The Glycosaminoglycan-Binding Domain of Decoy Receptor 3 Is Essential for Induction of Monocyte Adhesion¹

Decoy receptor 3 (DcR3)³/TR6/M68 is a soluble receptor belonging to the TNFR superfamily, and functions as decoy receptor for Fas ligand (FasL), LIGHT (homologous to lymphotoxins shows inducible expression and competes with HSV glycoprotein D for herpes virus entry mediator, a receptor expressed by T lymphocytes), and TL1A (1, 2, 3) to neutralize their cytotoxic and regulatory functions. DcR3 is overexpressed in malignant tumors arising from lung, colon, glioma, and gastrointestinal track; virus-associated lymphoma (1, 4, 5, 6, 7); as well as the PBMC of silicosis patients (8). In addition, high serum levels of DcR3 have been detected in many cancer patients (9) and are associated with poor prognoses (6). This suggests that overexpression of DcR3 may produce some advantage for tumor growth and survival.

Growing evidence has shown that members of the TNFR superfamily are capable of inducing reverse signals after engagement with their surface receptors (10). Our recent studies suggested that DcR3 also can directly modulate the activities of many cell types. DcR3 can regulate dendritic cell differentiation and down-regulate several costimulatory molecules, leading to Th2 polarization (11). In addition, DcR3 induces actin reorganization and adhesion of monocytes and THP-1 cells (12) as well as reduces phagocytic activity and proinflammatory cytokine production in macrophages (13). Furthermore, osteoclast formation is promoted by addition of DcR3 to monocyte/macrophage lineage precursor cells (14). DcR3 also sensitizes some lymphoma cells to TRAIL-induced apoptosis (15). Finally, DcR3 increases monocyte adhesion to endothelial cells via NF-κB activation, leading to the transcriptional up-regulation of adhesion molecules and IL-8 in endothelial cells (16). All the evidence suggests that DcR3 is not only a decoy receptor, but is also an effector molecule able to modulate physiological and/or pathological functions. However, the pleiotropic effects of DcR3 are independent of FasL, LIGHT, and TL1A (11, 12, 13, 15). In this study we have investigated the underlying mechanism of DcR3-mediated modulatory effects.

Heparan sulfate proteoglycans (HSPGs) are surface and extracellular matrix macromolecules that comprise a core protein to which heparan sulfate glycosaminoglycan (GAG) chains are attached (17, 18). Cell surface HSPGs can be classified into several groups based on their core protein structure, including full-time HSPGs (e.g., syndecans and glypicans) and part-time HSPGs that may bear heparan sulfate chains in some proportion or under specific conditions (e.g., β-glycans and v3 exon-containing CD44 isoform). HSPGs have been demonstrated to regulate numerous biological functions, including organization of the cytoskeleton, migration, inflammation, metastasis, and synapse formation, via interacting with various effector molecules, including growth factors, extracellular matrix components, cell adhesion molecules, chemokines, cytokines, enzymes, and pathogens (17, 19). It has been proposed that cell surface HSPGs can modulate the actions of several ligands by acting as receptors directly or as coreceptors in concert with ligand-specific receptors (20). One of the most well-documented examples is the effect of surface HSPGs on the binding and biological activity of fibroblast growth factor 2 (21). In addition, syndecans have been reported to interact with extracellular matrix to cause actin rearrangement even in the absence of integrin occupancy (22). Among these HSPG-binding proteins, common heparin-binding motifs, such as BBXB, BBBXXB (B represents a basic residue), and a 20-Å spacing of basic residues, are found (23). Osteoprotegerin (OPG), a member of the TNFR superfamily, has been demonstrated to bind to surface HSPGs, thus affecting its protein turnover rate (24, 25). Within the TNFR superfamily, OPG and DcR3 share the highest level of amino acid sequence identity; therefore, we asked whether DcR3 also binds to surface HSPGs to exert its broad-spectrum modulatory effects.

