Adhesion of Human Skin Fibroblasts to Cyr61 Is Mediated through Integrin α6β1 and Cell Surface Heparan Sulfate Proteoglycans*

The angiogenic inducer Cyr61 is an extracellular matrix-associated heparin-binding protein that can mediate cell adhesion, stimulate cell migration, and enhance growth factor-stimulated DNA synthesis in both fibroblasts and endothelial cells in culture. In vivo, Cyr61 induces neovascularization and promotes tumor growth. Cyr61 is a prototypic member of a highly conserved family of secreted proteins that includes connective tissue growth factor, nephroblastoma overexpressed, Elm-1/WISP-1, Cop-1/WISP-2, and WISP-3. Encoded by an immediate early gene, Cyr61 synthesis is induced by serum growth factors in cultured fibroblasts and in dermal fibroblasts during cutaneous wound healing. We previously demonstrated that Cyr61 mediates adhesion of vascular endothelial cells and activation-dependent adhesion of blood platelets through direct interaction with integrins αVβ3 and αIIbβ3, respectively. In this study, we show that the adhesion of primary human skin fibroblasts to Cyr61 is mediated through integrin α6β1 and cell surface heparan sulfate proteoglycans (HSPGs), which most likely serve as co-receptors. Either destruction of cell surface HSPGs or prior occupancy of the Cyr61 heparin-binding site completely blocked cell adhesion to Cyr61. A heparin-binding defective mutant of Cyr61 was unable to mediate fibroblast adhesion through integrin α6β1 but still mediated endothelial cell adhesion through integrin αVβ3, indicating that endothelial cell adhesion through integrin αVβ3 is independent of the heparin-binding activity of Cyr61. These results identify Cyr61 as a novel adhesive substrate for integrin α6β1 and provide the first demonstration of the requirement for HSPGs in integrin-mediated cell attachment. In addition, these findings suggest that Cyr61 might elicit disparate biological effects in different cell types through interaction with distinct integrin receptors.

Analyses of the angiogenic inducer Cyr61 have revealed a multifunctional protein with a remarkable diversity of activities. Originally identified in mouse fibroblasts as a secreted, cysteine-rich protein encoded by a growth factor-inducible immediate early gene, Cyr61 was found to be an extracellular matrix-associated, heparin-binding protein that is capable of diverse functions in different cell types (1)(2)(3). Purified recom-binant Cyr61 protein mediates cell adhesion, stimulates cell migration, and augments growth factor-induced DNA synthesis in both fibroblasts and endothelial cells (4,5). Cyr61 promotes chondrogenesis in mouse limb bud mesenchymal cells in micromass cultures, in agreement with its expression in condensing prechondrocytic mesenchyme of both neuroectodermal and mesodermal origins during embryogenesis (6,7). Cyr61 has pro-angiogenic activities in culture and induces neovascularization in vivo, consistent with its localization in angiogenic cell types during development (8). Furthermore, expression of cyr61 enhances the tumorigenicity of human tumor cells in immunodeficient mice by increasing tumor size and vascularization (8).
Cyr61 is a prototypic member of the CCN protein family, which also includes connective tissue growth factor (CTGF), 1 nephroblastoma overexpressed, Elm-1/WISP-1, Cop-1/WISP-2, and WISP-3 (3,9). These highly conserved proteins share four modular domains with sequence similarities to insulin-like growth factor-binding proteins, von Willebrand factor, thrombospondin, and growth factor cysteine knots. The biochemical and cellular activities of Cyr61 and CTGF are very similar in vitro, and both proteins induce angiogenesis in vivo (8,10). Both cyr61 and CTGF are co-induced in fibroblasts by growth factors implicated in tissue repair and matrix remodeling, including fibroblast growth factor, platelet-derived growth factor, and transforming growth factor-␤1 (1,11,12). Consistent with these observations, both cyr61 and CTGF are induced in granulation tissue during cutaneous wound healing (13,14). Thus, Cyr61 and CTGF may participate in wound healing by acting as angiogenic factors upon endothelial cells, and by acting as chemotactic, mitogenic, and matrix remodeling factors upon fibroblasts.
