Down-regulation of High Mobility Group-I(Y) Protein Contributes to the Inhibition of Nitric-oxide Synthase 2 by Transforming Growth Factor- b 1*

,

The inducible isoform of nitric-oxide synthase (NOS2) 1 catalyzes the production of nitric oxide (NO), which participates in the physiologic and pathophysiologic regulation of multiple organ systems (1)(2)(3)(4)(5). One disease process that is particularly affected by the overproduction of NO is sepsis (6,7). In sepsis, a severe underlying infection triggers a cascade of events that may lead to intractable hypotension, multiple organ system failure, and death (8,9). Within this cascade, inflammatory cytokines and vasoactive mediators promote the relaxation of vascular smooth muscle cells and a decrease in vascular tone. NO is a vasodilatory gas generated late in the sepsis cascade contributing to the hypotension. The overproduction of NO during sepsis is generated through the NOS2 pathway.
NOS2 is a highly inducible gene regulated at the level of gene transcription (1,10). Transactivation of the NOS2 gene is mediated by members of the nuclear factor (NF)-B family of transcription factors, which bind to a site in the downstream portion of the NOS2 5Ј-flanking sequence (Ϫ85 to Ϫ76) (11). Recently, we demonstrated that high mobility group (HMG)-I(Y) protein, an architectural transcription factor that binds to an AT-rich octamer site (Ϫ61 to Ϫ54) close to the NF-B binding site in the promoter of the NOS2 gene, acts in concert with p50 and p65 to facilitate NOS2 gene transactivation (12). Architectural transcription factors typically function by modifying the conformation of DNA, and thus provide a framework for the transcriptional machinery to operate. HMG-I(Y) does not drive transcription itself, but it facilitates the assembly of a functional nucleoprotein complex. Our studies revealed that binding of both HMG-I(Y) and NF-B subunits in the downstream 5Ј-flanking sequence of the NOS2 gene is essential for the most potent activation of the promoter (12). Moreover, we also demonstrated that HMG-I(Y) itself was up-regulated (both at the mRNA and protein levels) by inflammatory cytokines (13). Taken together, these studies suggest that HMG-I(Y) contributes to the transcriptional regulation of NOS2 by inflammatory mediators.
Previously, we and others have demonstrated that transforming growth factor (TGF)-␤1, a pleiotropic growth factor involved in a number of physiologic processes (14 -16), inhibited NOS2 expression in vitro (17)(18)(19) and in vivo (20). The down-regulation of NOS2, but not NOS3, in a rodent model of endotoxemia suggested that NOS inhibition by TGF-␤1 is more selective for the inducible isoform of NOS (NOS2) in an experimental model of sepsis (20). In vascular smooth muscle cells, inhibition of NOS2 by TGF-␤1 occurs at the level of gene transcription (19). However, the mechanism by which transcription of the NOS2 gene is inhibited by TGF-␤1 has not been fully elucidated.
We have shown previously that HMG-I(Y) is involved in the induction of NOS2 by inflammatory mediators (12,13). However, the ability of HMG-I(Y) to be regulated by inhibitors of NOS2 expression is not known. Furthermore, little is known about the down-regulation of HMG-I(Y) and its subsequent effect on NOS2 gene transcription. The present study was designed to determine (a) whether TGF-␤1, an inhibitor of NOS2 expression, also down-regulates the expression of HMG-I(Y), and (b) the functional importance of this TGF-␤1-mediated down-regulation of HMG-I(Y) on promoter activity of the NOS2 gene. Cell Culture-Rat aortic smooth muscle cells (RASMC) were harvested from male Harlan Sprague-Dawley rats (200 -250 g) by enzymatic dissociation according to the method of Gunther et al. (21). The cells were cultured in Dulbecco's modified Eagle's medium (JRH Biosciences, Lenexa, KS) supplemented with 10% fetal bovine serum (FBS, HyClone, Logan, UT), penicillin (100 units/ml), streptomycin (100 g/ ml), and 25 mM HEPES (pH 7.4) (Sigma) in a humidified incubator at 37°C. RASMC were passaged every 4 -5 days, and experiments were performed on cells 4 -6 passages from the primary culture. RAW 264.7 cells (mouse macrophage cell line from American Type Culture Collection) were grown in Dulbecco's modified Eagle's medium with supplements as described for RASMC, except HEPES was omitted. Peritoneal macrophages were obtained from C57/BL6 male mice (10 weeks of age) and cultured as described previously by Aggarwal and Mehta (22).

