Transforming Growth Factor-β1 Inhibition of Macrophage Activation Is Mediated via Smad3*

Activated macrophages are critical cellular participants in inflammatory disease states. Transforming growth factor (TGF)-β1 is a growth factor with pleiotropic effects including inhibition of immune cell activation. Although the pathway of gene activation by TGF-β1 via Smad proteins has recently been elucidated, suppression of gene expression by TGF-β1 remains poorly understood. We found that of Smad1–Smad7, Smad3 alone was able to inhibit expression of markers of macrophage activation (inducible nitric-oxide synthase and matrix metalloproteinase-12) following lipopolysaccharide treatment in gene reporter assays. Transient and constitutive overexpression of a dominant negative Smad3 opposed the inhibitory effect of TGF-β1. Domain swapping experiments suggest that both the Smad MH-1 and MH-2 domains are required for inhibition. Mutation of a critical amino acid residue required for DNA binding in the MH-1 of Smad3 (R74A) resulted in the loss of inhibition. Transient overexpression of p300, an interactor of the Smad MH-2 domain, partially alleviated the inhibition by TGF-β1/Smad3, suggesting that inhibition of gene expression may be due to increased competition for limiting amounts of this coactivator. Our results have implications for the understanding of gene suppression by TGF-β1 and for the regulation of activated macrophages by TGF-β1.

Transforming growth factor (TGF)-␤1 1 is a growth factor with multiple effects on cell differentiation, growth, deposition of extracellular matrix, and immune modulation (11,12). With respect to immune modulation, definitive evidence for a role in immune regulation stems from targeted disruption of TGF-␤1 in mice. TGF-␤1-deficient mice exhibit a wasting syndrome accompanied by a multifocal, mixed inflammatory cell response and tissue necrosis (13,14). The number of circulating monocytes is elevated in these animals (13), and inflammatory infiltrates include large macrophages (14). Previous studies support a role for TGF-␤1 in inhibiting macrophage activation (15,16) as evidenced by the suppression of a number of activation markers including inducible nitric-oxide synthase (iNOS) (15,17), tumor necrosis factor ␣ (18), interleukin-1␤, scavenger receptor (16), and matrix metalloproteinases (19). However, the mechanism(s) by which TGF-␤1 inhibits macrophage activation have not been elucidated.
Recently, the Smad proteins have been identified as principal intracellular mediators of TGF-␤ signaling (20 -23). Three classes of Smad proteins, pathway restricted, common, and inhibitory, have been identified to date. The pathway restricted (r)Smads (e.g. Smad1, 2, 3, and 5) are serine/threonine kinaseactivated proteins that interact in an unphosphorylated state with a TGF-␤ superfamily receptor. Upon ligand binding they are phosphorylated by the receptor and released. These activated rSmads then hetero-oligomerize with Smad4, the only common Smad identified in mammals to date, translocate to the nucleus, and activate specific target genes. The rSmads phosphorylated by the TGF-␤ receptor include Smad2 and Smad3, whereas Smads 1 and 5 are substrates for the BMP receptor. Smad6 and Smad7 constitute the third group of Smads termed inhibitory or "anti-Smads." They diverge structurally from other members of the family and have been shown to act as inhibitors of Smad signaling pathways by interfering with the activation of rSmads (24 -26).
Most studies to date examining the function of Smad proteins in TGF-␤1 signaling have focused on genes that are transactivacted by TGF-␤1, including plasminogen activator inhibitor (PAI)-1 (27), collagen-I (28) and collagen-VII (29,30), and the cyclin-dependent kinase inhibitors p15 (31) and p21 (32,33). In contrast to TGF-␤1-mediated gene transactivation, the mechanism by which TGF-␤1 inhibits gene transcription is not well characterized. Because of the importance of activated macrophages in inflammatory disease states and the known suppressive effect of TGF-␤1 on macrophage activation, we sought to understand the mechanism by which TGF-␤1 suppresses the expression of genes critical for macrophage activation. We hypothesized that Smad proteins are the effectors of this TGF-␤1 function.
In this study, we chose iNOS and matrix metalloproteinase (MMP)-12 as markers of macrophage activation. iNOS promotes cytotoxic effects on invading microorganisms and is upregulated in response to endotoxin (LPS) (17,34). MMP-12 is a macrophage-specific metalloproteinase that is critical for the penetration of basement membranes by the macrophage (35) and is thought to play an important role in a number of chronic inflammatory disease states (6, 36 -39). We provide evidence that Smad3 is a critical effector responsible for the inhibition of macrophage gene activation by TGF-␤1. Our data suggest that competition for essential coactivators such as p300 may be an important determinant in regulating the degree of macrophage activation and inhibition by TGF-␤1.

