Tenascin-C is an adhesion-modulatory extracellular matrix molecule that is highly expressed in tumors. To investigate the effect of tenascin-C on tumor cells, we analyzed its antiadhesive nature and effect on tumor cell proliferation in a fibronectin context. Glioblastoma and breast carcinoma cell adhesion was compromised by a mixed fibronectin/tenascin-C substratum, which concomitantly caused increased tumor-cell proliferation. We identified the antiadhesive mechanism as a specific interference of tenascin-C with cell binding to the HepII/syndecan-4 site in fibronectin through direct binding of tenascin-C to the 13th fibronectin type III repeat (FNIII13). Cell adhesion and proliferation levels were restored by the addition of FNIII13. Overexpression of syndecan-4, but not syndecan-1 or -2, reverted the cell adhesion defect of tenascin-C. We characterized FNIII13 as the binding site for syndecan-4. Thus we describe a novel mechanism by which tenascin-C impairs the adhesive function of fibronectin through binding to FNIII13, thereby inhibiting the coreceptor function of syndecan-4 in fibronectin-induced integrin signaling.

Tenascin-C is an adhesion-modulatory ECM3 molecule for a variety of cell types (1). It is prominently expressed in the stroma of most solid tumors (2) and found around newly formed blood vessels (3). In addition, tenascin-C expression precedes the manifestation of mammary neoplasia in stromelysin-1/MMP3 transgenic mice, suggesting that tenascin-C is involved in the early steps of tumorigenesis (4). Despite its intriguing expression pattern, the nature of adhesion modulation by tenascin-C and its role in tumorigenesis are not known.

The ECM has an important regulatory function in tissue homeostasis and, together with oncogenes and tumor suppressor genes, is critically involved in tumorigenesis (reviewed in Refs. 5 and 6). Enforced interaction of tumor cells with fibronectin can block proliferation in cell culture and decrease tumor growth in nude mice (7). Tenascin-C was shown to disrupt the interaction of cells with fibronectin and potentially enhance tumor cell proliferation. Chiquet-Ehrismann et al., (8) were the first to show that tenascin-C binds to fibronectin, blocks cell attachment to fibronectin, and increases proliferation of rat breast adenocarcinoma cells (2). Tenascin-C binds fibronectin in an RGD-independent manner and, thus, does not block the RGD cell-binding site in fibronectin (8). The mechanism by which tenascin-C blocks cell attachment to fibronectin is unknown.

Upon cell adhesion, signals from the ECM are coupled to the cytoskeleton through specific cell surface receptors (reviewed in Ref. 9). Cell adhesion to fibronectin involves integrins and proteoglycans that orchestrate the assembly of adhesion complexes and rearrangement of the actin cytoskeleton. This triggers the cytoplasmic signaling that determines cell behavior, e.g., survival and proliferation (reviewed in Ref. 10). Fibroblasts can attach to the cell-binding site of fibronectin (RGD and synergy site) but full spreading, including focal contact and actin stress-fiber formation, requires the additional activation of syndecan-4 (11). Cell binding of syndecan-4 was shown to be mediated by the heparin-binding site II (HepII site) in fibronectin (12). Upon clustering, syndecan-4 initiates cytoplasmic signaling pathways in conjunction with activated integrins (13).

Several integrins, syndecan and other sulfated glycosaminoglycans have been characterized as cell surface receptors for tenascin-C (reviewed in Ref. 1). Whether binding of tenascin-C to any of these receptors plays a role in the tenascin-C-induced adhesion modulation on fibronectin that affects tumor cell proliferation is unknown.

Investigating how tenascin-C alters cell adhesion and the proliferation of tumor cells under defined cell culture conditions, we found that tenascin-C specifically bound to the FNIII13 of the HepII site, thereby blocking cell-binding to fibronectin through syndecan-4. This caused enhanced proliferation of human glioblastoma and breast carcinoma cells that could be neutralized by the addition of recombinant FNIII13 as well as by overexpression of syndecan-4.

Reagents.

The following chemicals were obtained from Sigma Chemical Co. (St. Louis, MO): soybean trypsin inhibitor, PDGF-BB, insulin, ITS (1 mg/ml insulin, 0.55 mg/ml transferrin, 0.5 ng/ml selenium, 0.05 mg/ml BSA, and 0.47 μg/ml oleic acid), BSA, 1,4-diazobicyclo-(2.2.2.)-octane, and glutathione-Sepharose. Other reagents were purchased as indicated: Geneticin, DMEM, and αMEM (Life Technologies, Inc., Paisley, Scotland); [3H]thymidine (Moravek, Brea, CA); precasted 4–20% Tris-glycine gels (Invitrogen, Groningen, the Netherlands; Life Technologies, Inc.); Mowiol (Calbiochem, Schwalbach, Germany); FCS and horse serum (Bioconcept, Allschwil, Switzerland); Fugene6 (Roche, Mannheim, Germany); and protein G-Sepharose (Pharmacia, Wikström, Sweden).

DNA Constructs.

Plasmids and recombinant proteins were obtained as indicated: pCEP-Pu vector (14); FNIII13 (15); FNIII7-10 (16); FNIII4-6 (Luciano Zardi); FNIII12-15+CS (17); cDNAs for mouse syndecan-1 (18); syndecan-2 (19); and syndecan-4 (20).

Antibodies.

The following antibodies were used for ELISA, Western blot, and immunofluorescence: mouse antichicken tenascin-C M1 and 60 (21); mouse anti-syndecan-1 clone DL101 (22); mouse anti-syndecan-4 (8C7) (23); syndecan-4 clone150.9 and chicken anti-syndecan-2 (11); mouse anti-vinculin (St. Louis, MO); rabbit anti-fibronectin (24); rabbit anti-FAK (5592; Dr. David Schlaepfer, The Scripps Research Institute, La Jolla, CA); rabbit anti-phospho-Y397 FAK (BioSource International, Camarillo, CA); TRITC-coupled phalloidin and secondary FITC- and TRITC-coupled goat antimouse and goat antirabbit antibodies (Alexa, Molecular Probes, Eugene, OR).

Preparation of Tenascin-C, Fibronectin, and Recombinant Fibronectin Proteins.

Full-length chicken tenascin-C TN260 was cloned by insertion of all known extra fibronectin type III repeats of tenascin-C into construct pCDNA/TN 190 (25), subcloned into the pCEP-Pu vector (14) and transfected into human embryonic HEK-293 cells. Stable expressors were selected with puromycin. Recombinant tenascin-C was custom made by “4C” (Computer Cell Culture Center, Seneffe, Belgium) according to our protocol. Briefly, cells were grown to two-thirds confluence in 10% FCS-containing medium, the medium was replaced by serum-free DMEM, and conditioned medium was collected after 2 days. Conditioned medium was collected up to six times at intervals of 18 h, with cells kept in serum-containing medium between cycles. Recombinant tenascin-C was purified by immunoaffinity chromatography as described (25). Fibronectin was prepared by gelatin agarose chromatography as described (24, 26).

Cell Lines, Cell Culture, and Transfection.

All cell lines were originally obtained from American Type Culture Collection if not indicated otherwise: human KRIB osteosarcoma, MDA-MB435 breast carcinoma, T98G glioblastoma, and Chinese hamster CHO-K1 and derivatives (27). Cells were cultured in DMEM or αMEM with 10% FCS and antibiotics (0.36 mg/ml penicillin and 1 mg/ml streptomycin). Transfections were carried out with Fugene6 according to the manufacturer’s protocol. For selection of sTable syndecan overexpressors, T98G cells were grown with G418, and expression was analyzed by immunofluorescence. Clonal lines were derived by limited dilution.

Adhesion Assay.

Microtiter plates (60-well, Nunc, Roskilde, Denmark) were coated with 10 μg/ml ECM molecules to give 1 μg/cm2 (fibronectin and tenascin-C) and 4 μg/cm2 (FNIII13) for 1 h at 37°C. ECM proteins were coated separately, first with fibronectin and then with tenascin-C and FNIII13. The noncoated plastic surface was blocked with 1% heat-inactivated BSA in PBS giving rise to 10 μg/cm2 protein. Similarly, mixed substrata of collagen I and laminin1 with tenascin-C were prepared and tenascin-C was detected by ELISA. Efficient fibronectin and tenascin-C coating was determined by ELISA with an anti-fibronectin and anti-tenascin-C antibody (25), respectively, and by PAGE analysis of lysed surface-bound ECM material combined with Coomassie Brilliant Blue staining (data not shown).

