THE JOURNAL OF BIOLOGICAL CHEMISTRY © 1998 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 273, No. 43, Issue of October 23, pp. 28116 –28121, 1998 Printed in U.S.A.

Dermatan Sulfate Released after Injury Is a Potent Promoter of Fibroblast Growth Factor-2 Function* (Received for publication, December 3, 1997, and in revised form, June 9, 1998)

Stanley F. Penc‡§, Bohdan Pomahac¶, Thomas Winkler¶, Robert A. Dorschner‡, Elof Eriksson¶, Mary Herndoni, and Richard L. Gallo‡§** From the ‡Division of Developmental and Newborn Biology, Boston’s Children’s Hospital, the §Department of Dermatology, Harvard Medical School, the ¶Department of Plastic Surgery, Brigham and Women’s Hospital, and the iDepartment of Experimental Pathology, Beth Israel Deaconess Medical Center, Boston, Massachusetts 02115

Proteoglycans have been shown in vitro to bind multiple components of the cellular microenvironment that function during wound healing. To study the composition and function of these molecules when derived from an in vivo source, soluble proteoglycans released into human wound fluid were characterized and evaluated for influence on fibroblast growth factor-2 activity. Immunoblot analysis of wound fluid revealed the presence of syndecan-1, syndecan-4, glypican, decorin, perlecan, and versican. Sulfated glycosaminoglycan concentrations ranged from 15 to 65 mg/ml, and treatment with chondroitinase B showed that a large proportion of the glycosaminoglycan was dermatan sulfate. The total glycosaminoglycan mixture present in wound fluid supported the ability of fibroblast growth factor-2 to signal cell proliferation. Dermatan sulfate, and not heparan sulfate, was the major contributor to this activity, and dermatan sulfate bound FGF-2 with Kd 5 2.48 mM. These data demonstrate that proteoglycans released during wound repair are functionally active and provide the first evidence that dermatan sulfate is a potent mediator of fibroblast growth factor-2 responsiveness.

Proteoglycans are glycosaminoglycan (GAG)1-containing molecules characterized by core protein structure and the size and type of associated GAG(s) (1– 4). Heparan sulfate GAGs bind a plethora of molecules including several growth factors, cytokines, cell adhesion molecules, matrix proteins, proteases, and antiproteases (2, 5–9). Although the biological significance of these binding interactions remains unclear, the unique ability to interact with a vast array of ligands enables proteoglycans to influence several cell behaviors. This influence becomes of particular importance during inflammation and the response to injury when heparan sulfate-containing proteoglycans are thought to control the interaction between numerous cells, matrix components, and soluble effectors (10, 11). Dermatan sulfate may also interact with several molecules primarily thought of as heparan sulfate-binding proteins, e.g. fibroblast growth factor-2 (FGF-2) (12), hepatocyte growth factor/scatter

* 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. ** To whom correspondence should be addressed: Enders 950, Children’s Hospital, 300 Longwood Ave., Boston, MA 02115. Tel.: 617-3557678; Fax: 617-713-4340; E-mail [email protected]. 1 The abbreviations used are: GAG, glycosaminoglycan; FGF, fibroblast growth factor; FGFR1, FGF receptor-1; BSA, bovine serum albumin; ACE, affinity co-electrophoresis; PMSF, phenylmethylsulfonyl fluoride; MOPSO, 3-(N-morpholino)-2-hydroxypropanesulfonic acid; WFGAG, wound fluid GAG; mAb, monoclonal antibody.

factor (13), heparan cofactor II (14, 15), platelet factor 4 (16), fibronectin (17), and protein C inhibitor (18). Thus, multiple types of GAG are likely important during inflammation, the response to injury, and related processes. Proteoglycan expression is regulated during wound repair. Abundant syndecan-1 and -4 are induced in response to injury on endothelia and fibroblasts, cell types that normally express only low amounts of these proteoglycans (10, 19, 20). Soluble syndecan-1 and -4 extracellular domains have also been identified in wounds (21). The presence of both soluble and cellsurface proteoglycans provides multiple sites where cell behaviors required for wound healing may be regulated. However, although proteoglycan function has been extensively studied in vitro, little is known about native proteoglycans or GAG released in vivo. To determine whether soluble proteoglycan and/or GAG isolated from an in vivo source can function to support cell behaviors mediated by GAG-binding ligands, we studied the ability of purified human wound fluid GAG (WFGAG) to support FGF2-mediated cell proliferation. Fibroblast growth factors (FGFs) are an important family of at least nine structurally related GAG-binding molecules (11). FGF-2, the best characterized member of the FGF family, is present in wounds and can function as a mitogen that signals mesenchymal cell migration, proliferation, and differentiation (22–25). The interaction between FGF-2 and heparin, or heparan sulfate, has been well characterized, and function has been reported to depend on binding to these GAGs (26 –28). A model has been proposed in which cell-surface heparan sulfate functions as a low affinity co-receptor for FGF-2 and presents the growth factor to its high affinity signaling receptors (27–29). We report the presence in wounds of abundant soluble GAG, largely in the form of dermatan sulfate, that supports the ability of FGF-2 to signal cell proliferation. Elimination of dermatan sulfate from WFGAG resulted in an 85% reduction in FGF-2 activity. These findings directly demonstrate that proteoglycans released from in vivo sources can support cell behaviors required for wound healing. Furthermore, this study demonstrates that in addition to the ability of heparan sulfate to support FGF-2 activity, soluble dermatan sulfate also supports FGF-2-mediated cell proliferation and is likely an important molecule involved in the regulation of wound repair. EXPERIMENTAL PROCEDURES