Reagents
Northern Blot Analysis-To assess the effect of cytokine treatment on HMG-I(Y) and NOS2 mRNA levels, cells were grown to 80% confluence in 100-mm Petri dishes (Becton Dickinson, Franklin Lakes, NJ) and fed with fresh medium containing either 10 ng/ml IL-1␤, 500 ng/ml LPS, 100 units/ml IFN␥, with or without 1 to 10 ng/ml TGF-␤1 (IL-1␤ treatment required a decrease in FBS to 2% to prevent cytokine inactivation). Dose-response experiments were performed to select the optimal concentrations of inflammatory mediators to induce or suppress NOS2 mRNA levels in RASMC or macrophages (data not shown). Cells were harvested after 24 h. The studies performed to assess the in vivo effects of LPS and TGF-␤1 treatments, along with experiments to select the appropriate doses, were described in detail elsewhere (20). Briefly, rats received LPS alone (4 mg/kg, intraperitoneally) or LPS followed by TGF-␤1 (20 g/kg, intraperitoneally). The rats were sacrificed 10 h later for organ harvest.
Total RNA was extracted from rat spleen and cultured peritoneal macrophages by guanidinium isothiocyanate extraction and centrifugation through cesium chloride gradient (23). Total RNA extraction from cultured RASMC and RAW 264.7 cells was performed with the Qiagen RNeasy midi kit (Qiagen, Valencia, CA). RNA was fractionated on a 1.3% formaldehyde-agarose gel and transferred to nitrocellulose filters. The filters were hybridized at 68°C for 2 h with 32 P-labeled mouse HMG-I(Y) (13) and rat NOS2 (19) probes. The hybridized filters were then washed in 30 mM sodium chloride, 3 mM sodium citrate, and 0.1% SDS solution at 55°C, and autoradiographed with Kodak XAR film at Ϫ80°C for 8 -12 h or stored on phosphor screens for 2-4 h. To correct for differences in RNA loading, the filters were washed in a 50% formamide solution at 80°C and rehybridized with a 32 P-labeled oligonucleotide probe complementary to 28 S ribosomal RNA. Images were displayed, and radioactivity was measured on a PhosphorImager running the ImageQuaNT software (Molecular Dynamics, Sunnyvale, CA).
Western Blot Analysis-RASMC and RAW 264.7 cells were plated in 150-mm tissue culture dishes (Becton Dickinson). When the cells were 80 -90% confluent, 10% FBS medium was replaced with 0.4% bovine serum medium. After 48 h, either vehicle, IL-1␤ (10 ng/ml, RASMC), or LPS (0.5 g/ml, RAW 264.7 cells) was added to the cells, in the presence or absence of TGF-␤1 (10 ng/ml). Cells were then collected in phosphatebuffered saline solution 48 h later. HMG-I(Y) protein was obtained from RASMC by acidic extraction followed by trichloroacetic acid precipitation (24). The resulting protein pellet was resuspended in 30 l of high pressure liquid chromatography-grade water and stored at Ϫ20°C until it was assayed. SDS-polyacrylamide gel electrophoresis (25) was performed by mixing 10 l of the resuspended acidic extract with 10 l of Laemmli sample buffer (2ϫ). After 5 min of boiling, the samples were resolved on a 15% polyacrylamide gel and transferred to polyvinylidene difluoride (Immobilon-P, Millipore, Bedford, MA) overnight (26). The membrane was blocked with 10% skim milk in 10 mM Tris-HCl, 150 mM NaCl, pH 8.0, 0.05% Tween-20 (TBS-T) for 1 h at room temperature. It was then incubated with an anti-HMG-I(Y) antibody (N19, Santa Cruz Biotechnology, Santa Cruz, CA) at 0.1 g/ml in 5% skim milk in TBS-T for 2 h at room temperature. After a wash in TBS-T, the membrane was incubated with anti-goat, horseradish peroxidase-conjugated IgG (Santa Cruz) under the same conditions used for the primary antibody. After a final wash in TBS-T, the blot was developed with an enhanced chemiluminescence kit (Amersham Pharmacia Biotech, Arlington Heights, IL). Because the extraction also led to the isolation of histone H1, under transfer conditions insufficient for removal of H1 from the gel, we used the histone H1 band to correct for differences in protein loading. After scanning, the intensities of the HMG-I(Y) band and the Coomassie Blue-stained histone H1 band were measured with IMAGE software (National Institutes of Health), and the ratio of the two values was used to represent the normalized intensity of the HMG-I(Y) band.
