Dermatan sulfate binds and potentiates activity of keratinocyte growth factor (FGF-7).

FGF-7 is induced after injury and induces the proliferation of keratinocytes. Like most members of the FGF family, the activity of FGF-7 is strongly influenced by binding to heparin, but this glycosaminoglycan is absent on keratinocyte cell surfaces and minimally present in the wound environment. In this investigation we compared the relative activity of heparan sulfate and chondroitin sulfate B (dermatan sulfate), glycosaminoglycans that are present in wounds. A lymphoid cell line (BaF/KGFR) containing the FGF-7 receptor (FGFR2 IIIb) was treated with FGF-7 and with various glycosaminoglycans. FGF-7 did not support cell proliferation in the absence of glycosaminoglycan or with addition of heparan sulfate or chondroitin sulfate A/C but did stimulate BaF/KGFR division in the presence of dermatan sulfate or highly sulfated low molecular weight fractions of dermatan. Dermatan sulfate also enabled FGF-7-dependent phosphorylation of mitogen-activated protein kinase and promoted binding of radiolabeled FGF-7 to FGFR2 IIIb. In addition, dermatan sulfate and FGF-7 stimulated growth of normal keratinocytes in culture. Thus, dermatan sulfate, the predominant glycosaminoglycan in skin, is the principle cofactor for FGF-7.

FGF-7 is induced after injury and induces the proliferation of keratinocytes. Like most members of the FGF family, the activity of FGF-7 is strongly influenced by binding to heparin, but this glycosaminoglycan is absent on keratinocyte cell surfaces and minimally present in the wound environment. In this investigation we compared the relative activity of heparan sulfate and chondroitin sulfate B (dermatan sulfate), glycosaminoglycans that are present in wounds. A lymphoid cell line (BaF/KGFR) containing the FGF-7 receptor (FGFR2 IIIb) was treated with FGF-7 and with various glycosaminoglycans. FGF-7 did not support cell proliferation in the absence of glycosaminoglycan or with addition of heparan sulfate or chondroitin sulfate A/C but did stimulate BaF/KGFR division in the presence of dermatan sulfate or highly sulfated low molecular weight fractions of dermatan. Dermatan sulfate also enabled FGF-7-dependent phosphorylation of mitogen-activated protein kinase and promoted binding of radiolabeled FGF-7 to FGFR2 IIIb. In addition, dermatan sulfate and FGF-7 stimulated growth of normal keratinocytes in culture. Thus, dermatan sulfate, the predominant glycosaminoglycan in skin, is the principle cofactor for FGF-7.
Glycosaminoglycans (GAGs) 1 and proteoglycans are emerging as key regulators of a variety of cellular behaviors involved in development, homeostasis, and disease. Heparin, heparan sulfate (HS), and hyaluronic acid are GAGs that have been studied extensively in relation to their roles in anticoagulation, growth factor signaling, and connective tissue support. Less well studied, dermatan sulfate (DS) is the predominant GAG expressed in the skin and is released at high concentrations during wound repair, making it a particularly interesting topic for evaluation in relation to growth factors active in wounds. Wound fluid-derived DS, as well as physiologic concentrations of commercially purified DS, has been shown to promote FGF-2 function (1). DS and DS proteoglycans also bind to and influence the activity of other heparin-and HS-binding proteins including hepatocyte growth factor/scatter factor (2), thrombin (3), heparin cofactor II (4,5), fibronectin (6), platelet factor-4 (7), regulated on activation normal T cell expressed and secreted (RANTES) (8), and interferon-␥ (9, 10). Thus, although less well understood, DS-protein interactions are likely to be important events in control of several cellular behaviors. FGF-7 or keratinocyte growth factor is a polypeptide mitogen that belongs to the family of fibroblast growth factors. Whereas some FGF family members bind to multiple FGF receptors (FGFRs 1-4) (11) and stimulate proliferation in a variety of cell types, FGF-7 binds only to a splice variant of FGFR2 (FGFR2 IIIb) and is a highly specific paracrine growth factor for epithelial cells (12,13). FGF-7 and its receptor are believed to be important for normal wound healing. The expression of FGF-7 and FGFR2 IIIb are induced in wounds (14,15), and the application of FGF-7 to wounds has been shown to promote healing (16,17). Heparin and HS can support FGF-7 signaling, like that of other FGF family members. These GAGs are believed to act as stabilizing cofactors for the ligand-receptor interaction (18 -20). Interestingly, heparin can also inhibit FGF-7 receptor binding and activity at higher concentrations in some cells (21)(22)(23). Because DS and FGF-7 are both induced in wounds, and the FGF-7 signaling pathway is regulated differentially by heparin compared with other FGF ligands, we have asked whether DS serves as a cofactor for FGF-7.