Materials—Recombinant human FGF-2 was the generous gift of Dr. M. Klagsbrun, Harvard Medical School (Boston). Bovine serum albumin (BSA), chondroitin sulfate B lyase (chondroitinase B), porcine intestinal heparin, chondroitin sulfate ABC, and heparan sulfate were from Sigma. Unless indicated, porcine skin chondroitin sulfate B (dermatan sulfate) with molecular mass range 11–25 kDa and determined pure by infrared spectrophotometry was from Seikagaku America Inc. (Rockville, MD). The total nitrogen, sulfur, galactosamine, and iduronic acid

28116

This paper is available on line at http://www.jbc.org

Glycosaminoglycans and FGF-2 Responsiveness content of this dermatan sulfate preparation was 2.82, 6.77, 32.9, and 39.0%, respectively. The Blyscan Proteoglycan and GAG Assay System was purchased from Accurate Chemical and Scientific Corp. (Westbury, NY). Complete EDTA-free Proteases Inhibitor Mixture was from Boehringer Mannheim. IODO-BEADS were purchased from Pierce. Immunochemicals—Monoclonal antibody DL-101 specific for human syndecan-1 ectodomain was kindly provided by Dr. M. Bernfield, Harvard Medical School (Boston). Monoclonal antibodies 10H4 (30) and 1C7 (31) specific for human syndecan-2 and -3 ectodomains, respectively, and monoclonal antibodies against human perlecan (matrix mix) (32) and glypican (S1) (33) were generously provided by Dr. G. David, University of Leuven (Leuven, Belgium). Monoclonal antibody 5G9 was against human syndecan-4 ectodomain (10). Monoclonal antibody LF136 (34), previously known as LF-30, specific for human decorin was the gift of Dr. L. Fisher, NIH (Bethesda). Monoclonal antibody mAb-17 specific for human epican (35) was provided by Dr. L. Milstone, Yale University School of Medicine (New Haven, CT). Monoclonal antibody 12C5 (36) against human versican was purchased from the Developmental Studies Hybridoma Bank, University of Iowa (Iowa City, IA). Horseradish peroxidase-conjugated goat anti-rat IgG and goat antirabbit IgG were purchased from Jackson ImmunoResearch Laboratories (West Grove, PA) or Amersham Pharmacia Biotech. Glycosaminoglycan (GAG) Purification and Quantitation—Human wound fluids routinely collected within 24 h of surgery were centrifuged at 300 3 g for 10 min to remove cells and debris. The use of this discarded material was approved by the Human Research Committee of the Brigham and Women’s Hospital (Boston), protocol number 92-541601. Sulfated GAG in the form of proteoglycans or free GAG chains was purified from wound fluid samples by anion exchange using QAESephadex A-25 beads (Amersham Pharmacia Biotech) pre-equilibrated in buffer A (150 mM NaCl, 2 M urea, 0.5 M EDTA, 1 mM PMSF, 0.1% Triton X-100, 50 mM sodium acetate, pH 4.5). Wound fluid was mixed 1:1 with buffer A and incubated with buffer A pre-equilibrated QAESephadex beads overnight at 4 °C. Following incubation, beads were washed sequentially with buffer A, buffer B (300 mM NaCl, 2 M urea, 0.5 M EDTA, 1 mM PMSF, 0.1% Triton X-100, 50 mM sodium acetate, pH 4.5), and bound wound fluid GAG (proteoglycans or free GAG) eluted with buffer D (2 M NaCl, 2 M urea, 0.5 M EDTA, 1 mM PMSF, 0.1% Triton X-100, 50 mM sodium acetate, pH 4.5). Eluted material was precipitated twice with 3 volumes of 95% ethanol containing 1.3% K1 acetate, reconstituted with distilled H2O, and stored at 270 °C. In some WFGAG isolations, proteoglycans were further purified by boiling 10 min in 4 M guanidine HCl buffer containing 1% Triton X-100, 50 mM sodium acetate, pH 4.5, followed by cesium chloride density gradient separation. The proteoglycan-containing fractions were then precipitated twice with 3 volumes of 95% ethanol containing 1.3% K1 acetate, reconstituted in distilled H2O, and frozen at 270 °C. Proteoglycan was isolated from discarded human skin from anonymous donors by homogenization in detergent extraction buffer (50 mM Tris, pH 8.0, 7 M urea, 0.2 M NaCl, 1% Nonidet P-40, 1% Tween 20, 0.5% Triton X-100, 1 mM PMSF, containing protease inhibitor mixture) for 2.5 h at 4 °C. Extracted material was centrifuged at 15,000 3 g, and proteoglycans were purified from supernatants or, alternatively, the conditioned media of the A431 keratinocyte cell line by a modified version of the above anion exchange procedure. Briefly, supernatants or conditioned media were mixed 1:1 with buffer A and incubated with buffer A pre-equilibrated QAE-Sephadex A-25 beads overnight at 4 °C. Bound material was washed 3 3 for 5 min with buffer B, and proteoglycans were eluted with buffer D. Eluted material was precipitated with 3 volumes of 95% ethanol containing 1.3% K1 acetate, reconstituted in distilled H2O, and stored at 270 °C. Sulfated GAG was measured in WFGAG samples using the sulfatebinding cationic dye, dimethylmethylene blue, according to the Blyscan Proteoglycan and GAG Assay System manufacturer’s instructions (Accurate Chemical and Scientific Corp.). In some assays, accuracy of GAG quantitation was verified by carbazole assay which does not rely on GAG sulfation (37). To measure the amount of dermatan sulfate present in wound fluid, WFGAG was digested 1 h with 2000 milliunits of chondroitinase B in 50 mM Tris containing 50 mM NaCl, 4 mM CaCl2, pH 8.0. Undigested material was then precipitated with 3 volumes of 95% ethanol containing 1.3% K1 acetate and reconstituted in distilled H2O. GAG remaining following digestion was quantitated using the Blyscan Proteoglycan and GAG Assay System. For measurement of the contribution of dermatan sulfate and heparan sulfate to FGF-2 responsiveness, WFGAG was digested for 1 h under identical conditions with chondroitinase B or 6 milliunits of heparitinase in 50 mM Tris containing 4 mM CaCl2, pH 7.0. F32 Cell Culture and Proliferation Assay—F32 cells, which express

