The significance of differences in means was determined using a two-tailed Student's t test. Variance is described as ϮS.E.

RESULTS
FGF-2 is needed to support endothelial cell growth and survival in culture. To determine whether it retains this activity when bound to fibrinogen, cells were cultured in medium containing 1.5-3 nM FGF-2 in the presence or absence of 30 nM fibrinogen. Because our previous studies demonstrated that FGF-2 binds to fibrinogen with an apparent k D of 1.3 nM, the concentration of free FGF-2 under these conditions would be less than 0.01 nM and insufficient to support cell growth. Cells proliferated well in the presence of fibrinogen and FGF-2 and appeared normal microscopically (Fig. 1A). Apoptosis, which is known to be a consequence of growth factor deprivation, did not occur when cells were cultured with fibrinogen and FGF-2 ( Fig.  1B) but was observed with fibrinogen but no FGF-2 (Fig. 1C). Endothelial cells were plated on gelatin-coated wells and incubated in McCoy's 5A medium with 20% FBS and 100 g/ml heparin and allowed to adhere for 6 h. The medium was then removed, cells were washed twice, and then serum-free medium was replaced with 1% Nutridoma and 10 g/ml fibrinogen in the presence (panels A and B) and absence (panel C) of 25 ng/ml FGF-2. The cells were then cultured for an additional 24 h and photographed live using phase contrast (panel A). For panels B and C, the cells were fixed and stained using the TUNEL method, counterstained with propidium iodide, and viewed under dual wavelength excitation. Normal nuclei exhibit orange fluorescence (propidium iodide; panel B), and apoptotic nuclei typically appear condensed with green fluorescence (panel C). Bar in panel A ϭ 1 mm and bar in panels B and C ϭ 100 m.
These initial microscopic observations indicated that the presence of fibrinogen did not block the biologic activity of FGF-2.
The effect of fibrinogen on FGF-2-induced proliferation of endothelial cells was further examined by cell counting and determination of population doubling times. Cells were grown in the presence or FGF-2 with or without fibrinogen in the medium. Hirudin was included to inhibit the low levels of thrombin present in fetal bovine serum and prevent conversion of fibrinogen to fibrin. Population doubling times were calculated using endothelial cells between passages 2 and 8, which were cultured using 25 ng/ml FGF-2 in the presence or absence of 10 g/ml fibrinogen. A population doubling time of 5.6 Ϯ 0.2 days (n ϭ 7) was determined in the absence of fibrinogen, whereas it was shortened to 2.8 Ϯ 0.3 days (n ϭ 7) (p Ͻ 0.0001) when fibrinogen was present in the culture medium. Passage 1 was not used because cells in primary culture and first passage retain the capacity for proliferation independently of exogenously added growth factors. Doubling times were prolonged at later passages probably due to the onset of senescence. The maximum cell density achieved at confluence decreased with time in culture (not shown), but was not influenced by presence or absence of fibrinogen. No morphologic differences were observed microscopically between cultures with or without fibrinogen.
The effect of fibrinogen on FGF-2-mediated endothelial cell proliferation was also evaluated by [ 3 H]thymidine incorporation (Fig. 2). In the absence of FGF-2, there was no increase in [ 3 H]thymidine uptake between 6 and 48 h, indicating little or no cell proliferation. Also, little incorporation resulted from the addition of 10 g/ml fibrinogen to the medium in the absence of FGF-2. As expected, FGF-2 alone stimulated proliferation, and 3 H-thymidine incorporation was increased over control at all times points. Addition of fibrinogen potentiated proliferation mediated by FGF-2, with the greatest effect evident at 24 h (p Ͻ 0.04). There was little additional proliferation at later times, reflecting contact inhibition (not shown). Microscopic examination confirmed slower growth in the absence of fibrinogen, and a longer time was needed to reach confluence.
Cell proliferation was dependent on FGF-2 concentration in the presence and absence of fibrinogen (Fig. 3), with a maxi-mum 6.4-fold increase in [ 3 H]thymidine incorporation at 25 ng/ml in the absence of fibrinogen. There was, however, greater proliferation at all concentrations of FGF-2 in the presence of fibrinogen than in its absence. The effect of fibrinogen was particularly evident at FGF-2 concentrations below 1 ng/ml (Fig. 3, inset). For example, at 0.1 ng/ml FGF-2, there was 1.2 Ϯ 0.5-fold increase over baseline in the absence of fibrinogen but 2.8 Ϯ 0.5-fold in its presence (p Ͻ 0.001). The FGF-2 concentration dependence of endothelial cell proliferation was more complex in the presence of fibrinogen than in its absence. Maximum proliferation in the presence of both fibrinogen and FGF-2 occurred at 25 ng/ml, with a 13.7 Ϯ 2.7-fold increase in comparison with 5.5 Ϯ 0.2-fold in the absence of fibrinogen at the same FGF-2 concentration (p Ͻ 0.03). In the presence of fibrinogen, proliferation decreased at FGF-2 concentrations over 25 ng/ml, declining to 8.8 Ϯ 0.6-fold over baseline at 100 ng/ml and 8.1 Ϯ 0.5-fold at 150 ng/ml. At both latter concentrations, however, the [ 3 H]thymidine incorporation remained higher than in the absence of fibrinogen.
