Macrophages produce an array of proinflammatory mediators at sites of inflammation and contribute to the development of inflammatory responses. Important roles for cytokines, such as IL-1 or TNF-α, and bacterial products, such as LPS, in this process have been well documented; however, the role for the extracellular matrix proteins, such as collagen, remains unclear. We previously reported that discoidin domain receptor 1 (DDR1), a nonintegrin collagen receptor, is expressed during differentiation of human monocytes into macrophages, and the interaction of the DDR1b isoform with collagen facilitates their differentiation via the p38 mitogen-activated protein kinase (MAPK) pathway. In this study, we report that the interaction of DDR1b with collagen up-regulates the production of IL-8, macrophage inflammatory protein-1α, and monocyte chemoattractant protein-1 in human macrophages in a p38 MAPK- and NF-κB-dependent manner. p38 MAPK was critical for DDR1b-mediated, increased NF-κB trans-activity, but not for IκB degradation or NF-κB nuclear translocation, suggesting a role for p38 MAPK in the modification of NF-κB. DDR1b-mediated IκB degradation was mediated through the recruitment of the adaptor protein Shc to the LXNPXY motif of the receptor and the downstream TNFR-associated factor 6/NF-κB activator 1 signaling cascade. Taken together, our study has identified NF-κB as a novel target of DDR1b signaling and provided a novel mechanism by which tissue-infiltrating macrophages produce large amounts of chemokines during the development of inflammatory diseases. Intervention of DDR1b signaling may be useful to control inflammatory diseases in which these proteins play an important role.
Macrophages play an important role in the development of inflammatory responses by secreting an array of cytokines and chemokines in a tissue microenvironment. Proinflammatory cytokines, such as IL-1 and TNF-α, are potent activators of macrophages and up-regulate the expression and production of cytokines/chemokines. At sites of inflammation, they also interact with the components of the extracellular matrix (ECM)3 through receptors, such as integrins, and are activated for increased release or production of cytokines/chemokines. For instance, activation of monocytes with collagen, the most abundant protein in the ECM, induced IL-1 release (1). Adherence of PBMC to type I collagen also enhanced the production of IL-8 in response to LPS, IL-1β, or TNF-α (2). Collagen-induced IL-1 release was only partially inhibited by an Ab against α2β1 integrin, a classic cell-surface collagen receptor (1), suggesting the presence of an as yet unidentified receptor(s) involved in collagen-induced cytokine/chemokine production by macrophages.
Discoidin domain receptor 1 (DDR1) is a nonintegrin collagen receptor (receptor tyrosine kinase) with a unique extracellular domain homologous to discoidin 1 of Dictyostelium discoideum (3). DDR1 is constitutively expressed in epithelial cells of normal tissues, such as lung, kidney, colon, and brain, and also in tumor cells of epithelial origins, such as mammary, ovarian, and lung carcinomas (3). Five DDR1 isoforms (a, b, c, d, and e) can be generated by alternative splicing of the DDR1 gene (3, 4). Disruption of the DDR1 gene in mice resulted in viable animals that were significantly smaller than their littermates, and female DDR1-null mice showed defects in blastocyst implantation and mammary gland development (5). Primary vascular smooth muscle cells isolated from the DDR1-null mice showed decreased proliferation, collagen attachment, and migration in vitro (6, 7). In contrast, primary mesangial cells isolated from the kidney of the DDR1-null mice showed enhanced proliferation (8). These previous observations have indicated that DDR1 plays a role in cell attachment, migration and proliferation; however, the underlying molecular mechanisms of DDR1 activation remain unclear.
We previously reported that the expression of two DDR1 isoforms, DDR1a and DDR1b, could be induced in human leukocytes, including neutrophils, monocytes, and lymphocytes in vitro. In vivo, tissue-infiltrating mononuclear cells, especially macrophages, were positive for DDR1 mRNA (9). The DDR1a and DDR1b isoforms differ from each other by an in-frame insertion of 111-bp coding for an additional 37-aa peptide in the proline-rich juxtamembrane region. The 37-aa insertion in DDR1b contains the LXNPXY motif corresponding to the consensus-binding motif of the Shc phosphotyrosine binding domain (3). To study the biological role for DDR1 expressed in leukocytes, we produced DDR1-overexpressing cell lines by transducing the human monocytic THP-1 cells with retroviruses expressing either DDR1a or DDR1b. Interestingly, overexpression of DDR1a, but not DDR1b, promoted the migration of THP-1 cells through three-dimensional collagen lattices (9).