28117

FGF receptor-1 (FGFR1), require interleukin-3 or heparin and FGF-2 for proliferation, and lack detectable heparan sulfate (38), were cultured in T75 tissue culture flasks (Falcon; Becton Dickinson Labware, Bedford, MA) with RPMI 1640 media (Cellgro; Mediatech Inc., Herndon, VA) containing 10% calf serum, 10% P3X63 cell-conditioned media as a source of interleukin-3, and supplemented with L-glutamine and penicillin/streptomycin. Prior to proliferation assays cells were washed 3 times with the above media lacking interleukin-3. Cells were then added at a concentration of 2 3 104 cells/well to 96-well tissue culture plates (Falcon) in the absence or presence of 100 pM FGF-2 and test agents for 48 h at 37 °C in a 5% CO2, 95% air incubator. Following incubation the cells were pulsed either 6 h or overnight with 1 mCi/well [3H]thymidine and harvested using a TOMTEC Mach3–96 plate harvester (Wallac, Gaithersburg, MD.). [3H]Thymidine incorporation into cells was then measured using a Microbeta 1450 plate reader running the 1450 Microbeta Workstation software package (Wallac). Dot Blot Immunoassay—One mg of BSA, 0.5 mg of chondroitin sulfate, 0.5 mg of heparin, 0.5 mg of proteoglycan extracted from human skin, 0.5 mg of proteoglycan purified from A431 cell-conditioned media, or 2 mg of WFGAG samples were loaded onto cationic polyvinylidenebased membranes (Immobilon-N, Millipore, Bedford, MA) on a dot blot apparatus (Bio-Rad). Test wells were washed twice with 500 ml of buffer A and membranes blocked at room temperature for 1 h with Blotto blocking reagent (3% Carnation instant nonfat dry milk, 0.5% BSA, 0.3% Tween 20, 0.15 M NaCl in 10 mM Tris, pH 7.4). Membranes were then incubated overnight at 4 °C with primary antibodies, washed 3 3 for 5 min with Tris-buffered saline containing 0.3% Tween 20, and incubated with horseradish peroxidase-conjugated secondary antibodies for 30 min at room temperature. Primary and secondary antibodies were diluted in Blotto containing 0.3% Tween 20. Membranes were washed as above, and horseradish peroxidase was detected using enhanced chemiluminescence reagent (1.25 mM 3-aminophthalhydrazide, 0.2 mM coumaric acid, 0.3 mM hydrogen peroxide in 0.1 M Tris, pH 8.5). Membranes were then exposed on Kodak X-Omat AR x-ray film (Eastman Kodak) and results scanned using the ScanJet 4c/T running the DeskScan II 2.3 software package (Hewlett-Packard). Affinity Co-electrophoresis (ACE)—Binding of dermatan sulfate (Sigma) and heparin to FGF-2 was determined by ACE following derivatization of heparin with tyramine, radiolabeling, and chromatography on Sephadex G-100 to produce a low molecular weight fraction (Mr #6000) as described previously (39). Dermatan sulfate was prepared by alkaline cleavage to remove attached peptides (1 mg of GAG in 100 ml of 10% ethanol containing 0.17 M KOH for 1.5 h at 45 °C followed by adjustment of pH to 7.0), exchanged into water using a G-25 Sephadex spin column, and derivatized with tyramine as described for heparin. Tyraminated dermatan sulfate and heparin were then radiolabeled using the IODO-GEN method (Pierce). Labeled GAGs were radioprotected by addition of ethanol to 2% and stored at 280 °C until ACE analysis. The purity of 125I-labeled GAGs was confirmed by susceptibility to, and resistance to, digestion with appropriate GAG lyases followed by polyacrylamide gel electrophoresis analysis. ACE analysis was performed as described previously (40) with 50 mM MOPSO, pH 7.0, 125 mM sodium acetate as the gel preparation and running buffers. Gels were dried, exposed to phosphor screens (Molecular Dynamics; Sunnyvale, CA), GAG mobility measured, and converted to retardation coefficients as described previously (39). Data were fit using a non-linear least squares approach (Kaleidagraph, Synergy Software) to the equation R 5 R(`)/(1 1 (Kd/(Ptot2)), where R 5 retardation coefficient and (Ptot) 5 protein concentration in a given gel lane. Because GAGs were labeled to high specific activity and only trace amounts were used in ACE gels, their concentrations did not factor in calculation of Kd. RESULTS