Plasmids-To evaluate the effect of IL-1␤ and TGF-␤1 on HMG-I(Y) transcription in vascular smooth muscle cells, plasmid ⌬180 was used (27). This construct, cloned into pCAT-Basic, contained the second transcription start site of HMG-I(Y) (27,28). Plasmid pGL2-Control contained the firefly luciferase gene driven by an SV40 promoter and enhancer.
To evaluate NOS2 promoter activity in either vascular smooth muscle cells or RAW 264.7 cells, we inserted 1485 bp of the 5Ј-flanking region of the mouse NOS2 gene and the first 31 bp after the transcription start site into pGL2-Basic to make iNOS(Ϫ1485/ϩ31), as described previously (29). To assess the effect of HMG-I(Y) overexpression on NOS2 after treatment with TGF-␤1 in both RASMC and RAW 264.7, we constructed the HMGIYpcDNA3 expression vector by cloning the coding sequence of mouse HMG-I(Y) cDNA into the EcoRI site of pcDNA3 (Invitrogen). DNA mass was normalized by adding the control plasmids as needed. Luciferase activity was normalized by cotransfecting pCMV-␤gal plasmid (CLONTECH).
Transfections-RASMC were transfected by a diethylaminoethyl (DEAE)-dextran method (29). In brief, 500,000 cells were plated onto 100-mm tissue culture dishes and allowed to grow for 48 -72 h (until 80 -90% confluent). Then plasmids ⌬180 and pGL2-Control (to correct for differences in transfection efficiency) were added (5 g each) to the RASMC in a solution containing 500 g/ml DEAE-dextran. RASMCs were subsequently shocked with 10% dimethyl sulfoxide solution for 1 min and then allowed to recover in medium containing 10% heatinactivated FBS. 12 h after transfection, RASMCs were placed in 2% FBS. RASMCs were then stimulated with vehicle, IL-1␤ (10 ng/ml) alone, or IL-1␤ plus TGF-␤1 (10 ng/ml). After 24 h, cell extracts were prepared by detergent lysis (Promega), and CAT assays were performed as described (30,31). Luciferase activity was measured with an EG&G AutoLumat LB953 luminometer (Gaithersburg, MD) and the Promega Luciferase Assay system to assess efficiency of transfection. The ratio of CAT to luciferase activity in each sample served as a measure of normalized CAT activity.
To investigate the effect of HMG-I(Y) overexpression on TGF-␤mediated inhibition of NOS2 promoter activity by inflammatory stimuli, RASMC or RAW 264.7 were plated in 6-well cell culture plates at 100,000/well and 70,000/well, respectively. The next day, cells were transfected as described above using 1 g/well iNOS(Ϫ1485/ϩ31), 0.8 g/well pCMV-␤gal, and increasing amounts of HMGIYpcDNA3 (1, 2, or 3 g/well). The corresponding empty vector was added when needed to keep the DNA amount constant throughout the experiment. 18 h after transfection, RAMC were treated with IL-1␤ (10 ng/ml) alone or IL-1␤ plus TGF-␤1 (0.5 ng/ml). RAW 264.7 were treated with LPS alone (0.5 g/ml) or LPS plus TGF-␤1 (1 ng/ml). After an additional 24 h, cells were harvested by detergent lysis (Promega). In the resulting cell extract, luciferase and ␤-galactosidase activity was assayed as described previously (32).
Electrophoretic Mobility Shift Assay (EMSA)-For the preparation of the nuclear protein extract (29), RAW 264.7 cells were grown to 80% confluence then stimulated with either vehicle, LPS (1 g/ml), or LPS (1 g/ml) plus TGF-␤1 (10 ng/ml). LPS-stimulated nuclear extract was subjected to electrophoretic mobility assay analysis using a doublestranded oligonucleotide probe encoding region Ϫ87 to Ϫ52 of the NOS2 5Ј-flanking sequence (TGGGGACTCTCCCTTTGGGAACAGTTAT-GCAAAATA). Binding reactions were performed in a 25-l volume containing 20,000 cpm of labeled probe, 10 g of nuclear extract, 100 ng of poly(dG-dC)⅐poly(dG-dC) (Sigma), 25 mM Hepes (pH 7.9), 50 mM KCl, 0.1 mM EDTA, 1 mM dithiothreitol, and 10% glycerol, with or without oligonucleotide competitors. Reactions were incubated for 20 min at room temperature, and DNA-protein complexes were analyzed by electrophoresis on a 5% native polyacrylamide gel in 0.25 ϫ Tris borate-EDTA (TBE) buffer at 4°C. To characterize specific DNA-binding proteins, we incubated nuclear extracts with anti-HMG-I(Y) affinitypurified antibody (4 g) (33) for 2 h at 4°C before the addition of probe.

