Matricellular Protein CCN1 Activates a Proinflammatory Genetic Program in Murine Macrophages

Although the extracellular matrix (ECM) was classically viewed as an inert scaffold onto which cells are organized into organs and tissues, studies in recent decades have established the ECM as dynamic structures that can profoundly influence diverse aspects of cellular behavior and function (1). Aside from providing signals that regulate cell adhesion, shape, migration, differentiation, proliferation, and survival in parenchymal cells, the ECM also affects the functions of inflammatory cells that infiltrate the parenchyma. Recent studies have recognized that fragments of ECM proteins generated due to tissue injury and inflammation, and dynamic changes in the ECM in these contexts, can significantly regulate the inflammatory response (2, 3). Normally soluble matrix proteins from the blood plasma, including fibrinogen, plasma fibronectin (FN), and vitronectin, enter inflamed tissues as a result of vessel damage and become immobilized at sites of injury. Fragments of these and other ECM proteins, generated as products of elevated levels of matrix-degradative enzymes, have been shown to induce chemotaxis for inflammatory cells, enhance phagocytic functions, stimulate immune responses, and induce gene expression changes (2, 3). Induced expression of matricellular proteins, a group of matrix-associated signaling proteins that play diverse regulatory roles, also contributes to dynamic changes in the ECM at sites of inflammation and injury repair (4). Such matricellular proteins as osteopontin and thrombospondin are known to play regulatory roles in inflammation (5–8).

CCN1 (also named CYR61) is an angiogenic matricellular protein of the CCN family (CCN is the acronym of the first 3 members: Cyr61, CTGF, and Nov), and is essential for cardiovascular development (9–11). Correspondingly, Ccn1-null mice are embryonic lethal and suffer from atrioventricular septal defects and vascular hemorrhage. In the adult, CCN1 is normally expressed at a low level in most tissues but becomes highly expressed as a result of bacterial or viral infections (12–15) or in tissue repair (16–21). Consistent with a role in injury repair, CCN1 regulates the expression of genes involved in wound healing in fibroblasts (19). Recent studies have shown that CCN1 cooperates with and modulates the activities of several inflammatory cytokines, including TNF-α, Fas ligand, and TRAIL, further indicating that CCN1 can play a critical role in the inflammatory response (22–24).

In this study, we show that CCN1 supports macrophage adhesion through integrin α_Mβ₂ and the cell surface heparan sulfate proteoglycan (HSPG) syndecan-4 as coreceptors. Furthermore, CCN1 activates NF-κB signaling in macrophages, leading to the expression of multiple proinflammatory cytokines and chemokines characteristic of classically activated M1 macrophages. These results reinforce the notion that the ECM microenvironment can play critical roles in inflammation and highlight CCN1 as a novel player in regulating the inflammatory response of macrophages through integrin signaling.