EXPERIMENTAL PROCEDURES
Cell Culture and Reagents-RAW264.7 cells (American Type Culture Collection) were grown as described previously (19). Cells were seeded at a density of 2.5 ϫ 10 4 cells/ml. All cytokines were dissolved according to the instructions of the manufacturer and stored at Ϫ20°C until use. LPS was purchased from Sigma, and recombinant TGF-␤1 was from R & D Systems (Minneapolis, MN).
RNA Extraction and RNA Blot Analysis-Total RNA was isolated from cultured cells by guanidinium isothiocyanate extraction and centrifugation through cesium chloride (40). RNA was fractionated on a 1.3% formaldehyde-agarose gel and transferred to nitrocellulose filters. The filters were hybridized with 32 P-labeled, random-primed cDNA probes, washed, and exposed as described previously (19,41,42).
Western Blot Analysis-Cultured cells were lysed in RIPA sample buffer (40), and electrophoresis was performed under reducing conditions according to the method of Laemmli (43). Samples were resolved through 8% SDS-polyacrylamide gels and transferred to membranes according to the method of Towbin et al. (44). Blots were immunoblotted with 2 g/ml rabbit polyclonal anti-SMAD3 antibody (Zymed Laboratories Inc. Laboratory, Inc., San Francisco, Ca) and horseradish peroxidase-conjugated donkey anti-rabbit secondary antibody (1:4000; Amersham Pharmacia Biotech). Smad3 signal was visualized by enhanced chemiluminescence method (Amersham Pharmacia Biotech).
Generation of Reporter and Site-directed Mutagenesis Constructs and Probes-All promoter constructs were cloned into luciferase reporter plasmids. The generation of the pGL3 MMP-12 promoter construct (Ϫ1046/ϩ39) and of the pGL2 iNOS reporter construct (Ϫ1485/ϩ31) have been described previously (19,45).
To generate a SMAD3 construct with a mutation of Arg-74 to alanine (Smad3 R74A), we performed directed mutagenesis using polymerase chain reaction-based methods as described (40). In brief we first generated a 5Ј fragment of Smad3 containing the mutation in the antisense primer (underlined) 5Ј-ACCTGCAACGCGCCATCCAGGGACC-3Ј and a 3Ј fragment containing the mutation in the sense primer 5Ј-CGCGTT-GCAGGTGTCCCATCGGAAG-3Ј. These fragments were annealed, extended by Taq polymerase, and then used as a template in a second polymerase chain reaction using 5Ј-terminal and 3Ј-terminal primers of Smad3 to generate a full-length expression plasmid with the following mutation: Arg-74 3 Ala-74 (CGG 3 GCG).
Transient Transfections-RAW264.7 cells were transfected with Fugene TM 6 Transfection Reagent (Roche Molecular Biochemicals) on 6-well plates as described by the manufacturer. In brief, 2.5-3 g of total plasmid DNA was used in the experiments. Cells were treated with LPS (15 ng/ml), TGF-␤1 (2.5 to 10 ng/ml), or a combination of the two reagents 12 h after transfection. 24 h after stimulation cells were harvested for assays of luciferase and ␤-galactosidase. Luciferase activity was normalized to ␤-galactosidase activity (to correct for differences in transfection efficiency) by cotransfecting pCMV-␤gal plasmid (CLONTECH) (300 ng) in all experiments. All transfections were performed in triplicate from at least three independent experiments.
Generation of Stable Clones-To generate clones that stably expressed Smad3 constructs, RAW 264.7 cells were grown to approximately 50% confluency on 100-mm dishes (Falcon) and transfected with 10 g of DNA (SMAD3⌬c or vector, respectively) using Fugene TM 6. To select for transfectants, cells were treated with 500 g/ml G418 (Life Technologies, Inc.) starting 72 h after transfection. G418-resistant colonies emerged approximately 10 days after transfection and were pooled and expanded under reduced G418 levels (200 g/ml) for further analysis. Expression of mRNA and protein were confirmed by Northern analysis and Western analysis, respectively.