Cells perceive vitally important information from their three-dimensional tissue contexts through adhesion receptors. These receptors support various aspects of cellular functions, including proliferation, differentiation, migration, and survival (15)(16)(17). One of the activities of Cyr61 as an extracellular matrix-associated protein is to support cell adhesion (4,5). Although devoid of a RGD sequence motif, Cyr61 is a ligand of, and binds directly to, the adhesion receptors integrins ␣ V ␤ 3 and ␣ IIb ␤ 3 (18,19). A number of the activities of Cyr61 can be attributed to its interaction with integrin receptors. For example, endothelial cell adhesion and migration to Cyr61 are dependent on integrin ␣ V ␤ 3 , and activation-dependent adhesion of blood platelets to Cyr61 is mediated through interaction with integrin ␣ IIb ␤ 3 (8,18,19). To understand the mechanism of Cyr61 action in dermal fibroblasts, we sought to identify the cellular adhesion receptor(s) through which Cyr61 may function. In this study, we show that both integrin ␣ 6 ␤ 1 and cell surface HSPGs are indispensable for the adhesion of normal human skin fibroblasts to Cyr61. A heparin-binding defective Cyr61 mutant protein is unable to support fibroblast adhesion, yet retains the ability to support endothelial cell adhesion through integrin ␣ V ␤ 3 . These findings identify Cyr61 as a novel adhesion substrate for integrin ␣ 6 ␤ 1 , underscore the importance of HSPGs in adhesion through this integrin, and demonstrate the utilization of distinct adhesive receptors for Cyr61 in different cell types.
Construction of Cyr61 Heparin-binding Mutants-Cyr61 mutants harboring alanine substitutions within the putative heparin-binding motifs between aa 280 -290 and aa 306 -312 were constructed by sitedirected mutagenesis using a two-step PCR procedure as described (20). For the H1 mutant, the internal primers were 5Ј-GCGGCATGCAGC-GCGACCGCGAAATCCCCAGAACCAGTC and 5Ј-TCGCGCTGCATG-CCGCGCCCGCTTTTAGGCTGCTGTACACTG. For the H2 mutant, the internal primers were 5Ј-GTCGCGGCATACGCGCCCAAATACTGCG-GCTC and 5Ј-GCGCGTATGCCGCGACACTGGAGCATCCTGC. The outside primers used in the PCR for both mutants were 5Ј-CAGAC-CACGTCTTGGTCC (corresponding to nucleotides 887-894 of the cyr61 cDNA sequence) (1) and 5Ј-GAATAGGCTGTACAGTCGG (nucleotides 1284 -1302). Mouse Cyr61 cDNA was used as template in each of the two first step PCRs using one internal primer with its corresponding outside primer. Resulting products were gel-purified, combined, and used as template for the second step PCR using the outside primers. The final PCR product was digested with BsrGI, which cuts at two sites flanking the mutated sequences. The BsrGI fragment of the full-length cyr61 cDNA in pSG5 vector was then substituted with the mutated PCR product. The orientations and sequences of the resulting constructs were confirmed by DNA sequencing. Finally, the mutant cyr61 cDNAs were cloned in baculovirus expression vector pBlueBac 4.5 (Invitrogen) as described previously (4). To construct the double mutant (DM Cyr61), the H2 mutant cDNA was used as template, and PCR mutagenesis was performed using internal primers for H1 mutant.
Cyr61 and Matrix Proteins-Recombinant mouse Cyr61 and mutant Cyr61 proteins were synthesized in a baculovirus expression system using Sf9 insect cells and purified from serum-free conditioned medium by chromatography on a Sepharose S column as described (4). Purity and concentration of the protein were determined by SDS-polyacrylamide gel electrophoresis followed by Coomassie Blue staining and immunoblotting. Human fibronectin, human vitronectin, rat tail type-I collagen, and mouse laminin were obtained from Collaborative Research.
To assess the binding affinities of wild type and mutant Cyr61 proteins to heparin, purified protein preparations (5 g each) were loaded on 0.5 ml CL-6B heparin-Sepharose (Amersham Pharmacia Biotech) columns. After washing with 20 bed volume of radioimmune precipitation buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS), bound proteins were eluated with radioimmune precipitation buffer containing increasing amount of NaCl (0.5 ml at each NaCl concentration). An equal amount of eluate from each fraction was analyzed by SDS-polyacrylamide gel electrophoresis followed by immunoblotting with anti-Cyr61 antibodies.