TGF-␤1 Prevents the Induction of HMG-I(Y) and NOS2
mRNA by IL-1␤ in RASMC-Vascular smooth muscle cells are an important target of proinflammatory cytokines during sepsis, and their relaxation is an important determinant of the hypotension accompanying septic shock. Thus, we initially assessed the effect of TGF-␤1 on NOS2 and HMG-I(Y) mRNA levels after IL-1␤ stimulation in RASMC. Fig. 1 shows that TGF-␤1 not only prevented the induction of NOS2 mRNA by IL-1␤ (white bars), but it also prevented HMG-I(Y) induction (black bars). IL-1␤ induction of HMG-I(Y) mRNA was reduced by 96% after 24 h of TGF-␤1 treatment.
To investigate the mechanism by which TGF-␤1 decreased HMG-I(Y) mRNA levels after their induction by IL-1␤, we assessed HMG-I(Y) mRNA half-life and promoter activity (Fig.  2). To evaluate HMG-I(Y) mRNA half-life, RASMC were cultured to 70% confluence. Cells were then treated with either IL-1␤ alone or IL-1␤ plus TGF-␤1 for 24 h. Actinomycin D was then added to the cells, and total RNA was extracted as described at 0, 4, 8, 12, and 24 h. As reported previously (13), the half-life of HMG-I(Y) mRNA after IL-1␤ stimulation was ϳ12 h. The addition of TGF-␤1 did not reduce HMG-I(Y) mRNA half-life ( Fig. 2A). The effect of TGF-␤1 on promoter activity of the HMG-I(Y) gene was studied by transiently transfecting RASMC with plasmid ⌬180. Cells were then stimulated with vehicle, IL-1␤, or IL-1␤ plus TGF-␤1 for 24 h. TGF-␤1 completely prevented the induction of HMG-I(Y) promoter activity by IL-1␤ (Fig. 2B). Taken together, these results suggest that TGF-␤1 decreased the levels of HMG-I(Y) mRNA by acting at the level of gene transcription.
TGF-␤1 Prevents the Induction of HMG-I(Y) mRNA by LPS and IFN␥ in Macrophages-During sepsis, macrophages are activated by endotoxin and other bacterial products to produce a battery of mediators and proinflammatory cytokines (34). One of the main cytokines produced by macrophages is IFN␥ (34), which may act in an autocrine or a paracrine fashion. Both LPS and IFN␥ are potent inducers of NOS2 in macrophages (Fig. 3, white bars). For this reason we wanted to investigate the effects of LPS and IFN␥ on the expression of HMG-I(Y) in a mouse macrophage cell line (RAW 264.7 cells) and in resident peritoneal macrophages. LPS stimulation of RAW 264.7 cells (Fig. 3A) and peritoneal macrophages (Fig. 3B) increased HMG-I(Y) mRNA 4-and 11-fold, respectively (black bars). This increase in HMG-I(Y) message by LPS was decreased by TGF-␤1 in both cell types. NOS2 mRNA was also increased by LPS (14-fold in both RAW 264.7 and peritoneal macrophages) and suppressed by TGF-␤1 treatment (white bars). IFN␥ stimulation of RAW 264.7 cells induced HMG-I(Y) mRNA by 11-fold (Fig. 3C, black bars) and NOS2 mRNA by 24-fold (white bars). In peritoneal macrophages, the same cytokine induced a 5.5fold increase in HMG-I(Y) mRNA (Fig. 3D, black bars) and a 19-fold increase in NOS2 mRNA (white bars). TGF-␤1 treatment was able to abolish the induction of HMG-I(Y) and NOS2 mRNA by IFN␥ in both cell types.