TGF-␤ Inhibits the Expression of Markers of Macrophage
Activation-We first assessed the mRNA expression of iNOS and MMP-12 in RAW264.7 cells in response to cytokine activation (Northern analysis; Fig. 1A). Both genes were not expressed in unstimulated cells and were induced by LPS treatment. Pretreatment with TGF␤-1 (30 min) potently inhibited this induction. Similar results were observed with 12-O-tetradecanoylphorbol-13-acetate as the stimulating agent (data not shown). We chose LPS as the inducing agent in subsequent experiments. Previous work has shown that activation of these genes occurs principally at the level of transcription (19,42,49). Consistent with these results, transient transfection assays with the iNOS and MMP-12 promoters showed an induction of transcriptional activity with LPS that was inhibited by pretreatment with TGF-␤1 (30 min; Fig. 1B).
Smad3 Mediates Inhibition of Macrophage Activation-To determine whether Smad proteins could mediate TGF-␤1 inhibition of the iNOS and the MMP-12 promoter, we performed transient transfection assays with a panel of Smad expression plasmids. Cotransfection with Smad3, but not Smad2 and Smad4, reproduced the inhibitory effect observed with TGF-␤1. This effect was seen with both the iNOS and the MMP-12 promoters. No effect was observed with Smad1. Consistent with the previously identified role as anti-Smads, Smad6 and Smad7 increased the transactivation of both promoters above the level achieved with LPS alone ( Fig. 2A and data not shown). To confirm proper function of our expression plasmids, we used the 3TPlux and the PAI-1 promoter as positive controls. These have previously been shown to be transactivated by TGF-␤ related Smads in other cell types (27). As expected, TGF-␤1, Smad3, Smad4, and a combination of Smad2 and Smad4 transactivated the 3TPlux promoter in RAW 264.7 cells (Fig. 2B). Similar results were observed with the PAI-1 promoter (data not shown).
To further establish the role of Smad3 in TGF-␤1 signaling, we performed experiments combining TGF-␤1 treatment with transfection of Smad3 expression vectors. Additional treatment with TGF-␤1 enhanced the suppressive effect of Smad3 on the iNOS and the MMP-12 promoter significantly (Fig. 3A). Previous studies have shown that a downstream NF-B site in the iNOS promoter and a proximal AP-1 site in the MMP-12 promoter are essential for the inducibility of the respective promoters by LPS (19,50). Promoter deletion constructs retaining these crucial elements and concatamers of NF-B and AP-1 binding sites showed the same pattern of inhibition by Smad3 and a combination of Smad3 and TGF-␤1 ( Fig. 3B and data not shown).
Dominant Negative Smad3 Rescues TGF-␤1 Inhibition of the iNOS Promoter-To establish more firmly that Smad3 was the downstream effector mediating inhibition by TGF-␤1, we performed transient and constitutive overexpression studies using a dominant negative Smad3 (Smad3⌬c). This construct lacks the the C-terminal 39 amino acids and thus the C-terminal phosphorylation and activation site (47). We hypothesized that overexpression of Smad3⌬c would antagonize the inhibitory effect of TGF-␤1.
In transient transfection assays, Smad 3⌬c partially antagonized the inhibitory effect of TGF-␤1 in the iNOS promoter (Fig. 4A). To examine the effect of Smad3⌬c on the expression of the endogenous gene, we generated RAW 264.7 cells that stably expressed Smad3⌬c. We first confirmed the expression of our construct by Northern analysis (data not shown) and Western analysis (Fig. 4b). Next we stimulated the stable clones with LPS in the presence and absence of TGF-␤1 pretreatment and examined mRNA expression of iNOS. LPS strongly induced iNOS expression in Smad3⌬c as well as vector-transfected cells. However, in contrast to the vector transfected cells, the inhibition of this induction by TGF-␤1 pretreatment was markedly attenuated in the Smad3⌬c-expressing cells (Fig. 4C). Taken together, these data suggest that Smad3 is a critical downstream effector in the inhibition of macrophage activation by TGF-␤1.
To confirm the ability of Smad3⌬c to interfere with TGF-␤1 signaling, we examined the mRNA expression of PAI-1 following TGF-␤1 treatment. TGF␤-1 induces PAI-1 expression in many cell types including macrophages (51), and a dominant negative Smad3 has been shown to interfere with the induction of the PAI-1 promoter and the PAI-1 derived 3TPlux concatamer in transient transfection assays (47). As expected, the induction of PAI-1 mRNA by TGF␤-1 was markedly reduced in the Smad3⌬c expressing clones in comparison to the vector transfected cells (Fig. 4D).
The MH-1 Domain of SMAD3 Is Essential for the Inhibitory Effect of the Molecule-To elucidate the mechanism by which Smad3, in contrast to Smad2, could inhibit macrophage activation, we next sought to identify critical protein domains for the inhibitory effect of Smad3. rSmads have two distinct functional domains, the MAD homology MH-1 domain at the N terminus and the MH-2 domain at the C terminus of the molecule.