In this study we report the identification of a domain comprising a cluster of basic residues in the C terminus of DcR3 and show that DcR3 mutants lacking GAG-binding ability were unable to induce cell adhesion. The recombinant DcR3_HBD.Fc fusion protein, comprising the GAG-binding motif of DcR3 and the Fc portion of human IgG1 (hIgG1), was found to mimic functions of DcR3.Fc, such as the activation of protein kinase C (PKC) and FAK and the induction of cell adhesion. Moreover, in CD44v3 and syndecan-2 knockdown THP-1 cells, the percentage of adhesive cells after DcR3.Fc treatment was significantly reduced compared with DcR3.Fc-treated wild-type THP-1 cells. This suggests that DcR3.Fc-induced cell adhesion involves cross-linking of proteoglycans on the monocyte surface.

Chinese hamster ovary K1 (CHO-K1), CHO-pgsB 618, CHO-pgsD 677, THP-1, and Jurkat cells were obtained from American Type Cell Collection. Heparin, chondroitin, dermatan sulfate (chondroitin sulfate B), and chondroitin sulfate C were obtained from Sigma-Aldrich, whereas heparan sulfate and keratan sulfate were purchased from Seikagaku. Heparinase (EC 4.2.2.7) and chondroitinase ABC (EC 4.2.2.4) were obtained from Sigma-Aldrich. FITC-conjugated mAb against CD44 (clone F10-44-2) and PE-conjugated goat anti-hIgG Abs were purchased from Serotec.

To determine which GAGs could compete for DcR3.Fc binding to THP-1 cells, cells were incubated with DcR3.Fc^mut, whose Fc portion is mutated to prevent binding to the FcR (13, 26). Six GAGs, including heparin, heparan sulfate, chondroitin, dermatan sulfate, chondroitin sulfate C, and keratan sulfate, were used in this competition assay. Briefly, 10⁶ cells were incubated with DcR3.Fc^mut (1 μg) in the absence or the presence of different concentrations of GAG ranging from 0.1–100 μg/ml at 4°C for 20 min. After washing, cells were incubated with PE-conjugated goat anti-hIgG Abs. The fluorescence intensity of stained cells was determined by flow cytometry using a FACSCalibur (BD Biosciences).

The purification procedure of CD14⁺ primary monocytes was described previously (13). THP-1 cells (10⁵/100 μl/well) and monocytes (2 × 10⁵/100 μl/well) were treated with hIgG1, DcR3.Fc, DcR3_HBD.Fc, or various DcR3 mutants for 24 and 1 h, respectively. Cells were then washed twice with PBS before fixation with methanol for 10 s, followed by addition of 0.1% (v/v) crystal violet in PBS for 5 min. After washing with double-distilled H₂O five times, acetic acid (33%) was added to solubilize the cell lysate. The dissolved lysates from adherent THP-1 cells and monocytes were measured using an ELISA plate reader at 570 nm. To test whether GAGs can inhibit DcR3.Fc-induced cell adhesion, THP-1 cells and monocytes were pretreated with GAGs for 10 min before addition of DcR3.Fc, followed by adhesion assay, as described above.

The deletion mutants DcR3(1–246).Fc and DcR3(1–255).Fc were generated by PCR to remove the C-terminal basic regions, i.e., residues 247–300 and 256–300, respectively. The PCR primers used to amplify cDNA fragments encoding DcR3(1–246) and DcR3(1–255) are as follows: DcR3(1–246): sense, 5′-GGAATTCAAGGACCATGAGGGCGCTG-3′, and antisense 5′-GAATTCTGGTGTCGGACCCCAGC-3′; and DcR3(1–255): sense, 5′-GGAATTCAAGGACCATGAGGGCGCTG-3′, and antisense 5′-GAATTCCAGCTGCAAGGCCGCG-3′. The amplified fragments were ligated in-frame with the Fc portion of hIgG1 in pcDNA3 to produce Fc fusion proteins. The DcR3(A²⁵⁶A²⁵⁸A²⁵⁹).Fc mutant was produced by overlapping extension PCR to replace the basic residues K²⁵⁶, R²⁵⁸, and R²⁵⁹ with alanines. The primers for mutagenesis are as follows, with the mutated bases underlined: sense, 5′-GCCTTGCAGCTGGCGCTGGCTGCGCGGCTCACG-3′, and antisense, 5′-CGTGAGCCGCGCAGCCAGCGCCAGCTGCAAGGC-3′. Another two primers were designed to introduce the KpnI and SacI sites at the 5′ and 3′ ends, respectively, for subcloning: sense, 5′-ACCAGGGTACCAGGAGCT-3′; and antisense, 5′-CCCCAGGAGCTCCGTG-3′. The amplified cDNA fragment was ligated into the pCR2.1 vector (Invitrogen Life Technologies), and the fidelity of the sequence was verified by autosequencing (MB Mission Biotech). The KpnI/SacI fragment was then used to replace the wild-type DcR3 in pcDNA3.