In this study, we used a previously described GAG-and FGFR-deficient cell line engineered to express the FGF-7 receptor (FGFR2 IIIb) (21) to ask whether the addition of exogenous DS affects FGF-7 binding to and signaling through its receptor. Our results show that DS is superior to HS and approaches the effectiveness of heparin in its ability to promote FGF-7-FGFR2 IIIb binding and cellular proliferation. This effect is specific for FGF-7 as the activity of another ligand of FGFR2 IIIb, FGF-1, is not supported by the addition of exogenous DS. These findings strongly support the premise that DS is the physiologically relevant cofactor for FGF-7.

Materials
Heparin sodium salt from porcine intestine, heparan sulfate from bovine kidney, heparan sulfate from porcine intestine, and chondroitin sulfate A/C (70% A and 30% C) from bovine trachea were purchased from Sigma. Dermatan sulfate (chondroitin sulfate B) with a molecular mass range from 11 to 25 kDa was purchased from Seikagaku America (Falmouth, MA). Unfractionated dermatan sulfate (MW 35 ϩ 5) and DS fractions DO (MW 50 ϩ 5) and DT (MW 25 ϩ 5) were purchased from Celsus Laboratories, Cincinnati, OH. Human recombinant FGF-7 and FGF-1 were from R & D Systems. Disuccinimidyl suberate was from Pierce. Carrier free Na 125 I was from Amersham Biosciences. Unless indicated, all other chemicals were purchased from Sigma. HiTrap™ heparin HP columns and QAE-Sephadex A-25 beads were purchased from Amersham Biosciences. Penicillin/streptomycin, L-glutamine, trypsin, and RPMI 1640 were obtained from Invitrogen. Medium 154CF and human keratinocyte growth supplement for NHK culture were purchased from Cascade Biologicals. Fetal calf serum was purchased from Hyclone. The Cell Titer 96 non-radioactive cell proliferation kit was purchased from Promega, and the Blyscan proteoglycan and GAG assay system was obtained from Accurate Chemical Scientific Corp.

Wound Fluid Glycosaminoglycan Isolation and Quantitation
Human wound fluid was collected, and GAGs were isolated and quantified as described previously (1). Briefly, GAGs were isolated from post-surgical wound fluid collected within 24 h by anion exchange chromatography using QAE-Sephadex A-25 beads (Amersham Biosciences). Sulfated GAG was measured with the sulfate-binding cationic dye, dimethylene blue, according to the manufacturer's instructions for the Blyscan proteoglycan and GAG assay system (Accurate Chemical Scientific Corp.).

Cell Culture and Proliferation Assays
BaF/KGFR Cells-Preparation and culture of mouse lymphocyte BaF3 cells stably transfected with FGFR2 IIIb and designated BaF/ KGFR has been described previously (21). Cells were selected routinely in heparin (5 g/ml) and FGF-1 (10 ng/ml) in RPMI containing Lglutamine (2.92 mg/ml), penicillin (100 units/ml), streptomycin (50 g/ ml) for at least 1 week and then cultured in RPMI containing 10% interleukin-3 conditioned media, 10% FCS, L-glutamine, and penicillin/ streptomycin. Interleukin-3 conditioned media was prepared from WEHI-3B cells grown in RPMI ϩ10% FCS (24). Prior to GAG treatment, cells were rinsed in additive-free RPMI three times to remove any traces of cytokine. Cells were seeded in 96-well plates at 2 ϫ 10 5 cells/well in a final volume of 100 l/well. GAG and ligand dilutions were made in RPMI containing 10% FCS, penicillin/streptomycin, and L-glutamine. Cells were cultured for 48 to 72 h, and cell proliferation was determined.