Before plating, cells were serum starved for 18 h in DMEM and trypsinized. Trypsin was blocked with 100 μg/ml soybean trypsin inhibitor in PBS, and cells were resuspended in serum-free medium and counted. Approximately 500 cells/well were plated for the indicated time periods, fixed by the addition of glutaraldehyde (2% final concentration) for 15 min and stained with 0.1% crystal violet in 20% methanol for 30 min. Cells were observed under a Nikon microscope (Nikon Diaphot) equipped with a Nikon camera.

DNA Replication and Proliferation Assay.

Plates (96-well; Falcon, Franklin Lakes, NJ) were coated as described above. Cells were serum starved overnight and trypsinized as described. Cells (104) were transferred onto the coated plates in the presence of the indicated mitogens. Cells were labeled 14 h later with [3H]thymidine (0.5 μCi/well) for 4 h at 37°C. Incorporated [3H]thymidine was precipitated with 10% trichloroacetic acid and determined with a Beckman scintillation counter after cell lysis in 0.3 n NaOH and 2% SDS. For long-term cell proliferation assays, 2 × 103 MDA-MB435 cells were plated into ECM-coated 96-well plates in the presence of 100 ng/ml insulin and incubated for the indicated time periods in a humidified chamber at 37°C in a CO2-incubator. Fresh medium (50%) with growth factor was added every 24 h. Cells were trypsinized and counted at the indicated time points.

In vitro Binding Assay (ELISA).

ELISA 96-well plates were coated with the indicated ECM proteins for 1 h at 37°C and blocked with 1% skim milk and 0.05% Tween 20 in PBS. ECM proteins were added at the indicated concentrations in blocking solution for 1 h and detected with anti-fibronectin or anti-tenascin-C antibodies as described (8).

Immunofluorescence Microscopy.

Cells (104) were transferred onto 4-well Cellstar plastic plates (Greiner, Frickenhausen, Germany) coated with ECM proteins as described. Cells were fixed with 4% paraformaldehyde in 50 mm phosphate buffer and 5 mm EDTA in PBS for 15 min, blocked with 3% BSA, 0.5% Tween 20 in PBS, and incubated with primary and secondary antibodies in blocking solution. Slides were embedded in 10.5% Mowiol containing 2.5% 1,4-diazobicyclo-(2.2.2.)-octane as an antifade agent. Expression of syndecans was determined by immunofluorescence. Cells were fixed in methanol and incubated with anti-syndecan-1, -2 or -4 antibodies at a dilution of 1:50 each. Cells were analyzed by microscopy. Scale bars represent 50 μm (Fig. 5,B) or 25 μm (Figs. 4,C; 5, D and E; and 7 C).

FNIII13 Binding to Syndecan-4.

Attached T98G and T98G:S4* cells (1.5 × 106) were serum-starved 1 h before the addition of 100 μg/ml FNIII13 in PBS at 4°C. As a control, cells were incubated with 100 μg/ml FNIII13 plus 0.5 mg/ml heparin. Unbound FNIII13 was removed by three washes with copious amounts of PBS. Cells were lysed in a modified GST-fish buffer [0.2 mm Na2S2O5, 2 mm MgCl2, 2 mm DTT, 50 mm Tris-HCl (pH7.5), 1% NP40, 10% glycerol, and protease inhibitors], and cell surface-bound FNIII13 was purified by a pull-down experiment with glutathione-Sepharose in modified GST-fish buffer and detected by Western blot with an anti-His antibody. For detection of a syndecan-4/FNIII13 interaction, overnight serum-starved T98G:S4* cells were detached with 0.02% EDTA, washed, and incubated in PBS with 100 μg/ml FNIII13 for 15 min at room temperature. Cells were washed with copius amounts of PBS, lysed in 0.1% NP40 buffer [50 mm Tris-HCl (pH 7.5), 5 mm NaF, 250 mm NaCl, 5 mm EDTA, 0.1% NP40, and protease inhibitors] and 5 μg anti-syndecan-4 (8C7) antibody, and glutathione beads were used for immunoprecipitation and FNIII-GST pull-down, respectively.

Determination of Protein Affinity by Biosensor.

Solid-phase binding of FNIII13 to tenascin-C was determined in PBS by using a Biosensor (Fisons, Cambridge, United Kingdom). The kd was calculated using the LIBFIT program.

Tenascin-C Blocks Cell Attachment and Compromises Cell Spreading on Fibronectin.

Tumor cell lines were plated on adhesive substrata such as fibronectin, collagen I, and laminin1 or on the large tumor-expressed form of tenascin-C (Ref. 28; Table 1 and data not shown). Whereas most cell lines adhered to fibronectin, collagen I, and laminin1, <10% of cells attached to tenascin-C and remained rounded for 1 h after plating. Thus, tenascin-C was not adhesive for all cell lines tested. In fact, tenascin-C was persistently antiadhesive: the majority of cells tested failed to attach 20 h after plating in serum-free medium (Table 1). Mitogen addition (PDGF-BB, insulin, epidermal growth factor, lysophosphatidic acid, transforming growth factor β) did not reduce this antiadhesiveness of tenascin-C (Table 1 and data not shown).

In the tissue context, cells encounter tenascin-C in combination with other ECM molecules. Therefore, we investigated whether tenascin-C is antiadhesive when cells are allowed to attach to an adhesive ECM such as fibronectin, laminin1, or collagen I. Tenascin-C compromised T98G cell attachment at 1 h or 20 h after plating cells on a substratum containing equimolar amounts of tenascin-C and fibronectin (Fig. 1, A and B, and Table1). In contrast to a mixed fibronectin/tenascin-C substratum, the adhesiveness of collagen I or laminin1 with equimolar amounts of tenascin-C was the same as with no tenascin-C. Similar numbers of T98G cells (Fig. 1,C) attached and spread on these mixed substrata and on the individual ECM molecules. However, only 50% of T98G cells were attached and partially spread after 20 h on the mixed fibronectin/tenascin-C substratum (Fig. 1 B). Taken together, these results show that tenascin-C blocks cell attachment and interferes with proper cell spreading on fibronectin but not on collagen I or laminin1.

Increased α5β1 Integrin Expression Does Not Overcome Compromised Cell Attachment of Tenascin-C.

To test whether increased expression of the fibronectin-binding integrin α5β1 alleviates tenascin-C-induced compromised attachment on fibronectin, we compared adhesion of CHO-K1 cells, which express moderate levels of this integrin (29), with cells expressing essentially no fibronectin-binding integrins (CHO-B2) or overexpressing α5β1 (CHO-B2α27; Ref. 30). Cell adhesion on the mixed substratum was blocked similarly in all cell lines, irrespective of α5β1 integrin expression (Table 1). Furthermore, overexpression of α5β1 integrin in human HT29 colon carcinoma cells did not support cell attachment and spreading on the mixed fibronectin/tenascin-C substratum (data not shown). Thus, α5β1 integrin is probably not a direct target of tenascin-C action.

Increased Tumor Cell Proliferation on a Mixed Fibronectin/Tenascin-C Substratum.

Weak binding to the ECM correlates with enhanced proliferation of many tumor cells (31). Because tenascin-C and fibronectin are coexpressed in the tumor stroma (32), and tenascin-C weakens cell binding to fibronectin (8, 33, 34), we examined whether tenascin-C enhances cell proliferation and DNA synthesis of MDA-MB435 breast carcinoma, T98G glioblastoma, and CHO carcinoma cells grown on either fibronectin, tenascin-C, or a mixed fibronectin/tenascin-C substratum. As shown in Fig. 2,A, 24 and 33% more MDA-MB435 cells were counted on fibronectin/tenascin-C than on fibronectin after 51 h and 75 h of culture, respectively. In addition, the DNA replication indices (cpm/cell) of MDA-MB435 and T98G cells were approximately 2- and 3-fold higher on the mixed fibronectin/tenascin-C substratum than on fibronectin alone (Fig. 2,B; Table 2). The possibility that rounded cells took up more [3H]thymidine than attached cells could be ruled out, because the total radioactivity in cells plated on the different substrata was similar after 4 h when T98G cells had not yet entered S phase (data not shown). Increased DNA synthesis levels were also observed in the other tumor cell lines tested (Table 2). Thus, the mixed fibronectin/tenascin-C substratum triggered more cells to enter S-phase than fibronectin alone. Because all tumor cell lines tested were equally compromised in cell adhesion on a mixed fibronectin/tenascin-C substratum, we focused on two representative cell lines, the MDA-MB435 and T98G cells.