Spleen is a macrophage-rich organ that responds to LPS administration in vivo with a dramatic up-regulation of NOS2 expression (Refs. 29 and 35, and Fig. 4A). We decided to use the spleen as a model to investigate the regulation of HMG-I(Y) in macrophages in vivo, in response to LPS or LPS plus TGF-␤1. LPS administration in rats caused a 4-fold induction of HMG-I(Y) transcript in the spleen (Fig. 4B). HMG-(Y) mRNA levels were markedly reduced in rats receiving LPS plus TGF-␤1 compared with levels in rats receiving LPS alone (Fig. 4B). This effect of TGF-␤1 on HMG-I(Y) mRNA levels was similar to the suppression of NOS2 mRNA levels in rats receiving LPS and TGF-␤1 (Fig. 4A).
TGF-␤1 Prevents the Induction of HMG-I(Y) Protein by IL-1␤ in RASMC and LPS in Macrophages-We next focused our alone (IL, 10 ng/ml), or IL-1␤ plus TGF-␤1 (ILϩT, 1 ng/ml). Total RNA was extracted 24 h after treatment, and Northern blot analysis was performed using 10 g of total RNA per lane. After electrophoresis, the RNA was transferred to nitrocellulose filters, which were hybridized with 32 P-labeled HMG-I(Y) and NOS2 probes. Filters were also hybridized with a 32 P-labeled oligonucleotide probe complementary to 28 S ribosomal RNA to control for differences in loading. The signal intensities were then plotted as -fold increases compared with vehicle signal (mean Ϯ S.E.) for HMG-I(Y) (black bars) and NOS2 (white bars). Each experiment was performed three times.

FIG. 2. Effect of TGF-␤1 on HMG-I(Y) mRNA half-life and promoter activity.
A, half-life of HMG-I(Y) mRNA. Cells were treated with IL-1␤ alone (IL, 10 ng/ml, filled circles) or IL-1␤ plus TGF-␤1 (ILϩT, 1 ng/ml, open circles) for 24 h. After this incubation period, actinomycin D (10 g/ml) was added to the cells, and total RNA was extracted at the indicated times (0, 4, 8, 12, and 24 h). Northern blot analysis was performed as described in Fig. 1. The normalized signal intensity was then plotted as a percentage of the 0-h value. The experiment was performed three times. B, effect of TGF-␤1 on IL-1␤-induced promoter activity of the HMG-I(Y) gene. A plasmid that contains the HMG-I(Y) promoter and drives a CAT reporter, ⌬180, was transfected transiently into RASMC. Cells were then stimulated with vehicle (V), IL-1␤ alone (IL, 10 ng/ml), or IL-1␤ plus TGF-␤1 (ILϩT, 1 ng/ml) for 24 h, after which the cell extracts were harvested. Normalized CAT activity is shown as the -fold induction from the activity of vehicletreated cells (mean Ϯ S.E., n ϭ 4 in each group). attention on the regulation of HMG-I(Y) protein by IL-1␤ in RASMC and by LPS in RAW 264.7 cells. We have shown previously that HMG-I(Y) protein is maximally induced after 48 h of cytokine stimulation in RASMC (13), thus this time point was used to assess the effect of TGF-␤1 on HMG-I(Y) protein induction. HMG-I(Y) protein levels were increased 7.5fold by IL-1␤ in RASMC (Fig. 5A) and 5.5-fold by LPS in RAW 264.7 cells (Fig. 5B). When TGF-␤1 was added, the induction of HMG-I(Y) protein levels was decreased by 83% (versus IL-1␤, Fig. 5A) in RASMC and by 70% (versus LPS, Fig. 5B) in RAW 264.7 cells.