To generate the DcR3_HBD.Fc fusion gene, complementary oligonucleotides (5′-AAGCTTGGGCTGAAGCTGCGTCGGCGGCTCGGGAAGCTT-3′ and 5′-AAGCTTCCCGAGCCGCCGACGCAGCTTCAGCCCAAGCTT-3′), corresponding to the GAG binding motif of DcR3, were annealed to generate a dsDNA fragment, followed by phosphorylation of the 5′-hydroxyl terminus by T4 polynucleotide kinase (New England Biolabs). The phosphorylated DNA fragment was subcloned into HincII-cut pBluescript II KS, followed by digestion with HindIII, then ligated in-frame with HindIII-cut pFlag-CMV1 (Sigma-Aldrich). The Flag-tagged DcR3_HBD was amplified by PCR and ligated in-frame with the Fc portion of hIgG1 in pcDNA3. The fidelity of the sequence was verified by autosequencing. The primers for amplification of Flag-tagged DcR3_HBD are: sense, 5′-GGATTCATTGATCTACCATGTCTG-3′, and antisense, 5-GAATTC CGCGGCCGCAAG-3′.

DcR3.Fc and its derived mutants, DcR3(1–246).Fc, DcR3(1–255).Fc, and DcR3(A²⁵⁶A²⁵⁸A²⁵⁹).Fc, as well as DcR3_HBD.Fc fusion proteins were produced using the FreeStyle 293 expression system (Invitrogen Life Technologies). Briefly, plasmids were transfected into 293-F cells according to the manufacturer’s instructions. Culture media collected from the transfected cells were incubated with protein A-Sepharose beads (Amersham Biosciences), and the bound proteins were then eluted with 0.1 M glycine buffer (pH 3.0), followed by dialysis against PBS.

To remove cell surface GAGs, 10⁶ CHO-K1 cells were detached from their culture vessels using PBS containing 2 mM EDTA and then incubated with 10 U/ml Flavobacterium heparinum heparinase I or 10 U/ml Proteus vulgaris chondroitinase ABC in PBS containing 0.1 mg/ml BSA for 2 h at 37°C. For trypsin treatment, cells were washed with PBS containing 2 mM EDTA and incubated for 4 min at 37°C. Control cultures were incubated for the same period in the same reaction buffer, but without enzyme addition. At the end of the incubation, the cells were washed twice with cold PBS, then assayed for DcR3.Fc binding as described above.

PKC activity was monitored using the PKC activity assay kit (Upstate Biotechnology) according to the manufacturer’s instructions. Briefly, THP-1 cells were washed twice in ice-cold PBS and incubated with lysis buffer (20 mM Tris-HCl (pH 7.5), 0.5 mM EGTA, 2 mM EDTA, 2 mM DTT, and 0.5 mM PMSF) and protease inhibitor mixture (Roche). Assays were then performed at 30°C in a total volume of 60 μl containing a specific substrate peptide for PKC and an inhibitor mixture that blocks the activity of PKA and calmodulin kinase. Reactions were initiated by addition of [γ-³²P]ATP and were terminated by transfer of the reaction mixture onto P81 phosphocellulose paper after 10-min incubation. Incorporated ³²P was quantified using a scintillation counter (Beckman Coulter).

The algorithm used to predict sequences that would lead to silencing of target genes, targeting vector construction, and lentivirus production as described previously (27). Briefly, the oligonucleotides 5′-GGATCAGGCATTGATGATG-3′ and 5′-GCTTCAGGAGTGTATCCTA-3′ were selected to knock down CD44v3 and syndecan-2, respectively. These complementary dsDNA fragments were subcloned into HpaI/XhoI-digested pLL3.7 (gift from Dr. L. Van Parijs, Center for Cancer Research, Massachusetts Institute of Technology, Boston, MA). The resulting targeting vector was cotransfected with packaging vectors into 293T cells, and the resulting supernatants were collected after 48 h. Viruses were recovered after ultracentrifugation for 1.5 h at 25,000 rpm in a Beckman SW28 rotor and were resuspended in PBS (500 μl). Titers were determined by infecting 293T cells with serial dilutions of concentrated lentiviruses.