The nonvirally immortalized splenic macrophage cell line I-13.35 (CRL-2471; American Type Culture Collection, Manassas, VA) originated from the tlr4 mutant strain C3H/HeJ (25). These cells were maintained in DMEM (Invitrogen, Carlsbad, CA) with 1.0 mM sodium pyruvate, 10% FBS (Intergen, Purchase, NY), and 20% conditioned media from the mouse bone marrow cell line LADMAC (ATCC, CRL-2420), which secretes M-CSF at 37°C with 10% CO₂. The RAW 264.7 cell line, established from a tumor induced by Abelson murine leukemia virus, was cultured in DMEM containing 10% FBS at 37°C with 5% CO₂. Both cell lines were obtained from the American Type Culture Collection. Wild-type CCN1 protein was produced in a baculovirus expression system using Sf9 cells and purified from serum-free insect-cell–conditioned medium on Sepharose S columns as described (26). FN, BSA, heparin (sodium salt, from porcine intestinal mucosa), LPS (from Escherichia coli 026:B6), and anisomycin were from Sigma-Aldrich (St. Louis, MO). Biotin-11-UTP was obtained from PerkinElmer (Wellesley, MA). Chemical inhibitors U0126, SB203580, BAY11-7082, and I-κB kinase (IKK) inhibitor X were from Calbiochem (San Diego, CA). Anti-CCN1 polyclonal Abs were raised in rabbits using GST fusion proteins linked to domain II (vWC domain) of CCN1 as an immunogen, and affinity-purified from with column containing GST-vWC protein cross-linked to CNBr-activated Sepharose (27). mAbs recognizing mouse total and phospho–NF-κB p65 (Ser⁵³⁶) were from Cell Signaling Technology (Beverly, MA), and function-inhibiting mAbs against integrin α_M (clone M1/70) were from Abcam (Cambridge, MA), integrin α_L (clone M17/4) from eBioscience (San Diego, CA), integrin α₆ (clone GoH3) and integrin β₂ (clone P4H9-A11) from Millipore (Bedford, MA), and TNF-R1 (clone 55R-170) from R&D Systems (Minneapolis, MN). Polyclonal Abs against syndecan-4 were from Santa Cruz Biotechnology (Santa Cruz, CA), and Abs neutralizing TNFα were from R&D Systems. Anisomycin caused a >98% inhibition of [³⁵S]methionine labeling in our experimental conditions (data not shown), showing its effectiveness.

Adhesion assays were performed as described previously (28). Briefly, I-13.35 cells were harvested in PBS by scraping and resuspended in serum-free DMEM containing 1% BSA. The protein under study was diluted to the desired concentration in PBS, applied to 96-well microtiter plates (50 μl per well), incubated at 4°C for 16 h, and blocked with 1% BSA at room temperature for 1 h. Where indicated, reagents including EDTA (2.5 mM), Ca²⁺ (5 mM), or heparin (1 μg/ml) were mixed with cells prior to plating. GRGDSP or GRGESP peptides (2 mM) were incubated with cells at room temperature for 30 min, whereas Abs against integrins α_M, α_L, α_6, or β₂ or syndecan-4 (50 μg/ml each) were incubated with cells for 1 h with normal IgG as a negative control before plating. To each well, 5 × 10⁵ macrophages were plated, and after incubation at 37°C for 15 min, wells were washed twice with PBS. Adherent cells were fixed with 10% formalin, stained with methylene blue, and quantified by dye extraction and measurement of absorbance at 620 nm.

The Oligo GEArray Mouse Inflammatory Cytokines and Receptors Microarray (catalog no. OMM-011) was purchased from SABiosciences (Frederick, MD). Mouse I-13.35 macrophages grown to near confluence were made quiescent by starvation in a low serum concentration (2% FBS) overnight. Cells then were washed and incubated in serum-free DMEM containing 1% BSA and stimulated with purified rCCN1 (10 μg/ml) for 6 h or as otherwise indicated. Total RNA isolation, cRNA synthesis, biotinylated UTP probe labeling, array hybridization, and chemiluminescent detection using alkaline phosphatase conjugated with streptavidin were performed following standard protocols. The hybridization signal was exposed to an X-ray film, scanned with a flatbed desktop scanner (Epson Perfection 1650, 600 dpi), and analyzed with a web-based image analysis program (GEArray Expression Analysis Suite, from SABiosciences), and normalized against internal controls (GAPDH) on the same blot. Microarray data are available from the National Center for Biotechnology Information Gene Expression Omnibus under accession number GSE19731 (www.ncbi.nlm.nih.gov/geo/).

The small interfering RNA (siRNA) against mouse integrin β₂ (ACCGUGGUAGGUGUCGUACUGAUUG) was purchased from Invitrogen. Transfection was done using the Lipofectamine RNAiMAX reagent (Invitrogen) with 10 nM siRNA following the manufacturer’s protocol. Cells were used for experiments 48 h after transfection.

