monocyte chemoattractant protein-1 (MCP-1) is a member of the CC chemokines and is found mainly in monocytes, endothelial cells, smooth muscle cells, and fibroblasts (23, 28). A bulk of evidence suggests that the inflammatory response in vascular injury involves recruitment and activation of monocytes through activation of the MCP-1 (17), a potent chemotactic factor for monocytes. The view that atherosclerosis is indeed a chronic inflammatory disease initiated by monocyte adhesion to activated endothelial cells (ECs) is now widely accepted (20, 42). In both human and experimental animal models, expression of MCP-1 and its mRNA level has been found to be markedly elevated in atherosclerotic lesions (60). In contrast, a reduced atherosclerotic lesion was found in the mice that are deficient in MCP-1 or its receptor CCR2 (4, 19). In vascular ECs, MCP-1 expression and release is activated by various stimuli such as oxidized low-density lipoprotein (oxLDL) (10, 42); inflammatory cytokines including tumor necrosis factor (TNF)-α (35), IL-1β (46), and IL-4 (41); some bacterial product such as lipopolysaccharide (LPS) (3), platelet-derived growth factor (PDGF) (26), and interferon-γ (IFN-γ) (51). Many studies have been devoted to understand how MCP-1 expression is regulated in activated ECs. Among them, redox regulation is probably the most important. It has been reported that reactive oxygen species (ROS), such as O₂^·− and H₂O₂ derived from Rac1-activated NADPH oxidase, upregulated TNF-α-induced MCP-1 expression, whereas the antioxidants, such as pyrrolidine dithiocarbanate and N-acetylcysteine, significantly inhibited the IL-4-induced MCP-1 mRNA expression in human vascular ECs (6, 25). Depletion of the endogenous Nox4 by transfection of small interfering RNA (siRNA) for Nox4 in human aortic ECs (HAECs) resulted in a failure to induce ROS generation and subsequent expression of MCP-1 in response to LPS stimulation (37). In addition, many other factors were also reported to modulate MCP-1 gene expression. For example, estrogen, such as estradiol, downregulates atherosclerosis by suppressing vascular MCP-1 expression in human coronary artery ECs (44) and in vivo (38). The insulin-sensitizing drug troglitazone significantly inhibited the TNF-α-induced MCP-1 secretion in human ECs (35), and transforming growth factor TGF-β1 inhibits LPS-mediated induction of MCP-1 in macrophages (16). However, no investigation on the possible role of thioredoxin (Trx)/thioredoxin reductase system, which plays an important role in regulation of cellular redox status, in regulating MCP-1 gene expression has been reported, except for the finding that extracellular Trx inhibits chemotactic effect of MCP-1 on monocytes (36) and neutrophils (31).

Trx, a small, ubiquitous thiol protein, is one of the most important regulators of reduction-oxidation balance and, thus, redox controls cell functions. It reduces oxidized cysteine groups in proteins through an interaction with its redox-active center Cys-Gly-Pro-Cys. The oxidized Trx is then reduced by thioredoxin reductase using NADPH as electron donor (58). Now, three isoforms of human Trx encoded by separate genes have been identified. Trx1 is a 104-amino acid protein that is typically found in the cytoplasm but has also been identified in the nucleus of normal endometrial stromal cells, tumor cells, and primary solid tumors (39). In contrast, Trx2 is a 166-amino acid protein that contains a 60-amino acid NH₂-terminal translocation sequence that directs it to the mitochondria (63). The third isoform SpTrx is a variant that is highly expressed in spermatozoa (29). The processes influenced by Trx include the control of cellular redox balance (1, 39), promotion of cell growth (40), inhibition of apoptosis (43), and modulation of inflammation (27). Not surprisingly, Trx certainly plays a role in a wide range of human diseases and conditions including cancer (40), viral disease (32), ischemia-reperfusion injury (48), cardiac conditions (62), aging (61), premature birth, and newborn physiology (12). In human coronary atherosclerotic specimens, it was interestingly found that Trx expression was enhanced throughout the vessel wall, and the greatest increases were observed in ECs and infiltrating macrophages within the neointimal plaques (47). The high expression of Trx in the atherosclerotic plaques suggests that Trx may play a role in the pathogenesis of atherosclerosis. However, up to now, there is no solid evidence to show whether Trx really protects the vascular endothelium from forming atherosclerotic plaque. For this reason, we established an EC mode, the human endothelial-like EA.hy 926 cells (15) either overexpressing Trx1 or having their endogenous Trx1 knocked down, and we investigated the questions of does and how the Trx1 modulate the expression of MCP-1 in the human endothelial-like cells.