THP-1 cells are known to differentiate along the monocytic lineage following exposure to PMA (10), and to acquire characteristics of macrophages, including loss of proliferation and expression of the MHC class II molecule, HLA-DR (11, 12). Thus, PMA-treated THP-1 cells provided a useful model to study the mechanisms of macrophage differentiation. Using this model, we previously demonstrated that collagen activation of DDR1b, but not DDR1a, facilitates PMA-induced differentiation of THP-1 cells via activation of the p38 mitogen-activated protein kinase (MAPK) pathway. Activation of neither DDR1a nor DDR1b with collagen affected THP-1 cell differentiation (13). We further identified that activation of DDR1b induces the formation of a protein complex containing TNFR-associated factor 6 (TRAF6), TGF-β-activated protein kinase 1 (TAK1)-binding protein 1β (TAB1β), and p38α MAPK, and subsequent autophosphorylation of p38α MAPK in PMA-treated THP-1 cells (14), indicating that the DDR1b signaling targets the p38 MAPK pathway through an alternative mechanism for p38 MAPK activation in these cells (15, 16). We confirmed our data with PMA-treated THP-1 cells using GM-CSF-induced, human monocyte-derived primary macrophages (GM macrophages) (13, 14). It is well known that TRAF6 mediates the signaling of other receptors, including IL-1R, Toll-like receptors (TLRs) TRANCE-R, and CD40 (17, 18). Thus, DDR1b is a new member of the receptor family that uses TRAF6 as a critical signaling molecule to transduce signals to its downstream pathways.
The p38 MAPK pathway plays an important role in the expression of several cytokine and chemokine genes, including IL-1β, IL-8, and MCP-1 (19, 20). Our previous finding that the interaction of DDR1b with collagen activates p38 MAPK in PMA-induced differentiated THP-1 cells, a model for macrophages, and in monocyte-derived human primary macrophages (13), led us to the hypothesis that activation of DDR1b also up-regulates the expression and release of cytokines and chemokines in macrophages. In the present study, we tested our hypothesis using PMA-induced, differentiated DDR1b-overexpressing THP-1 cells and GM macrophages. Collagen-activation of DDR1b markedly up-regulated the expression and release of IL-1β, IL-8, macrophage inflammatory protein-1α (MIP-1α), and monocyte chemoattractant protein-1 (MCP-1) in a p38 MAPK-dependent manner. Up-regulation of IL-8, MIP-1α, and MCP-1 release was also dependent on NF-κB, indicating that NF-κB is a novel target of DDR1b signaling. Interestingly, IκB degradation induced by DDR1b activation was regulated through a unique signaling cascade involving the adaptor protein Shc, TRAF6, and NF-κB activator 1 (Act1). Thus, DDR1b-collagen interaction up-regulates the production of chemokines by macrophages and it is likely to contribute to the development of inflammatory responses in a tissue microenvironment.
Materials and Methods
A mouse monoclonal anti-β1-integrin blocking Ab (DE9) was obtained from Upstate Biotechnology (Lake Placid, NY). Rabbit polyclonal Abs against human DDR1 (C-20) and Act1 (H-300), and mouse mAbs against human TRAF6 (D-10 and C-20) and IκBα were from Santa Cruz Biotechnology (Santa Cruz, CA). The production of the agonistic anti-DDR1 mouse mAb 513 (IgM) was previously described (13). Rabbit polyclonal Abs against p38 MAPK, phosphorylated p38 MAPK, or p38α MAPK were from Cell Signaling Technology (Beverly, MA). Sheep anti-mouse or anti-rabbit IgGs coupled with HRP were from Amersham Pharmacia Biotech (Piscataway, NJ). An anti-human IL-1β neutralizing Ab was from R&D Systems (Minneapolis, MN). GM-CSF was from PeproTech (Rocky Hill, NJ). PBS, RPMI 1640, G418, recombinant protein G-agarose and TRIzol reagent were from Invitrogen (Gaithersburg, MD). Protease inhibitor mixture tablets, Complete mini, were from Roche (Indianapolis, IN). FCS was from HyClone Laboratories (Logan, UT). Paraformaldehyde, formamide, PMA, anti-Flag Ab, and bovine collagen type I solution were from Sigma-Aldrich (St. Louis, MO). SB203580, PD98059, and caffeic acid phenethyl ester (CAPE) were from Biochem-Novabiochem (San Diego, CA). The NF-κB-Luc plasmid containing five copies of the NF-κB site of the human Ig κ-chain gene and the luciferase gene was from Stratagene (La Jolla, CA). Accu-prep was from Accurate Chemical & Scientific (Westbury, NY). [α-32P]dCTP and [γ-32P]ATP were from Amersham Biosciences (Piscataway, NJ). Human β-actin cDNA was from Clontech Laboratories (Palo Alto, CA).