NHK Cells-Normal human keratinocytes were isolated and cultured as described (25). Cells were maintained in 154CF medium with human keratinocyte growth supplement (0.2% (v/v) bovine pituitary extract, 5 g/ml bovine insulin, 0.18 g/ml hydrocortisone, 5 g/ml bovine transferrin, and 0.2 ng/ml human epidermal growth factor), penicillin (100 units/ml), streptomycin (50 g/ml), L-glutamine (2.92 mg/ml), and 0.06 mM Ca 2ϩ ). For proliferation assays, cells were seeded in 96-well plates at 2 ϫ 10 5 cells/well. When cell confluence reached ϳ30%, cells were rinsed three times in growth factor-free medium (154CF medium without human keratinocyte growth supplement, designated basal medium). GAGs and FGF-7 were diluted in basal medium for stimulation. Cells were cultured for 48 h, and cellular proliferation was determined. Proliferation assays were carried out according to the manufacturer's instructions for the Promega Cell Titer 96 non-radioactive cell proliferation kit. Relative cell number was determined as the quotient of the A 570 determined for experimental samples to A 570 of cells in identical wells grown in basal medium alone. In some experiments, results of proliferation assay were confirmed by manual cell counts using a hemocytometer.

Phosphorylation of p44 and p42 MAP Kinase
BaF/KGFR cells in culture were washed three times and then incubated for 18 h in RPMI containing 0.1% FCS, penicillin/streptomycin, and L-glutamine prior to addition of test reagents. After 2 h of exposure, 3 ϫ 10 5 cells were extracted in 0.05 ml of lysis buffer (50 mM Tris-HCL, pH 7.0, 150 mM NaCl, 1 mM EGTA, 100 mM NaF, 10% glycerol, 1.5 mM MgCl 2 , 1% Triton X-100, 1 mM sodium orthovanadate, 0.1 mM phenylmethylsulfonyl fluoride), and 15 l of this extract were separated on a 10% SDS-PAGE gel prior to transfer to nitrocellulose (Osmonics, Westborugh, MA). Western blot was performed with anti-phospho MAP kinase or anti-non-phospho MAP kinase per the manufacturer's instructions (Cell Signaling Technology, Inc.). Phosphorylated and nonphosphorylated p44/42 control proteins (Cell Signaling Technology, Inc.) were evaluated simultaneously to confirm sensitivity and specificity of Western blot.

Radioiodination of FGF-7, Cell Surface Receptor Binding, and Cross-linking
Recombinant FGF-7 was radioiodinated using the chloramine T method as described (26). Free iodine was separated from labeled FGF-7 using HiTrap™ heparin-Sepharose columns. The specific activity of the labeled ligand was low, ϳ3 ϫ 10 4 cpm/ng. FGF-7 binding to BaF/KGFR cells was performed as described (26). Briefly, 2.5 ϫ 10 6 cells were suspended in 500 l of binding buffer in the presence or absence of GAGs (1 g/ml heparin, 50 g/ml DS, 10 g/ml DT, 100 g/ml DO, and 100 g/ml HS) and I 125 -FGF-7 (5 ng/ml) and incubated for 2 h at 4°C. Chemical cross-linking was performed according to the manufacturer's instructions with disuccinimidyl suberate. Following cross-linking, cells were lysed in 50 l of buffer (10 mM Tris, pH 7, 1% (v/v) Nonidet P-40, 1 mM EDTA). Binding was quantified using a Genesys 5000 series multiwell ␥-counter. Receptor⅐ligand complexes were resolved on a 4 -20% gradient gel and visualized using a Molecular Dynamics PhosphorImager Screen and Storm Phosphorimager.