MATERIALS AND METHODS

Reagents.

Diphenyleneiodonium (DPI), 5,5′-dithiobis (2-nitrobenzoic acid) (DTNB), and phenylmethanesulfonyl fluoride (PMSF) were purchased from Sigma. 5-Carboxy-2′,7′-dichlorodihydrofluorescein diacetate (carboxy-H₂DCF-DA) and G-418 sulfate were obtained from Molecular Probes. Dihydroethidium (DHE) was obtained from Beyotime Institute of Biotechnology (Shanghai, China). The bovine thioedoxin reductase was a gift from Professor A. Holmgren (Karolinska Institute, Sweden). Rabbit polyclonal Trx1 (FL-105) antibody, goat anti-actin antibody, rabbit polyclonal c-Jun (N) antibody, and mouse monoclonal redox factor-1 (Ref-1) antibody were purchased from Santa Cruz Biotech. Monoclonal anti-human MCP-1 antibody and goat anti-human MCP-1 biotinylated antibody were obtained from R&D Systems. Recombinant human MCP-1 and rabbit polyclonal histone H3 antibody were purchased from Peprotech and Abcam, respectively. The pGL3-based human MCP-1 full-length promoter-luciferase reporter and pGL3 plasmid were kind gifts from Dr. Remick (Dept. of Pathology, University of Michigan Medical School). The Trx1 expression vector pcDNA3-Trx1 and the Nox1 expression vector pcDNA3-Nox1 were kindly provided by Dr. J. Yodoi (Institute for Virus Research, Kyoto University) and Dr. J. D. Lambeth (Emory University Medical School, Atlanta, GA), respectively. The AP-1 luciferase reporter pAP-1-Luc was obtained from Stratagene. LightShift Chemiluminescent Electrophoretic Mobility Shift Assay (EMSA) Kit, NE-PER Nuclear and Cytoplasmic Extraction Reagents, and BCA assay Kit were purchased from Pierce. All other reagents were of analytic grade.

Cell culture and construction of a stable Trx1-transfected cell line.

The EA.hy 926 cell line was provided by Dr. C. J. Edgell (University of North Carolina, Chapel Hill, NC). Bovine aortic ECs (BAECs) in third passage were scraped off with a razor blade from bovine aorta according to previous literature (54). Both EA.hy 926 and BAEC cells were cultured in DMEM [containing 10% fetal bovine serum (FBS), 100 U/ml penicillin, and 100 μg/ml streptomycin] in a CO₂ incubator at 37°C. The Trx1-overexpressing EA.hy 926 cells (referred as EA-Trx1 cells) and their controls (referred as EA-neo cells) were established by transfection of pcDNA3-Trx1 containing Trx1 plus a neomycin-resistant gene and pcDNA3 containing only a neomycin-resistant gene, respectively, into cells with VigoFect transfection reagent (Vigorous Biotechnology, Beijing, China), selection with 400 μg/ml G418 for 4 wk, and screening by immunoblot analysis for Trx1 expression.

Assay of Trx activity in cells.

Cellular Trx1 activity was determined by an insulin reduction-based assay (2). In brief, cells were lysed with lysis buffer (50 mM KH₂PO₄, 1% Triton X-100, 2 mM EDTA, and 1 mM PMSF, pH 7.5). After sonication on ice, the cell lysate was centrifuged at 15,000 g for 30 min at 4°C. The supernatant was heated at 55°C for 5 min and filtered in a centrifugal filter with a 10-kDa cut-off. The retentate was washed three times with PE buffer (50 mM KH₂PO₄ and 2 mM EDTA, pH 7.5) and used as a crude cell extract. Trx1 content was measured by Trx1-dependent insulin reduction, in which Trx1 in the cell extract was first reduced by bovine thioredoxin reductase 1 (TrxR1) and then in turn reduced insulin. Numbers of exposed sulfhydryl groups in the reduced insulin were quantified by 5,5-dithio-bis-2-nitrobenzoic acid (DTNB) reduction. In the assay, 15 μl of cell extract and 10 μl of bovine TrxR1 (50 A₄₁₂/ml·min) were added to 95 μl of reaction mixture containing 100 μM insulin, 0.2 mM NADPH, 25 mM KH₂PO₄, and 5 mM EDTA in 0.1 M HEPES buffer (pH 7.6). After incubation at 37°C for 10 min, the reaction was terminated with 500 μl of 8 M guanidine-HCl in 50 mM HEPES (pH 7.6) containing 1 mM DTNB. Trx1 content in cells was measured as absorbance of DTNB at 412 nm and quantified using a calibration curve obtained by the same procedure except for that purified human wild-type Trx1 of various concentrations replaced the cell extract. A blank prepared by adding cell extracts and TrxR1 to the reaction mixture that was preincubated with DTNB-guanidine-HCl solution was used for each measurement. Cellular content of Trx1 is expressed as net nanograms of Trx1 per microgram of protein.