Mock-transduced, DDR1a-, or DDR1b-overexpressing THP-1 cells
The production of DDR1a- or DDR1b-overexpressing THP-1 cells was previously described (9). Briefly, the human monocytic leukemic cell line, THP-1, was transduced with retrovirus expressing DDR1a or DDR1b, or control retrovirus containing the vector only (mock-transduced). THP-1 cells transduced by each virus were selected with G418, and cloned by limiting dilution. The cloned THP-1 cells were maintained in RPMI 1640 medium supplemented with 100 μg/ml gentamycin and 10% FCS (complete medium).
Preparation of monocyte-derived macrophages
Human PBMC were obtained from leukapheresis preparations obtained by the Blood Bank, Clinical Center, National Institutes of Health (Bethesda, MD). The leukocyte-rich preparation was overlaid on Accu-prep in 50-ml tubes and the tubes were centrifuged at 800 × g for 20 min at room temperature. PBMC fractions were collected, washed once with PBS at room temperature and twice with complete medium at 4°C, and resuspended in the complete medium. Monocytes were further purified by using iso-osmotic Percoll gradient. At this stage, the purity of monocytes was higher than 90% (21). The cells (5 × 105/ml) were allowed to adhere to the surface of six-well plates. Nonadherent cells were removed after a 5-h incubation at 37°C, and remaining adherent cells were cultured in the presence of GM-CSF (50 ng/ml) for an additional 5 days to produce monocyte-derived macrophages (GM macrophages).
THP-1 cells were cultured for various times at the cell density of 5 × 106 cells/ml in complete medium in the presence of 10 nM PMA in six-well tissue culture plates that were either collagen-coated or uncoated. Total RNA was extracted from each culture by using TRIzol reagent, and Northern blotting was performed as previously described (22). The cloning of human MCP-1 cDNA (23) and the sources of IL-8 and MIP-1α cDNAs (24) were previously described. IL-1β cDNA was obtained by RT-PCR from a human monocyte cDNA library using a primer pair, 5′-CGTATGGCAGGACAAATGCTTCTTC-3′ and 5′-TTCCCTCCAGGCTGCCATGAG-3′. Each cDNA was labeled with [α-32P]dCTP using the Rediprime II random prime labeling system (Amersham Biosciences).
The concentrations of IL-1β, IL-8, MIP-1α, and MCP-1 were measured in the Lymphokine Testing Laboratory, Clinical Services Program, Science Applications International Corporation-Frederick, National Cancer Institute-Frederick (Frederick, MD), by using ELISA kits (R&D Systems) specific for each human cytokine and chemokine. The sensitivity of the assay was 10.0 (IL-8), 10.0 (MIP-1α), 1.0 (IL-1β), and 5.0 pg/ml (MCP-1), respectively.
Cell lysates were prepared from 1 × 107 cells using a lysis buffer containing 50 mM HEPES, 150 mM NaCl, 1% Triton X-100, 10% glycerol, and a mixture of protease inhibitors. Each sample was mixed with double-strength sample buffer and boiled for 10 min. Electrophoresis was performed on 10% polyacrylamide gels by SDS-PAGE, and proteins were transferred electrophoretically to nitrocellulose membranes at 150 mA for 1 h in a semidry system. Membranes were incubated with an Ab specific to each protein, followed by sheep anti-rabbit or anti-mouse IgG coupled with HRP. Peroxidase activity was visualized by the Enhanced Chemiluminescence Detection System (Amersham Biosciences). To study the degradation of IκBα, mock-transduced or DDR1b-overexpressing THP-1 cells were incubated with PMA for 12 h, serum-starved for 12 h, and then activated with 50 μg/ml collagen for various times. To evaluate the association of Act1 with TRAF6, cell lysates were subjected to immunoprecipitation with anti-TRAF6 or anti-Act1Ab, and coimmunoprecipitation of Act1 or TRAF6 was examined with anti-Act1 or anti-TRAF6 Ab.
Expression of Y513F DDR1b mutant and dominant negative (DN)-TRAF6
Production of the Y513F DDR1b mutant expression vector was previously described (13). The numbering for the tyrosine residue is based on the amino acid sequence available from GenBank (accession number NP_054699). Parental THP-1 cells were transfected with either wild-type DDR1b or Y513F mutant using Effectene Transfection Reagent (Qiagen, Valencia, CA) for 12 h, rinsed with PBS, and incubated for an additional 12 h in complete medium. The transfected cells were treated with 10 nM PMA for 12 h, starved in RPMI 1640 for 10 h, rinsed three times with PBS, and subsequently activated with 50 μg/ml collagen for 3 h. One hundred million cells were lysed and the degradation of IκBα was examined by Western blotting.