I-13.35 and RAW 264.7 cells were starved as described above and treated as indicated, and total RNA was extracted using a RNeasy Kit (Qiagen, Valencia, CA). cDNA was generated from 0.5 μg total RNA using an iScript cDNA Synthesis Kit (Bio-Rad, Hercules, CA). Reaction product was mixed with iQ SYBR Green Supermix (Bio-Rad), and the real-time PCR reaction was carried out, detected, and analyzed using the iQ5 Real-Time PCR Detection System according to manufacturer’s instructions (Bio-Rad). The level of target gene expression was normalized against the internal control cyclophilin and expressed as relative mRNA levels. Primer sequence and annealing temperature are listed in Table I.

For semiquantitative RT-PCR, cDNA was synthesized from 1 μg total RNA using an M-MLV Reverse Transcriptase Kit (Promega), and PCR was performed using a Taq DNA Polymerase Kit from Takara (Shiga, Japan). Primer sequence and annealing temperatures are listed in Table II. Pilot experiments were performed to determine the appropriate number of PCR cycles so that all of the experiments were done in the exponential phase. Reaction products were resolved on 1% agarose gels, stained with ethidium bromide, and visualized under UV light. β-Actin mRNA level was used as a loading control

Reporter constructs for NF-κB activity (pNFκB-Luc) and transfection efficiency (pRL-CMV) were from Promega (Madison, WI) and transfected using Lipofectamine 2000 (Invitrogen) as described by the vendor. Luciferase activity was assayed using a dual-luciferase reporter assay kit (Promega). Activity of the firefly luciferase, which is encoded by pNFκB-Luc, was normalized against Renilla luciferase activity, which is encoded by the transfection efficiency control vector pRL-CMV.

Measurement of TNF-α and IL-6 protein levels in cell lysates and conditioned media was carried out using TNF-α and IL-6 ELISA kits (eBioscience) following the manufacturer’s protocols. Conditioned media were collected and cleared by centrifugation at 500 × g for 5 min at 4°C. Cell lysates were prepared by first rinsing cells with PBS followed by lysis on ice for 15 min using 0.3 ml of 10⁶ cells in lysis buffer (50 mM HEPES [pH 7.4], 130 mM NaCl, 1% Triton X-100, 1% SDS, 1% sodium deoxycholate, 2 mM EDTA, 2 mM EGTA, 0.1 M NaF, 10 mM sodium pyrophosphate, 10 mM β-glycerophosphate, 1 mM sodium orthovanadate, 1 mM PMSF, and a protease inhibitor mixture from Roche [Basel, Switzerland]).

Six normal 12-wk-old male wild-type or Itgam-null (α_M/CD11b knockout) mice on a C57BL/6 background (The Jackson Laboratory, Bar Harbor, ME) were housed in a barrier facility, and each was treated with 1 ml 3% Brewer thioglycollate medium (BD Biosciences, San Jose, CA) by i.p. injection. Four days later, mice were euthanized by CO₂ asphyxiation, and peritoneal macrophages were isolated according to a published protocol with minor modifications (29, 30). Briefly, peritoneal cells were flushed out by ice-cold DMEM containing 5% FBS, and macrophages were collected by plating cells on plastic culture dishes at 37°C for 2 h followed by washing away the floaters with normal saline. The adherent macrophages then were incubated in serum-free DMEM containing 1% BSA at 37°C and immediately used for experiments in the same day after isolation.