GAGs Isolated from Wound Fluid Support FGF-7-dependent
Cellular Proliferation-Human wound fluid contains abundant amounts of soluble GAGs, with DS and HS at concentrations up to 65 g/ml. At these physiologic concentrations both DS and HS can promote FGF-2-dependent proliferation (1). To determine whether wound fluid-derived glycosaminoglycan (WFGAG) stimulates FGF-7-dependent signaling, we used BaF3 cells, a GAG-deficient mouse lymphocyte cell line stably expressing the FGFR2 IIIb receptor (designated BaF/KGFR) that binds FGF-7 and FGF-1 (11,21,27,28). Cells were cultured with 50 g/ml WFGAG or 0.5 g/ml heparin with or without 5 ng/ml FGF-1 or FGF-7 for 72 h, and cellular proliferation was measured. Neither ligand nor GAG alone significantly stimulated cell growth above background levels (Fig. 1).
In the presence of heparin, FGF-7 stimulated growth by ϳ3.0fold. WFGAG resulted in an approximate doubling of the proliferative effect of FGF-7.
Dermatan Sulfate Acts as a Cofactor for FGF-7-but Not FGF-1-dependent Cellular Proliferation-We next asked whether the addition of pure DS, a major component of WFGAG, to BaF/KGFR cells would affect FGF-7 signaling. Cells were cultured with commercially available purified DS (at 50 g/ml). The absence of contaminating heparin or HS in the DS preparations was confirmed by independent analysis of monosaccharide composition by the University of California, San Diego Glycobiology Core Facility. Cellular proliferation increased in a dose-dependent manner with increasing amounts of FGF-7 (Fig. 2a). By contrast, FGF-1 was not as effective in promoting cell proliferation with DS. As expected based on prior reports (29), heparin (1 g/ml) enhanced the proliferative activity of both FGF-1 and FGF-7 (data not shown).
The concentration of sulfated GAGs in wound fluid has been reported to range between 15 and 65 g/ml (1). We saw more FIG. 1. WFGAGs support FGF-7-dependent proliferation in BaF/KGFR cells. Cells were seeded in 96-well plates as described with 5 ng/ml FGF-1 or FGF-7 and heparin (1 g/ml) or human wound fluid-derived GAG (WFGAG) (50 g/ml) as indicated. Following a 3-day incubation, cell number was determined as described under "Experimental Procedures." Relative cell number is expressed as a proportion of untreated control. Data are mean Ϯ S.E. of triplicates and are representative of five independent experiments. than a 2-fold stimulation when 50 g/ml DS was combined with 5 ng/ml or more of FGF-7 but only ϳ1.5-fold stimulation with FGF-1 in similar amounts. To determine the effect of a range of DS concentrations on FGF-1 and FGF-7, we compared approximately equipotent amounts of FGF-1 (5 ng/ml) and FGF-7 (2 ng/ml, the minimum concentration capable of inducing an increase in cell proliferation at 50 g/ml DS) (Fig. 2b). Both FGF-7 and FGF-1 showed a dose-dependent increase in cell proliferation in response to DS, but DS always potentiated a greater increase in cell proliferation by FGF-7. A 2-fold increase in cell number was induced by 2.5 g/ml DS and FGF-7 whereas 250 g/ml DS was required to stimulate a similar response in the presence of FGF-1.
Dermatan Sulfate Supports FGF-7-dependent Cellular Proliferation More Than Heparan Sulfate or Chondroitin Sulfate A/C-Heparin is frequently used in vitro as a required cofactor for FGF-dependent signaling and to study the three-dimensional structure of the GAG⅐FGFR⅐FGF complex believed to be required for this signaling to take place (30 -32). However, as mast cells (the sole source of cutaneous heparin) are present in very low numbers in the skin under physiologic conditions, we asked whether the abundant native GAGs (HS and chondroitin sulfates (CS)) support FGF-7-dependent cellular proliferation. Because GAG structure varies between tissues, and this may affect GAG-FGF binding, we tested HS from two sources (33,34). Porcine intestinal heparin, bovine kidney HS, porcine intestine HS, CS A/C (70% CS A and 30% CS C), and DS were tested with 5 ng/ml FGF-7 at concentrations ranging from 0 to 50 g/ml. Consistent with our previous results, heparin led to the highest increase in proliferation at low doses, stimulating it ϳ3-fold over control levels at a concentration of 10 g/ml (Fig.  3). At concentrations similar to those found in wounds, DS stimulated FGF-7-dependent proliferation to a maximum of 4.5-fold at 25 g/ml. Comparatively, HS from porcine origin and CS A/C had minimal stimulatory effects on FGF-7 although bovine kidney HS had activity at high concentrations (50 g/ml) that approached heparin.