HMG-I(Y) Overexpression Rescues the TGF-␤1-mediated Down-regulation of NOS2 Gene Transactivation in RASMC
and RAW 264.7 Cells-We hypothesized that down-regulation of HMG-I(Y) may play a role in mediating the inhibition of NOS2 expression by TGF-␤1 in response to inflammatory stimuli. To test this hypothesis, we transiently cotransfected RASMC and RAW 264.7 cells with promoter construct iNOS(Ϫ1485/ϩ31) and increasing concentrations of expression plasmid HMGIYpcDNA3 (0, 1, 2, 3 g/well as described under "Experimental Procedures"). Cells were then treated with either vehicle, IL-1␤ (RASMC, Fig. 6A), or LPS (RAW 264.7 cells, Fig. 6B), alone or in combination with TGF-␤1. Compared with cells receiving vehicle, the addition of IL-1␤ or LPS promoted a significant increase in promoter activity of the NOS2 gene in both RASMC (Fig. 6A) and RAW 264.7 cells (Fig. 6B), respectively. As expected, when vector-transfected cells (pcDNA3) were treated with TGF-␤1, NOS2 promoter activity in response to IL-1␤ (RASMC) and LPS (RAW 264.7 cells) was significantly reduced in RASMC (Fig. 6A) and RAW 264.7 cells (Fig. 6B). HMG-I(Y)-transfected cells showed a dose-dependent recovery of NOS2 promoter activity by both IL-1␤ (Fig. 6A) and LPS (Fig. 6B). At a dose of 2 g/well HMG-I(Y) expression plasmid, the responsiveness of the NOS2 gene promoter activity was almost completely restored in both RASMC (89% of IL-1␤ alone value, Fig. 6A) and RAW 264.7 cells (90% of the LPS alone value, Fig. 6B) despite the presence of TGF-␤1. To determine whether HMG-I(Y) could restore other TGF-␤1 down-regulated genes, we performed the same experiment using the promoter for the matrix metalloproteinase-12 gene (36). Overexpression of HMG-I(Y) did not rescue the TGF-␤1-mediated down-regulation of matrix metalloproteinase-12 promoter activity (data not shown). These data suggest that recovery of NOS2 promoter activity by HMG-I(Y) was a specific effect.
Effect of TGF-␤1 on HMG-I(Y) Binding to the Promoter of the NOS2 Gene-We investigated the binding of HMG-I(Y) protein to the downstream promoter sequence (region Ϫ87 to Ϫ52 of the NOS2 5Ј-flanking region, TGGGGACTCTCCCTTTGG-GAACAGTTATGCAAAATA) of the NOS2 gene by using nuclear extracts obtained from RAW 264.7 cells stimulated with LPS in the presence or absence of TGF-␤1 for 4, 24, and 48 h. Following LPS stimulation, we observed the appearance of a high molecular weight band that exhibited a clear induction at 24 and 48 h after LPS (band 1, Fig. 7A). A second band was also evident (band 2). Band 2 decreased in intensity in a time-dependent manner. The specificity of the two bands was evalu-  Fig. 1. The signal intensities were then plotted as -fold increases compared with vehicle signal (mean Ϯ S.E.) for HMG-I(Y) (black bars) and NOS2 (white bars). Each experiment was performed three times.

FIG. 4. Effect of TGF-␤1 on LPS-induced NOS2 and HMG-I(Y) mRNA in vivo.
Conscious male Sprague-Dawley rats were injected with vehicle (V, n ϭ 2), LPS alone (L, 4 mg/kg ip, n ϭ 2), or LPS plus TGF-␤1 (LϩT, 20 g/kg, n ϭ 2). The rats were killed 10 h after LPS administration, and total RNA was extracted from spleen tissue. Northern blot analysis was performed as described in Fig. 1. The signal intensities were then plotted as -fold increases compared with vehicle signal (mean Ϯ S.D.) for NOS2 (A) and HMG-I(Y) (B). ated by adding a 100-fold molar excess of either the unlabeled probe (identical competitor, IdC) or an unrelated oligonucleotide (nonidentical competitor, NIdC). As shown in Fig. 7B, band 1 was abolished by the identical but not the unrelated competitor. Band 2 was abolish by both competitors. These data indicated that band 1 was the only specific band. To investigate whether HMG-I(Y) was present in band 1, the nuclear extracts from cells stimulated with LPS for 48 h were submitted for EMSA in the presence of either a specific anti-HMG-I(Y) antibody or a control antibody (Fig. 7C). Band 1 supershifted in the presence of anti-HMG-I(Y) antibody, but not control antibody, indicating that HMG-I(Y) was present in the nucleoprotein complex binding to the NOS2 promoter.