CCN1 supports cell adhesion in a variety of cell types including fibroblasts, endothelial cells, smooth muscle cells, and peripheral blood monocytes (20, 21, 31). However, its cell adhesive property in differentiated macrophages has not been tested. To help to avoid issues with potential LPS contamination in laboratory reagents, we focused most of our studies on I-13.35 mouse macrophages. I-13.35 cells originated from C3H/HeJ mice which carry a missense mutation in the tlr4 gene and express a nonfunctional LPS receptor, TLR4 (32–34). I-13.35 cells were plated on microtiter wells coated with purified insect-cell–produced rCCN1 protein in serum-free medium and incubated for 15 min. As shown in Fig. 1A, macrophage adhesion to CCN1 was dose-dependent and saturable, reaching an optimal level when CCN1 was coated at a concentration of 12.5 μg/ml. The presence of Ca²⁺ or EDTA completely blocked macrophage adhesion to CCN1, consistent with the notion that CCN1 supports cell adhesion through certain integrins that are sensitive to Ca²⁺ (Fig. 1B). The RGD-containing peptide had little effect on cell adhesion to CCN1 but inhibited adhesion to FN, indicating that RGD-binding integrins (integrins α₅β₁, α₈β₁, α_v, and β₃) are unlikely to be involved (Fig. 1B). mAbs against integrin subunits α_M and β₂, but not α_L or α₆, effectively blocked macrophage adhesion to CCN1, indicating that adhesion to CCN1 is mediated through integrin α_Mβ₂, which is prominently expressed in macrophages (Fig. 1C). Because cell surface HSPGs are involved in CCN1-mediated cell adhesion in fibroblasts (31), we tested the role of heparin binding in CCN1-mediated cell adhesion. Soluble heparin added to the cell medium completely blocked macrophage adhesion to CCN1 (Fig. 1B), suggesting that CCN1 interaction with cell surface HSPGs may be indispensable for macrophage adhesion. Because the HSPG syndecan-4 has been found to serve as a coreceptor for CCN1 in some contexts, we tested the potential role of syndecan-4 (22, 35). Indeed, Abs against syndecan-4 effectively blocked adhesion to CCN1 but not to FN, showing that syndecan-4 is specifically required for macrophage adhesion to CCN1 (Fig. 1C).

The engagement of integrin α_Mβ₂ is known to regulate gene expression (36–38). To assess whether CCN1 can regulate gene expression in macrophages, I-13.35 cells were treated with BSA or soluble CCN1 protein in serum-free medium for 6 h, after which cellular RNAs were collected and used as probes to interrogate a microarray containing 113 key genes involved in the inflammatory response (Fig. 2). A total of 24 genes were identified as significantly regulated by CCN1, including those encoding inflammatory cytokines (TNF-α, IL-1α, IL-1β, IL-6, IL-12b, and IFN-γ) and chemokines (MCP-1, MCP-3, MIP-1α, MIP-1β, growth-related oncogene [GRO] 1, GRO2, and inflammatory protein 10 [IP-10]). Other CCN1-regulated genes encode the anti-inflammatory cytokine TGF-β1, inducible NO synthase (iNOS), complement protein C3, and signaling receptors involved in host defense and inflammation, including TLR4 and TLR7, TLR-interacting protein, IL-10Rβ, IL-6 receptor complex signaling subunit IL-6st (also known as gp130), and the high-affinity IgE receptor γ (FcεR1γ). To confirm the microarray hybridization results, expression of 15 genes was verified by both semiquantitative (Fig. 3A, and Table II) and quantitative real-time RT-PCR (Fig. 3B, Table I) analyses. CCN1-regulated gene expression was annihilated by neutralizing anti-CCN1 Abs (Fig. 3C), indicating that induced gene expression is an activity specific to the CCN1 polypeptide. Many of the genes, including TNF-α, IL-1β, IL-6, IL-12b, MIP-1α, MCP-3, GRO1, GRO2, IP-10, and iNOS, were upregulated by CCN1 by >30-fold compared with BSA-treated controls. Several other genes, such as TLR4, TGF-β1, and IL-10Rβ, were downregulated by ∼40%. These CCN1-regulated genes are known to play critical roles in the inflammatory response, suggesting that the presence of CCN1 in the ECM microenvironment may significantly regulate the genetic program for inflammation in macrophages.