Cell Culture and Adhesion Assay-Primary human foreskin fibroblast 1064SK (ATCC CRL-2076, passage 2) was kept in Dulbecco's modified minimal essential medium (4.5 mg/ml glucose, Life Technologies, Inc.) with 10% fetal bovine serum (Intergene). HUVECs were purchased from Cascade Biologics, Inc. and grown in the medium provided by the company. For the adhesion assay, cells were harvested in phosphate-buffered saline (137 mM NaCl, 2.7 mM KCl, 4.3 mM Na 2 HPO 4 , 1.4 mM KH 2 PO 4 , pH 7.3) with 2.5 mM EDTA and resus-pended in serum-free medium (F12K medium for HUVECs and Dulbecco's modified minimal essential medium for fibroblasts are from Life Technologies, Inc.) containing 1% bovine serum albumin as described (18). Protein under study was diluted to the desired concentration in phosphate-buffered saline, applied to 96-well microtiter plates (50 l/ well), incubated at 4°C for 16 h, and blocked with 1% bovine serum albumin at room temperature for 1 h. Where indicated, reagents (EDTA, heparin, peptides, etc.) were mixed with cells prior to plating. Antibodies were incubated with cells at room temperature for 30 min before plating. To each well 50 l of cell suspension was plated, and after incubation at 37°C for 30 min wells were washed twice with phosphate-buffered saline. Adherent cells were fixed with 10% formalin, stained with methylene blue, and quantified by dye extraction and measurement of absorbance at 620 nm as described (21). Inhibition of glycosaminoglycan sulfation was achieved by growing cells in medium containing the indicated amount of sodium chlorate for 24 h; cells were then detached and plated for cell adhesion assays as described above (22). To show the specificity of sulfation blockage, 10 mM sodium sulfate was included in the culture medium together with sodium chlorate. For enzymatic digestion of cell surface glycosaminoglycans, harvested cells were washed, resuspended in Dulbecco's modified minimal essential medium, and treated with heparinase I (

Primary Human Skin Fibroblast Adhesion to Cyr61
Is Mediated through Integrin ␣ 6 ␤ 1 -We have previously shown that immobilized Cyr61 can support the adhesion of endothelial cells, epithelial cells, 3T3 fibroblasts, as well as blood platelets (4,5,19). In view of the potential involvement of Cyr61 in cutaneous wound healing, we examined the ability of Cyr61 to mediate cell adhesion in normal human skin fibroblasts. Microtiter wells were coated with purified recombinant Cyr61 protein, onto which 1064SK primary human foreskin fibroblasts were allowed to adhere under serum-free conditions. Adhesion of 1064SK cells to Cyr61 was dose-dependent and saturable (Fig. 1A). Furthermore, affinity purified anti-Cyr61 antibodies blocked cell adhesion to Cyr61 but not to vitronectin (Fig. 1B), indicating that the ability to mediate fibroblast adhesion is an intrinsic property of the Cyr61 protein.
We previously showed that Cyr61 is a ligand of the integrins ␣ V ␤ 3 and ␣ IIb ␤ 3 and mediates the adhesion of endothelial cells and platelets, respectively, through these integrins (18,19). Therefore, we sought to determine whether adhesion of human skin fibroblasts to Cyr61 is also mediated through an integrin receptor. Preincubation of fibroblasts with LM609, a mAb specific for integrin ␣ V ␤ 3 (23), had no effect on fibroblast adhesion to Cyr61 (data not shown). Thus, integrin ␣ V ␤ 3 is unlikely to be involved in fibroblast adhesion to Cyr61. We next examined the effects of divalent cations in Cyr61-mediated cell adhesion. Fibroblast adhesion to Cyr61 was completely blocked by 2.5 mM EDTA and was restored by the addition of 5.0 mM Mg 2ϩ (Fig.  1C). As expected, similar effects were observed for fibroblasts plated on collagen or vitronectin (data not shown). However, whereas Mg 2ϩ or Mn 2ϩ supported cell adhesion, the presence of Ca 2ϩ abolished cell adhesion to Cyr61 completely (Fig. 1C). The presence of Mn 2ϩ , but not Mg 2ϩ , was able to abrogate the inhibitory effect of Ca 2ϩ on cell adhesion to Cyr61, suggesting that Mn 2ϩ may bind the Cyr61 adhesion receptor with higher affinity than Ca 2ϩ (Fig. 1C). This divalent cation sensitivity profile of Cyr61-mediated cell adhesion is consistent with and indicative of an integrin being the adhesion receptor for Cyr61 (24,25).
A number of integrins expressed on fibroblasts recognize the RDG sequence motif in their protein ligands and are sensitive to inhibition by RGD-containing peptides (26). The peptide GRGDSP, but not the control peptide GRGESP, completely abolished 1064SK cell adhesion to vitronectin at concentrations of 0.2 mM or higher ( Fig. 2A and data not shown). However, the GRGDSP peptide had no effect on fibroblast adhesion to Cyr61 or type I collagen even at 2 mM ( Fig. 2A), indicating that the adhesion receptor for Cyr61 is insensitive to RGD-containing peptides. Because the major RGD-sensitive integrins in fibroblasts are the vitronectin receptors (integrins ␣ V ␤ 3 and ␣ V ␤ 5 ) and the fibronectin receptor (integrin ␣ 5 ␤ 1 ), this observation indicated that fibroblast adhesion to Cyr61 is unlikely to be mediated through either one of the ␣ V integrins or integrin ␣ 5 ␤ 1 . To investigate this notion further, we challenged 1064SK cells by preincubation with function-blocking mAbs against the ␣ V and ␣ 5 integrin subunits. Whereas cell adhesion to vitronectin and fibronectin was blocked by mAbs against ␣ V and ␣ 5 , respectively, these antibodies had no effect on cell adhesion to Cyr61 (Fig. 2B). These results excluded the possibility that either one of the ␣ V integrins or integrin ␣ 5 ␤ 1 may function as the Cyr61 adhesion receptor.