Knockdown of Trx1 by siRNA.

The small interfering RNA (siRNA) against Trx1 was used as 5′-AUGACUGUCAGGAUGUUGCdTdT-3′ (50). The scramble oligonucleotide 5′-UUCUCCGAACGUGUCACGUTT-3′ was used as negative control (mock RNA). Cells were seeded in six-well plates and cultured overnight, transfected with 150 nM siRNA or scramble oligonucleotide, respectively, for 6 h using VigoFect, and then incubated with fresh medium for 54 h. All experiments with Trx1-knockdown cells were performed 60 h after transfection.

Preparation of oxidized human low-density lipoprotein.

Human LDL was prepared from healthy blood plasma as described previously (9). The LDL was oxidized (oxLDL) by CuSO₄ according to previous literature (45).

ELISA assay of MCP-1 protein release from cells.

EA.hy 926 cells (1×10⁵) were seeded in each well of a 24-well plate (Corning) with 0.5 ml medium. The cells were starved in low-serum DMEM (0.2% FBS) overnight before stimulation. After stimulation with oxLDL for 6 h, supernatant of each culture medium was collected, and then the MCP-1 concentration in the supernatant was measured by using an enzyme-linked immunosorbent assay (ELISA). In brief, each well of a 96-well EIA/RIA plate was coated with 50 μl of the monoclonal anti-human MCP-1 antibody (1 μg/ml in PBS) and incubated for 48 h at 4°C. After aspiration and three washes with washing buffer (0.125% Triton X-100/PBS) were completed, the plate was blocked by adding 50 μl of casein in PBS (Pierce) in each well and then incubated for 1 h. After additional aspiration and wash again, fivefold diluted supernatants were added to the plate's wells and incubated for 2 h. After further aspiration and wash, 50 μl of biotinylated goat anti-human MCP-1 antibody was added to each well and incubated again for 1.5 h. After repeated aspiration and wash, the plate contents were incubated with horseradish peroxidase-conjugated streptavidin for 0.5 h. After three final washes, 100 μl substrate buffer containing 0.9 mM H₂O₂ and 0.5 mg/ml tetramethylbenzidine were added to each well to develop color and then stopped by adding 100 μl of 1.5 M H₂SO₄. The optical density of each well was determined within 30 min with a microplate reader (Thermo Labsystems) at 450 nm. The MCP-1 concentration was finally quantitated using calibration curve obtained with recombinant human MCP-1.

Immunoblotting.

Cells were lysed in lysis buffer (20 mmol/l Tris, 150 mmol/l NaCl, 1 mmol/l EDTA, 1 mmol/l EGTA, 1% Triton X-100, 2.5 mmol/l sodium pyrophosphate, 1 mmol/l h-glycerolphosphate, 1 mmol/l Na₃VO₄, 1 μg/ml leupeptin, and 1 mmol/l phenylmethylsulfonyl fluoride, pH 7.5) on ice for 15 min. Cell debris was removed by centrifugation at 18,000 g at 4°C for 15 min, and the nuclear protein extracts were obtained from the cell lysates with NE-PER Reagents. Protein contents of the cell lysates or nuclear extracts were determined using Bio-Rad protein assay kit. Cell lysates or nuclear extracts with equal protein content were then loaded and separated by 12% SDS-PAGE. The protein bands were electrotransferred onto nitrocellulose membrane and blocked with 5% nonfat milk in TBS/T buffer (20 mmol/l Tris base, 135 mmol/l NaCl, 0.1% Tween 20, pH 7.6) for 1 h. After incubation of the membrane with the appropriate antibodies, i.e., anti-Trx1, anti-MCP-1, anti-c-Jun, anti-Ref-1, and anti-histone H3 antibody, for at least 4 h, specific protein bands were visualized with SuperSignal West Pico Chemiluminescent Substrate. As an internal control, the actin contents in the samples were also immunoblotted using polyclonal anti-actin antibody as primary antibody.