Human Wound Fluids Contain Abundant Dermatan Sulfate—Sulfated glycosaminoglycans (GAG) were purified from eight wound fluid samples and quantitated using dimethylmethylene blue dye. GAG concentrations were high in all samples tested with a mean value of 31.7 mg/ml and a range from 15 to 65 mg/ml (Table I). In contrast, negligible GAG was detected in normal human sera from six individuals (data not shown), suggesting that wound fluid GAG (WFGAG) was generated during the response to injury. The concentration of WFGAG was independent of both the nature of surgery performed and volume of wound fluid generated. The ability of dimethylmethylene blue to measure accurately the sulfated

28118

Glycosaminoglycans and FGF-2 Responsiveness

TABLE I GAG composition of human wound fluids Wound fluid collected from eight randomly selected patients within 24 h of surgery was measured and GAG composition determined as described under “Experimental Procedures.” The proportion of the total measured GAG represented by dermatan sulfate was determined by selective digestions with chondroitinase B. Patient

Surgery

Wound fluid volume ml

1 2 3 4 5 6 7 8

Radical mastectomy Areola reconstruction Radical mastectomy Radical mastectomy Radical mastectomy Neck revision Simple mastectomy Radical mastectomy

80 180 180 45 650 35 365 100

GAG was verified by carbazole assay which yielded similar results (data not shown). To determine if the GAG that predominates in skin, dermatan sulfate, was also present in a soluble form in wound fluid, WFGAG was treated with chondroitinase B to remove dermatan sulfate. Table I shows that as compared with the total amount of GAG measured in wound fluid, dermatan sulfate was always abundant and represented from 36 to 78% of the total GAG. Data shown represent maximal digestion under the above conditions as determined by separate dose response and time course determinations on these WFGAG preparations and on parallel samples of commercially purified GAG. Separate treatment of WFGAG with heparitinase demonstrated that the remaining sulfated GAG was predominantly heparan sulfate (data not shown). WFGAG Is Associated with Multiple Proteoglycans—To evaluate which core proteins may be associated with the large quantity of GAG present in wound fluid, immunoblot analysis of WFGAG was performed with monoclonal antibodies specific for proteoglycan core proteins known to be expressed in skin (Fig. 1). Nonspecific binding of antibodies was evaluated with an excess of BSA, chondroitin sulfate ABC, or heparin. The cell-surface proteoglycans syndecan-1 and -4 were detected in GAG extracts from normal skin and conditioned media from cultured keratinocytes. These proteoglycans also appeared to be abundant in WFGAG, an observation consistent with prior reports of induction of these proteoglycans at cell surfaces in wounds (10, 19, 20). Although syndecans-2 and -3 were easily detected in normal skin extracts and keratinocyte-conditioned media, respectively, they were only faintly detected in wound fluid. The cell-surface proteoglycan, glypican, was strongly detected in WFGAG. In contrast, the keratinocyte-derived cellsurface proteoglycan, epican, was not detected. To determine whether proteoglycans normally found associated with extracellular matrix were also present in a soluble form in wound fluid, immunoblots were probed with monoclonal antibodies against decorin, perlecan, or versican. Decorin and perlecan were strongly detected in both normal skin extracts and WFGAG (Fig. 1). Interestingly, versican, like syndecan-1 and -4, was easily detected in WFGAG but not extracts from normal skin. WFGAG Supports FGF-2-mediated Cell Proliferation—To determine if soluble GAG produced during wound repair supports FGF-2-mediated cell proliferation, [3H]thymidine incorporation into F32 lymphoid cells, which lack cell surface heparan sulfate and express the FGF receptor, FGFR1 (38), was measured in the presence of FGF-2 and increasing amounts of WFGAG. The results show a dose-dependent increase in cell proliferation when cells were incubated with 100 pM FGF-2 and increasing amounts of WFGAG pooled from 10 patients (Fig. 2). Maximum proliferation occurred in the presence of ;3 mg/ml

Sulfated GAG

mg/ml

39.0 6 4.2 25.2 6 2.6 28.6 6 4.4 65.4 6 8.9 33.2 6 2.5 15.2 6 3.2 23.2 6 1.7 23.6 6 2.4

Dermatan sulfate %

57 6 10.2 63 6 8.1 55 6 4.3 36 6 6.0 38 6 10.6 56 6 44 78 6 3.8 53 6 9.5

FIG. 1. Multiple proteoglycans are released into human wound fluid. Soluble GAG purified from 10 pooled human wound fluid samples was applied to Immobilon-N membranes and analyzed by dot blot as described under “Experimental Procedures.” Data represent results for detection of soluble BSA, chondroitin sulfate ABC, heparin, and proteoglycans isolated from human skin, A431 cell-conditioned media (CM), or wound fluid GAG. Identical membranes probed with monoclonal antibodies specific for human syndecan-1 (DL-101), syndecan-2 (10H4), syndecan-3 (1C7), syndecan-4 (5G9), decorin (LF-136), perlecan (matrix mix), epican (mAb 17), glypican (mAb s1), or versican (12C5) are shown.