Construction of a mammalian expression vector for the Flag-DN-TRAF6 (289–522) fusion protein (25) was previously described (14). DDR1b-overexpressing THP-1 cells were treated with 10 nM PMA for 12 h and then transfected with the vector with or without insert, using Effectene Transfection Reagent (Qiagen) for 24 h, rinsed with PBS, and incubated for an additional 12 h in RPMI 1640 containing 1% FCS, and subsequently activated with 50 μg/ml collagen for 60 min or 3 h. One hundred million cells were lysed and the cell lysates were subjected to Western blotting.
Mock-transduced or DDR1b-overexpressing THP-1 cells were incubated with 10 nM PMA on collagen-coated plates for 6 h, nuclear extracts were prepared as previously described (26), and aliquots were frozen at −80°C. For EMSA, end-labeled 32P-oligonucleotide probes corresponding to the NF-κB binding site of the Ig κ-chain gene (5′-AGTTGAGGGGACTTTCCCAGGC-3′) was incubated with 5 μg of nuclear extracts in a 20-μl binding mixture (50 mM Tris-HCl, pH 7.4, 25 mM MgCl2, 0.5 mM DTT, 50% glycerol) at 4°C for 15 min. The DNA-protein complexes were resolved on a 5% polyacrylamide gel. Gels were dried and then exposed to x-ray films.
Mock-transduced or DDR1b-overexpressing THP-1 cells were transfected with 10 μg of the NF-κB-Luc, pGLM-ENH, or pGLM-MA1MA2 (26) and 5 μg of the pSV-β-galactosidase (Promega, Madison, WI) per 100-mm tissue culture plate, using Effectene Transfection Reagent (Qiagen). After a 12-h incubation, cells were rinsed with PBS, and then incubated with 10 nM PMA on collagen-coated plates for an additional 12 h. Cells were collected, lysed using the Reporter Lysis Buffer (Promega), and luciferase and β-galactosidase activities were measured according to the protocol provided by the manufacturer. The resulting luciferase activity was corrected, based on the β-galactosidase activity in the same cell extract.
A mixture of four small interfering RNA (SMARTPool siRNA) specific for Act1 was purchased from Dharmacon (Lafayette, CO). DDR1b-overexpressing THP-1 cells were cultured at a density of 5 × 106 cells/ml in complete medium in six-well plates in the presence of 10 nM PMA for 12 h. Cells were washed three times with PBS and transfected with the siRNA at the final concentration of 100 nM by using OligofectamineTM Reagent (Invitrogen) according to the manufacturer’s protocol. After a 48-h incubation, cells were rinsed with PBS, starved in RPMI 1640 for 10 h, plated onto collagen-coated plates at 5 × 106 cells/ml, and incubated for 3 h. After three washes with PBS, 1 × 108 cells were lysed and used for Western blotting.
Collagen-activation of DDR1b up-regulates the expression and production of IL-1β, IL-8, MIP-1α, and MCP-1 in PMA-induced differentiated THP-1 cells
To examine whether collagen-activation of DDR1b affects the expression of proinflammatory cytokines and chemokines by macrophages, we first performed a cDNA array analysis using PMA-treated, differentiated, mock-transduced, and DDR1b-overexpressing THP-1 cells. The expression of IL-1β, IL-8, MIP-1α, and MCP-1 mRNA was markedly up-regulated in DDR1b-overexpressing, but not mock-transduced, THP-1 cells that were incubated in the presence of PMA on collagen-coated plates for 12 h (data not shown).
To confirm our data obtained by the cDNA analysis, we performed Northern blotting (Fig. 1). Parental, mock-transduced, and DDR1a-overexpressing THP-1 cells expressed IL-1β, IL-8, MIP-1α, and MCP-1 mRNA after a 16-h incubation in the presence of PMA on collagen-coated plates. The slow kinetics of MCP-1 mRNA expression by these cell lines on collagen-coated plates were identical to those previously detected by PMA-treated parental THP-1 cells on uncoated plates (26). Therefore, collagen did not appear to have any effect on the expression of the cytokine/chemokine mRNA by these three cell lines. The mechanisms regulating the delayed up-regulation of the cytokine/chemokine mRNA expression in PMA-treated THP-1 cells remain uncharacterized. In contrast, the expression of all four mRNA was rapidly induced in collagen-activated, PMA-treated DDR1b-overexpressing cells on collagen-coated plates, and the expression level of each mRNA in these cells was much higher than that in other cell lines at 12 h, consistent with the cDNA array result. The expression level of IL-1β rapidly decreased at 20 h, whereas the expression of IL-8, MIP-1α, and MCP-1 mRNA was sustained up to 24 h.