Finally, we studied the binding of HMG-I(Y) to the promoter of the NOS2 gene in RAW 264.7 cells that were treated with LPS alone or LPS plus TGF-␤1 (Fig. 7D). TGF-␤1 did not influence the binding at 4 and 24 h, as shown by the comparable intensities of the bands. In addition, the anti-HMG-I(Y) antibody was still able to supershift band 1 in the TGF-␤1treated nuclear extracts (data not shown), indicating that TGF-␤1 was not affecting HMG-I(Y) DNA binding properties at these time points. In contrast, the intensity of band 1 was dramatically decreased after 48 h of treatment with LPS plus TGF-␤1 compared with LPS alone. Interestingly, the Western blot analysis showed that, at the same time point (48 h), HMG-I(Y) protein content was dramatically decreased by TGF-␤1 (Fig. 5B). These data suggest that the decreased intensity of band 1 after 48 h of TGF-␤1 was related to a decrease in HMG-I(Y) protein not an alteration in binding affinity of the HMG-I(Y) protein. DISCUSSION HMG-I(Y) is a nonhistone chromosomal protein that, in conjunction with transcription factors, including NF-B, ATF-2/c-Jun, and IRF-1, assembles nucleoprotein complexes important for the transcriptional regulation of genes (37,38). In particular, HMG-I(Y) has been shown to be involved in the regulation of cytokine-responsive genes that are critical for the mediation of inflammation (39 -45). Previously, we showed that HMG-I(Y) cooperates with NF-B in the transcriptional regulation of NOS2 by inflammatory cytokines and endotoxin in vascular smooth muscle cells and macrophages (12). Moreover, HMG-I(Y) is up-regulated by IL-1␤ and endotoxin in vitro and in vivo, respectively (13). These data suggest that inflammatory mediators increase the expression of HMG-I(Y) protein, while promoting nuclear translocation of NF-B at the same time, resulting in maximal activation of NOS2 gene transcription. Thus, up-regulation of HMG-I(Y) appears to play an important role in the expression of NOS2 during inflammation.
The inflammatory response is a key component of host defense, but excessive activation of the immune system and subsequent release of vasoactive mediators (such as occurs in sepsis) may be fatal (8,46). The inflammatory process must therefore be regulated tightly in vivo. The effects of proinflammatory cytokines are counterbalanced by the production of anti-inflammatory mediators. Among these anti-inflammatory factors, TGF-␤1 has been shown to decrease hypotension and LPS-induced mortality in rats when administered exogenously (20). TGF-␤1 is also known to decrease macrophage responsiveness to LPS (47), and TGF-␤1 can prevent NOS2 expression in response to proinflammatory cytokines (5,17,19,48). The effects of TGF-␤1 are mediated, in part, through Smad proteins that are essential components in the signaling pathways of TGF-␤1 receptors (49). However, the effector(s) downstream of the TGF-␤1 receptors responsible for transcriptional downregulation of the NOS2 gene is(are) not known.
In this study, TGF-␤1 prevented the IL-1␤ induction of . The corresponding empty vector (pcDNA3) was added to keep the total DNA content constant. After the transfection, RASMC were treated with IL-1␤ (10 ng/ml) in the presence (ϩ) or absence (Ϫ) of TGF-␤1 (0.5 ng/ml). RAW 264.7 cells were treated with LPS (0.5 g/ml) in the presence (ϩ) or absence (Ϫ) of TGF-␤1 (1 ng/ml). Normalized luciferase activity was plotted as the -fold induction from the activity in cells receiving no IL-1␤/LPS, no TGF-␤1, and no HMG-I(Y). Each experiment was performed at least twice, with a total n ϭ 6.
HMG-I(Y) and NOS2 mRNA in RASMC (Fig. 1). The effect on HMG-I(Y) mRNA levels did not involve a modification of transcript stability ( Fig. 2A) but was related to a decrease in HMG-I(Y) promoter activity in the presence of TGF-␤1 (Fig. 2B). This effect of TGF-␤1 on HMG-I(Y) mRNA levels was not restricted to vascular smooth muscle cells, because TGF-␤1 down-regulated the HMG-I(Y) message (similar to the NOS2 message) in macrophages stimulated with inflammatory cytokines and endotoxin in vitro (Fig. 3) and in vivo (Fig. 4). In agreement with mRNA levels, TGF-␤1 also down-regulated HMG-I(Y) protein in vascular smooth muscle cells (Fig. 5A) and macrophages (Fig. 5B) induced by IL-1␤ and endotoxin, respectively. These results provide the first evidence that TGF-␤1 can decrease HMG-I(Y) expression driven by inflammatory stimuli.