The capacity to potentiate FGF-2 stimulated cell proliferation was also characterized over a range of fibrinogen concentrations (Fig. 4). An increase in cell proliferation was observed at 0.25 g/ml (0.75 nM), and there was progressive enhancement of activity to a maximum of 2.9-fold over baseline at 5 g/ml (15 nM), representing a molar ratio of fibrinogen to FGF-2 of 2:1. At higher concentrations of fibrinogen, no increased effect on cell proliferation was observed.
Following tissue injury, thrombin converts fibrinogen to fibrin, an insoluble polymer that forms the initial matrix required for cell adhesion and wound healing. To determine whether FGF-2 was active when bound to fibrinogen or fibrin presented as an adhesive substrate, we prepared surfaces coated with either fibrin or fibrinogen with or without added FGF-2. Endothelial cells were cultured on these surfaces and viewed microscopically. Cells grown on surfaces of fibrinogen (Fig. 5A) or fibrin (Fig. 5B) in the absence of FGF-2 were consistently sparse, but incorporation of FGF-2 into the matrix After 24 h of incubation, nonadherent cells were removed, and the isotope incorporated into DNA was extracted with trichloroacetic acid. Precipitates were collected by vacuum filtration, and incorporated isotope was measured by scintillation counting; cpm per well was determined and normalized to the results obtained in the absence of FGF-2 and fibrinogen. resulted in marked proliferation (Fig. 5, C and D). Gelatin was used as an alternative adhesive substrate, and wells coated with a solution of gelatin to which FGF-2 had been added did not adequately support proliferation (Fig. 5E). In control wells, however, endothelial cells grew well on a coating of gelatin if FGF-2 was included in soluble form in the culture medium (Fig.  5F). As quantitated by [ 3 H]thymidine incorporation, proliferation was minimal on a surface of either fibrinogen or fibrin in the absence of FGF-2 (Fig. 6), but was significantly enhanced with FGF-2 immobilized with either fibrinogen or fibrin (p Ͻ 0.04 for both).
Samples of culture medium containing fibrinogen were collected following the 24 h incubation and analyzed by SDSpolyacrylamide gel electrophoresis. The migration pattern of A␣, B␤, and ␥ chains was unchanged and showed no fibrinopeptide A or fibrinopeptide B cleavage, indicating that there was no proteolytic degradation or conversion of fibrinogen into fibrin (data not shown). To determine whether other adhesive glycoproteins also stimulated cell proliferation with FGF-2, endothelial cells were also cultured in the presence of vitronectin (10 g/ml) or fibronectin (10 g/ml). Neither fibronectin nor vitronectin increased proliferation significantly, indicating the specificity of fibrinogen in enhancing the effect of FGF-2.

DISCUSSION
The findings presented demonstrate that fibrin-or fibrinogen-bound FGF-2 retains biological activity. Cells cultured in medium containing 1.5 nM FGF-2 and 30 nM fibrinogen supported proliferation and did not undergo apoptosis, which is known to occur under conditions of growth factor deprivation (6). At the concentrations of fibrinogen and FGF-2 used in these experiments, the amount of free FGF-2 would be insufficient to support cell growth. FGF-2 associated with surface-immobilized fibrinogen or fibrin was also active as indicated by its ability to support growth (Fig. 5). Cell proliferation was quantitated by [ 3 H]thymidine incorporation, determination of population doubling times, and cell counting. The findings with each of these methods indicated that the proliferative potential of FGF-2 is enhanced in the presence of fibrinogen. The population doubling time was shorter with fibrinogen, the proliferative rate was greater (Fig. 2), and the cells responded to a lower concentration of FGF-2 when presented in combination with fibrinogen (Fig. 3). At all FGF-2 concentrations, fibrinogen increased its proliferative capacity (Fig. 3), and the maximum effect was observed at a molar ratio of fibrinogen to FGF-2 of 2:1 (Fig. 4), corresponding to the maximum molar binding ratio (32).