DS Supports FGF-7-dependent Phosphorylation of MAP Kinase-To directly evaluate rapid downstream events after FGF-R2 occupancy, the phosphorylation of p44 and p42 MAP kinases (extracellular signal-regulated kinases 1 and 2) was studied in BaF/KGFR cells. Cells were serum-starved in 0.1% FCS for 18 h and then exposed for 2 h to fresh medium alone, FGF-7, DS, or the combination of FGF-7 and DS. Phosphorylation of p44/42 MAP kinases was only seen when DS and FGF-7 were present together (Fig. 4).
DS Fractions Have Differential Effects upon FGF-7-dependent Cellular Proliferation-The ability of certain GAGs to bind to and stimulate their protein partners depends upon their structure, size (number of disaccharides), and sulfation pattern. These elements combine to form discrete preferentially bound cassettes within the heterogeneous GAG molecule (35)(36)(37)(38)(39). To begin to investigate this possibility for DS, we used unfractionated and fractionated DS species from Celsus Laboratories (designated DO and DT, respectively) that differ by size and sulfation (Table I). These preparations of DS, which differentially affect the anticoagulant effect of activated protein C (40), were added to BaF/KGFR cells with 5 ng/ml FGF-7. Native, unfractionated DS and the low molecular weight fraction (DT) at 50 g/ml resulted in a 4.5-and 5-fold induction in cellular proliferation, respectively (Fig. 5). The high molecular weight fraction (DO) induced cell growth by only 1.6-fold at this concentration. Increasing the concentration of DO by 10-fold (to FIG. 2. FGF-7-dependent cellular proliferation is enhanced in a dose-dependent manner by increasing either the dermatan sulfate or the ligand. a, dose response to FGF-7 or FGF-1. Cells were seeded in 96-well plates as described in the presence of 50 g/ml dermatan sulfate and treated with increasing concentrations of FGF-1 or FGF-7. Cells were cultured for 72 h, and cellular proliferation was determined. Relative cell number is expressed as a proportion of control in the absence of dermatan sulfate. Data are mean Ϯ S.E. of triplicates and are representative of three independent experiments. b, dose response to dermatan sulfate. BaF/KGFR cells were seeded in a 96-well plate as described in the presence of 5 ng/ml of FGF-1 or 2 ng/ml FGF-7. DS was added at increasing concentrations (3.9 to 250 g/ml), cells were cultured for 72 h, and cellular proliferation was determined. Relative cell number is expressed as a proportion of control in the absence of dermatan sulfate. Data are mean Ϯ S.E. of triplicates and are representative of two independent experiments. 500 g/ml) did not result in an effect equal to that of the other species. Interestingly, DT supports FGF-7-dependent cellular proliferation at a 10-fold lower dose (5 g/ml) than unfractionated DS.
Dermatan Sulfate Enhances the Binding of FGF-7 to Its Cell Surface Receptor-Based upon our assumption that the proliferative effects of DS stem from its ability to promote the binding of FGF-7 to its receptor, we next attempted to demonstrate this effect. Radioiodinated FGF-7 was incubated with BaF/ KGFR cells in the presence of heparin (1 g/ml), DS (50 g/ml), DT (10 g/ml), DO (100 g/ml), and HS (100 g/ml), and a cross-linking reaction was performed. The GAG concentrations were chosen based upon the results obtained in our proliferation assays. DS increased the binding of labeled FGF-7 by 6.9-fold compared with control (Fig. 6a, no GAG). Heparin increased binding by 9.1-fold. In agreement with the proliferation assay results, DT exceeded the amount of binding supported by unfractionated DS and approached that of heparin (8-fold increase over control). In all cases binding was observed as two bands resolved by SDS-PAGE that correspond to the expected sizes of a monomer and dimer of the receptor⅐ligand complex (Fig. 6b). DO had no significant effect on binding, and HS did not support binding above background. All reactions were carried out using molar concentrations of GAGs equivalent to those used in proliferation assays.