Detection of the ROS generation in cells.

Two methods were used to detect ROS generation in cells. For Trx1-overexpressing cells, carboxy-H₂DCFDA was used to detect the intracellular ROS generation (49). Briefly, 1×10⁶ cells were seeded in each well of a six-well plate. The cells were cultured overnight and then loaded with 10 μM carboxy-H₂DCFDA for 40 min. After the ROS indicator in medium was washed off, the cells were either left unstimulated or stimulated with 100 μg/ml oxLDL or 100 μg/ml LDL for 30 min. After three additional washes, the cells were lysed in 1 M NaOH on ice for 5 min. The lysates were then cleared up by centrifugation at 14,000 g at 4°C for 10 min. Fluorescence of the oxidized carboxy-H₂DCF (DCF) in the supernatants was determined at the excitation of 480 nm and emission of 520 nm on a Fluorescence Spectrophotometer F-4500 (Hitachi) as the measure of the intracellular ROS generation during the 30-min culture with or without the indicated stimulator.

DHE was used to detect the intracellular ROS generation (30) in Trx1-knockdown cells, and the cells were transfected with scramble oligonucleotide (mock RNA). Briefly, after 100 μg/ml oxLDL stimulation, the cells were washed with serum-free culture medium and incubated with 5 μM DHE at 37°C for 30 min. After washing with PBS was completed, the ethidium formed by oxidation of DHE in 1.5×10⁴ cells was analyzed on a BD FACS Caliber flow cytometer.

EMSA of the DNA-binding activity of AP-1.

Nuclear extracts were isolated using NE-PER Reagents, and the protein contents in nuclear extracts were determined with BCA assay Kit. Biotinated double-stranded oligonucleotide containing the AP-1 binding site 5′-biotin-TTCCGGCTGACTCATCAAGCG-3′ was used as probe for specific binding of AP-1. Unlabeled double-stranded oligonucleotide was used as competitor. In performing EMSA, 10 μg nuclear extracts were added 10 min early, and then 130 fmol of probe were added in 15 μl of the Pierce-provided 1× binding buffer containing additional 5 mM MgCl₂, 2.5% (vol/vol) glycerol, 50 ng of poly dI/dC, and 0.05% Nonidet-40 and then kept on ice for 30 min. In competition experiments, molar excess (40 pmol) of the unlabeled oligonucleotide was added in the reaction mixture before adding biotin-labeled probe. The formed protein-DNA complexes were separated by 6.5% native polyacrylamide gel and electrotransferred onto nylon membrane. The DNA was then cross-linked to the membrane by UV illumination. The biotin-labeled DNA bands were visualized using LightShift Chemiluminescent EMSA Kit.

Assay of transcription with luciferase reporter.

The luciferase reporter pMCP-1-luc containing full-length MCP-1 promoter and AP-1 luciferase reporter pAP-1-luc containing 7× AP-1 binding sites was used for transcription activity assay of MCP-1 promoter and transcriptional activity of AP-1, respectively, in cells under various conditions. In the assay, EA.hy 926 cells (for effect of ROS on MCP-1 transcription) or BAEC cells (for effect of Trx1 on AP-1 activity) were plated on 12-well plates at a density of ∼80% confluence and transfected in the same medium using VigoFect (for EA.hy 926) or jetPEI (Polyplus) (for BAECs). In each transfection, cells were transfected with various combinations of a luciferase reporter plasmid and desired expression vectors (The amount of each expression vector was equal if it is not specified) as well as 100 ng of CMV-β-galactosidase expression vector as internal control. Total DNA quantity was kept constant with pcDNA3 or relevant empty vectors. Twelve hours after transfection, the cells were washed and cultured in fresh medium for 24 h and then harvested. Both luciferase and β-galactosidase activities were measured with a home-made luminometer and a Bio-Rad plate Reader, respectively. The luciferase activity normalized against the β-galactosidase activity represents the transcription efficiency of the reporter gene. All transfections throughout the investigation were performed in triplicate, and each assay was repeated three times.