To address the possibility that tenascin-C alone and in ECM contexts other than fibronectin might also enhance proliferation, we investigated DNA synthesis with a pure tenascin-C substratum, on which cells remained rounded (Fig. 1,B), and a mixed substrata of collagen I or laminin1 containing tenascin-C, on which cells spread normally (Fig. 1,C). The DNA replication indices of all cell lines tested increased, with levels enhanced as much as 11-fold on pure tenascin-C (Table 2). However, on adhesive substrata containing tenascin-C (mixtures with collagen I or laminin1), rates of DNA synthesis were identical to those on the single ECM molecules (Fig. 2,B). To investigate whether tenascin-C shortened the S phase in tumor cells, we measured DNA synthesis in T98G cells every 2 h starting 11 h after plating (Fig. 2 C). On the mixed substratum, the kinetics of DNA replication were similar to those on fibronectin. In summary, these observations suggest that tenascin-C stimulates a subset of cells to enter the S phase of the cell cycle by interfering with fibronectin-specific cell adhesion signaling.

Tenascin-C Binds to the FNIII13 of the HepII Cell-binding Site in Fibronectin.

Because the antiadhesive and proliferation stimulatory effect of tenascin-C was specific for cells on a mixed fibronectin/tenascin-C substratum, we considered the possibility that tenascin-C blocks cell attachment on fibronectin by masking one of the cell-binding sites on fibronectin (Fig. 3,E). Although there are several reports that fibronectin binds tenascin-C (8, 34, 35, 36, 37), the binding site on fibronectin has not been characterized. We first confirmed by ELISA that fibronectin binds to substratum-immobilized tenascin-C in a dose-dependent manner (Fig. 3,A), and then we investigated the location of the binding site on fibronectin. Recombinant fragments FNIII4-6 (Fig. 3,B), FN12 15+CS (Fig. 3,B) and FNIII13 (Fig. 3,C), which are part of the heparin and cell-binding sites HepIII and HepII in fibronectin, respectively, bind to tenascin-C in a dose-dependent manner, reaching saturation (FNIII13, Fig. 3,B). In addition, tenascin-C and FNIII13 were found to form complexes in coimmunoprecipitation experiments (data not shown). FNIII13 competed with the binding of tenascin-C to surface-immobilized intact fibronectin in a concentration- dependent fashion (Fig. 3,D). Binding of FNIII7-10 to tenascin-C, including the RGD and synergy sites, was not detected by ELISA (Fig. 3 B). Thus, binding of tenascin-C to FNIII13 in the HepII site of fibronectin was specific. We also determined the affinity of FNIII13 for tenascin-C in a solid-phase assay as a kd of 128 ± 6.4 nm (data not shown). This binding affinity is in the same range as that obtained for the interaction of FNIII12-15 with the heparan sulfate chains extracted from syndecan-4 (63 ± 10 nm; Ref. 12). This newly characterized interaction of tenascin-C with FNIII13 may prevent cellular access to the HepII site in fibronectin.

The FNIII13 Restores the Tenascin-C-induced Cell Spreading Defect on Fibronectin.

If tenascin-C blocks spreading on fibronectin by preventing cells from binding to the HepII site of fibronectin, we postulated that addition of recombinant fragment FNIII13 would restore cell spreading. Indeed, results from attachment assays revealed that the majority of T98G cells (Fig. 4), MDA-MB435 cells (data not shown), and rat embryo fibroblasts REF52 (data not shown) spread on a triple matrix of fibronectin/tenascin-C/FNIII13. Thus, FNIII13 restored the cell-spreading defect caused by tenascin-C in the three cell lines tested. The fragment did not provide a spreading signal by itself, as T98G cells poorly attached and remained rounded on an FNIII13 substratum (Fig. 4,A). Cell spreading was also restored when cells were added to a fibronectin/tenascin-C substratum together with soluble FNIII13 (Fig. 4,B). In contrast to FNIII13, FNIII4-6 did not restore cell spreading on fibronectin/tenascin-C, either when offered immobilized or in solution (data not shown). A role for FNIII13 in the tenascin-C-specific spreading defect was confirmed by analysis of the cytoskeleton and adhesion structures of cells plated on the different substrata. With T98G cells plated for 2 h on fibronectin or on the mixed substratum and stained for vinculin, focal contacts formed on fibronectin but not on fibronectin/tenascin-C (Fig. 4,C, panels b and d). In addition, polymerization of actin into stress fibers was blocked by tenascin-C, as recently described for a substratum of tenascin-C plus fibronectin and fibrinogen (38). However, in contrast to the filopodia that form on the fibrinogen-based tenascin-C-containing substratum (38), we found no specific structures of polymerized actin in T98G cells (Fig. 4,C, panel c). Because the FNIII13 fragment restored cell spreading on fibronectin/tenascin-C, we investigated whether actin stress fibers and focal contacts were restored by FNIII13. Indeed, in T98G (Fig. 4 C, panels e and f) and REF52 cells (data not shown) actin stress fibers and focal adhesions were largely restored upon addition of the FNIII13 fragment to a fibronectin/tenascin-C substratum.

Thus, tenascin-C interferes with cell spreading, focal contact and actin stress fiber formation on fibronectin and this effect can be neutralized by addition of FNIII13.

FNIII13 Neutralizes the Stimulatory Effect of Tenascin-C on Cell Proliferation.

Because detachment by tenascin-C correlated with enhanced tumor cell proliferation and FNIII13-restored cell spreading, we investigated whether FNIII13 also reduced proliferation of cells on the mixed substratum to levels found on fibronectin. Similar numbers of MDA-MB435 breast carcinoma cells were counted on the triple matrix containing FNIII13 and on fibronectin alone 75 h after plating (Fig. 4,D). With T98G cells, similar DNA synthesis levels were determined on fibronectin/tenascin-C/FNIII13 and on fibronectin (Fig. 4 E).

Overexpression of Syndecan-4 Restores Cell Spreading and Proliferation on a Mixed Fibronectin/Tenascin-C Substratum.

The finding that tenascin-C binds to FNIII13 localized in the HepII site of fibronectin and that this interaction competes with cell binding to fibronectin suggests that tenascin-C competes with a cell surface receptor binding to the HepII site in fibronectin. Syndecan-4, a member of the syndecan family of transmembrane heparan sulfate proteoglycans (39, 40) is a potential receptor candidate because it is reported to bind to the HepII site in fibronectin and is required for full cell spreading on fibronectin (11, 41). Thus, we examined whether tenascin-C competes with syndecan-4 for binding to the HepII site in fibronectin. We tested whether activation of syndecan-4 through overexpression rescues the tenascin-C-induced spreading defect on fibronectin. Pools of T98G cells (T98G:S4) overexpressing syndecan-4 (Fig. 5,B) attached and the majority spread on the mixed fibronectin/tenascin-C substratum (Fig. 5, A and C). This was in contrast to T98G cells with lower endogenous syndecan-4 expression levels (Fig. 5,B), which spread very poorly (3.5%) on the mixed substratum (Figs. 1,B and 5,A). Also spreading of the syndecan-4-overexpressing clone T98G:S4* was completely restored on the mixed fibronectin/ tenascin-C substratum 2 h after plating (Fig. 5,D). In addition, overexpression of syndecan-4 in T98G pools (data not shown) and clonal T98G:S4* cells (Fig. 5,E) also completely restored focal contact and actin stress fiber formation on the mixed fibronectin/tenascin-C substratum (Fig. 5,D, panels b and d). This observation indicates a function of syndecan-4 in cell spreading linked to Rho-mediated actin stress fiber formation, as recently suggested (42). Activation of FAK by autophosphorylation at Y397 is an early step in cell adhesion signaling. This is compromised in T98G parental cells (data not shown) and in T98G:S4* cells on a fibronectin/tenascin-C substratum (Fig. 5,F). In contrast, plating on a substratum of fibronectin/ tenascin-C that contains FNIII13 largely restored FAK autophosphorylation, indicating that activation of syndecan-4 by FNIII13 is linked to the restoration of cell adhesion signaling by FNIII13 (Fig. 5,F). Moreover, overexpression of syndecan-4 also reduced DNA replication levels on a mixed fibronectin/tenascin-C substratum to that on fibronectin (Fig. 5 G). In summary, restoration of cell adhesion to fibronectin by overexpression of syndecan-4 or by addition of FNIII13 neutralized the tenascin-C effect on cell proliferation.