The ␤ 1 integrins constitute the major types of integrins expressed in skin fibroblasts, although ␤ 3 and ␤ 4 integrins are also expressed at low levels. Because the only ␤ 3 integrin in fibroblasts is ␣ V ␤ 3 , this integrin can be excluded inasmuch as fibroblast adhesion to Cyr61 was insensitive to RGD-containing peptides and was not inhibited by the mAb LM609. Likewise, anti-␤ 4 mAb did not block adhesion to Cyr61, thus excluding the ␤ 4 integrins (data not shown). By contrast, inhibitory anti-␤ 1 mAb (JB1A) effectively blocked adhesion to Cyr61 by Ͼ80%. In comparison, adhesion to type I collagen, which binds to several ␤ 1 integrins (␣ 1 ␤ 1 , ␣ 2 ␤ 1 , and ␣ 3 ␤ 1 ), was inhibited about ϳ70% by the anti-␤ 1 mAb (Fig. 3A). Adhesion to vitronectin, which is mediated through ␣ V integrins (␣ V ␤ 3 and ␣ V ␤ 5 ), was unaffected. These results indicate that adhesion of fibroblasts to Cyr61 is mediated through a ␤ 1 integrin other than ␣ 5 ␤ 1 , consistent with the RGD insensitivity and Ca 2ϩ sensitivity of this process.
Adhesion of Human Skin Fibroblasts to Cyr61 Requires Cell Surface HSPGs-Because Cyr61 is known to bind heparin with high affinity (2), we endeavored to determine the role of heparin binding in the activities of Cyr61. We added various amounts of soluble heparin to suspensions of 1064SK fibroblasts prior to plating onto either Cyr61-or fibronectin-coated wells (Fig. 4A). Remarkably, as little as 1-10 ng/ml heparin was sufficient to elicit a detectable inhibition of cell adhesion to Cyr61, whereas 0.1 g/ml heparin completely inhibited Cyr61mediated fibroblast adhesion. In comparison, the addition of up to 10,000 times more heparin had little or no effect on fibroblast adhesion to fibronectin, itself also a heparin-binding protein. Similar inhibitory effects were observed when heparin was added onto Cyr61-coated microtiter wells and washed before cells were permitted to adhere, suggesting that heparin can bind to immobilized Cyr61 and block cell adhesion (data not shown). The same inhibitory effect was observed with low molecular mass heparin (ϳ3 kDa), thus arguing against the possibility that heparin binding might prevent Cyr61 from inter- , they were blocked with affinity purified anti-Cyr61 antibodies for 1 h at 37°C before 1064SK cells were plated, and adhesion assay was performed as above. C, cells were plated on microtiter wells coated with 2 g/ml Cyr61; EDTA (2.5 mM); Ca 2ϩ , Mg 2ϩ , or Mn 2ϩ (5.0 mM each) were added alone or in combination as indicated, and cell adhesion assay was performed as above. Data shown for all panels are mean Ϯ S.D. of triplicate determinations and are representative of three experiments.
FIG. 2. Fibroblast adhesion to Cyr61 is independent of fibronection and vitronectin receptors. Cell adhesion assays were performed with 1064SK fibroblasts plated on microtiter wells coated with either Cyr61 (2 g/ml), type I collagen (Col. I, 2 g/ml), vitronectin (VN, 0.5 g/ml), or fibronectin (FN, 1 g/ml). A, GRGDSP or GRGESP peptide (2 mM) was added to the cell suspension, and cell adhesion assay was performed. B, cells were preincubated with 40 g/ml mAbs against integrin ␣ 5 or integrin ␣ V where indicated prior to adhesion assay. Data shown for all panels are mean Ϯ S.D. of triplicate determinations and are representative of three experiments.
acting with cell adhesion molecules on account of steric hindrance (data not shown). These data are consistent with at least two alternative interpretations. First, heparin binding may alter the conformation of Cyr61, or mask its integrinbinding site, thus render Cyr61 unable to interact with integrin receptors; and second, occupancy of the Cyr61 heparin-binding site precludes Cyr61 from interacting with cell surface HSPGs, and this interaction may be indispensable for Cyr61-mediated cell adhesion.