Statistics.

All data obtained in this study are presented as means ± SD of three independent measurements and were subjected to Student's t-test. The statistical analyses of data are shown as single asterisk (P < 0.05) and double asterisks (P < 0.01) in corresponding figures.

RESULTS

Trx1 downregulates MCP-1 release and expression.

To know whether Trx1 plays any role in regulation of MCP-1 secretion in human ECs, the MCP-1 release and expression were determined in the cells either overexpressing a recombinant human Trx1 or having their endogenous Trx1 knocked down. Immunoblot analysis showed that the protein level of Trx1 in the stable Trx1-transfected cells (EA-Trx1 cells) is about twofold of that in the cells expressing only a neomycin resistance gene (EA-neo cells) and in the wild-type EA.hy 926 cells (referred as EA-wt) (see Fig. 1A). Insulin reduction-based assays showed similar results; i.e., the Trx1 activity in EA-Trx1 cells was nearly doubled compared with that in two controlled cell lines (see Fig. 1B). These data indicate that the Trx1-transfected cells EA-Trx1 do overexpress functional Trx1.

Basal and the oxLDL-stimulated MCP-1 release from the Trx1-overexpressing cells and their controls were measured by ELISA. As Fig. 1C shows, overexpression of Trx1 significantly suppressed the basal and the oxLDL-stimulated MCP-1 secretion from ECs. To confirm whether such depression of MCP-1 release is due to decreased protein expression of this enzyme, the cellular expression of MCP-1 in wild-type, neo-, and Trx1-overexpressing cells was analyzed by immunoblotting. As Fig. 1D shows, the MCP-1 expression was also suppressed by overexpression of Trx1.

To avoid any possible side effect caused by introducing an exogenous Trx1 gene in cells, the effect of Trx1 on MCP-1 release was also investigated in the cells having their endogenous Trx1 knocked down. As an immunoblot analysis showed, expression of Trx1 in the cells transfected with 150 and 200 nM Trx1 siRNA was reduced approximately by 70% and 80%, respectively, whereas protein level of Trx1 in cells transfected with scramble oligonucleotides remained almost unchanged (see Fig. 1E). The insulin reduction-based assay also showed that Trx1 activity in the cells transfected with 150 nM Trx1 siRNA was reduced by about 30% when compared with that in the cells transfected with equal amount of scramble oligonucleotide (see Fig. 1F). ELISA experiments showed that both basal and the oxLDL-stimulated MCP-1 secretion significantly increased in the cells transfected with Trx1 siRNA (see Fig. 1G). Thus the results of RNA interference experiment appear to be well consistent with previously observed depression of MCP-1 secretion in Trx1-overexpressing cells. Furthermore, the MCP-1 expression in Trx1-knockdown cells was also determined by immunoblotting (see Fig. 1H). It showed that MCP-1 expression was obviously enhanced in the cells having endogenous Trx1 depleted. Both depression of MCP-1 release/expression by overexpression of recombinant Trx1 and enhancement of the release/expression by knockdown of endogenous Trx1 strongly suggest that Trx1 functions as a suppressor for MCP-1 secretion and expression in ECs.

Trx1 downregulates MCP-1 release by depressing intracellular ROS generation.

Since ROS generation has been reported to be associated with increased MCP-1 expression under various stimulation (7, 22) and Trx can remove intracellular ROS by providing electron to the peroxiredoxin-catalyzed reduction of ROS (5), the effect of Trx1 on ROS generation in oxLDL-stimulated endothelial-like cells and the regulatory role of ROS in MCP-1 release were investigated for knowing whether Trx1 downregulates MCP-1 release through elimination of ROS. As measured by DCF fluorescence, the intracellular generation of ROS in Trx1-overexpressing cells was reduced by about 20% compared with that in controls (EA-neo cells). In contrast, stable transfection of the redox-inactive mutant Trx1(C32/35S), in which the residue Cys32 and Cys35 in its redox-active center was mutated to serines, resulted in no inhibition of ROS generation but even notably enhanced the cellular ROS level (see Fig. 2A). In addition, the ROS generation was also measured in Trx1-knockdown cells using DHE as indicator. It was observed that depletion of endogenous Trx1 apparently enhanced ROS generation in the cells (see Fig. 2B). These results clearly indicate that Trx1 suppresses cellular ROS generation.