The concentrations of IL-1β, IL-8, MIP-1α, and MCP-1 in the culture supernatants of each cell line were quantified by ELISA (Fig. 2). The concentration of IL-1β in the supernatants of DDR1b-overexpressing cells was significantly higher than that in the culture supernatants of other cell lines at 24 h, but there was no significant difference at 48 h. This agrees with the mRNA expression data showing that activation of DDR1b caused only a shift in the kinetics (Fig. 1). The concentrations of IL-8, MIP-1α, and MCP-1 were significantly higher in the culture supernatants of differentiated DDR1b-overexpressing cells at 24 and 48 h in comparison with other cell lines. Because IL-1β is known to up-regulate the expression of IL-8, MIP-1α, and MCP-1 (27, 28), we added anti-IL-1β neutralizing Ab to the cultures and determined whether early production of IL-1β was responsible for the increased production of these chemokines. The release of these chemokines was not affected by the neutralizing Ab, indicating that the up-regulated production of these chemokines by differentiated DDR1b-overexpressing cells was independent of IL-1β released by the same cells (data not shown).
Activation of DDR1 up-regulates the release of IL-8, MIP-1α, and MCP-1 from human primary macrophages in p38 MAPK- and NF-κB-dependent manners
Using in vitro-generated GM macrophages, we next evaluated the effect of DDR1 activation on the production of these proinflammatory mediators by primary macrophages. GM macrophages are reported to show characteristics that resemble those of alveolar macrophages (29), and alveolar macrophages are a well-known source of cytokines and chemokines in vivo (28). GM macrophages also express a high level of endogenous DDR1, predominantly DDR1b (13). Thus, GM macrophages provided an excellent model to study the role for DDR1b-collagen interaction in the production of cytokines and chemokines. As shown in Fig. 3, GM macrophages released only low levels of IL-1β, IL-8, MIP-1α, and MCP-1 in complete medium. Activation of DDR1 with agonistic anti-DDR1 Ab (513) slightly but significantly increased the release of these proteins. Activation with LPS increased the release of all four proteins, and DDR1 activation markedly augmented LPS-induced release of these proteins. Control mouse IgM had no effect. Activation with collagen also augmented LPS-induced release of these proteins. The effects of collagen on the release of IL-1β, IL-8, and MIP-1α were higher than that of 513 Ab, and addition of anti-β1 integrin blocking Ab reduced the levels of these three proteins to those obtained with 513 Ab, indicating that the effect of collagen on the release of IL-1β, IL-8, and MIP-1α was also regulated through activation of β1 integrins. In contrast, collagen-mediated up-regulation of MCP-1 release was totally dependent on DDR1. Pretreatment of cells with the p38 MAPK inhibitor SB203580 or NF-κB inhibitor CAPE, but not DMSO, inhibited collagen-induced and 513 Ab-induced up-regulation of IL-8, MIP-1α, and MCP-1 release by GM macrophages, whereas collagen- or 513 Ab-mediated up-regulation of IL-1β release was inhibited by only SB203580. These results indicated that activation of DDR1, most likely DDR1b, up-regulates the production of these chemokines not only in differentiated THP-1 cells, a model for macrophages, but also in primary macrophages in both p38 MAPK- and NF-κB-dependent manners.
Activation of DDR1b up-regulates trans-activity of NF-κB in differentiated THP-1 cells
To examine whether activation of DDR1b increases NF-κB trans-activity in macrophages, we performed a luciferase assay by transfecting differentiated DDR1b-overexpressing THP-1 cells with reporter constructs containing the sequences of NF-κB binding sites of two different NF-κB-dependent gene promoters, the human Ig κ-chain (NF-κB-Luc) or MCP-1 (pGLM-ENH) gene promoter. As shown in Fig. 4,A, there was no increase in luciferase activity in response to PMA, collagen, or PMA plus collagen in mock-transduced cells. However, in DDR1b-overexpressing cells, luciferase activity was markedly increased in response to PMA plus collagen, and it was dose-dependently inhibited by SB203580. CAPE also completely inhibited the increased luciferase activity induced with PMA plus collagen. Almost identical results were obtained with pGLM-ENH, as shown in Fig. 4 B. There was no increase in luciferase activity when pGLM-MA1MA2 that contained mutations in the NF-κB sequences was used. These results indicated that collagen activation of DDR1b increases NF-κB activity in differentiated DDR1b-overexpressing THP-1 cells in a p38 MAPK-dependent manner.