To identify the critical domain for the inhibitory effect of Smad3, we first used chimeric constructs of Smad2 and Smad3 that combine the MH-1 domain of Smad2 with the MH-2 domain of Smad3 (Smad2/3) and vice versa (Smad3/2) (46). In transient transfection assays the Smad3/2 construct inhibited transactivation of the iNOS promoter to a similar degree as Smad3. In contrast, Smad2/3 had no inhibitory effect, like Smad2 (Fig. 5a). This result suggests that the MH-1 domain of Smad3 is critical for the inhibitory effect of the molecule.
The MH-1 domain of Smad3 mediates direct binding of Smad3 to DNA (27,52,53). Smad2, in contrast, cannot directly bind to DNA efficiently (20). To evaluate whether DNA binding of Smad3 is important for its inhibitory effect, we generated a mutated Smad3 construct replacing the arginine at position 74 in the MH-1 domain with an alanine (Smad3-R74A). This Arg-74 residue has previously been shown to be critically involved in DNA binding (52). Transient transfection assays with the mutated Smad3 construct showed that in comparison to wild-type Smad3, Smad3 R74A lost the ability to inhibit the iNOS promoter as well as to transactivate the 3TPlux promoter (Fig. 5, B and D). These data indicate that the ability of Smad3 to bind DNA through its MH-1 domain is critical for the inhibitory effect.
We next evaluated the possibility that the Smad3 MH-1 domain alone might confer the inhibitory effect. Transient transfection assays using an expression plasmid for the MH-1 domain of Smad3 showed that the MH-1 domain of Smad3 alone was unable to inhibit the iNOS promoter (Fig. 5C). The experiments were also performed with an MH-1 construct tagged with a nuclear localization signal with similar results (data not shown). Taken together, these results suggest that DNA binding through the MH-1 domain of Smad3 is essential but not sufficient for its inhibitory effect. The presence of an MH-2 domain either from Smad2 or Smad3 is requisite.
Coactivator Competition as a Mechanism for Smad3-medi-ated Suppression-We next sought to understand the requirement for the MH-2 domain of either Smad2 or Smad3 in mediating this effect. Previous studies have shown that NF-B and AP-1 driven promoters can be inhibited by limiting amounts of coactivators if these are recruited to other promoters (54,55). The MH-2 domains of both Smad 2 and 3 can bind to the coactivators p300/CREB-binding protein (56). To test whether this might be a potential mechanism in the case of TGF-␤1/Smad3-mediated inhibition of the iNOS promoter, we performed transient transfection assays overexpressing p300. Inhibition of the iNOS promoter both by TGF-␤1 and Smad3 was partially rescued by coexpression of p300 in a dose-dependent manner (Fig. 6A). The reverse experiment was also performed. As shown in Fig. 6B, cotransfection of p300 increased the transactivation of the iNOS promoter above the level achieved with LPS alone. This transactivation was inhibited by TGF-␤1 or Smad3 in a dose-dependent manner. These results indicate that recruitment of limiting coactivators such as p300 is important for the inhibitory effect of TGF-␤1/Smad3 on markers of macrophage activation.
To test the specificity of this mechanism, we generated a p300 expression plasmid containing only amino acids 1840 -1960 (p300⌬). This region has previously been shown to mediate the interaction of p300/CREB-binding protein with Smad3 (56). We hypothesized that this construct should specifically interact with activated Smad3, thus liberating endogenous p300 for transactivation of the inflammatory promoters. Indeed, we found that p300⌬ was able to alleviate the suppressive effect of TGF-␤1 and Smad3 on the iNOS promoter to a comparable degree to full-length p300 (Fig. 6C). As expected, this construct was not able to transactivate the 3TPlux promoter (data not shown).

DISCUSSION
Infiltration of activated macrophages into tissues is a key pathogenetic event in a number of inflammatory disease states. As such, identification of mechanisms limiting activation of this cell type is of interest. Previous studies have supported an important role for TGF-␤1 in inhibiting the activation of immune cells, including macrophages. However, the mechanism by which TGF-␤1 is able to carry out this function has not been elucidated.
In this report we provide evidence that Smad3 is a critical effector molecule for TGF-␤1-mediated inhibition of macrophage activation. In support, we found that only Smad3 was able to inhibit the expression of several markers of macrophage activation (Fig. 2). The addition of TGF-␤1 enhanced Smad3mediated inhibition (Fig. 3). Furthermore, transient and constitutive overexpression of a dominant negative Smad3 alleviated the inhibitory effect of TGF-␤1 (Fig. 4).