To characterize the regulation of macrophage gene expression by CCN1, we monitored the time course and dose dependence of the response. Upregulation of proinflammatory cytokines and chemokines was noticeable by 3–6 h after incubation with CCN1 and sustained for >12 h (Fig. 4A, left panel). Among the genes studied, maximal TNF-α and IP-10 upregulation occurred at ∼6 h, whereas other genes peaked at ∼9 h or later. IL-1β expression reached the highest level by 24 h (Fig. 4A). The upregulation of TNF-α, IL-6, MIP-1α, MCP-3, and IP-10 was observed at ∼0.4 μg/ml CCN1 (10 nM), whereas IL-1β and GRO1 induction required a higher concentration of CCN1 (2–10 μg/ml) (Fig. 4A, right panel). To examine whether the regulation of inflammatory genes by CCN1 occurs in macrophages generally, we tested the effects of CCN1 on gene expression in RAW264.7 cells, another macrophage cell line derived from Abelson murine leukemia virus-induced tumors in mice, and in freshly isolated peritoneal macrophages from wild-type C57BL/6 mice. CCN1 treatment upregulated the expression of a number of representative cytokines and chemokines in RAW264.7 cells with similar time courses and dosage requirements to those in I-13.35 cells (Fig. 4B) and increased TNF-α, IL-6, IP-10, and IL-12b expression in peritoneal macrophages (Fig. 4C). These results show that CCN1 can regulate inflammatory gene expression in macrophages generally. Next, we measured the levels of proteins produced by I-13.35 cells and peritoneal macrophages in response to CCN1 treatment. After incubation with CCN1 for 6 h, TNF-α protein concentrations in I-13.35 cell lysates were elevated by >7-fold (basal, 35 pg/ml; CCN1, 270 pg/ml), and IL-6 concentrations in conditioned media increased by >65-fold (basal, 4 pg/ml; CCN1, 270 pg/ml) as determined by ELISA (Fig. 4D). In peritoneal macrophage-conditioned media, CCN1 increased TNF-α concentrations by >5-fold (basal, 47 pg/ml; CCN1, 260 pg/ml) and IL-6 concentrations by >25-fold (basal, 17 pg/ml; CCN1, 430 pg/ml) (Fig. 4D). These results confirmed that CCN1 increased cytokine production at the protein level.

Real-time RT-PCR analysis showed that CCN1 upregulated TNF-α and IP-10 expression even in the presence of the protein synthesis inhibitor anisomycin, indicating that this process occurs without requiring de novo protein synthesis (Fig. 5A). By contrast, although anisomycin alone elevated IL-1β and IL-6 mRNA levels, CCN1 did not cause a further increase, suggesting that de novo protein synthesis is necessary for their upregulation by CCN1 (Fig. 5A). These results indicate that CCN1 can induce TNF-α and IP-10 gene expression directly through activation of intracellular signaling, but upregulation of IL-1β and IL-6 may require the synthesis of intermediary protein factors.

Because CCN1 can support macrophage cell adhesion, we tested whether CCN1 can regulate gene expression as a cell adhesion substrate. Thus, we allowed I-13.35 cells to adhere for 3 h to surfaces coated with poly-l-lysine or rCCN1 protein in serum-free medium and tested the expression of TNF-α and IL-1β by real-time RT-PCR. Cell adhesion to CCN1 caused 4-fold higher TNF-α expression and 10-fold higher IL-1β expression when compared with those of cells plated on the inert substrate poly-l-lysine (Fig. 5B). These results show that CCN1 can regulate gene expression as both a soluble and an immobilized adhesive factor.