Syndecan-4 Binds the FNIII13 of Fibronectin.

To investigate FNIII13 as a potential ligand of syndecan-4, we tested for an interaction of syndecan-4 with FNIII13. Upon addition of FNIII13 to T98G and T98G:S4* cells, either when attached or in solution, we detected FNIII13 in a syndecan-4 immunoprecipitation and syndecan-4 in a FNIII13-GST pull-down experiment (Fig. 6, A and B).

FNIII13 bound to the cell surface of T98G:S4* at higher levels than to parental T98G cells (Fig. 6,B), but not to T98G:S1 or T98G:S2 cells (data not shown; see below). This interaction was largely impaired by heparin (Fig. 6, A and B), indicating that syndecan-4 primarily bound FNIII13 in a glycosaminoglycan-dependent manner. This formally proves that FNIII13 is a ligand for syndecan-4. To test whether overexpression of syndecan-1 and -2 also rescues the tenascin-C-specific cell-spreading defect, we generated T98G cells that stably overexpressed syndecan-1 (T98G:S1) and -2 (T98G:S2; Fig. 7,C) and plated these cells on fibronectin/ tenascin-C. In contrast with syndecan-4, neither syndecan-1 nor syndecan-2 overexpression allowed cells to spread on the mixed substratum (Fig. 7, A and B). As with T98G:S4*, syndecan-1 and -2 overexpressors attached neither to FNIII13 nor tenascin-C (Fig. 7 B).

In summary, rescue of cell spreading was only accomplished by activation of syndecan-4 and not by syndecan-1 or -2. We showed that FNIII13 serves as a ligand for syndecan-4 and that tenascin-C specifically competes with binding of syndecan-4 to FNIII13, thereby preventing cell spreading on fibronectin (Fig. 8).

Tenascin-C Enhances Tumor Cell Proliferation by Blocking Fibronectin Adhesion Signaling.

Cell adhesion to fibronectin plays an important role in tumorigenesis and angiogenesis, with an inverse correlation between tumorigenesis and adhesion of tumor cells to fibronectin (reviewed in Refs. 43 and 44). In particular, blocking α5β1 integrin enhances DNA replication (45), and on the contrary, overexpression of integrin α5β1 decreases proliferation and tumorigenesis of CHO cells in nude mice (31, 46). Although the single animal model analyzed does not support a tumorigenesis-enhancing effect of tenascin-C (47), a wealth of immunohistochemical studies (48, 49) and cell culture experiments (2, 8) suggest a role for tenascin-C in tumorigenesis, probably by enhancing proliferation of cancer cells in situ.

Here we show that tenascin-C enhanced proliferation of a variety of tumor cell lines, including glioblastoma and breast carcinoma cells. Thus in the tissue context, tenascin-C may increase tumor mass by elevating the number of tumor cells. There is no evidence for an altered apoptosis rate,4 but we found that a higher proportion of cells on the fibronectin/tenascin-C mixture entered S phase than counterparts grown attached to fibronectin. Apparently, cells not yet determined to enter S-phase on fibronectin were triggered to start DNA synthesis upon tenascin-C-compromised adhesion to fibronectin. It will be interesting to see whether enhanced proliferation on the mixed fibronectin/tenascin-C substratum is linked to increased cyclin E-cdk2 activity attributable to reduced p27Kip1 inhibitor levels, as was recently shown to occur in detached myeloma cells, in comparison with fibronectin-attached counterparts (50).

Detached cells largely fail to arrest in the G1 phase upon exposure to genotoxic levels of radiation (51) and are thus probably more prone to accumulate mutations. We speculate that tenascin-C challenges genomic stability through its antiadhesive and proliferation stimulatory properties and by that may contribute to tumorigenesis.

Tenascin-C Interferes with Fibronectin Adhesion Signaling.

Tenascin-C specifically interfered with cell attachment and spreading on fibronectin, suggesting that integrin signaling was abrogated. This is supported by our observation that tenascin-C compromised focal contact and actin stress fiber formation, two hallmarks of integrin-mediated cell adhesion. In addition, β1 integrins did not localize to focal contacts and neither integrin activation by MnCl24 nor overexpression of α5β1 integrin reverted the tenascin-C phenotype. Our data suggest that tenascin-C affects integrin function by an indirect mechanism (see below).

Tenascin-C Masks the Cell-binding Site in the FNIII13 of Fibronectin.

We investigated the possibility that tenascin-C blocks cell-binding sites in fibronectin and confirmed that tenascin-C binds to fibronectin (8). We characterized this interaction as specific binding of tenascin-C to the HepII and HepIII cell-binding sites in fibronectin. The fibronectin type III repeats 1–5 of tenascin-C are documented to bind fibronectin (34). Also for FNIII13 we find strongest binding to the fibronectin type III repeats of tenascin-C.4 Whereas FNIII4-6 does not affect cell spreading of T98G cells,4 FNIII13 neutralized the spreading defect and restored actin stress fiber and focal contact formation and FAK-associated adhesion signaling on a mixed fibronectin/tenascin-C substratum. FNIII13 also neutralized the tenascin-C effect on tumor cell proliferation. An important function of FNIII13 in cell spreading was demonstrated recently (15): the addition of FNIII13 led to full cell spreading on the major cell-binding site of fibronectin (FNIII7-11). We conclude that tenascin-C efficiently blocks cell access to fibronectin by binding directly to the cell-binding site in FNIII13, thereby inhibiting full cell spreading.

Syndecan-4 Overexpression Neutralizes the Tenascin-C Effect on Cell Spreading and Proliferation.

Because syndecan-4 is required for full cell spreading through interaction with the HepII site in fibronectin (41, 42), it was a candidate for inactivation by tenascin-C. Indeed, overexpression of syndecan-4 in T98G cells rescued the spreading defect and the prevention of actin stress fiber formation on a mixture of fibronectin and tenascin-C. This was specific for syndecan-4. Both addition of FNIII13 and overexpression of syndecan-4 restored tenascin-C-compromised cell spreading on fibronectin and neutralized the effect of tenascin-C on cell proliferation. We conclude that FNIII13 neutralization of the tenascin-C effect is mediated through syndecan-4 and that binding of syndecan-4 occurs through FNIII13 within the characterized syndecan-4 recognition sequence in fibronectin (12) because we now formally proved FNIII13 as ligand for syndecan-4. Because glycosylation-deficient cells with endogenous syndecan-4 levels are impaired in full spreading (42), and we found that heparin interfered with FNIII13 binding to T98G:S4* cells, we suggest that the interaction of syndecan-4 with FNIII13 is primarily mediated through glycosaminoglycans. This is in agreement with crystallographic, mutagenesis and sequence conservation data that identified a sequence in FNIII13 as the major heparin binding site in fibronectin (17). Additional support for a syndecan-4-binding site in FNIII13 derives also from experiments by Bloom et al.(15), who reported that any mutation in FNIII13 abrogating heparin binding also reduced the ability of FNIII13 to induce actin stress fibers, which was shown to be a property of syndecan-4 activation (42). Taken together, our results indicate that syndecan-4 binds to the same site in fibronectin as tenascin-C (FNIII13) and that the interaction of tenascin-C with FNIII13 competes with syndecan-4 binding. Another potential syndecan-4-binding site in fibronectin (FNIII14; Ref. 52) is apparently not relevant for the tenascin-C-induced cell spreading defect, because addition of FNIII13 alone was sufficient to restore cell spreading. Whether other cell surface receptors are influenced by tenascin-C through competitive binding to FNIII13 needs to be addressed in future experiments.