WFGAG, severalfold less than the amount of soluble GAG measured in vivo. Results were confirmed with WFGAG further purified by boiling in 4 M guanidine HCl followed by cesium chloride gradient separation (data not shown). FGF-2mediated cell proliferation was not seen in the absence of WFGAG or exogenously added heparin. Similarly, WFGAG was not able to support proliferation in the absence of FGF-2. Thus, F32 proliferation in response to wound fluid GAG was FGF-dependent, and FGF-2 action could be mediated by GAG(s) present in wound fluid. Dermatan Sulfate Binds and Supports FGF-2-mediated Cell Proliferation—The large amount of dermatan sulfate present in WFGAG led us to test whether dermatan sulfate binds FGF-2 and supports FGF-2-mediated cell proliferation. To evaluate quantitatively the ability of dermatan sulfate to bind directly FGF-2 under physiological conditions of pH and ionic strength, ACE analysis was performed. Fig. 3 depicts electro-

Glycosaminoglycans and FGF-2 Responsiveness

FIG. 2. Human wound fluid-derived GAG supports FGF-2-mediated cell proliferation. F32 cells were incubated for 48 h with 100 pM FGF-2 in the presence of the indicated amounts of wound fluid GAG purified from 10 patients as described under “Experimental Procedures.” Proliferation of F32 cells was then measured by 6 h of [3H]thymidine incorporation (1 mCi/well). Mean values of triplicate determinations (6 S.D.) representative of three experiments performed on WFGAG derived from 30 individual wound fluid samples are shown. [3H]Thymidine incorporation in the presence of all GAG concentrations (in the absence of FGF-2) was similar to that seen with FGF-2 alone (31.1 cpm). FGF-2 (100 pM)-mediated proliferation in the presence of heparin (500 ng/ml) was 7641 cpm.

FIG. 3. Dermatan sulfate binds FGF-2. 125I-Labeled GAGs were subjected to electrophoresis through zones containing indicated concentrations (in mM) of purified FGF-2. A, electrophoretogram demonstrating FGF-2 binding to dermatan sulfate. The direction of electrophoresis was from top to bottom. Dermatan sulfate was progressively shifted with increasing concentrations of FGF-2. B, low molecular weight heparin binding to FGF-2 as tested in A. Derived dissociation constants are 2.48 mM for dermatan sulfate and 344 nM for heparin, and described under “Experimental Procedures.”

phoretograms from two ACE gels, with binding to dermatan sulfate shown in Fig. 3A and heparin (as control) in Fig. 3B. Dissociation constants derived from these analyses showed that dermatan sulfate bound FGF-2 with Kd 5 2.48 mM, about 7-fold weaker than FGF-2 binding to the more highly sulfated low molecular weight form of heparin (Kd 5 344 nM). The ability to bind FGF-2 suggested dermatan sulfate may also support FGF-2 bioactivity. To determine the specificity of this FGF-2 response to dermatan sulfate, commercially available pure dermatan sulfate with a molecular mass range of 11–25 kDa was used. This preparation was found to be 100% dermatan sulfate as determined by infrared spectrophotometry and contained 39.0% iduronic acid, 32.9% galactosamine, 6.77% total sulfur, and 2.82% total nitrogen. Measurement of F32 cell proliferation in the presence of pure dermatan sulfate and FGF-2 showed that the cells responded in a dose-depend-

28119

FIG. 4. Pure dermatan sulfate supports FGF-2-mediated cell proliferation. F32 cells were incubated for 48 h with 100 pM FGF-2 in the presence of the indicated amounts of dermatan sulfate (closed circles) or heparan sulfate (open circles), as described under “Experimental Procedures.” Proliferation of F32 cells was then measured by 6 h [3H]thymidine incorporation (1 mCi/well). Mean values of triplicate determinations (6S.D.) representative of three experiments are shown. GAG-mediated proliferation in the absence of FGF-2 was similar to proliferation mediated by FGF-2 alone (983 cpm). FGF-2 (100 pM)mediated proliferation in the presence of heparin (500 ng/ml) was 13,520 cpm.

ent manner to dermatan sulfate (Fig. 4). Maximum proliferation in the presence of heparan sulfate was reached at a heparan sulfate dose of ;3 mg/ml. Pure dermatan sulfate also supported FGF-2 activity at this dose although it was slightly less active. Chondroitin sulfate ABC had minimal activity at these concentrations (data not shown), although all GAGs supported FGF-2 activity at concentrations above 25 mg/ml. GAGs did not induce cell proliferation in the absence of FGF-2. To determine if dermatan sulfate in WFGAG supports FGF-2 activity, WFGAG was treated with chondroitinase B to specifically eliminate dermatan sulfate. Following confirmation of digestion by dimethylmethylene blue, measurement of F32 cell proliferation in the presence of FGF-2 demonstrated that digestion with chondroitinase B resulted in an ;85% reduction in cell proliferation (Fig. 5). In contrast, treatment with heparitinase showed only a ;29% reduction in proliferation. DISCUSSION