Values are expressed as the mean ± SD of at least three experiments. Student’s t test from the PRISM software package (GraphPad) was used to assess the statistical significance of the differences, with p < 0.05 considered statistically significant.

In our previous study we found that DcR3.Fc activates PKC to induce actin reorganization and enhance cell adhesion (12, 16) independently of its three ligands: FasL, LIGHT, and TL1A. To explore the nature of the DcR3-binding partner responsible for this phenomenon, different GAGs were added during the staining procedure to test their abilities to inhibit DcR3.Fc binding to THP-1 cells. To prevent DcR3.Fc binding through FcRs on THP-1 cells, DcR3.Fc^mut was used in place of DcR3.Fc for staining. We found that both heparin and heparan sulfate significantly diminished the percentage of DcR3.Fc^mut-bound cells in a dose-dependent manner, whereas dermatan sulfate only had an inhibitory effect at the highest concentration tested (100 μg/ml). In contrast, neither chondroitin nor chondroitin sulfate C was able to inhibit DcR3.Fc^mut binding to THP-1 cells (Fig. 1,A). This suggests that binding of DcR3.Fc^mut to THP-1 cells might be via an interaction with specific GAGs on the cell surface. We next examined whether heparin or heparan sulfate could block DcR3.Fc-induced THP-1 cell adhesion. We found that the GAGs themselves do not induce THP-1 cell adhesion (data not shown), whereas DcR3.Fc-triggered cell adhesion was abolished by heparin and was reduced significantly by heparan sulfate (Fig. 1,B). Similar observations were made for CD14⁺ primary monocytes (Fig. 1 C). These results suggest that DcR3 may interact with a heparin/heparan sulfate-conjugated molecule to trigger monocyte/THP-1 cell adhesion.

To confirm whether cell surface GAGs account for DcR3.Fc binding, CHO-K1 cells and CHO-K1-derived mutants (CHO-pgsB 618 and CHO-pgsD 677) defective in proteoglycan synthesis were stained with DcR3.Fc. CHO-pgsB 618 lacks galactosyltransferase and is defective in the synthesis of GAGs, whereas CHO-pgsD 677 lacks both N-acetylglucosaminyl- and glucuronosyl-transferase activities and loses the ability to synthesize heparan sulfate (28, 29). In contrast to parental CHO-K1 cells, DcR3.Fc showed a complete lack of binding to CHO-pgsB 618 cells and only minimal binding to CHO-pgsD 677 (Fig. 2,A). To confirm whether GAGs are responsible for DcR3.Fc binding to the cell surface, CHO-K1 cells were treated with enzymes that selectively degrade either heparan sulfate or chondroitin sulfate. After heparinase treatment, which specifically removes heparan sulfate, the binding of DcR3.Fc to CHO-K1 cells was dramatically reduced, whereas treatment with chondroitinase ABC did not affect DcR3.Fc binding (Fig. 2,B). In addition, treatment of CHO-K1 cells with trypsin, which cleaves the extracellular domain of membrane proteins, also prevented detection of CHO-K1-bound DcR3.Fc (Fig. 2,B). These observations are consistent with the hypothesis that DcR3 binding to CHO-K1 cells is via interaction with GAGs on the cell surface (Fig. 2 B).