In contrast to syndecan-1, which has been linked to tumor growth, especially in association with the Wnt signaling pathway (53), only limited data imply a role for syndecan-4 in tumor cell proliferation. Zvibel et al.(54) showed that the expression of growth-promoting erb-B2 and erb-B3 is increased upon addition of cell surface-shed syndecan-4 to colon cancer cells that may compete with the function of membrane-bound syndecan-4. Our data support and extend this novel link between tumor cell proliferation and syndecan-4 by showing that blocking syndecan-4 function through tenascin-C enhances tumor cell proliferation, including glioblastoma and breast carcinoma cells. Furthermore, this suggests that fibronectin signaling through syndecan-4 and integrins attenuates tumor cell proliferation.

The blocking of integrin function by competition mechanisms is an emerging topic in cell adhesion modulation as, for example, high molecular kininogen masking the αvβ3 integrin-binding site in vitronectin (55) and the melanoma inhibitory activity competing with α4β1 integrin binding to the 14th fibronectin type III repeat in fibronectin (56).

In conclusion, we have established a mechanistic link between not yet understood observations that deregulated fibronectin signaling can stimulate tumor cell growth, that tenascin-C is highly expressed in tumor tissue, and that tenascin-C blocks cell adhesion to fibronectin. Here we described a mechanism by which tenascin-C impairs the adhesive properties of fibronectin by blocking the coreceptor function of syndecan-4 in integrin signaling, thereby triggering tumor cell proliferation.

Fig. 1.

Compromised cell adhesion on a mixed fibronectin/tenascin-C substratum. All experiments were performed in triplicates. ECM molecules were coated at 1 μg/cm2 (10 μg/ml). T98G glioblastoma cells were plated on the indicated substrata, fixed for 1 h (A) or 20 h (B and C) after plating and were stained. Round () and spread cells (▪) were counted (A and B).

Fig. 1.

Compromised cell adhesion on a mixed fibronectin/tenascin-C substratum. All experiments were performed in triplicates. ECM molecules were coated at 1 μg/cm2 (10 μg/ml). T98G glioblastoma cells were plated on the indicated substrata, fixed for 1 h (A) or 20 h (B and C) after plating and were stained. Round () and spread cells (▪) were counted (A and B).

Close modal
Fig. 2.

Tenascin-C enhances proliferation of tumor cells. A, MDA-MB435 breast carcinoma cells were grown on fibronectin (▪) or on an equimolar mixture of fibronectin and tenascin-C (▴) in the presence of 100 ng/ml insulin. The cells were counted at the indicated time points. B, T98G glioblastoma cells were cultured on fibronectin, collagen I, laminin1, and on mixtures of these ECM molecules with tenascin-C as indicated, in the presence of 40 ng/ml PDGF-BB. [3H]thymidine incorporation per attached cell is expressed relative to that on fibronectin. C, T98G glioblastoma cells were grown on fibronectin (▪) or on a mixture of fibronectin and tenascin-C (▴) in the presence of 40 ng/ml PDGF-BB, labeled with [3H]thymidine for 1 h, and harvested and counted at the indicated time points.

Fig. 2.

Tenascin-C enhances proliferation of tumor cells. A, MDA-MB435 breast carcinoma cells were grown on fibronectin (▪) or on an equimolar mixture of fibronectin and tenascin-C (▴) in the presence of 100 ng/ml insulin. The cells were counted at the indicated time points. B, T98G glioblastoma cells were cultured on fibronectin, collagen I, laminin1, and on mixtures of these ECM molecules with tenascin-C as indicated, in the presence of 40 ng/ml PDGF-BB. [3H]thymidine incorporation per attached cell is expressed relative to that on fibronectin. C, T98G glioblastoma cells were grown on fibronectin (▪) or on a mixture of fibronectin and tenascin-C (▴) in the presence of 40 ng/ml PDGF-BB, labeled with [3H]thymidine for 1 h, and harvested and counted at the indicated time points.

Close modal
Fig. 3.

ELISA assay for determination of binding of recombinant fibronectin fragments to tenascin-C. Native fibronectin (A) or the indicated recombinant fibronectin fragments (B and C) were added to tenascin-C-coated wells, and bound fragments were detected with a polyclonal anti- fibronectin antibody. D, tenascin-C was preincubated with increasing concentrations of FNIII13 before addition to fibronectin-coated wells. Bound tenascin-C was detected with a monoclonal anti-tenascin-C antibody. E, in this scheme of fibronectin, narrow boxes, ovals, and wide boxes represent fibronectin type I, II, and III repeats, respectively. The heparin-binding sites (HepI-III), RGD, and the synergy site are marked and aligned with interacting cell surface receptors (integrins; e.g., α5β1), syndecan-4 (S4), and chondroitin sulfate proteoglycan (CSPG). Localization of the recombinant fragments used and binding to tenascin-C are depicted above the model.

Fig. 3.

ELISA assay for determination of binding of recombinant fibronectin fragments to tenascin-C. Native fibronectin (A) or the indicated recombinant fibronectin fragments (B and C) were added to tenascin-C-coated wells, and bound fragments were detected with a polyclonal anti- fibronectin antibody. D, tenascin-C was preincubated with increasing concentrations of FNIII13 before addition to fibronectin-coated wells. Bound tenascin-C was detected with a monoclonal anti-tenascin-C antibody. E, in this scheme of fibronectin, narrow boxes, ovals, and wide boxes represent fibronectin type I, II, and III repeats, respectively. The heparin-binding sites (HepI-III), RGD, and the synergy site are marked and aligned with interacting cell surface receptors (integrins; e.g., α5β1), syndecan-4 (S4), and chondroitin sulfate proteoglycan (CSPG). Localization of the recombinant fragments used and binding to tenascin-C are depicted above the model.

Close modal
Fig. 4.

FNIII13 reverts tenascin-C-induced cell morphology and proliferation. T98G glioblastoma cells were plated on the indicated substrata for 1 h (A and B) or 2 h (C), fixed, and stained with crystal violet (A) or vinculin and TRITC-coupled phalloidin (C). Adherent cells were photographed (A and C), and spread cells were counted (B). A 6-fold molar excess of FNIII13 (4 μg/cm2) was either bound to a mixed fibronectin/tenascin-C substratum before the addition of the cells (FN/TN/III13) or was added in solution together with the cells (FN/TN+FNIII13). Proliferation of MDA-MB435 cells (D) and DNA replication in T98G glioblastoma cells (E) were determined as described (compare Fig. 2).

Fig. 4.

FNIII13 reverts tenascin-C-induced cell morphology and proliferation. T98G glioblastoma cells were plated on the indicated substrata for 1 h (A and B) or 2 h (C), fixed, and stained with crystal violet (A) or vinculin and TRITC-coupled phalloidin (C). Adherent cells were photographed (A and C), and spread cells were counted (B). A 6-fold molar excess of FNIII13 (4 μg/cm2) was either bound to a mixed fibronectin/tenascin-C substratum before the addition of the cells (FN/TN/III13) or was added in solution together with the cells (FN/TN+FNIII13). Proliferation of MDA-MB435 cells (D) and DNA replication in T98G glioblastoma cells (E) were determined as described (compare Fig. 2).

Close modal
Fig. 5.

Activation of syndecan-4 restores cell spreading, actin stress fibers, FAK autophosphorylation, and DNA replication. Parent T98G cells, pools (T98G: S4), or a selected clone (T98G: S4*) of T98G glioblastoma cells stably overexpressing syndecan-4 were plated on the indicated substrata for 2 h (A, C, and D), 18 h (B, E, and G), or 30 min (F); fixed, stained with crystal violet (A and C), photographed (A), and counted (C) or stained with TRITC-labeled phalloidin and vinculin (D). Arrowheads point at focal adhesions. Syndecan-4 was detected by immunofluorescence with an anti-syndecan-4 antibody (B and E). There was no staining with an unspecific anti-His monoclonal antibody (E; control) or with the secondary antibody alone (data not shown). F, FAK autophosphorylation at Y397 and total FAK expression levels were assessed by immunoblotting with specific antibodies. G, DNA replication was determined as described (compare Fig. 2).