Proteoglycans have been shown in vitro to influence multiple cell behaviors by binding physiologic ligands including several growth factors, cytokines, matrix proteins, cell-surface molecules, proteases, and protease inhibitors (2, 5–9). In the current study, we demonstrate that abundant proteoglycans are released during wound repair in vivo, and a large proportion of the soluble GAG in wounds is chondroitin sulfate B (dermatan sulfate). We find that wound fluid GAG (WFGAG) supports the ability of FGF-2 to signal cell proliferation and that this activity resides predominantly in dermatan sulfate. These observations support the hypothesis that proteoglycans released in response to injury function as essential components in wound healing and identify a unique role for dermatan sulfate in supporting cell proliferation in response to FGF-2. Abundant Dermatan Sulfate Is Released into Wounds following Injury—Measurement of soluble GAG in wounds using dimethylmethylene blue dye demonstrated that a large amount of sulfated GAG was present in human wound fluids. Consistent with the large amounts of dermatan sulfate expressed in skin, treatment of WFGAG from individual patients with chondroitinase B showed that much of the soluble GAG was derma-

28120

Glycosaminoglycans and FGF-2 Responsiveness

FIG. 5. Wound fluid-derived dermatan sulfate supports FGF-2mediated cell proliferation. F32 cells were incubated for 48 h with 100 pM FGF-2 in the presence of 1 mg of untreated wound fluid GAG (A), or equal amounts of wound fluid GAG treated with either chondroitinase B (B) or heparitinase (C) for 1 h, as described under “Experimental Procedures.” Proliferation of F32 cells was then measured by 6 h of [3H]thymidine incorporation (1 mCi/well). Mean values of triplicate determinations (6 S.D.) representative of duplicate experiments are shown. FGF-2-mediated proliferation in the absence of GAG or the presence of heparin (500 ng/ml) was 143 and 42,246 cpm, respectively.

tan sulfate. Since dimethylmethylene blue is a relatively insensitive assay for soluble GAG determinations, measurements of total GAG may be underestimated due to the presence of nonsulfated GAG. Thus, we verified measurements with carbazole, a uronic acid binding molecule that does not rely on GAG sulfation level (37). Measurements by carbazole assay confirmed the large amounts of GAG released from wounds after injury. In addition, because measurements were similar by carbazole assay there was likely little, if any, nonsulfated soluble hyaluronic acid in these wound fluid GAG preparations. Since hyaluronic acid is not sulfated, this lack of detection may have reflected loss during purification by anion exchange, or alternatively, there may be negligible release into the wound environment. Several Proteoglycans Are Present in Wound Fluid—Purification of WFGAG by anion exchange can yield intact proteoglycans or free GAG. Measurement of GAG using dimethylmethylene blue, or carbazole, does not distinguish between these GAG forms, and both may be present in wound fluid. Immunoblot analysis of purified WFGAG using monoclonal antibodies against a panel of proteoglycans likely to be found in cutaneous wounds showed that WFGAG contains several types of proteoglycan core proteins. Since these core proteins were detected in wound fluid that was purified by anion exchange, they remain associated with GAG. Therefore, multiple core proteins in the form of proteoglycans are released into the wound environment. The mechanism of this proteoglycan release into wounds remains unknown, but thrombin, growth factor receptor activation, and the action of calcium ionophores or protein kinase C activators have been reported to release cell surface-associated molecules, including proteoglycans, from cells in culture (21, 41– 43). Thus, multiple mechanisms may generate soluble proteoglycan and may represent the mechanism of proteoglycan release in wounds. Proteoglycans associated with cell surfaces or extracellular matrix were both found in a soluble form in wound fluid. Furthermore, there was specificity of proteoglycan release in wounds; not all proteoglycans were released to the same extent after injury. This selectivity of proteoglycan release is also consistent with prior observations of induction of syndecans on cell surfaces after injury in vitro (20) and in vivo (10), as well as