To distinguish between these possibilities, we asked whether damaging the integrity of cell surface glycosaminoglycans can affect fibroblast adhesion to Cyr61. 1064SK fibroblasts were cultured in the presence of sodium chlorate, an inhibitor of 3-phosphoadenosine 5Ј-phosphosulfate synthesis, to block sulfation of proteoglycans (Fig. 4B). Adhesion to Cyr61 was inhibited in a dose-dependent manner when cells were treated with increasing amounts of sodium chlorate, and Ͼ80% inhibition was achieved at 40 mM sodium chlorate. By contrast, adhesion of the same cells to fibronectin, collagen, vitronectin, or laminin was not significantly affected (Fig. 4B). The inhibitory effect of sodium chlorate on cell adhesion to Cyr61 was reversed by the inclusion of 10 mM Na 2 SO 4 in the culture medium (Fig. 4C), thus verifying that this inhibitory effect is mediated through a sulfation block (22).
Cell surface-sulfated proteoglycans include heparan sulfate and chondroitin sulfate proteoglycans. To test whether chondroitin sulfate proteoglycans might play a role in Cyr61-mediated adhesion, we examined the effect of various chondroitin sulfates (Fig. 5A). The presence of chondroitin sulfate A, B, or the small chondroitin sulfate proteoglycan decorin, but not chondroitin sulfate C, blocked Cyr61-mediated cell adhesion. However, the concentrations of chondroitin sulfates needed for this inhibition were 2-4 orders of magnitude higher than the effective concentration of heparin (compare 0.1 g/ml heparin with 1 mg/ml chondroitin sulfate A), suggesting that the affinity of Cyr61 for chondroitin sulfate is relatively weak compared with that for heparan sulfate.
To confirm the importance of cell surface sulfated proteoglycans on fibroblast adhesion to Cyr61, we treated 1064SK fibroblasts with heparinase I and chondroitinase ABC. As shown in Fig. 5B, heparinase I treatment rendered cells unable to adhere to Cyr61, but had no effect on cell adhesion to vitronectin or fibronectin. In contrast, chondroitinase ABC had no effect on cell adhesion to Cyr61, indicating that cell surface chondroitin sulfates do not contribute significantly to human skin fibroblast adhesion to Cyr61. Together, these results show that the presence of cell surface HSPGs are necessary for fibroblast adhesion to Cyr61 and that Cyr61 must be available to bind heparan sulfate to mediate cell adhesion through integrin ␣ 6 ␤ 1 .
Adhesive Properties of Heparin-binding Defective Cyr61 Mutants-Although cell surface heparan sulfates are essential for Cyr61-mediated adhesion of fibroblasts (Figs. 4 and 5), whether Cyr61 must interact with heparan sulfate proteoglycans through its heparin-binding activity remained to be determined. To address this question, we created mutant Cyr61 proteins that are deficient in heparin binding. Two putative heparin-binding motifs could be recognized within the Cyr61 carboxyl-terminal domain that conformed to the consensus XB-BXBX sequence for heparin binding, where B denotes basic amino acid residues such as lysine or arginine (27; Fig. 6A). We performed site-directed mutagenesis to replace the lysine and arginine residues in these heparin-binding motifs with alanine, although maintaining the integrity of the cysteine residues to avoid disruption of any potential disulfide bonds. Two Cyr61 mutants, H1 and H2, were thus created, each altering one of the heparin-binding motifs (Fig. 6A). In addition, both motifs were mutated in another Cyr61 variant, DM. All three mutations were created in the context of the full-length Cyr61, and the full-length mutant proteins were expressed and purified from the baculovirus system in insect cells ("Experimental Procedures"). The recombinant mutant and wild type Cyr61 proteins were bound to a heparin-Sepharose column and eluted with increasing concentrations of salt. Whereas wild type Cyr61 eluted from a heparin-Sepharose column at 0.8 to 1.0 M NaCl as previously reported (2), H1 and H2 eluted at lower salt FIG. 4. Cell surface glycosaminoglycans are indispensable for human skin fibroblast adhesion to Cyr61. A, 1064SK fibroblasts were plated on microtiter wells coated with Cyr61 (2 g/ml) or fibronectin (FN, 1 g/ml). Increasing amounts of heparin were included in the cell suspension as indicated before plating and adhesion assay performed. B, cells were cultured in medium containing the indicated amount of sodium chlorate for 24 h, washed, harvested, and then plated on Cyr61 (2 g/ml), fibronectin (FN, 1 g/ml), type I collagen (Col. I, 2 g/ml), vitronectin (VN, 0.5 g/ml), or laminin (LN, 5 g/ml). C, cells were treated with 40 mM sodium chlorate as above, either in the presence or absence of 10 mM Na 2 SO 4 . Data shown for all panels are mean Ϯ S.D. of triplicate determinations and are representative of three experiments.