Dermatan Sulfate Enhances the Proliferative Effect of FGF-7 in Normal Human Keratinocytes-We next extended our results to a more physiologically relevant cell system. FGF-7 is a potent mitogen for keratinocytes and has also been shown to modulate differentiation (13,41). Primary cultures of human keratinocytes in basal medium were treated with 5 ng/ml FGF-7 in the presence of heparin or DS. Cellular proliferation was assessed after 48 h in culture. DS at 5 g/ml supported FGF-7-dependent growth in keratinocytes to a level that exceeded heparin and FGF-7 or FGF-7 alone (Fig. 7). Interestingly, this effect was seen at a 10-fold lower concentration than was required for FGF-7-stimulated proliferation in the BaF/ KGFR system but decreased at higher DS concentrations. A similar trend was noted for FGF-10, another FGFR2 IIIb ligand that has been shown to promote proliferation in keratinocytes (data not shown) (42). As expected, no growth was observed for cells treated with GAG alone.   5. Dermatan sulfate fractions differentially affect FGF-7dependent cellular proliferation. BaF/KGFR cells were seeded in 96-well plates as described with 5 ng/ml FGF-7 with the following glycosaminoglycans: heparin 0.5 g/ml, dermatan sulfate 50 g/ml, large MW dermatan sulfate (DO) at 5, 50, and 500 g/ml, and low molecular weight dermatan sulfate (DT) at 5, 50, and 500 g/ml. Cells were cultured for 72 h, and cellular proliferation was determined. Relative cell number is expressed as a proportion of control for each GAG in the absence of FGF-7. Data are mean Ϯ S.E. of triplicates and are representative of three independent experiments.

DISCUSSION
In the present work we have demonstrated how a physiologically relevant WFGAG, DS, affects FGF-7. Most FGF family members require the presence of GAGs to bind to and signal through their tyrosine kinase receptors (24,43,44). These effects presumably result from either (a) the stabilization of the receptor⅐ligand complex (45,46) or (b) by acting as an extracellular stockpile for growth factors (47,48). Heparin and HS have been the traditional focus of GAG-FGF interactions. These GAGs share structural aspects, but their availability to cells in vivo is not likely to be equivalent. HS is ubiquitous in the extracellular matrix, cell surface, and basement membrane. DS is located primarily in the extracellular matrix of skin. Heparin, however, is sequestered in mast cells and only becomes available to the tissue upon recruitment and degranulation of this cell type. DS is more similar structurally to HS than the other chondroitin sulfates as both DS and HS contain iduronate residues. The core proteins to which DS and HS are attached show altered expression patterns during development, pathogenesis, and wounding (49 -52). No such regulation has been shown for serglycin, the primary core protein on which heparin is found. These observations suggest that heparin may have limited relevance to the functional significance of GAG and growth factor interactions seen in vitro to an in vivo system such as wound repair where FGF-7 is expressed abundantly.
Previous work has demonstrated that FGF-2-dependent activity is enhanced by DS derived from wound fluid (1). The present data show that FGF-7-dependent proliferation is also enhanced by a wound fluid-derived GAG mixture (Fig. 1) and that this effect can be reproduced with the use of purified DS (Fig. 2). One explanation for the observed effect of DS upon FGF-7 signaling is contamination of the DS reagent with heparin or HS. Analysis performed by the Glycobiology Core Facility at our institution (by acid hydrolysis followed by high pH anion exchange chromatography with pulsed amperometric detection) found that the DS used in our assays contains N-acetyl galactosamine, iduronate, and glucuronate. N-Acetyl glucosamine was not detected significantly, confirming the absence of heparin and HS contamination. Furthermore, the proliferative effect of DS was likely to be directly through the interaction of FGF-7 with its receptor, because downstream phosphorylation of MAP kinase was observed within 2 h. Within this time period secondary effects on cell growth that could manifest themselves within the 48 -72-h time periods of the proliferation assays become less likely.