prior detection of syndecan-1 and -4 in wound fluids (21). Together, these data demonstrate that several proteoglycans from various compartments are present in wound fluids where they may act to affect cell behaviors required for wound repair. In addition, proteoglycans known to contain heparan sulfate or chondroitin sulfates (including dermatan sulfate) were both detected in wound fluid. Thus, multiple soluble proteoglycans are located at sites where their presence can influence cell responsiveness to GAG-binding ligands that function during wound repair. Dermatan Sulfate in Wounds Supports FGF-2 Function— FGF-2 is an important GAG-binding growth factor that signals the proliferation of cells that function in response to injury (44). Several heparan sulfate proteoglycans derived from tissue culture cells including syndecans, glypican, and perlecan have been shown to bind and support FGF activity (45– 48). The presence of a large amount of soluble GAG in wounds suggested that native proteoglycans released after injury in vivo may also influence FGF function. To test this hypothesis, the ability of WFGAG to support FGF-2-mediated cell proliferation was measured. F32 cells were chosen because they contain transfected FGF receptor, FGFR1, and lack endogenous heparan sulfate (38). Thus, using this model, FGF-2 activity supported by GAG purified from an in vivo source could be studied directly. The results of these studies demonstrated that WFGAG supported FGF-2-mediated cell proliferation at GAG concentrations well below the physiological levels we have measured in wound fluid. The ability of WFGAG to support FGF-2-mediated proliferation was not the result of growth factor contamination of WFGAG preparations because WFGAG boiled in 4 M guanidine and further purified by cesium chloride density gradient separation also supported FGF-2 activity. Likewise, WFGAG did not support cell proliferation in the absence of added FGF-2. Thus, soluble proteoglycans generated in wounds support FGF-2 function. Our finding that large amounts of soluble dermatan sulfate were present in wounds led us to evaluate whether dermatan sulfate could directly influence FGF function. Affinity co-electrophoresis showed that dermatan sulfate bound FGF-2, and pure dermatan sulfate supported FGF-2-mediated cell proliferation. This effect was observed at concentrations at, or below, that of dermatan sulfate in wounds. To determine if dermatan sulfate in wounds also supported FGF-2 bioactivity, WFGAG was treated with chondroitinase B to specifically eliminate dermatan sulfate. This treatment decreased FGF-2 activity by ;85%. The nature of the FGF-2 binding domains within dermatan sulfate was not determined. However, these data confirm that both pure dermatan sulfate and dermatan sulfate present in wound fluid support FGF-2-mediated cell proliferation. Future studies will be needed to map the GAG sequences that confer this activity. The observation that dermatan sulfate supported FGF-2 activity is consistent with report of the ability of dermatan sulfate, and chondroitin sulfates A and C, to bind FGF-2 (12) and dermatan sulfate to inhibit 125I-FGF-2 binding to extracellular matrix-coated tissue culture wells (49). Furthermore, dermatan sulfate has been found to interact with other molecules previously known to interact only with heparan sulfate (13, 15–18). Thus, when considered in context with the large quantity of dermatan sulfate present in wounds, and the affinity we measured of dermatan sulfate for FGF-2 (Kd 5 2.48 mM), it appears that dermatan sulfate is a major contributor to FGF-2 responsiveness in the wound environment. This study demonstrates that abundant functionally active proteoglycan is present in human wound fluids. This places GAGs at sites where they can act as soluble effectors to influ-

Glycosaminoglycans and FGF-2 Responsiveness ence cell behaviors required for wound repair. Soluble GAG was derived both from cell surfaces and extracellular matrix. Thus, insoluble GAG serves as a source for release in response to injury and by release has paracrine potential at distant sites. The novel finding that dermatan sulfate released following injury supported FGF-2 activity further suggests that nonheparan sulfate proteoglycans also participate in the regulation of growth factor responsiveness. The specific proteoglycans that are responsible for supporting FGF-2 activity have not been identified, but demonstration of the presence of several core proteins indicates that numerous proteoglycans have the potential to function as a source of biologically active GAG in vivo. Further study of dermatan sulfate GAGs and their established ligands will contribute to our understanding of wound repair and related processes. REFERENCES 1. Ruoslahti, E. (1989) J. Biol. Chem. 264, 13369 –13372 2. Kjelle´n, L., and Lindahl, U. (1991) Annu. Rev. Biochem. 60, 443– 475 3. Wight, T. N., Heinegard, D. K., and Hascall, V. C. (1991) in Cell Biology of Extracellular Matrix (Hay, E. D., ed) pp. 45–78, Plenum Publishing Corp., New York 4. Hardingham, T. E., and Fosang, A. J. (1992) FASEB J. 6, 861– 870 5. Salmivirta, M., Elenius, K., Vainio, S., Hofer, U., Chiquet-Ehrismann, R., Thesleff, I., and Jalkanen, M. (1991) J. Biol. Chem. 266, 7733–7739 6. Jackson, R. L., Busch, S. J., and Cardin, A. D. (1991) Physiol. Rev. 2, 481– 485 7. Bernfield, M., Kokenyesi, R., Kato, M., Hinkes, M. T., Spring, J., Gallo, R. L., and Lose, E. J. (1992) Annu. Rev. Cell Biol. 8, 365–398 8. David, G. (1993) FASEB J. 7, 1023–1030 9. Stringer, S. E., and Gallagher, J. T. (1997) Int. J. Biochem. Cell. Biol. 29, 709 –714 10. Gallo, R. L., Kim, C., Kokenyesi, R., Adzick, N. S., and Bernfield, M. (1996) J. Invest. Dermatol. 107, 667– 683 11. Clark, R. A. F. (1996) in The Molecular and Cellular Biology of Wound Repair (Clark, R. A. F., ed) Plenum Publishing Corp., New York 12. Turnbull, J. E., Fernig, D. G., Ke, Y., Wilkinson, M. C., and Gallagher, J. T. (1992) J. Biol. Chem. 267, 10337–10341 13. Lyon, M., Deakin, J. A., Rahmoune, H., Fernig, D. G., Nakamur, T., and Gallagher, J. T. (1998) J. Biol. Chem. 273, 271–278 14. Maimone, M. M., and Tollefsen, D. M. (1990) J. Biol. Chem. 265, 18263–18271 15. Mascellani, G., Liverani, L., Bianchini, P., Parma, B., Torri, G., Bisio, A., Guerrini, M., and Casu, B. (1993) Biochem. J. 296, 639 – 648 16. Cella, G., Boeri, G., Saggiorato, G., Paolini, R., Luzzatto, G., and Terribile, V. I. (1992) Angiology 43, 59 – 62 17. Walker, A., and Gallagher, J. T. (1996) Biochem. J. 317, 871– 877 18. Priglinger, U., Geiger, M., Bielek, E., Vanyek, E., and Binder, B. R. (1994) J. Biol. Chem. 269, 14705–14710 19. Elenius, K., Vainio, S., Laato, M., Salmivirta, M., Thesleff, I., and Jalkanen, M.

