Fibroblast growth factors (FGFs) are potent mitogens for vascular and capillary endothelial cells in vitro and can stimulate the formation of blood capillaries (angiogenesis) in vivo. A crucial event in this process is the invasion of the perivascular extraceflular matrix by sprouting endothelial cells. Using a recently developed in vitro model of angiogenesis, we show here that highly purified basic pituitary FGF can induce capillary endothelial cells to invade a three-dimensional collagen matrix and to organize themselves to form characteristic tubules that resemble blood capillaries. We also show that basic FGF concomitantly stimulates endothelial cells to produce a urokinase-type plasminogen activator, a protease that has been implicated in the neovascular response. The results demonstrate that basic FGF can stimulate processes that are characteristic of angiogenesis in vivo, including endothelial cell migration, invasion, and production of plasminogen activator.The formation of new blood capillaries (angiogenesis) occurs in a wide range of important biological processes in response to angiogenic factors released by either normal or tumoral cells (1). A crucial step in the sequence of events that leads to the angiogenic response is the invasion of the perivascular extracellular matrix by sprouting endothelial cells (2). The process includes endothelial cell migration, proliferation, and production of enzymes capable of modifying the extracellular matrix. We have recently shown that the invasiveness of capillary endothelial cells can be induced experimentally in vitro by well-defined chemical signals (3). Cells grown on three-dimensional collagen gels and treated with the tumor promoter 4/3-phorbol 12-myristate 13-acetate (PMA) infiltrate the underlying collagen matrix and organize into vessellike tubular structures (3). Although phorbol esters are not physiologically occurring substances, they have bedn shown to mimic, in many instances, the effects of endogenous mediators, such as hormones or growth factors (4-7). It was therefore important to establish whether phenomena similar to those induced by PMA could also be triggered by physiological angiogenic factors.In this study, we have examined the effect of highly purified basic fibroblast growth factor (FGF) on MATERIALS AND METHODSIsolation of FGF. Basic FGF was purified to homogeneity from bovine pituitaries by successive steps ofammonium sulfate precipitation, ion exchange chromatography, and heparin-Sepharose affinity chromatography (8). Purity of the growth factor was established by reverse-phase high-performance liquid chromatography, amino acid analyses, NaDodSO4/ polyacrylamide gel electrophoresis, and amino-terminal sequence analyses (10).Cell Culture. Three-dimensional gels of reconstituted collagen fibrils were prepared as described (3). Cloned capillary endothelial cells derived from the bovine adrenal cortex (14) were a generous gift of M. B. Furie and S. C. Silverstein (Columbia University, New York). The cells were routinely subcul...

The two major mitogenic polypeptides for endothelial cells have been purified to homogeneity. The complete primary structure of bovine pituitary basic fibroblast growth factor (FGF) and the amino-terminal amino acid sequence of bovine brain acidic FGF have been established by gas-phase sequence analyses. Homogeneous preparations of these polypeptides are potent mitogens (basic FGF, ED50 60 pg/ml; acidic FGF ED50 6000 pg/ml) for many diverse cell types including capillary endothelial cells, vascular smooth muscle cells, and adrenocortical and granulosa cells; in vivo, basic FGF is a powerful angiogenic agent in the chick chorioallantoic membrane assay. The available protein sequence data demonstrate the existence of significant structural homology between the two polypeptides.

Basic fibroblast growth factor (FGF) has been found to increase neuronal survival and neurite e on in a highly purified population of fetal rat hippocampal neurons under well-defined serum-free cell culture conditions. In the presence of FGF, neuronal survival after 7 days in culture on a simple plastic substrate is increased 4-fold, to 54% of the Initial population. Survival is increased 2-fold to 40% on polyornithine-laminin. When FGF was bound to plastic or heparin substrates, neurite outgrowth was sgnflcantiy increased to lengths comparable to those seenwith lamin; however, FGF produced no further increase in neurite outgrowth on laminin. Half-maximal survival was observed at FGF concentrations of about 15 pg/nl (1 pM); half-maximal process outgrowth occurred at about 375 pg/ml (20 pM). The responsive cells were identified as neurons by their labeling with tetanus toxin and by antibodies to neurofilAments and to the neuron-speciflc enolase. Astrocytes, identified by the presence of glial fibrillary acidic protein, constituted about 10% of cells present at 1 week both in the presence and in the absence of FGF. These results strongly suggest that, in addition to its known mitogenic effects on nonneuronal cells, FGF possesses neurotrophic activity for hippocampal neurons.The discovery that during normal development substantial numbers of nerve cells in most regions of the central and peripheral nervous systems die and that the proportion of neurons that survive in any given region is closely related to the size of their target field has led to the hypothesis that all nerve cells are dependent for their growth and maintenance on the availability of specific trophic (growth) factors (1-3). The first neurotrophic factor to be identified was "nerve growth factor" (NGF), and the critical role it plays in supporting the survival and growth of sympathetic and certain sensory neurons has been well documented (4,5 Until recently very few systems have been described in which it is possible to reproducibly obtain adequate cultures of virtually pure neuronal (as opposed to mixed neuronal and glial) cultures. One such system involves the use of dissociated hippocampal neurons from late fetal rats (9-12). The main advantage of this system is that the cell population is unusually homogeneous, the overwhelming majority of the cells at embryonic day 18 being pyramidal neurons (9), and with appropriate controls the proportion of nonneuronal cells in the cultures can be kept to less than 10%1 (10,11). Previous studies have established that virtually all the neurons degenerate within 72-96 hr ifthey are maintained at low density and in a chemically defined medium (10-12), but they can be maintained for periods of several weeks in the presence of explants of hippocampal tissue or a feeder layer of astrocytes, or in medium conditioned by astrocytes (9, 10, 12). We have used this system to examine the effects of highly purified basic fibroblast growth factor (FGF) on the survival of the isolated neurons and on their capacit...

Abstract. To survive and proliferate in pure culture, human melanocytes require basic fibroblast growth factor (bFGF) and cAMP. Without these factors, even in the presence of serum, the cells die. Melanocytes cultured in the presence of keratinocytes, however, survive for weeks without added bFGF and cAMP. We show here that the growth factor for melanocytes produced by human keratinocytes is bFGF because its activity can be abolished by neutralizing antibodies to bFGF and by a bFGF synthetic peptide that inhibits the binding of the growth factor to its receptor. The melanocyte mitogen in keratinocytes is cell associated and increases after irradiation with ultraviolet B. Northern blots reveal bFGF gene transcripts in keratinocytes but not melanocytes. These studies demonstrate that bFGF elaborated by keratinocytes in vitro sustains melanocyte growth and survival, and they suggest that keratinocyte-derived bFGF is the natural growth factor for normal human melanocytes in vivo.

Normal human melanocytes, unlike pigment cells from metastatic melanomas, do not survive in culture in routine, serum-supplemented media. The search for natural growth factors for melanocytes has shown that mitogenic activity is ubiquitous in several tissues and in melanomas. Of several known growth factors tested, basic fibroblast growth factor (bFGF) was the only one mitogenic for melanocytes but only in the presence of cyclic-adenosine-monophosphate (cAMP) stimulators. The mitogenic activity toward melanocytes in tissues and melanoma cell extracts had high affinity for heparin and antibodies to bFGF synthetic peptides. These results suggest that one of the growth factors for melanocytes might be bFGF or a bFGF-like polypeptide and that autocrine production of bFGF-like molecules by melanoma cells may contribute to the malignant phenotype of melanocytes. Because acidic FGF (aFGF) did not stimulate growth, the receptors for bFGF on melanocytes might be significantly different from those for a FGF.