signaling. Extrapolating from current structural models, we suggest that FGFR dimerization and autophosphorylation is supported by cooperative "heparin-like end structures," and that cell surface association and concentration compensate for the relative scarcity of such end structures in native HSPGs. In this model, "proteolytic" shedding of heparan sulfate would act as a diluting, down-regulatory mechanism, while "heparanolytic" shedding might act as an up-regulatory mechanism, by increasing the concentration of these end structures.
Fibroblast growth factors (FGFs)1 bind not only to their cognate receptors (FGFRs) but also to heparan sulfate proteoglycans (HSPGs). HSPGs are associated with the cell surface of many, if not most, cell types. Most known HSPG functions are contributed by the interactions of the heparan sulfate (HS) chains of these molecules. Besides growth factors, these HSPGs interact with various adhesion molecules, protease inhibitors, and enzymes, modifying the spatial distributions and activities of these ligands. Among FGFs, the interaction with FGF2 has been studied most intensively, and it is now generally accepted that HSPGs play important roles in FGF2 signaling. Initially, the association of FGF2 with HS has been proposed to protect this FGF from proteolysis and thermal denaturation (1, 2) and to serve as a reservoir of growth factor that can be released by enzymes that degrade the proteoglycans (1). Later, HSPGs were identified as co-receptors for FGF2, strongly promoting FGF-FGFR binding and the subsequent activation of the receptor (3, 4). Recently, genetic studies in Drosophila provided compelling evidence that HSPGs are essential for FGF signaling in vivo (5).Although the importance of HSPG in FGF-signaling is well documented, the nature of the "co-receptor" and the precise mechanisms at work are less well characterized. Distinctive core protein structures define two major families of cell surfaceassociated HSPGs: syndecans and glypicans (6). Prior work from our laboratory showed that syndecans and glypican-1 stimulate FGF2-FGFR1 interaction and signaling in K562 cells, at least when co-expressed with receptor in these cells (4). This agrees with the work of other groups, showing that cell surface syndecan-1 from Raji cells acts as a positive regulator of FGF2 binding and signaling (7), that syndecan-2 on human macrophages promotes FGF2-mediated proliferation (8), and that glypican-1 can stimulate FGF2 signaling (9, 10). Meanwhile, there are also other, contradictory reports. Syndecans and glypican-1 purified from human lung fibroblast extracts are unable to promote high affinity binding of FGF2 to FGFR1 (11), overexpression of syndecan-1 in NIH 3T3 cells inhibits FGF2-induced proliferation (12), and HSPGs purified from endothelial secretions prevent FGF2 binding to vascular smooth muscle cells and inhibit FGF2-induced mitogenesis (13). One possible explanation for this discrepancy is that the HSPGs were from different sources and might have had different compositions. It is well k...