The ability of DcR3.Fc to interact with GAGs was more directly demonstrated using Jurkat cells expressing various CD44 isoforms, CD44H, CD44_{v3, 8–10}, and CD44_{v3, 8–10ΔHS}. It has been reported that the v3 exon of CD44_{v3, 8–10} contains a motif, SGSG, which directs O-glycosylation with heparan sulfate, whereas a point mutation in CD44_{v3, 8–10ΔHS} abolishes the conjugation of heparan sulfate (30). Neither anti-CD44 mAb (x-axis) nor DcR3.Fc (y-axis) could bind to pRcCMV-transfected Jurkat cells (Fig. 2,C, upper left panel). Jurkat cells expressing CD44_{v3, 8–10} gained robust and specific binding of DcR3.Fc (Fig. 2,C, lower left panel). In contrast, DcR3.Fc lost the ability to bind to Jurkat cells expressing CD44 (Fig. 2,C, upper right panel) or CD44_{v3, 8–10ΔHS} (lower right panel). Moreover, the binding of DcR3.Fc to CD44_{v3, 8–10} was inhibited by heparin, but not by chondroitin (Fig. 2 C, second row, lower left). Taken together, our data strongly suggest that the DcR3-interacting partners are sulfated GAG side chains, rather than the polypeptide portions of cell surface proteoglycans.

Many growth factors and chemokines have been shown to interact with HSPGs through positively charged sites in a C-terminal α-helix (18, 23, 31). We searched for a cluster of residues with a net positive charge that fulfills the criteria to form a HSPG-binding site in DcR3. A putative HSPG-binding motif, K²⁵⁶L²⁵⁷R²⁵⁸R²⁵⁹, which resembles the proposed consensus sequence for heparin binding. BXBB (B for basic and X for any amino acid), was found (Fig. 3,A). To identify the minimal domain/amino acid residues essential for GAG binding, deletion mutants DcR3(1–246).Fc and DcR3(1–255).Fc were generated, and their ability to bind GAGs was tested (Fig. 3, B and C). As shown in Fig. 3,D, neither DcR3(1–246).Fc nor DcR3(1–255).Fc could bind to CHO-K1 cells. This observation indicated that the GAG-binding domain is located within residues 256–300 at the C terminus of DcR3, which includes the putative GAG-binding motif, K²⁵⁶L²⁵⁷R²⁵⁸R²⁵⁹. To further map the positively charged amino acid residues essential for GAG binding, a DcR3.Fc triple mutant, DcR3(A²⁵⁶A²⁵⁸A²⁵⁹).Fc, was generated by site-directed mutagenesis (Fig. 3,B). In contrast to wild-type DcR3.Fc (Fig. 3,D), DcR3(A²⁵⁶A²⁵⁸A²⁵⁹).Fc was incapable of binding to CHO-K1 cells or to CD44_{v3, 8–10}-expressed Jurkat cells (Fig. 3,E). This suggests that the motif K²⁵⁶L²⁵⁷R²⁵⁸R²⁵⁹ of DcR3 is essential for DcR3 to bind GAGs. To rule out the possibility that the loss of GAG-binding ability resulted from a conformational destruction in the DcR3 deletion mutants, the GAG-binding motif was fused with the Fc fragment of hIgG1 (DcR3.HBD.Fc) to test its modulatory effects (Fig. 3 B). We found that DcR3_HBD.Fc has the same binding ability as DcR3.Fc, and its binding to the cell surface can be inhibited by heparin/heparan sulfate, but not chondroitin/chondroitin sulfate (data not shown). This observation provides direct evidence that the heparin-binding domain of DcR3 is sufficient to bind surface HSPGs.

An adhesion assay was used to test whether the loss of GAG-binding ability of DcR3 mutants correlated with their ability to induce THP-1 cell and monocyte adhesion. As shown in Fig. 4,A, DcR3.Fc induced THP-1 cell adhesion in a dose-dependent manner, whereas neither the deletion mutants, DcR3(1–246).Fc and DcR3(1–255).Fc, nor the point mutation mutant, DcR3(A²⁵⁶A²⁵⁸A²⁵⁹).Fc, could induce THP-1 cell adhesion under the same conditions. The importance of the GAG-binding site of DcR3.Fc was also investigated by addition of DcR3^252–261 peptide comprising aa residues 252–261 of DcR3. Even though DcR3^252–261 peptide itself did not induce THP-1 cell adhesion, it could compete with DcR3 binding to THP-1 cells as well as to reduce THP-1 cell adhesion in a dose-dependent manner (Fig. 4 B). This suggests that DcR3.Fc-triggered cell adhesion is via interaction with surface GAGs.