concentrations from 0.4 to 0.8 M NaCl (Fig. 6B). DM bound to the heparin-Sepharose poorly, and nearly all protein was found in the wash fraction or eluted at 0.25 M NaCl. These elution profiles indicated that H1 and H2 exhibited reduced heparinbinding affinities compared with the wild type, whereas DM was severely deficient in heparin binding.
Various concentrations of mutant Cyr61 proteins were coated onto microtiter wells, and 1064SK fibroblasts were allowed to attach (Fig. 7A). Both H1 and H2 were able to support fibroblast adhesion similar to wild type Cyr61 (see Fig. 1A), although maximal adhesion was reached at a slightly higher concentration of mutant protein (2.5-5.0 g/ml) compared with wild type Cyr61 (1.5-2 g/ml). Moreover, fibroblast adhesion to both H1 and H2 exhibited the same characteristics as adhesion to wild type Cyr61 with regard to sensitivities to EDTA, divalent cations, RGD-containing peptides, and heparin, suggesting that H1 and H2 supported adhesion through integrin ␣ 6 ␤ 1 and HSPGs as wild type Cyr61 (data not shown). Importantly, fibroblast adhesion to H1 and H2 was blocked by mAb against the integrin subunit ␣ 6 , confirming that cell adhesion to these mutants is mediated through integrin ␣ 6 ␤ 1 as wild type Cyr61 (Fig. 7B). These results suggest that the integrin ␣ 6 ␤ 1 -binding sites of Cyr61 are distinct from the heparin-binding sites mutated in H1 and H2. The ability of H1 and H2 to support integrin ␣ 6 ␤ 1 -mediated cell adhesion is consistent with their ability to bind heparin, albeit at a lower affinity than wild type Cyr61 (Fig. 6B).
By contrast, DM was unable to support 1064SK fibroblast adhesion at any concentration tested, indicating that the intrinsic heparin-binding activity of Cyr61 is essential for mediating fibroblast adhesion (Fig. 7A). In a similar cell adhesion assay, DM was able to support adhesion of HUVECs (Fig. 8), as demonstrated for the wild type Cyr61 protein (4,18). Furthermore, HUVEC adhesion to the Cyr61 mutant DM was also mediated through integrin ␣ V ␤ 3 like the wild type, because this adhesion was inhibited by agents that disrupt ␣ V ␤ 3 -ligand interaction such as EDTA, RGD-containing peptide, and the anti-␣ V ␤ 3 mAb LM609. These results show that the Cyr61 mutant DM is conformationally intact and is able to interact with integrin ␣ V ␤ 3 . Cyr61 interaction with integrin ␣ V ␤ 3 is independent of its ability to bind heparin, whereas its interaction with integrin ␣ 6 ␤ 1 to mediate fibroblast adhesion is absolutely dependent on its binding to cell surface HSPGs. DISCUSSION The major findings of this study are that human skin fibroblasts adhere to Cyr61 through integrin ␣ 6 ␤ 1 , and this adhesion process requires the interaction of Cyr61 with cell surface HSPGs. These observations identify both the adhesion receptor for Cyr61 and a novel adhesive substrate for integrin ␣ 6 ␤ 1 and demonstrate the importance of HSPGs in cell adhesion through this integrin. Furthermore, whereas the heparin-binding activity of Cyr61 is crucial for ␣ 6 ␤ 1 -mediated fibroblast adhesion, it is not required for ␣ V ␤ 3 -mediated adhesion of endothelial cells. These results show a differential requirement for HSPGs in the interaction of Cyr61 with different integrins and suggest that Cyr61 may utilize and signal through different receptors in different cell types, thereby eliciting disparate responses.