DDR1b-mediated degradation of IκB is independent of p38 MAPK
We next analyzed the mechanisms of DDR1b-mediated NF-κB activation using differentiated DDR1b-overexpressing THP-1 cells. As shown in Fig. 5,A, there was no change in the level of IκBα in differentiated mock-transduced cells in response to collagen. However, in differentiated DDR1b-overexpressing cells, the level of IκBα rapidly decreased in response to collagen, and then returned to the original level. This is typical for the kinetics of IκB degradation. Although SB203580 inhibited the trans-activation of NF-κB-dependent gene promoters as shown in Fig. 4, it had no effect on DDR1b-mediated IκB degradation (Fig. 5,B). Collagen activation of DDR1b also induced nuclear translocation of NF-κB, and SB203580 had no effect, as determined by EMSA (Fig. 5 C). These results indicated that DDR1b-mediated IκB degradation and subsequent translocation of NF-κB were independent of p38 MAPK.
DDR1b-mediated degradation of IκB is regulated by Shc, TRAF6, and Act1
The juxtamembrane domain of DDR1b contains the LXNPXY motif that corresponds to the consensus binding motif for the Shc phosphotyrosine binding domain (30, 31). As described above, we previously demonstrated that the activation of DDR1b induced the recruitment of Shc to the LXNPXY sequence and it was necessary for the activation of p38 MAPK in differentiated THP-1 cells (13). To test the involvement of Shc recruitment in DDR1b-mediated IκB degradation, we transfected parental THP-1 cells with the wild-type or a mutant DDR1b cDNA in which tyrosine at the residue 513 in the LXNPXY motif is substituted to phenylalanine (Y513F) (13), and then examined the degradation of IκB degradation in these cells. As shown in Fig. 6,A, collagen activation of the wild-type DDR1b caused the degradation of IκB (Fig. 6,A, lanes 1 and 2). However, in Y513F-transfected cells, collagen activation did not cause IκB degradation (Fig. 6 A, lanes 3 and 4), indicating that DDR1b-mediated IκB degradation is regulated through the recruitment of Shc.
We have recently demonstrated that TRAF6 is a signaling molecule downstream of Shc and regulates DDR1b-mediated p38 MAPK activation (14). Therefore, we evaluated the role of TRAF6 in DDR1b-mediated IκB degradation by overexpressing DN-TRAF6 in DDR1b-overexpressing THP-1 cells (Fig. 6,B). As shown in Fig. 6,C, overexpression of DN-TRAF6 almost completely abrogated collagen-induced IκB degradation in these cells (Fig. 6 C, lanes 3 and 4), indicating that TRAF6 also plays a critical role in DDR1b-mediated IκB degradation in differentiated THP-1 cells.
Previous studies demonstrated that TAK1 is a downstream signaling molecule of TRAF6 and is involved in the activation of NF-κB during IL-1R or TLR4 signaling (32, 33). TAK1 is activated by its binding to TAB1 (34, 35); however, in our previous study the protein complex formed in response to the activation of DDR1b contained TAB1β, a splicing variant of TAB1 lacking the domain necessary for TAK1 binding (16), resulting in the absence of TAK1 in this protein complex (14). Therefore, it is highly unlikely that TAK1 is involved in DDR1b-mediated NF-κB activation. Recently, Kanamori et al. (36) detected the capacity of a NF-κB activator, Act1, to bind to TRAF6, and proposed that TRAF6 functions as an anchor for Act1 in unstimulated cells and that activation of cells with IL-1 releases Act1, resulting in direct binding of Act1 to IκB kinase (IKK)γ and subsequent activation of IKKs (36). As shown in Fig. 7,A, Act1 was equally expressed in both differentiated mock and DDR1b-overexpressing THP-1 cells. To explore a potential involvement of Act1 in DDR1b-mediated IκB degradation, we immunoprecipitated TRAF6 from the lysates of collagen-activated, differentiated mock or DDR1b-overexpressing THP-1 cells, and the association of Act1 with TRAF6 was evaluated by Western blotting with anti-Act1 Ab. As we expected, Act1 was coimmunoprecipitated with TRAF6 from the lysates of nonactivated differentiated DDR1b-overexpressing THP-1 cells (Fig. 7,B, lanes 1 and 7) and also from collagen-activated differentiated mock-transduced cells (Fig. 7,B, lanes 2 and 6). However, Act1 was no longer coimmunoprecipitated with TRAF6 from the lysates of differentiated DDR1b-overexpressing cells 30 min after collagen activation (Fig. 7,B, lanes 9–12). Almost identical results were obtained when anti-Act1 Ab was used for immunoprecipitation (Fig. 7 C). These results indicated that Act1 and TRAF6 associate in nonactivated cells, but the activation of DDR1b induces the dissociation of the TRAF6-Act1 complex.