WFGAG and DS stimulated FGF-7-dependent cell proliferation to a much greater extent than FGF-1 (see Fig. 1 and Fig.  2). Even at high, and likely supraphysiologic, concentrations of DS (Fig. 2b) or FGF-1 (Fig. 2a), FGF-1-dependent cellular proliferation was not enhanced, suggesting that the effect of DS is selective for FGF-7. This may represent the physiological situation or reflect diminished activity of the recombinant growth factors used in this study. However, this differential effect of GAGs upon FGF-dependent binding and activity has been shown previously. Glypican-1, a heparan sulfate proteoglycan, enhances FGF-1 receptor binding and cellular proliferation but does not support either the binding or activity of FGF-7 in BaF/KGFR cells and mouse keratinocytes (21,53). Our data support these results. HS had less effect upon FGF-7-dependent cellular proliferation over a range of concentrations (Fig. 3) and did not support cell surface receptor binding (Fig. 6). The differential potentiation and binding of FGF-1 and FGF-7 are supported by, and in the case of HS likely explained by, their dependence upon unique oligosaccharides that differ by size and sulfation patterns (29). In addition, we have shown that the size and sulfation level of DS affect both proliferation and receptor binding (see Fig. 5 and Fig. 6). Proliferation and cell surface binding studies performed with equivalent molar concentrations of GAG and ligand showed similar results, thus supporting the hypothesis that the action of DS is similar to mechanisms described previously for heparin. The highest molecular weight DS fraction, DO (charge to mass ratio 0.12-0.13), had minimal effectiveness in promoting FGF-7-dependent proliferation and did not support FGF-7 binding to FGFR2 IIIb. By contrast, the lowest molecular weight fraction, DT (charge to mass ratio 0.23-0.29), has a higher charge to mass density than either DO or unfractionated DS and was more active and potent than native DS. This fraction approached the effectiveness of heparin in promoting cellular proliferation. This derivative is also more active than either unfractionated DS or DO in enhancing activated protein C-dependent anticoagulation (40). DT is enriched in 4,6-di-O-sulfated N-acetyl galactosamine residues, but no other structural information is available currently. Unfractionated heparin exhibits differential effects upon FGF-7, supporting receptor binding and cellular proliferation at low concentrations but inhibiting both events at higher doses (21,22). We did not observe an inhibitory effect by heparin upon either binding FGF-7 to FGFR2 IIIb or FGF-7-dependent proliferation in our BaF/KGFR studies. This is likely because of the relatively low concentration of heparin (1 g/ml) used in our assays. Prior studies have reported that heparin enhances FGF-dependent cellular proliferation at this concentration (21,29). We did however observe inhibition of FGF-7-dependent proliferation at increasing concentrations of DS in normal human keratinocytes. Thus, as has been reported previously with heparin, facilitation of FGF-7 activity by DS is highly dependent on the concentration of DS available to the cell.
One question raised by our findings is whether the nature of DS in wound fluid differs (in terms of oligosaccharide size) from that present normally in the extracellular matrix. Other molecules in the extracellular matrix and the skin microenvironment, such as the perlecan core protein, can also serve as ligands for FGF-7 (54). Thus, the structural requirements for FGF-7 binding and potentiation of receptor activation are complex. Further investigation of DS structure is needed to characterize how DS size and sulfation affect FGF-7:FGFR2 IIIb FIG. 7. DS enhances FGF-7-dependent growth in normal human keratinocytes. Normal primary human keratinocytes were maintained in defined growth medium supplemented with bovine pituitary extract, bovine transferrin, bovine insulin, human epidermal growth factor, and hydrocortisone prior to being "stepped down" into basal medium lacking growth supplements. 5 ng/ml FGF-7 was added to the culture medium alone or in the presence of 5 to 250 g/ml dermatan sulfate or 1 g/ml heparin. Cells were maintained in culture for 48 h when cellular proliferation was determined. Data are mean Ϯ S.E. of triplicates and are representative of three independent experiments. binding and the affinity constants of these interactions. As more structural information becomes available from the study of the various GAGs, it will become clear that particular subdomains within GAGs (e.g. highly sulfated regions rich in iduronate residues), or other ligands, will prove to differentially affect FGF family member signaling. Such information will not only advance our understanding of the basic mechanisms of GAG function but will also help us understand the basic science of wound healing. The current observations illustrate that cellular synthesis and availability of GAGs must be considered when extending models of GAG and growth factor interactions from the test tube to the tissue. In the case of events associated with the skin, DS may exceed HS for many heparin-dependent processes characterized previously.