The conclusion that human dermal fibroblast adhere to Cyr61 through integrin ␣ 6 ␤ 1 is supported by both biochemical Site-directed mutagenesis created aa substitutions as shown by the bold letters, resulting in mutants H1 and H2. Mutant DM combines the mutations in H1 and H2, resulting in the sequence as shown. B, equal amount of wild type (WT) or mutant Cyr61 protein was loaded on a CL-6B heparin-Sepharose column. After washing, bound protein was eluted with buffer containing the indicated concentration of sodium chloride. An equal amount of eluate from each fraction was analyzed on SDS-polyacrylamide gel electrophoresis followed by immunoblotting with anti-Cyr61 antibodies. L, purified proteins loaded on the heparin-Sepharose column; FT, flow through; W, buffer wash. and immunological evidence. Cell adhesion to Cyr61 was abolished by EDTA and Ca 2ϩ but was supported by Mn 2ϩ and Mg 2ϩ and was resistant to inhibition by RGD-containing peptide ( Figs. 1 and 2). These observations, taken in the context of fibroblasts, are consistent with and indicative of the involvement of ␤ 1 -type integrins, although ␣ 5 ␤ 1 and ␣ V ␤ 1 can be ruled out. Survey of a panel of highly specific, function-blocking mAbs directed against various integrin subunits revealed that only mAbs against integrins ␣ 6 and ␤ 1 were able to block fibroblast adhesion to Cyr61 (Fig. 3). The specificity and effectiveness of inhibition by these mAbs identify integrin ␣ 6 ␤ 1 as the primary adhesion receptor for Cyr61 in fibroblasts. In contrast to endothelial cells, integrin ␣ V ␤ 3 appears to play no role in the adhesion of fibroblasts to Cyr61.
An unexpected finding in this study is the absolute requirement for Cyr61 to interact with cell surface heparan sulfate proteoglycan in the adhesion process (Figs. 4 -7). This conclusion was reached based on several lines of evidence. First, the presence of heparin in the assay medium effectively blocked cell adhesion to Cyr61, whereas adhesion to other substrates was unaffected; second, perturbation of heparan sulfate moieties on the cell surface by either heparinase treatment or by metabolic inhibition of proteoglycan sulfation abrogated cell adhesion to Cyr61; and third, a Cyr61 mutant protein unable to bind heparin no longer supported fibroblasts adhesion. Collectively, these data argue that the Cyr61 heparin-binding site(s) must be available and competent to interact with cell surface heparan sulfate proteoglycan for it to support fibroblast adhesion.
Increasing evidence now indicates that HSPGs can function as co-receptors with integrins in cell-matrix interactions (28,29). For example, syndecans not only act as co-receptors for fibroblast growth factors (22,29) they also cooperate with integrins in cell adhesion to result in Rho-dependent assembly of focal adhesion plaques and actin stress fibers (30,31). Induction of ␤-casein expression by laminin in mammary epithelial cells is mediated through both integrin ␣ 6 ␤ 4 and cell surface HSPGs, again indicating a role in signal transduction for the cooperation between integrins and HSPGs (32). To our knowledge, however, the present study provides the first demonstration of an absolute requirement of HSPGs for integrin-mediated cell attachment. Although adhesion to other heparinbinding extracellular matrix molecules such as fibronectin and thromospondin also involves cell surface HSPGs, these interactions contribute to adhesion but are not essential for the attachment of cells to substrates (33,34; also see Fig. 4B).
The specific sequence in Cyr61 that interacts with integrin ␣ 6 ␤ 1 is currently unknown. Aside from laminin-1, integrin ␣ 6 ␤ 1 is known to bind invasin (35), a bacterial protein involved in the invasion of mammalian tissues, and fertilin (36), a sperm surface protein of the ADAM family involved in the binding of sperm to egg. However, there is no clearly discernible sequence similarity between the integrin-binding fragments or peptides derived from these protein ligands and Cyr61 (37)(38)(39)(40). Furthermore, three peptides that can bind integrin ␣ 6 ␤ 1 have been identified through a nonsaturating phage display library screening, and these peptides do not share sequence similarity with any of the known protein ligands of integrin ␣ 6 ␤ 1 or with Cyr61 (41). These findings suggest that integrin ␣ 6 ␤ 1 may recognize a broad range of sequences or may recognize structural or conformational features presented by its protein ligands.
Because the Cyr61 mutants H1 and H2 still mediated ␣ 6 ␤ 1dependent fibroblast adhesion (Fig. 7), it follows that the sequences altered in H1 and H2 are not essential for binding to this integrin. The mutant DM, which combined alterations in H1 and H2, lost the ability to mediate cell adhesion concomitant with the complete loss of heparin-binding activity. These results support the conclusion that the heparin-binding activity of Cyr61 is indispensable for integrin ␣ 6 ␤ 1 -mediated fibroblast adhesion. However, endothelial cell adhesion to Cyr61 through integrin ␣ V ␤ 3 was unaffected by the heparin-binding status of Cyr61, suggesting fundamental differences between the interaction of Cyr61 with integrins ␣ V ␤ 3 and ␣ 6 ␤ 1 (Fig. 8).