To identify the functional role of Act1, we inhibited the expression of Act1 in differentiated DDR1b-overexpressing THP-1 cells by RNA interference. Transfection of the cells with siRNA reduced the level of Act1 by >90%, but it did not change the level of DDR1 or the control protein actin (Fig. 7,D, lane 2). As shown in Fig. 7,E, Act1 was no longer coimmunoprecipitated with TRAF6 in the cells (lane 4). Interestingly, inhibition of Act1 expression markedly reduced collagen-induced IκB degradation in differentiated DDR1b-overexpressing cells without affecting p38 MAPK activation (Fig. 7 F, lane 4). Thus, Act1 plays a critical role in DDR1b-mediated IκB degradation in differentiated THP-1 cells.
We also examined the association of Act1 with TRAF6 in GM macrophages. As shown in Fig. 7,H, Act1 was coimmunoprecipitated with TRAF6 in nonactivated GM macrophages (Fig. 7,H, lane 1). However, Act1 was no longer coimmunoprecipitated with TRAF6 30 min after activation of the cells with 513 Ab when TRAF6 formed a protein complex with p38α MAPK (Fig. 7 H, lanes 3–6). Identical results were obtained after activation of GM macrophages with collagen (data not shown). These results strongly suggest that Act1 regulates DDR1b-mediated degradation of IκB in primary macrophages.
We previously reported that the expression of two DDR1 isoforms, DDR1a and DDR1b, could be induced in monocytes (9), and activation of DDR1b facilitates their differentiation through activation of the p38 MAPK pathway (13). Activation of p38 MAPK is dependent on the recruitment of the adaptor protein Shc to the LXNPXY motif in the juxtamembrane domain of this receptor (13), and it was due to the autophosphorylation of p38α MAPK mediated through the TRAF6/TAB1β/p38α cascade (14). In the present study, we examined the effects of DDR1b-collagen interaction on the production of cytokines and chemokines by macrophages using PMA-induced differentiated DDR1b-overexpressing THP-1 cells and monocyte-derived human primary macrophages. Activation of DDR1b up-regulated the production of IL-1β, IL-8, MIP-1α, and MCP-1 in a p38 MAPK-dependent manner. Production of IL-8, MIP-1α, and MCP-1 was also dependent on NF-κB. DDR1b-mediated activation of NF-κB was regulated through Shc and a novel signaling cascade involving TRAF6 and Act1. Thus, we have found that DDR1b plays a novel role in the production of proinflammatory mediators, an important function of macrophages, and that NF-κB is a novel target of the DDR1b signaling.
NF-κB is a ubiquitous, pleotropic, multisubunit transcription factor activated in response to inflammatory and noninflammatory exogenous stimuli. In most cells, NF-κB exists as an inactive heterodimer, the predominant of which is composed of p50 and p65 (Rel A) subunits, and is sequestered within the cytoplasm by association with the inhibitory protein, IκB. Interaction of IκB with the NF-κB dimer prevents the nuclear translocation of NF-κB. Phosphorylation by IKKs leads to the ubiquitination and degradation of IκB, resulting in the nuclear translocation of NF-κB and subsequent activation of downstream target genes, including proinflammatory cytokines and chemokines (37, 38). As described above, we have determined Shc and TRAF6 to be crucial upstream molecules regulating the degradation of IκB during the DDR1b signaling. Act1, also known as connection to IκB kinase and stress-activated protein kinase/c-Jun N-terminal kinase (CIKS), was originally cloned as a protein capable of binding to IKKγ, the regulatory unit of IKK, and forced expression of Act1/CIKS in 293 cells or HeLa cells induced constitutive activation of NF-κB in the cells (39, 40). A recent study by Kanamori et al. (36) demonstrated that Act1 binds to TRAF6 and Act1-mediated NF-κB activation can be inhibited by a DN-TRAF6, suggesting an involvement of Act1 in IL-1R/TLR-mediated signaling. Therefore, we examined the role of Act1 in DDR1b-mediated IκB degradation. Act1 was associated with TRAF6 in nonactivated, differentiated mock-transduced or DDR1b-overexpressing THP-1 cells, as well as in nonactivated primary macrophages; however, activation of DDR1b caused dissociation of Act1 from TRAF6. Furthermore, inhibition of Act1 expression by RNA interference markedly reduced DDR1b-mediated IκB degradation. Thus, we have demonstrated, for the first time, that Act1 is a physiological regulator of IκB degradation.