We have suggested previously that induction of HMG-I(Y), and subsequent transactivation of the NOS2 gene, may contribute to a reduction in vascular tone during sepsis and other inflammatory disease processes (12,13). In addition, we have shown in the present study that HMG-I(Y) and NOS2 expression are down-regulated by TGF-␤1 in a comparable manner after stimulation with inflammatory mediators. Thus, we hypothesized that HMG-I(Y) content may be a limiting factor for NOS2 induction by endotoxin and proinflammatory cytokines. To test this hypothesis, we overexpressed HMG-I(Y) in vascular smooth muscle cells and macrophages in the presence of TGF-␤1 to determine whether HMG-I(Y) could restore inflammatory cytokine and endotoxin inducibility of the NOS2 gene promoter. In both RASMC (Fig. 6A) and RAW 264.7 cells (Fig.  6B), overexpression of HMG-I(Y) was able to restore cytokine inducibility of the NOS2 promoter that was suppressed by TGF-␤1. Taken together, these data suggest that factors (such as TGF-␤1) that decrease the amount of endogenous HMG-I(Y) in cells limit the inducibility of NOS2 by inflammatory mediators.
HMG-I(Y) has the potential to be post-translationally modified. In vitro experiments have shown that phosphorylation of HMG-I(Y) can decrease its affinity for DNA (50,51). Thus, to determine whether HMG-I(Y) binding to the promoter of the NOS2 gene was altered by TGF-␤1, we performed EMSA. LPS induced the formation of a specific, high molecular weight complex in macrophages (Fig. 7, A and B). Supershift experiments using an anti-HMG-I(Y) antibody revealed the presence of HMG-I(Y) in this complex (Fig. 7C). We have shown previously that NF-B binding to this portion of the NOS2 5Јflanking sequence (12) is important for NOS2 gene transactivation; however, a direct effect of TGF-␤1 on NF-B binding has not been demonstrated (29). Thus, we wanted to determine whether an alteration in HMG-I(Y) binding by TGF-␤1 might be responsible for the suppression in NOS2 gene transactivation. TGF-␤1 treatment did not modify the intensity of the specific HMG-I(Y) binding complex after 4 and 24 h (Fig. 7D,  lanes 2 and 4), suggesting that DNA binding of HMG-I(Y) was not modified by TGF-␤1. In contrast, the dramatic decrease in intensity of this complex after 48 h of TGF-␤1 treatment (Fig.  7D, lane 6) corresponded with a decreased in HMG-I(Y) protein content as shown by Western blot analysis (Fig. 5). Taken together, these data suggest that a down-regulation in HMG-I(Y) protein, not a disruption in HMG-I(Y) binding, is accountable for the decreased nucleoprotein complex responsible for transactivation of the NOS2 gene by inflammatory mediators.
The present study provides further insight into the mechanism by which TGF-␤1 is able to down-regulate NOS2 gene transactivation by inflammatory mediators. Our data underscore the importance of HMG-I(Y) in contributing to nucleoprotein complex formation for driving transcription of the NOS2 gene. Moreover, we show that inhibition of HMG-I(Y) transcription and subsequent decrease in HMG-I(Y) protein by TGF-␤1 significantly decreases NOS2 promoter activity, which can be restored by exogenous administration of HMG-I(Y). These data confirm the importance of HMG-I(Y) in regulating NOS2 gene transactivation by inflammatory mediators in vascular smooth muscle cells and macrophages, cell types involved in the pathophysiology of sepsis and other systemic inflammatory diseases. FIG. 7. Effect of TGF-␤1 on LPS-induced nuclear protein complex containing HMG-I(Y). EMSA were performed with a 32 P-labeled oligonucleotide probe from region Ϫ87 to Ϫ52 of the NOS2 5Ј-flanking sequence, and nuclear extracts from RAW 264.7 cells. A, cells were treated with vehicle (Ϫ) or LPS (ϩ, 0.5 g/ml) for increasing periods of time (as indicated) prior to extraction of nuclear extract. Arrowheads denote two bands (1 and 2) consistent with DNA-protein complexes. B, EMSA was performed on nuclear extract from LPS-treated (0.5 g/ml) cells for 48 h. Unlabeled competitors were added at a 100-fold molar excess. IdC, identical competitor; and NidC, unrelated competitor. Band 1 denotes a specific DNA-protein complex. C, nuclear extracts from cells stimulated with LPS (0.5 g/ml) for 48 h. EMSA was performed in the presence (ϩ) of an HMG-I(Y) antibody [␣-HMG-I(Y)] or a control antibody. The arrowhead denotes specific band 1. The asterisk indicates supershifted complex. D, nuclear extracts were harvested from cells stimulated with LPS (0.5 g/ml) for increasing periods of time as indicated in the presence (ϩ) or absence (Ϫ) of TGF-␤1 (10 ng/ml).