Given the high degree of sequence homologies (ϳ40 -50% aa FIG. 7. Heparin-binding defective Cyr61 mutant fails to support adhesion of 1064SK fibroblasts. A, cells were plated on wells coated with the indicated amounts of mutant Cyr61 proteins H1, H2, or DM. Adhesion assay was performed as described before. B, cells were preincubated with 20 g/ml mAb against integrin ␣ 6 at room temperature for 30 min and then plated on wild type Cyr61 (2 g/ml), H1 mutant (2.5 g/ml), H2 mutant (2.5 g/ml), or vitronectin (VN, 0.5 g/ml). Data shown are mean Ϯ S.D. of three determinations and representative of at least three experiments. BSA, bovine serum albumin. sequence identity) among members of the CCN protein family (3,9), we anticipate that other CCN proteins might also exhibit similar properties with respect to fibroblast adhesion. Members of this protein family share four conserved structural domains, each encoded by a separate exon (3,42). The heparinbinding motifs (Fig. 6) are well conserved, suggesting that other members of this protein family may interact with HSPGs and mediate fibroblast adhesion in a manner similar to Cyr61. A notable exception in the protein family is Cop-1/WISP-2, which is unique in that it lacks precisely the carboxyl-terminal domain where the heparin-binding motifs lie (43,44). Based on our results, we can hypothesize that Cop-1/WISP-2 would be unable to support integrin ␣ 6 ␤ 1 -mediated fibroblast adhesion, even though it can mediate adhesion of osteoblasts (45). It is also possible that Cop-1 may counteract the activities of other CCN proteins in a dominant negative manner (46). These possibilities can be experimentally tested to help unravel the distinct functions of each CCN protein.
Whereas Cyr61 has multiple functions, its roles in fibroblasts are likely related to granulation tissue formation during cutaneous wound healing. The cyr61 gene is transcriptionally activated in cultured fibroblasts by growth factors implicated in wound healing, including fibroblast growth factor, plateletderived growth factor, and transforming growth factor-␤1 (1,12,11). In addition, cyr61 expression level is minimal in normal dermis and becomes highly induced in dermal fibroblasts in granulation tissue during wound healing (14). A protein closely related to Cyr61, CTGF, is also induced in granulation tissues during wound repair (13). Expression of CTGF has been associated with many fibrotic diseases (47), and subcutaneous injection of CTGF into NIH Swiss mice results in the formation of granulation tissue (48). Furthermore, CTGF regulates collagen and fibronectin gene expression in fibroblasts, suggesting that it may play a role in matrix remodeling in wound healing (13,48). Thus, considering the biochemical and functional similarities of Cyr61 and CTGF (3), these proteins may play roles in wound healing by inducing angiogenesis upon endothelial cells and stimulating chemotaxis, proliferation, and matrix remodeling upon fibroblasts. In this context, it is interesting to note that whereas laminin is a major component of the basement membrane, its relatively low abundance in the stromal extracellular matrix suggests that its role in wound healing may be limited. Therefore, Cyr61 may serve as the major adhesive substrate for integrin ␣ 6 ␤ 1 in fibroblasts during dermal wound repair.
Targeted disruption of the integrin ␣ 6 and ␤ 4 genes share the same phenotype, neonatal death due to epidermolysis bullosa, suggesting that this phenotype is primarily because of the lack of integrin ␣ 6 ␤ 4 in hemidesmosomes between epithelial cells and the basement membrane rather than because of the lack of integrin ␣ 6 ␤ 1 (49,50). Examination of the ␣ 6 knockout mice also revealed abnormality in the lamination of the developing cerebral cortex and retina (51) where Cyr61 is also expressed (5,14). Comparison of the expression patterns of Cyr61 and integrin ␣ 6 during mouse embryonic development shows that they are also co-localized in such tissues as the skin epidermis, the nervous system (ganglia, nerves, and spinal cord), and the cardiovascular system (heart and endothelia), suggesting that the potential interaction between Cyr61 and integrin ␣ 6 ␤ 1 may play a role in the development of these organs and tissues (5,52). Integrin ␣ 6 ␤ 1 has also been implicated in regulating tumor invasion and metastasis (53)(54)(55), whereas Cyr61 can promote tumor growth and vascularization (8). It is thus possible to envisage an interaction between Cyr61 and integrin ␣ 6 ␤ 1 during tumorigenesis. The possibility that Cyr61 plays a role in development, wound healing, and tumorigenesis through unique signaling events resulting from its interactions with integrin ␣ 6 ␤ 1 and HSPGs is intriguing and merits further investigation.