Because CCN1 can act as a macrophage adhesion substrate via integrin α_Mβ₂ and syndecan-4 (Fig. 1), we hypothesized that the same receptors may mediate CCN1 effects on inflammatory gene expression. Preincubation of I-13.35 cells with function-blocking mAb against integrin β₂ significantly reduced TNF-α and IL-1β upregulation by >50%, but nonspecific murine IgG had no effect, indicating a critical role for integrin β₂ in mediating CCN1 regulation of gene expression (Fig. 6A). The addition of soluble heparin to the culture medium effectively inhibited TNF-α and IL-1β upregulation by CCN1 (Fig. 6A), consistent with the interpretation that heparin can bind to CCN1 and prevent CCN1 interaction with cell surface HSPGs to regulate gene expression. Indeed, pretreatment of cells with anti-syndecan-4 Abs reduced the upregulation of TNF-α and IL-1β expression by CCN1 by 35 and 60%, respectively, indicating that syndecan-4 plays an important role in CCN1-regulated gene expression (Fig. 6A). To pinpoint the role of integrin α_Mβ₂ in CCN1 function further, we isolated and examined peritoneal macrophages from integrin α_M knockout mice on a C57BL/6 background. These integrin α_M-null macrophages were completely refractory to CCN1-induced gene expression (Fig. 6B), whereas peritoneal macrophages from syngeneic wild-type C57BL/6 mice were highly responsive to CCN1 (Fig. 4C), indicating that integrin α_M is critical for CCN1-regulated gene expression. Likewise, siRNA against integrin β₂ effectively knocked down integrin β₂ expression in I-13.35 cells (Fig. 6C, left panel) and resulted in inhibition of CCN1-regulated TNF-α and IP-10 expression (Fig. 6C, right panel). Control siRNA, by contrast, had no effect. These results show that CCN regulates macrophage gene expression through integrin α_Mβ₂ and syndecan-4 plays an important role in this process.

Because I-13.35 cells express nonfunctional TLR4 (32–34), it is unlikely that effects of CCN1 are mediated through TLR4. Furthermore, inhibition of TLR4 signaling by the selective inhibitor TAK-242 (39) had no effect on CCN1 upregulation of TNF-α and IL-6 mRNA (Fig. 6D), indicating that CCN1 signaling does not involve TLR4. Taken together, our results show that CCN1-regulated gene expression is an activity specific to the CCN1 polypeptide (Fig. 3C) and is mediated through integrin α_Mβ₂ and syndecan-4, whereas TLR4 plays no role in this process.

Engagement of integrin α_Mβ₂ can lead to the activation of NF-κB, a transcription factor capable of regulating a multitude of genes involved in inflammation, cell survival, and cell proliferation (36–38, 40, 41). CCN1 thus may activate NF-κB signaling through binding to integrin α_Mβ₂. Because CCN1 can induce the expression of TNF-α, which is a potent activator of NF-κB, it may further activate NF-κB through the induction of TNF-α. To evaluate whether CCN1 can activate NF-κB signaling, we monitored NF-κB–dependent transcription and phosphorylation of p65 NF-κB in I-13.35 cells. CCN1 treatment enhanced NF-κB–dependent transcription by ∼5-fold within 3 h, as judged by luciferase activity in cells transiently transfected with an NF-κB–luciferase reporter construct (Fig. 7A), and rapidly upregulated p65 NF-κB phosphorylation within 30 min (Fig. 7B, upper left). Effects of CCN1 on p65 phosphorylation and NF-κB–dependent transcription lasted for at least 6 h (Fig. 7A, 7B). Consistent with a direct role for CCN1 in NF-κB activation, the protein synthesis inhibitor anisomycin did not prevent p65 phosphorylation (Fig. 7B, upper right). Pretreatment of cells with Abs blocking integrin β₂ or syndecan-4 effectively annihilated NF-κB p65 phosphorylation (Fig. 7B, lower panel), showing that CCN1 acts through these receptors to activate NF-κB. Addition of the IKK complex inhibitor BAY-117082 or an IKK inhibitor [N-(6-chloro-9H-β-carbolin-8-yl) nicotinamide] significantly attenuated CCN1-induced TNF-α and IL-1β mRNA, indicating that their upregulation is dependent on NF-κB (Fig. 7C). To test whether CCN1-induced TNF-α may further enhance NF-κB activation, we treated cells with CCN1 for 6 h in the presence of polyclonal Abs that neutralize TNF-α activity and examined NF-κB p65 phosphorylation. Results show that the Abs attenuated p65 phosphorylation by ∼30% compared with control IgG (Fig. 7D). Taken together, we demonstrate that CCN1 can bind to integrin β₂ and syndecan-4 to activate NF-κB to induce the synthesis of TNF-α and IL-1β. TNF-α in turn acts in an autocrine/paracrine mechanism to further enhance NF-κB activation.