To test whether the GAG-binding domain of DcR3 is entirely responsible for the DcR3.Fc-mediated effect, we compared the abilities of DcR3.Fc and DcR3_HBD.Fc to trigger THP-1 adhesion. When THP-1 cells were treated with DcR3_HBD.Fc, cells exhibited a dose-dependent adhesive response like that of DcR3.Fc (Fig. 5,A). Similarly, DcR3_HBD.Fc was able to induce monocyte adhesion, whereas none of the DcR3.Fc mutants could induce cell adhesion under equivalent conditions (Fig. 5,B). Because DcR3.Fc-induced THP-1 adhesion occurs via activation of PKC and FAK (12), we examined whether DcR3_HBD.Fc can also mediate this effect. As shown in Fig. 5,C, both DcR3.Fc and DcR3_HBD, but not the DcR3.Fc mutants, can activate PKC activity. Moreover, levels of the tyrosine-phosphorylated form of FAK increased after DcR3_HBD.Fc treatment for 10 min (Fig. 5 D). Therefore, we concluded that DcR3_HBD.Fc, like DcR3.Fc, is able to activate PKC and FAK to induce cell adhesion.

To identify the candidate HSPGs responsible for DcR3.Fc-induced cell adhesion, we examined the levels of HSPGs that are abundantly expressed on immune cells, including CD44v3 and syndecans 1 and 2 (32, 33). Immunofluorescence analysis of human primary monocytes and THP-1 cells revealed the presence of significant levels of syndecan-2; CD44v3 is expressed in monocytes, but only a small amount in THP-1 cells (Fig. 6,A). To confirm the roles of these HSPGs in DcR3.Fc-induced cell adhesion, THP-1 cells were infected with lentiviruses expressing shRNAs against CD44v3, and syndecan-2, respectively. Taking advantage of the presence of enhanced GFP in the lentivirus-based vector (pLL3.7), the infected cells were sorted for cell adhesion analysis (27). Infection of THP-1 cells with shRNA-containing lentiviruses down-regulated the mRNA expression of HSPGs for both CD44v3 (95% reduction) and syndecan-2 (90% reduction; Fig. 6,B), and we examined the effects of these changes on DcR3-induced THP-1 cell adhesion. As shown in Fig. 6 C, adhesion of THP-1 cells upon DcR3.Fc stimulation was significantly reduced after the knockdown of syndecan-2 and CD44v3 compared with that of mock-infected cells. These data are consistent with surface HSPGs being responsible for DcR3.Fc-induced cell adhesion.

In the present study we found that DcR3 is unique among TNFR superfamily members in terms of the ability of DcR3.Fc to trigger THP-1 cell adhesion. The other TNFRs (represented by Fas.Fc, LTβR.Fc, DR3.Fc, TNFR1.Fc, DR4.Fc, and RANK.Fc) are unable to mediate this process. Moreover, this unique feature of DcR3.Fc is unrelated to its interaction with FasL, LIGHT, or TL1A (11, 12, 13). In this study we demonstrated that DcR3 can bind to HSPGs via a cluster of positively charged amino acids at its C terminus. In addition, DcR3 mutants lacking GAG-binding activity lost their ability to induce monocyte adhesion (Fig. 4) and were unable to activate downstream signaling pathways, such as those involving PKC and FAK (Fig. 5). This suggests that DcR3.Fc-induced monocyte adhesion might occur via cross-linking of cell surface HSPGs.