TGF-␤1 activates both Smad2 and Smad3, which, upon dimerization with Smad4, carry out effector functions (20,22). Smad2 and Smad3 are highly homologous proteins (83% amino acid identity). Thus it is noteworthy that Smad3, but not Smad2, was cabable of inhibiting promoters (iNOS and MMP-12) in macrophages. To investigate this difference we used chimeric constructs of Smad2 and Smad3 (Fig. 5A) and found that the MH-1 domain of SMAD3 was essential for inhibition. However, the MH-1 alone was not sufficient for this effect (Fig.  5B). The MH-2 domain of either Smad2 or Smad3 is also required for inhibition.
The principal function of the MH-1 domain is to mediate DNA binding. Crystal structure analyses have demonstrated that three highly conserved amino acid residues directly bind to nucleotides of the Smad consensus binding site (Arg-74, Gln-76, and Lys-81 for Smad3) (52). Previous studies have demonstrated that mutation of Arg-81 in Smad4 (the analogous residue to Arg-74 in Smad3) to an alanine abrogates DNA binding (52). Induction of the same point mutation at Arg-74 in Smad3 almost eliminated its inhibitory effect in our study, suggesting that DNA binding is important for Smad3-mediated inhibition.
Our data suggest that although the MH-1 domain of Smad3 FIG. 6. The coactivator p300 is critical for TGF-␤1 and Smad3mediated inhibition of transcriptional activity. RAW cells were transfected with the iNOS promoter and treated with LPS (15 ng/ml). A, addition of p300 expression plasmid (0.37-0.75 g) is able to partially rescue the transcriptional inhibitory effect of TGF-␤1 and Smad 3. *, p Ͻ 0.001 versus LPSϩ TGF-␤1 or LPS ϩ Smad3. B, TGF-␤1 (2.5 and 10 ng/ml) and Smad3 (0.25-1 g) are able to inhibit LPS-mediated induction of transcriptional activity in the presence of p300 in a dosedependent manner. *, p Ͻ 0.001 versus LPS ϩ p300. **, p Ͻ 0.01 versus LPS ϩ p300 ϩ low dose TGF-␤1 (2.5 ng/ml). C, a p300 fragment (p300⌬) containing only the p300 interaction domain with Smad3 was able to rescue the inhibitory effect of TGF-␤1 and Smad3 to a degree similar to full-length p300. *, p Ͻ 0.05 versus LPS ϩ TGF-␤1 or LPS ϩ Smad3. is essential for inhibition it is not sufficient for this effect (Fig.  5B). Thus, we investigated the mechanism by which the MH-2 domain may contribute to the inhibitory function. NF-B-, AP-1-, and Smad3-driven promoters all require p300/CREB-binding protein for their transactivation (56 -59). Previous studies have suggested that NF-B-and AP-1-driven promoters can be inhibited by competitive recruitment of coactivators such as p300/CPB to other unrelated promoters (54,55). We hypothesized that NF-B and AP-1 compete with Smad3 for limiting quantities of p300. This hypothesis predicts that added p300 should alleviate TGF-␤1/Smad3-mediated inhibition of inflammatory genes. Conversely, increasing doses of TGF-␤1/Smad3 would compete away even overexpressed p300 from NF-B/AP-1-driven promoters. Our findings support both of these predictions (Fig. 6, A and B). Consistent with this hypothesis, overexpression of a p300 fragment containing only the domain that mediates the interaction with Smad3 rescued the suppressive effect of TGF-␤1/Smad3 on inflammatory promoters in a similar degree as full-length p300 (Fig. 6C).
Unlike Smad2 and Smad4, Smad3 has an expression pattern that varies with tissue types, with the highest levels of expression in spleen and thymus (60), suggesting a role for Smad3 in immune regulation. This was confirmed in gene targeting experiments that show mild chronic inflammatory disease, impaired immune response, and activated T-cells (60 -62) in Smad3-deficient animals. In addition, a recent study reported that the chemotactic response to TGF-␤1 was affected in Smad3-deficient macrophages (48). Our study suggests that in addition to mediating chemotactic responses in macrophages, Smad3 may also be essential for the inhibition of activated macrophages. This expands on the established role of Smad3 in the regulation of immune cells and suggests that dysregulated macrophages may contribute to the inflammatory phenotype of Smad3 deficient mice.