Fig. 5.

Activation of syndecan-4 restores cell spreading, actin stress fibers, FAK autophosphorylation, and DNA replication. Parent T98G cells, pools (T98G: S4), or a selected clone (T98G: S4*) of T98G glioblastoma cells stably overexpressing syndecan-4 were plated on the indicated substrata for 2 h (A, C, and D), 18 h (B, E, and G), or 30 min (F); fixed, stained with crystal violet (A and C), photographed (A), and counted (C) or stained with TRITC-labeled phalloidin and vinculin (D). Arrowheads point at focal adhesions. Syndecan-4 was detected by immunofluorescence with an anti-syndecan-4 antibody (B and E). There was no staining with an unspecific anti-His monoclonal antibody (E; control) or with the secondary antibody alone (data not shown). F, FAK autophosphorylation at Y397 and total FAK expression levels were assessed by immunoblotting with specific antibodies. G, DNA replication was determined as described (compare Fig. 2).

Close modal
Fig. 6.

FNIII13, a ligand for syndecan-4. Parental and syndecan-4-overexpressing T98G: S4* cells were incubated with 100 μg/ml FNIII13 for 1 h in the presence or absence of 0.5 mg/ml heparin. Syndecan-4-bound FNIII13 was detected by immunoprecipitation of syndecan-4 and immunoblotting for FNIII13 or by pull-down of the GST-tagged FNIII13 with glutathione-Sepharose and immunoblotting for syndecan-4 (A). Cell-surface-bound FNIII13 was detected by GST pull-down of FNIII13 and Western blot with an anti-His antibody (B).

Fig. 6.

FNIII13, a ligand for syndecan-4. Parental and syndecan-4-overexpressing T98G: S4* cells were incubated with 100 μg/ml FNIII13 for 1 h in the presence or absence of 0.5 mg/ml heparin. Syndecan-4-bound FNIII13 was detected by immunoprecipitation of syndecan-4 and immunoblotting for FNIII13 or by pull-down of the GST-tagged FNIII13 with glutathione-Sepharose and immunoblotting for syndecan-4 (A). Cell-surface-bound FNIII13 was detected by GST pull-down of FNIII13 and Western blot with an anti-His antibody (B).

Close modal
Fig. 7.

Overexpression of syndecan-1 or -2 does not rescue cell spreading. T98G cells overexpressing syndecan-1, -2, or -4 (S4*) were plated on the indicated substrata for 2 h, fixed, stained, photographed (A), and counted (B). The majority of T98G: S4* cells were spread on fibronectin/tenascin-C in contrast to T98G: S1 and T98G: S2, which remained rounded. The numbers of attached cells [round () and spread cells (▪)] on the individual substrata are expressed with respect to a fibronectin reference (100%). Expression levels of syndecan-1 and -2 in T98G: S1 and T98G: S2 were determined by immunofluorescence C, syndecan-2 expression was determined by incubation with chicken anti-syndecan-2 and then rabbit antichicken and FITC-coupled goat antirabbit immunostaining.

Fig. 7.

Overexpression of syndecan-1 or -2 does not rescue cell spreading. T98G cells overexpressing syndecan-1, -2, or -4 (S4*) were plated on the indicated substrata for 2 h, fixed, stained, photographed (A), and counted (B). The majority of T98G: S4* cells were spread on fibronectin/tenascin-C in contrast to T98G: S1 and T98G: S2, which remained rounded. The numbers of attached cells [round () and spread cells (▪)] on the individual substrata are expressed with respect to a fibronectin reference (100%). Expression levels of syndecan-1 and -2 in T98G: S1 and T98G: S2 were determined by immunofluorescence C, syndecan-2 expression was determined by incubation with chicken anti-syndecan-2 and then rabbit antichicken and FITC-coupled goat antirabbit immunostaining.

Close modal
Fig. 8.

Model of tenascin-C-mediated cell adhesion modulation. Cell adhesion to fibronectin that is mediated through the interactions of integrin α5β1 with FNIII9/10 (III9/10) and syndecan-4 (S4) with FNIII13 (III13) attenuates tumor cell proliferation. Tenascin-C binds to FNIII13 and thus prevents syndecan-4 activation and cell adhesion and spreading. In consequence, fibronectin adhesion signaling is impaired, and that results in enhanced tumor-cell proliferation.

Fig. 8.

Model of tenascin-C-mediated cell adhesion modulation. Cell adhesion to fibronectin that is mediated through the interactions of integrin α5β1 with FNIII9/10 (III9/10) and syndecan-4 (S4) with FNIII13 (III13) attenuates tumor cell proliferation. Tenascin-C binds to FNIII13 and thus prevents syndecan-4 activation and cell adhesion and spreading. In consequence, fibronectin adhesion signaling is impaired, and that results in enhanced tumor-cell proliferation.

Close modal

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.

1

G. O. was supported by long-term fellowships from the European Molecular Biology Organization and the Swiss Cancer League.

3

The abbreviations used are: ECM, extracellular matrix; HepII site, heparin-binding site II; FNIII13, 13th fibronectin type III repeat; TRITC, tetramethylrhodamine isothiocyanate; GST, glutathione S-transferase; His, histidine; CHO, Chinese hamster ovary; kd, affinity; FAK, focal adhesion kinase; RGD, arginine, glycine, aspartic acid; PDGF-BB, platelet-derived growth factor BB; MMP3, matrix metalloprotease 3.

4

W. Huang, R. Chiquet-Ehrismann, and G. Orend, unpublished data.

Table 1

Compromised cell adhesion by tenascin-C

Results of 1 h-attachment and 20 h-adhesion assays are summarized for the indicated tumor cell lines. Cells were plated in serum-free medium or in medium supplemented with 40 ng/ml PDGF-BB, 100 ng/ml insulin, and ITS. The numbers of attached cells are described as the percentage of cells attached on fibronectin, including SD. Numbers of a representative experiment are shown.
Cell lineTimeTNFN/TNBSA
T98G 1 h 9.3 ± 0.9 31.9 ± 5.2 8.1 ± 0.8 
MDA MB 435  8.2 ± 3.9 24.5 ± 2.8 7.4 ± 4.1 
T98Ga 20 h 14.6 ± 0.4 50.4 ± 2.7 11.1 ± 2.7 
MDA MB 435b  4.0 ± 0.2 62.2 ± 5.1 4.7 ± 2.1 
CHO-B2c  8.0 ± 2.9 31.5 ± 9.6 7.4 ± 2.4 
CHO-K1c  4.7 ± 0.2 37.9 ± 2.1 7.4 ± 1.4 
CHO-B2α27c  4.3 ± 0.7 45.1 ± 3.7 8.1 ± 2.1 
Results of 1 h-attachment and 20 h-adhesion assays are summarized for the indicated tumor cell lines. Cells were plated in serum-free medium or in medium supplemented with 40 ng/ml PDGF-BB, 100 ng/ml insulin, and ITS. The numbers of attached cells are described as the percentage of cells attached on fibronectin, including SD. Numbers of a representative experiment are shown.
Cell lineTimeTNFN/TNBSA
T98G 1 h 9.3 ± 0.9 31.9 ± 5.2 8.1 ± 0.8 
MDA MB 435  8.2 ± 3.9 24.5 ± 2.8 7.4 ± 4.1 
T98Ga 20 h 14.6 ± 0.4 50.4 ± 2.7 11.1 ± 2.7 
MDA MB 435b  4.0 ± 0.2 62.2 ± 5.1 4.7 ± 2.1 
CHO-B2c  8.0 ± 2.9 31.5 ± 9.6 7.4 ± 2.4 
CHO-K1c  4.7 ± 0.2 37.9 ± 2.1 7.4 ± 1.4 
CHO-B2α27c  4.3 ± 0.7 45.1 ± 3.7 8.1 ± 2.1 
a

Medium supplemented with PDGF-BB.

b

Medium supplemented with 100 ng/ml insulin.

c

Medium supplemented with ITS.