28121

(1991) J. Cell Biol. 114, 585–595 20. Gallo, R. L., Ono, M., Povsic, T., Page, C., Eriksson, E., Klagsbrun, M., and Bernfield, M. (1994) Proc. Natl. Acad. Sci. U. S. A. 91, 11035–11039 21. Subramanian, S. V., Fitzgerald, M. L., and Bernfield, M. (1997) J. Biol. Chem. 272, 14713–14720 22. Werner, S., Peters, K. G., Longaker, M. T., Fuller-Pace, F., Banda, M. J., and Williams, L. T. (1992) Proc. Natl. Acad. Sci. U. S. A. 89, 6896 – 6900 23. Cooper, D. M., Yu, E. Z., Hennessey, P., Ko, F., and Robson, M. C. (1994) Ann. Surg. 219, 688 – 692 24. Folkman, J., and Klagsbrun, M. (1987) Science 235, 442– 447 25. Klagsbrun, M. (1989) Prog. Growth Factor Res. 1, 207–235 26. Klagsbrun, M. (1990) Curr. Opin. Cell Biol. 2, 857– 863 27. Rapraeger, A. C., Krufka, A., and Olwin, B. B. (1991) Science 252, 1705–1708 28. Yayon, A., Klagsbrun, M., Esko, J. D., Leder, P., and Ornitz, D. M. (1991) Cell 64, 841– 848 29. Klagsbrun, M., and Baird, A. (1991) Cell 67, 229 –231 30. David, G., Bai, X. M., Van der Schueren, B., Marynen, P., Cassiman, J.-J., and Van den Berghe, H. (1993) Development 119, 841– 854 31. Lories, V., Cassiman, J. J., Van den Berghe, H., and David, G. (1989) J. Biol. Chem. 264, 7009 –7016 32. Heremans, A., Cassiman, J.-J., Van den Berghe, H., and David, G. (1988) J. Biol. Chem. 263, 4731– 4739 33. Lories, V., Cassiman, J., Van den Berghe, H., and David, G. (1992) J. Biol. Chem. 267, 1116 –1122 34. Fisher, L. W., Stubbs, J. T., and Young, M. F. (1995) Acta Orthop. Scand. Suppl. 266, 61– 65 35. Haggerty, J. G., Bretton, R. H., and Milstone, L. M. (1992) J. Invest. Dermatol. 99, 374 –380 36. Asher, R., Perides, G., Vanderhaeghen, J.-J., and Bignami, A. (1991) J. Neurosci. Res. 28, 410 – 421 37. Bitter, T., and Muir, H. M. (1962) Anal. Biochem. 4, 330 –334 38. Ornitz, D. M., Yayon, A., Flanagan, J. G., Svahn, C. M., Levi, E., and Leder, P. (1992) Mol. Cell. Biol. 12, 240 –247 39. San Antonio, J. D., Slover, J., Lawler, J., Karnovsky, M. J., and Lander, A. D. (1993) Biochemistry 32, 4746 – 4755 40. Lee, M. K., and Lander, A. D. (1991) Proc. Natl. Acad. Sci. U. S. A. 88, 2768 –2772 41. Pandiella, A., and Massague, J. (1991) J. Biol. Chem. 266, 5769 –5773 42. Arribas, J., and Massague, J. (1995) J. Cell Biol. 128, 433– 441 43. Arribas, J., Coodly, L., Vollmer, P., Kishimoto, T. K., Rose-John, S., and Massague´, J. (1996) J. Biol. Chem. 271, 11376 –11382 44. Abraham, J. A., and Klagsburn, M. (1996) in Molecular and Cellular Biology of Wound Repair (Clark, R. A. F., ed) Plenum Publishing Corp., New York 45. Steinfeld, R., Van Den Berghe, H., and David, G. (1996) J. Cell Biol. 133, 405– 416 46. Bonneh-Barkay, D., Shlissel, M., Berman, B., Shaoul, E., Admon, A., Vlodavsky, I., Carey, D. J., Asundi, V. K., Reich-Slotky, R., and Ron, D. (1997) J. Biol. Chem. 272, 12415–12421 47. Aviezer, D., Hecht, D., Safran, M., Eisinger, M., David, G., and Yayon, A. (1994) Cell 79, 1005–1013 48. Aviezer, D., Iozzo, R. V., Noonan, D. M., and Yayon, A. (1997) Mol. Cell. Biol. 17, 1938 –1946 49. Bashkin, P., Doctrow, S., Klagsbrun, M., Svahn, C. M., Folkman, J., and Vlodavsky, I. (1989) Biochemistry 28, 1737–1743