Because the effect of CCN1 on gene expression is dependent on NF-κB activation and the upregulation of IL-1β and IL-6 requires de novo protein synthesis (Figs. 5A, 7), we tested whether CCN1-induced TNF-α may be a necessary mediator for IL-1β and IL-6 induction. Preincubation of cells with either a mAb that inhibits TNFR1 function or polyclonal Abs that neutralize TNF-α had no effect on TNF-α induction and only modest effects on IP-10 expression, consistent with the idea that CCN1 can directly activate TNF-α and IP-10 expression without requiring de novo protein synthesis (Figs. 5A, 8A). In contrast, both Abs severely reduced IL-1β and IL-6 expression, showing that TNF-α is a critical mediator for the expression of these genes (Fig. 8B). Taken together, these results demonstrate that CCN1 can directly induce the expression of certain inflammatory genes, such as TNF-α and IP-10, whereas upregulation of other inflammatory genes, including those encoding IL-1β and IL-6, requires the induction of TNF-α as a mediator (Figs. 8C, 9).

This study shows that the matricellular protein CCN1 can induce a proinflammatory genetic program in macrophages, implicating CCN1 as a novel regulator of macrophage function. Upon presentation of inflammatory stimuli, IFN-γ in combination with the bacterial component LPS can activate macrophages in a process known as classical activation, resulting in macrophages that participate in Th1 responses (42–44). CCN1 induces robust expression of proinflammatory cytokines characteristic of classically activated M1 macrophages (Fig. 9), including upregulation of inflammatory cytokines and chemokines, microbicidal iNOS, and complement component C3. Consistent with a proinflammatory role, CCN1 downregulates the anti-inflammatory cytokine TGF-β1 and the receptor for IL-10, IL-10Rβ. CCN1 regulates proinflammatory genes in two macrophage cell lines of distinct origins as well as in mouse peritoneal macrophages, indicating that these effects are not cell line-specific. In contrast to this proinflammatory M1 response, macrophages also may undergo alternative activation by cytokines involved in Th2 responses, such as IL-4 and IL-13, or other anti-inflammatory factors, such as glucocorticoids, leading to various forms of M2 macrophages with phenotypic changes important for humoral immunity, immune suppression, allergic and antiparasitic responses, or immune regulation (42–44). CCN1 does not induce gene expression changes consistent with the M2 phenotype, suggesting that CCN1 may help to polarize macrophages to participate as inducers and effectors in polarized Th1 responses.

Several lines of evidence show that the CCN1 polypeptide acts through integrin α_Mβ₂ to activate this proinflammatory genetic program. First, most of the experiments in this study were conducted with I-13.35 cells isolated from the TLR4 mutant C3H/HeJ strain, which are unresponsive to signaling through the LPS receptor TLR4 (32–34). Furthermore, the TLR inhibitor TAK-242 was unable to curtail CCN1-regulated gene expression (Fig. 6D), indicating that TLR4 does not mediate CCN1 action. Second, neutralizing anti-CCN1 Abs abrogated CCN1-induced gene expression (Fig. 3C), showing that the activity is specific to the CCN1 polypeptide. Third, CCN1-induced gene expression was annihilated in peritoneal macrophages isolated from integrin α_M knockout mice, whereas macrophages from syngeneic wild-type mice were highly responsive (Figs. 4C, 6B). Moreover, both siRNA knockdown and Ab blockade of integrin β₂ greatly blunted CCN1-induced gene expression (Fig. 6A, 6C). Together, these results show that the CCN1 protein acts through integrin α_Mβ₂ to induce a proinflammatory program in macrophages and TLR4 signaling does not play a role in this process.