Table 2

Enhanced DNA synthesis by tenascin-C

DNA replication was determined by 3[H]-thymidine incorporation and is described as relative increase in cpm over cells plated on fibronectin [fold (cpm/cell)], including SD. Numbers of representative experiments are shown. Serum-starved cells were grown in serum-free medium supplemented with 40 ng/ml PDGF-BB, 100 ng/ml insulin, and ITS.
Cell line TN FN/TN 
T98Ga 3.0 ± 0.9 3.0 ± 0.2 
MDA MB 435b 2.7 ± 0.6 1.8 ± 0.6 
CHO-B2c 7.1 ± 0.2 2.6 ± 0.3 
CHO-K1c 5.9 ± 1.7 2.3 ± 0.2 
CHO-B2α27c 11 ± 3.9 2.5 ± 0.6 
DNA replication was determined by 3[H]-thymidine incorporation and is described as relative increase in cpm over cells plated on fibronectin [fold (cpm/cell)], including SD. Numbers of representative experiments are shown. Serum-starved cells were grown in serum-free medium supplemented with 40 ng/ml PDGF-BB, 100 ng/ml insulin, and ITS.
Cell line TN FN/TN 
T98Ga 3.0 ± 0.9 3.0 ± 0.2 
MDA MB 435b 2.7 ± 0.6 1.8 ± 0.6 
CHO-B2c 7.1 ± 0.2 2.6 ± 0.3 
CHO-K1c 5.9 ± 1.7 2.3 ± 0.2 
CHO-B2α27c 11 ± 3.9 2.5 ± 0.6 
a

Serum-free medium supplemented with 40 ng/ml PDGF-BB.

b

Serum-free medium supplemented with 100 ng/ml insulin.

c

Serum-free medium supplemented with ITS.

We thank the following people for reagents: Richard Hynes for FNIII13; Luciano Zardi for FNIII4-6; Martin Humphries for FNIII12-15+CS; Harold Erickson for FNIII7-10; Kevin Langford for syndecan-1 cDNA; Merton Bernfield and Guido David for anti-syndecan-1 and -4 antibodies; and Anne Woods for syndecan-2 and -4 cDNAs and antibodies. We also thank Josephine Adams, Claudia Bagutti, Uli Müller, and Wilhelm Krek for discussion and critical reading of the manuscript.