p38 MAPK regulates the transcription of many NF-κB-dependent genes (41, 42). In fact, inhibition of p38 MAPK with SB203580 almost completely abolished DDR1b-mediated IL-8, MIP-1α, and MCP-1 release by primary macrophages. SB203580 also inhibited the trans-activation of the Ig κ-chain or MCP-1 gene promoters in differentiated THP-1 cells. These results indicated that DDR1b-mediated release of these chemokines was dependent on p38 MAPK. However, inhibition of p38 MAPK had no effect on IκB degradation or nuclear translocation of NF-κB (Fig. 4, B and C). It has been shown that activation of NF-κB requires two separate mechanisms: the degradation of IκB and subsequent nuclear translocation of NF-κB, and modification of the RelA/p65 trans-activation subunit of NF-κB by p38 MAPK (43, 44). This provides a mechanism whereby blocking p38 MAPK activity inhibited DDR1b-mediated trans-activation of these two gene promoters without affecting the degradation of IκB.
Direct activation of DDR1 with agonistic anti-DDR1 Ab augmented LPS-induced IL-1β, IL-8, MIP-1α, and MCP-1 release from GM macrophages, although it also induced low-level release of these proteins without LPS activation. This suggests that the main role of DDR1 signaling is not to induce strong responses but to amplify the effects by other stimuli, such as proinflammatory cytokines or bacterial products. Because collagen is abundant and present in almost all tissues, this may be a very important mechanism to avoid unnecessary signaling through DDR1. An important question that remains to be answered is whether collagen in different forms equally activate DDR1. Normal tissues are composed of a fibrillar mesh of the ECM, including polymerized type I collagen. However, in inflammatory conditions, such as atherosclerosis, polymerized collagen fibrils are absent from intermediate stages of lesion development that instead contain thin and disordered collagen fibers and collagen fragments. Several previous reports have indicated a differential role between monomeric and polymerized collagen in arterial smooth muscle proliferation (45) and macrophage matrix metalloproteinase 9 production (46). It will be important to determine the optimal physical state of collagen for DDR1 binding and activation.
Our previous study demonstrated that DDR1b transduces signals only in PMA-treated differentiated THP-1 cells, but not in undifferentiated THP-1 cells (13). In primary macrophages, a detectable level of Shc recruitment to DDR1b occurs after 3-day incubation of monocytes with GM-CSF (W. Matsuyama and T. Yoshimura, unpublished data). Therefore, differentiation of THP-1 cells or monocytes is necessary for signaling by DDR1b. PMA and GM-CSF are potent activators of THP-1 cells and monocytes, respectively, and these agents could influence the signaling in these cells. However, we did not detect a significant basal level of either p38 MAPK phosphorylation or IκB degradation before DDR1 activation, indicating that the activation of p38 MAPK and NF-κB detected in this study was induced through activation of DDR1. The molecular mechanisms whereby DDR1b transduces signals only in differentiated THP-1 cells or macrophages remain unknown.
Macrophages are one of the major cell types infiltrating the sites of chronic inflammatory diseases, such as atherosclerosis, pulmonary fibrosis, and rheumatoid arthritis, and release an array of cytokines and chemokines, and contribute to the development of these diseases. Our data presented here indicate that DDR1b-collagen interaction markedly amplifies macrophage production of IL-8, MIP-1α, and MCP-1, providing a novel mechanism by which macrophages produce and release large amounts of these chemokines in a tissue microenvironment in the course of inflammatory responses. Intervention of DDR1b-collagen interaction or DDR1b signaling may be useful to control the development of inflammatory diseases in which these proteins play an important role.
We are grateful to Dr. Joost J. Oppenheim for his invaluable comments.
Abbreviations used in this paper: ECM, extracellular matrix; DDR1, discoidin domain receptor 1; MAPK, mitogen-activated protein kinase; TRAF6, TNFR-associated factor 6; TAK1, TGF-β-activated protein kinase 1; TAB1, TAK1 binding protein 1; MIP-1α, macrophage inflammatory protein-1α; MCP-1, monocyte chemoattractant protein-1; CAPE, caffeic acid phenethyl ester; TLR, Toll-like receptor; DN, dominant negative; IKK, IκB kinase; siRNA, small interfering RNA.