The leukocyte-specific integrin receptor α_Mβ₂ is broadly expressed in monocytes, granulocytes, macrophages, and NK cells and plays a critical role in regulating leukocyte adhesion, migration, phagocytosis, and cell-mediated cytotoxicity. Individuals with integrin α_Mβ₂ deficiency suffer recurrent and even fatal bacterial and fungal infections (45). CCN1 supports macrophage adhesion through integrin α_Mβ₂, and the HSPG syndecan-4 and induces macrophage gene expression by activating NF-κB in an integrin β₂- and syndecan-4–dependent manner (Figs. 1, 6). Engagement of integrin α_Mβ₂ through extracellular ligands or cell adhesion to certain ECM proteins is known to activate the transcription factor NF-κB, thereby inducing the expression of inflammatory cytokines (37, 46–48). Whereas activation of NF-κB through engagement of integrin α_Mβ₂ has been reported previously (46, 49), in this paper we show evidence that syndecan-4 also is involved, suggesting that syndecan-4 is a coreceptor of integrin α_Mβ₂ that can mediate NF-κB activation. Syndecan-4–null mice are more susceptible to endotoxin shock, suggesting a role in the inflammatory response (50). It is interesting to note that the integrin α_Mβ₂ binding site of CCN1 is closely juxtaposed to a HSPG binding site located in the carboxyl domain (27, 31), although it is currently unclear whether a single CCN1 molecule must simultaneously bind both integrin α_Mβ₂ and syndecan-4 to activate NF-κB or separate CCN1 molecules can bind integrin α_Mβ₂ and syndecan-4 independently to induce signaling in the same cell.

Mechanistically, genes regulated by CCN1 in I.13-35 cells fall into two categories: immediate-early genes that are induced with rapid kinetics (peaking at 3–6 h) without requiring de novo protein synthesis, including TNF-α and IP-10; and delayed response genes that are upregulated with delayed kinetics (peaking at 12–24 h) through the synthesis of protein mediators, represented by IL-1β and IL-6 (Figs. 4A, 5A). We show that CCN1 induces the synthesis of TNF-α as an immediate-early gene product (Fig. 4D), which in turn acts as the mediator for the upregulation of delayed response genes, such as IL-1β and IL-6, because function-blocking Abs against TNF-α or TNFR1 obliterated their response to CCN1 (Fig. 8). The different responses of immediate-early and delayed response genes may be due to their differential sensitivity to NF-κB activation. Both TNF-α and IP-10 promoters contain two or more copies of functional κB-like elements (51, 52), whereas the IL-1β and IL-6 promoters have only a single potential κB site (53, 54), suggesting a differential responsiveness to NF-κB activation among these genes. Because TNF-α is a potent activator of NF-κB, it may enhance IL-1β and IL-6 gene transcription by amplifying the activation of NF-κB (Fig. 7D). Alternatively, TNF-α also may enhance IL-1β and IL-6 transcription through a p38 MAPK/AP-1–dependent mechanism (47, 55).

Although our studies establish the role of CCN1 in modulating proinflammatory gene expression in macrophages, the significance of this activity in inflammation and injury repair in vivo remains to be determined. CCN1 is normally expressed at low levels in most adult tissues but is redeployed at a high level at sites of inflammation and injury repair (56), where its ability to regulate gene expression in macrophages may play an important role. Recent studies have shown that CCN1 enables the cytotoxicity of the inflammatory cytokine TNF-α and enhances the apoptotic activity of Fas ligand and TRAIL (22–24). Further, TNF-α- or Fas-mediated apoptosis is blunted in mice expressing an apoptosis-defective CCN1, indicating that CCN1 is an important physiologic modulator of the biological responses to inflammatory cytokines in vivo. On the basis of its induction of Th1 responses in macrophages, it is tempting to speculate that CCN1 might play a role in M1/M2 polarization to enhance the inflammatory response (42–44). In wound healing, CCN1 may provide a provisional matrix and may serve as a localized regulator of inflammatory cell recruitment and function (18, 19). Further investigations using models of infection and tissue injury repair will be required to elucidate the role of CCN1 in these biological contexts.

Disclosures The authors have no financial conflicts of interest.