1
Vollmer G. Biologic and oncologic implications of tenascin-C/hexabrachion proteins.
Crit. Rev. Oncol. Hematol.
,
25
:
187
-210,  
1997
.
2
Chiquet-Ehrismann R., Mackie E. J., Pearson C. A., Sakakura T. Tenascin: an extracellular matrix protein involved in tissue interactions during fetal development and oncogenesis.
Cell
,
47
:
131
-139,  
1986
.
3
Schnyder B., Semadeni R. O., Fischer R. W., Vaughan L., Car B. D., Heitz P. U., Winterhalter K. H., Odermatt B. F. Distribution pattern of tenascin-C in normal and neoplastic mesenchymal tissues.
Int. J. Cancer
,
72
:
217
-224,  
1997
.
4
Thomasset N., Lochter A., Sympson C. J., Lund L. R., Williams D. R., Behrendtsen O., Werb Z., Bissell M. J. Expression of autoactivated stromelysin-1 in mammary glands of transgenic mice leads to a reactive stroma during early development.
Am. J. Pathol.
,
153
:
457
-467,  
1998
.
5
Boudreau N., Bissell M. J. Extracellular matrix signaling: integration of form and function in normal and malignant cells.
Curr. Opin. Cell Biol.
,
10
:
640
-646,  
1998
.
6
Ruoslahti E. Fibronectin and its integrin receptors in cancer.
Adv. Cancer Res.
,
76
:
1
-20,  
1999
.
7
Akamatsu H., Ichihara-Tanaka K., Ozono K., Kamiike W., Matsuda H., Sekiguchi K. Suppression of transformed phenotypes of human fibrosarcoma cells by overexpression of recombinant fibronectin.
Cancer Res.
,
56
:
4541
-4546,  
1996
.
8
Chiquet-Ehrismann R., Kalla P., Pearson C. A., Beck K., Chiquet M. Tenascin interferes with fibronectin action.
Cell
,
53
:
383
-390,  
1988
.
9
Hynes R. O. Cell adhesion: old and new questions.
Trends Cell Biol.
,
9
:
M33
-M37,  
1999
.
10
Giancotti F. G., Ruoslahti E. Integrin signaling.
Science (Wash. DC)
,
285
:
1028
-1032,  
1999
.
11
Woods A., Couchman J. R. Syndecan-4 heparan sulfate proteoglycan is a selectively enriched and widespread focal adhesion component.
Mol. Biol. Cell
,
5
:
183
-192,  
1994
.
12
Woods A., Longley R. L., Tumova S., Couchman J. R. Syndecan-4 binding to the high affinity heparin-binding domain of fibronectin drives focal adhesion formation in fibroblasts.
Arch. Biochem. Biophys.
,
374
:
66
-72,  
2000
.
13
Rapraeger A. C. Syndecan-regulated receptor signaling.
J. Cell Biol.
,
149
:
995
-998,  
2000
.
14
Kohfeldt E., Maurer P., Vannahme C., Timpl R. Properties of the extracellular calcium binding module of the proteoglycan testican.
FEBS Lett.
,
414
:
557
-561,  
1997
.
15
Bloom L., Ingham K. C., Hynes R. O. Fibronectin regulates assembly of actin filaments and focal contacts in cultured cells via the heparin-binding site in repeat III13.
Mol. Biol. Cell
,
10
:
1521
-1536,  
1999
.
16
Redick S. D., Settles D. L., Briscoe G., Erickson H. P. Defining fibronectin’s cell adhesion synergy site by site-directed mutagenesis.
J. Cell Biol.
,
149
:
521
-527,  
2000
.
17
Sharma A., Askari J. A., Humphries M. J., Jones E. Y., Stuart D. I. Crystal structure of a heparin- and integrin-binding segment of human fibronectin.
EMBO J.
,
18
:
1468
-1479,  
1999
.
18
Liu W., Litwack E. D., Stanley M. J., Langford J. K., Lander A. D., Sanderson R. D. Heparan sulfate proteoglycans as adhesive and anti-invasive molecules. Syndecans and glypican have distinct functions.
J. Biol. Chem.
,
273
:
22825
-22832,  
1998
.
19
Klass C. M., Couchman J. R., Woods A. Control of extracellular matrix assembly by syndecan-2 proteoglycan.
J. Cell Sci.
,
113
:
493
-506,  
2000
.
20
Oh E. S., Woods A., Couchman J. R. Multimerization of the cytoplasmic domain of syndecan-4 is required for its ability to activate protein kinase C.
J. Biol. Chem.
,
272
:
11805
-11811,  
1997
.
21
Fischer D., Tucker R. P., Chiquet-Ehrismann R., Adams J. C. Cell-adhesive responses to tenascin-C splice variants involve formation of fascin microspikes.
Mol. Biol. Cell
,
8
:
2055
-2075,  
1997
.
22
Kainulainen, V., Wang, H. Schick, C., and Bernfield, M. Syndecans, heparan sulfate proteoglycans, maintain the proteolytic balance of acute wound fluids. J. Biol. Chem., 273: 11563–11569
23
Subramanian S. V., Fitzgerald M. L., Bernfield M. Regulated shedding of syndecan-1 and -4 ectodomains by thrombin and growth factor receptor activation.
J. Biol. Chem.
,
272
:
14713
-14720,  
1997
.
24
Ehrismann R., Chiquet M., Turner D. C. Mode of action of fibronectin in promoting chicken myoblast attachment. Mr = 60,000 gelatin-binding fragment binds native fibronectin.
J. Biol. Chem.
,
256
:
4056
-4062,  
1981
.
25
Fischer D., Chiquet-Ehrismann R., Bernasconi C., Chiquet M. A single heparin binding region within the fibrinogen-like domain is functional in chick tenascin-C.
J. Biol. Chem.
,
270
:
3378
-3384,  
1995
.
26
Fischer D., Brown-Ludi M., Schulthess T., Chiquet-Ehrismann R. Concerted action of tenascin-C domains in cell adhesion, anti-adhesion and promotion of neurite outgrowth.
J. Cell Sci.
,
110
:
1513
-1522,  
1997
.
27
Schreiner C. L., Bauer J. S., Danilov Y. N., Hussein S., Sczekan M. M., Juliano R. L. Isolation and characterization of Chinese hamster ovary cell variants deficient in the expression of fibronectin receptor.
J. Cell Biol.
,
109
:
3157
-3167,  
1989
.
28
Borsi L., Carnemolla B., Nicolo G., Spina B., Tanara G., Zardi L. Expression of different tenascin isoforms in normal, hyperplastic and neoplastic human breast tissues.
Int. J. Cancer
,
52
:
688
-692,  
1992
.
29
Bauer J. S., Schreiner C. L., Giancotti F. G., Ruoslahti E., Juliano R. L. Motility of fibronectin receptor-deficient cells on fibronectin and vitronectin: collaborative interactions among integrins.
J. Cell Biol.
,
116
:
477
-487,  
1992
.
30
Schreiner C., Fisher M., Hussein S., Juliano R. L. Increased tumorigenicity of fibronectin receptor deficient Chinese hamster ovary cell variants.
Cancer Res.
,
51
:
1738
-1740,  
1991
.
31
Giancotti F. G., Ruoslahti E. Elevated levels of the α 5 β1 fibronectin receptor suppress the transformed phenotype of Chinese hamster ovary cells.
Cell
,
60
:
849
-859,  
1990
.
32
Yoshida T., Matsumoto E., Hanamura N., Kalembeyi I., Katsuta K., Ishihara A., Sakakura T. Co-expression of tenascin and fibronectin in epithelial and stromal cells of benign lesions and ductal carcinomas in the human breast.
J. Pathol.
,
182
:
421
-428,  
1997
.
33
Pesheva P., Probstmeier R., Skubitz A. P., McCarthy J. B., Furcht L. T., Schachner M. Tenascin-R (J1 160/180) inhibits fibronectin-mediated cell adhesion—functional relatedness to tenascin-C.
J. Cell Sci.
,
107
:
2323
-2333,  
1994
.
34
Hauzenberger D., Olivier P., Gundersen D., Rüegg C. Tenascin-C inhibits β1 integrin-dependent T lymphocyte adhesion to fibronectin through the binding of its fnIII 1–5 repeats to fibronectin.
Eur. J. Immunol.
,
29
:
1435
-1447,  
1999
.
35
Hoffman S., Edelman G. M. A proteoglycan with HNK-1 antigenic determinants is a neuron-associated ligand for cytotactin.
Proc. Natl. Acad. Sci. USA
,
84
:
2523
-2527,  
1987
.
36
Chung C. Y., Zardi L., Erickson H. P. Binding of tenascin-C to soluble fibronectin and matrix fibrils.
J. Biol. Chem.
,
270
:
29012
-29017,  
1995
.
37
Ramos D. M., Chen B., Regezi J., Zardi L., Pytela R. Tenascin-C matrix assembly in oral squamous cell carcinoma.
Int. J. Cancer
,
75
:
680
-687,  
1998
.
38
Wenk M. B., Midwood K. S., Schwarzbauer J. E. Tenascin-C suppresses Rho activation.
J. Cell Biol.
,
150
:
913
-920,  
2000
.
39
Bernfield M., Kokenyesi R., Kato M., Hinkes M. T., Spring J., Gallo R. L., Lose E. J. Biology of the syndecans: a family of transmembrane heparan sulfate proteoglycans.
Annu. Rev. Cell Biol.
,
8
:
365
-393,  
1992
.
40
Rapraeger A. C., Ott V. L. Molecular interactions of the syndecan core proteins.
Curr. Opin. Cell Biol.
,
10
:
620
-628,  
1998
.
41
Tumova S., Woods A., Couchman J. R. Heparan sulfate chains from glypican and syndecans bind the Hep II domain of fibronectin similarly despite minor structural differences.
J. Biol. Chem.
,
275
:
9410
-9417,  
2000
.
42
Saoncella S., Echtermeyer F., Denhez F., Nowlen J. K., Mosher D. F., Robinson S. D., Hynes R. O., Goetinck P. F. Syndecan-4 signals cooperatively with integrins in a Rho-dependent manner in the assembly of focal adhesions and actin stress fibers.
Proc. Natl. Acad. Sci. USA
,
96
:
2805
-2810,  
1999
.
43
Akiyama S. K., Olden K., Yamada K. M. Fibronectin and integrins in invasion and metastasis.
Cancer Metastasis Rev.
,
14
:
173
-189,  
1995
.
44
Ruoslahti E. Integrins as signaling molecules and targets for tumor therapy.
Kidney Int.
,
51
:
1413
-1417,  
1997
.
45
Gong J., Ko T. C., Brattain M. G. Disruption of fibronectin binding to the α 5 β 1 integrin stimulates the expression of cyclin-dependent kinases and DNA synthesis through activation of extracellular signal-regulated kinase.
J. Biol. Chem.
,
273
:
1662
-1669,  
1998
.
46
Gong J., Wang D., Sun L., Zborowska E., Willson J. K., Brattain M. G. Role of α 5 β 1 integrin in determining malignant properties of colon carcinoma cells.
Cell Growth Differ.
,
8
:
83
-90,  
1997
.
47
Talts J. F., Wirl G., Dictor M., Muller W. J., Fassler R. Tenascin-C modulates tumor stroma and monocyte/macrophage recruitment but not tumor growth or metastasis in a mouse strain with spontaneous mammary cancer.
J. Cell Sci.
,
112
:
1855
-1864,  
1999
.
48
Tan M. I., Strunck E., Scholzen T., Gerdes J., Vollmer G. Extracellular matrix regulates steady-state mRNA levels of the proliferation associated protein Ki-67 in endometrial cancer cells.
Cancer Lett.
,
140
:
145
-152,  
1999
.
49
Jahkola T., Toivonen T., Nordling S., von Smitten K., Virtanen I. Expression of tenascin-C in intraductal carcinoma of human breast: relationship to invasion.
Eur. J. Cancer
,
34
:
1687
-1692,  
1998
.
50
Hazlehurst L. A., Damiano J. S., Buyuksal I., Pledger W. J., Dalton W. S. Adhesion to fibronectin via β1 integrins regulates p27Kip1 levels and contributes to cell adhesion mediated drug resistance (CAM-DR).
Oncogene
,
19
:
4319
-4327,  
2000
.
51
Dimitrijevic-Bussod M., Balzaretti-Maggi V. S., Gadbois D. M. Extracellular matrix and radiation G1 cell cycle arrest in human fibroblasts.
Cancer Res.
,
59
:
4843
-4847,  
1999
.
52
Woods A., McCarthy J. B., Furcht L. T., Couchman J. R. A synthetic peptide from the COOH-terminal heparin-binding domain of fibronectin promotes focal adhesion formation.
Mol. Biol. Cell
,
4
:
605
-613,  
1993
.
53
Alexander C. M., Reichsman F., Hinkes M. T., Lincecum J., Becker K. A., Cumberledge S., Bernfield M. Syndecan-1 is required for Wnt-1-induced mammary tumorigenesis in mice.
Nat. Genet.
,
25
:
329
-332,  
2000
.
54
Zvibel I., Brill S., Halpern Z., Moskovitz S., Yayon A., Papa M. Soluble and matrix-associated heparan sulfate proteoglycans increase expression of erb-B2 and erb-B3 in colon cancer cell lines.
Int. J. Cancer
,
91
:
316
-321,  
2001
.
55
Asakura S., Yang W., Sottile J., Zhang Q., Jin Y., Ohkubo I., Sasaki M., Matsuda M., Hirata H., Mosher D. F. Opposing effects of low and high molecular weight kininogens on cell adhesion.
J. Biochem.
,
124
:
473
-484,  
1998
.
56
Stoll R., Renner C., Zweckstetter M., Bruggert M., Ambrosius D., Palme S., Engh R. A., Golob M., Breibach I., Buettner R., Voelter W., Holak T. A., Bosserhoff A. K. The extracellular human melanoma inhibitory activity (MIA) protein adopts an SH3 domain-like fold.
EMBO J.
,
20
:
340
-349,  
2001
.