To date, only three members of the TNF and TNFR superfamilies have been reported to exhibit HSPG-binding ability. TNF-α has been demonstrated to interact with biglycan and decorin, which are small extracellular matrix proteoglycans substituted with chondroitin sulfate and dermatan sulfate, respectively (34). A proliferation-inducing ligand (APRIL) is able to bind to proteoglycan-positive cells and induce APRIL oligomerization, resulting in the triggering of transmembrane activator and calcium-modulating and cyclophilin ligand interactor- and/or B cell maturation Ag-mediated B cell activation (35). Moreover, OPG is able to bind to surface HSPGs via a highly positively charged, heparin-binding domain in its C terminus (24). This HSPG-binding ability accounts for the internalization and degradation of OPG by multiple myeloma cells and may explain the biological mechanism for the bone destruction that is commonly associated with multiple myeloma (25, 36). Based on the high sequence similarity between DcR3 and OPG and the presence of HSPG-binding domains in their C termini, we asked whether OPG.Fc can also induce THP-1 cell adhesion. We found that OPG.Fc was as potent as DcR3.Fc in its ability to trigger THP-1 cell adhesion, whereas a mutant lacking the HSPG-binding domain, OPG.Fc(Δ352–401), was unable to mediate this effect (data not shown). This supports the argument that cross-linking of HSPGs can trigger signaling cascades responsible for both DcR3.Fc- and OPG.Fc-induced THP-1 cell adhesion. Members of TNF have been shown to induce reverse signaling after engagement with its cognate receptors. DcR3 has been shown to induce reverse signaling via LIGHT to inhibit T cell chemotaxis (37). However, we demonstrated that DcR3.Fc modulates the activities of monocytes and THP-1 cells via triggering surface HSPGs and is independent of its three known ligands (11, 12, 13). Based on all available information, DcR3.Fc can execute its biological functions via at least three aspects: 1) to neutralize TL1A (3, 38), FasL (1), and LIGHT (2); 2) to induce reverse signaling via LIGHT to inhibit T cell chemotaxis (37); and 3) to activate monocytes via cross-linking HSPGs, as shown in this report.

Protein-GAG binding is believed to be dominated by electrostatic interactions between the basic residues of protein and the sulfate and carboxylate moieties of GAGs. In experiments using cells treated with chondroitinase ABC or heparinase (Fig. 2,B), DcR3.Fc exhibited a marked preference for binding to heparin/heparan sulfate, and additional analysis showed that this activity of DcR3 was critical for the induction of cell adhesion (Figs. 2–5). Therefore, we set out to identify the cell surface HSPGs responsible for DcR3.Fc-induced cell adhesion. The best-characterized cell surface HSPGs fall into three groups: syndecans, glypicans, and β-glycan family. Haman et al. (39) have reported that there are no glypicans expressed on monocyte-derived macrophages, and the expression of CD44v3 isoform and syndecan-2 could be detected on monocytes and THP-1 cells; thereafter, we used RNA interference to specifically knock down the expression of CD44v3 and syndecan-2 in THP-1 cells. DcR3.Fc-stimulated adhesion of THP-1 cells was significantly reduced after the knockdown of CD44v3 and syndecan-2. Syndecan-2 has been shown to participate in the regulation of cell adhesion and cytoskeleton rearrangement in several cell lines (40, 41, 42). Although little is known about syndecan-2 signaling pathways, some reports have demonstrated that the cytoplasmic regions of syndecan-2 are responsible for diverse interactions with the cytoskeleton and signaling molecules. For example, the cytoplasmic domain of syndecan-2 can be phosphorylated (43, 44), and the intracellular adaptor, calcium/calmodulin-dependent serine protein kinase, can bind to band 4.1 protein, thereby linking syndecan-2 to the actin cytoskeleton (45), syntenin (46), and ezrin (47). All these proteins might be implicated in the augmentation of cell adhesion and spreading. More and more evidence is accumulating to suggest that HSPGs may directly mediate signal transduction via cytoplasmic kinases and adaptor proteins as well as functioning as coreceptors to facilitate the formation of active ternary ligand-receptor complexes to generate the ligand’s physiological function. HSPGs are also reported to bind cell adhesion molecules (e.g., L-selectin, integrin α_m chain, and platelet endothelial cell adhesion molecule-1) (17); therefore, DcR3-induced signaling cascades might function via activating HSPG directly or via a yet undefined receptors associated with HSPG on the cell surface.

The interaction between soluble molecules and proteoglycans has important functional consequences in many biological systems (17, 31, 48). GAGs can immobilize and concentrate ligands close to the signaling receptors at the cell surface or promote ligand oligomerization. In this context, the profound modulatory effects of DcR3.Fc we have observed in other assay systems might also occur via interactions with cell surface HSPGs (11, 12, 13, 14, 15, 16). Therefore, it would be very interesting to test whether the effects of DcR3.Fc on the activation and differentiation of dendritic cells (12), macrophages (14), osteoclasts (15), and TRAIL-induced cell death (16) also occur via cross-linking of cell surface HSPGs.

The authors have no financial conflict of interest.