Fibrinogen is an adhesive substrate for endothelial cells, but it is unlikely that the enhanced proliferation observed was due to an effect of fibrinogen on adhesive properties. Cells were fully spread before FGF-2 exposure, and no change in cell spreading was observed during incubation with fibrinogen. No significant increase in FGF-2-induced cell proliferation was observed with fibronectin or vitronectin, indicating the fibrinogen effect was specific. Also, enhanced cell proliferation was observed with fibrinogen even in medium containing 20% fetal bovine serum, which is rich in adhesive proteins, confirming the specificity of the enhancement by fibrinogen. Proliferation was evaluated using several methods because of the experimental limitations of each. Cultures were examined microscopically to evaluate morphologic characteristics including spreading and apoptosis in addition to proliferation. [ 3 H]Thymidine incorporation was measured as an overall index of DNA synthesis, recognizing that proliferation may be underestimated with prolonged exposures. Population doubling time was determined as a direct measure of cell proliferation during prolonged exposure to FGF-2 and fibrinogen with multiple cell passages. Each of these methods indicated significant enhancement of FGF-2 proliferative capacity in the presence of fibrinogen.
The mechanism by which fibrinogen potentiates FGF-2 activity is not known but may involve receptor clustering or coordination of cell signaling. Because fibrinogen is a dimeric molecule, binding of two or more FGF-2 molecules could increase cell activation through receptor dimerization analogous to that observed with heparin (41). Also, binding of a complex of fibrinogen and FGF-2 could result in co-localization of integrin and FGF receptors at the focal adhesion complex, contributing to signal integration (42). Fibrinogen may protect FGF-2 from inactivation by serum or cell-associated proteases, thereby prolonging and increasing its activity. Such protection from proteolytic degradation has been observed for FGF-2 bound to extracellular matrix (43,44). Additionally, a recent report indicates that FGF-2 can bind to ␣ v ␤ 3 (45), an endothelial cell integrin receptor that also binds fibrinogen, suggesting that the adhesive and proliferative activities of FGF-2 and fibrinogen may be coordinated through a single receptor.
Endothelial cell responses to injury and angiogenesis are dependent on both growth factor stimulation and interactions with matrix components. The importance of endothelial cellmatrix interactions in angiogenesis is evident from the binding of FGF-2 to extracellular matrix heparan sulfate proteoglycans. Although of lower affinity than the binding to specific tyrosine kinase receptors, the association with heparan sulfates is physiologically important in protecting FGFs from proteolytic degradation (43,44,46) and by providing a local reservoir of growth factor that can be released by enzymes that degrade proteoglycans (47,48). Additionally, heparan sulfates increase the binding affinity of FGFs for specific receptors and facilitate presentation to transmembrane signaling receptors (41). The mechanisms by which heparan sulfate proteoglycans modulate FGF function remain under investigation, but they may act to reduce the dimensionality of ligand diffusion to a plane from three dimensions (48). Endothelial cells are exposed to fibrin both pathologically and in response to vessel injury where fibrin forms the initial matrix necessary for cell organization and healing. As such, fibrin could play a role similar to that of the extracellular matrix in binding FGF-2, which both localizes and prolongs its action. FIG. 4. Fibrinogen concentration dependence of FGF-2-mediated endothelial cell proliferation. Endothelial cells were plated on gelatin-coated wells in McCoy's 5A medium supplemented with 20% FBS, 50 g/ml ECGS, and 100 g/ml heparin and allowed to adhere for 6 h. The medium was then removed. Cells were washed twice, and then serum-free medium was added containing 1% Nutridoma, and 25 ng/ml FGF-2, and 1 Ci/ml [ 3 H]thymidine in the presence of different concentrations of fibrinogen The molar ratios of FGF-2 to fibrinogen varied from 0.1-2. After 24 h of incubation, nonadherent cells were removed, and isotope incorporated into DNA was extracted with trichloroacetic acid. Precipitates were collected by vacuum filtration, and incorporated isotope was measured by scintillation counting.
The potential to manipulate angiogenesis therapeutically is now being realized in initial clinical trials, and strategies to either inhibit or stimulate new vessel growth appear promising (49,50). The association of FGF-2 with fibrin(ogen) is relevant to the development of these therapeutic strategies. Because of its high plasma concentration, binding to fibrinogen will affect distribution of FGF-2 if administered systemically. Fibrin binding may also be important therapeutically. For example, in the successful initial trial in coronary artery disease (51), FGF-1 was injected locally near the site of vascular anastomosis where fibrin formation would be expected. Binding of the growth factor to fibrin would serve to localize and possibly increase its effect, contributing to the observed neovascularization. The binding of FGF-2 to fibrin(ogen) and the effects on endothelial cell proliferation suggest a new level of coordination between the hemostatic system and cell regulatory growth factors in the vascular response to injury and angiogenesis.