To know how the cellular ROS affects MCP-1 release and expression, a special NADPH oxidase inhibitor DPI (33) was used to treat Trx1-overexpressing cells and their controls during stimulation with or without 100 μg/ml oxLDL. As Fig. 2C shows, the NADPH oxidase inhibitor (10 μM) significantly suppressed both basal and the oxLDL-stimulated MCP-1 release from both Trx1-overexpressing cells (EA-trx1) and their control (EA-neo ells). To confirm these results at a transcriptional level, a MCP-1 (full-length promoter) luciferase reporter was transfected into the ECs to see how intracellular ROS regulates the transcription of MCP-1 by either cotransfection with Nox1 gene, the homologues of the catalytic subunit of NADPH oxidase (gp91^phox) that generates a low-level ROS (24), or addition of the NADPH oxidase inhibitor in cell suspension. The assays with MCP-1 luciferase reporter also showed that DPI significantly depressed, whereas cotransfection with Nox1 markedly enhanced the transcription of MCP-1 (see Fig. 2D). The observed suppression of intracellular ROS generation by Trx1 and regulation of MCP-1 release/transcription by ROS suggest that, at least as one of the mechanisms, Trx1 downregulates MCP-1 secretion and expression through suppression of intracellular ROS generation.

Trx1 suppresses oxLDL-stimulated nuclear translocation of AP-1 and its activation mediator Ref-1.

Since AP-1 is one of the major nuclear factors for transcription of MCP-1 under oxLDL stimulation (65), it was investigated whether AP-1 is involved in the downregulation of MCP-1 by Trx1. AP-1 consists of a variety of dimers composed of members of the Jun and Fos families of proteins. At first, we looked at whether Trx1 could affect nuclear translocation of c-Jun, which is the most important component of AP-1, forms heterodimer with c-Fos and enters the nucleus to activate gene transcription (8). The nuclear content of c-Jun was probed by immunoblotting after the ECs were stimulated with 100 μg/ml of oxLDL for various periods of time. It was found that oxLDL stimulation induced an increase of nuclear c-Jun with a maximum appearing at about 1 h (see Fig. 3A). Therefore, all immunoblot analysis on nuclear proteins of interest was performed when cells were stimulated with oxLDL for 1 h. As immunoblot analysis of nuclear extracts showed, overexpression of Trx1 considerably reduced nuclear content of c-Jun in the oxLDL-stimulated cells (relative reduction is about 60%, see Fig. 3B), whereas knockdown of endogenous Trx1 increased c-Jun content in the nuclei of the oxLDL-stimulated cells to about fourfold (see Fig. 3C). The results suggest that Trx1 strongly suppresses oxLDL-induced translocation of AP-1 from cytoplasm to nucleus.

In addition to look at nuclear translocation of AP-1, the regulatory effect of Trx1 on nuclear translocation of Ref-1 was also investigated, since DNA-binding activity of c-Jun and c-Fos requires Ref-1 to reduce their cysteine residues (21). Thus the nuclear content of Ref-1 was also determined in the cells having their endogenous Trx1 knocked down. It was found that depletion of endogenous Trx1 by its siRNA resulted in a twofold increase of nuclear Ref-1 under oxLDL stimulation (see Fig. 3D). These results suggest that Trx1 downregulates MCP-1 release and expression likely through suppression of oxLDL-stimulated nuclear translocation of both AP-1 and its activation mediator Ref-1.

Trx1 downregulates MCP-1 expression by repression of transcriptional activity of AP-1.

As previous results showed suppression of nuclear translocation of AP-1 and Ref-1 by Trx1, does the suppression lead to depression of transcriptional activity of AP-1? To answer this question, DNA-binding activities of AP-1 in the nuclear extract from both the Trx1-overexpressing and Trx1-knockdown cells were determined by EMSA and compared with those in the nuclear extracts from their respective controls. As Fig. 4 shows, the shifted band was significantly reduced in the presence of anti-c-Jun antibody in the reaction mixture of nuclear extract with DNA probe (comparing lanes 6, 12 with lanes 4, 11), indicating that the shifted probe-containing bands are complexes of AP-1 with its DNA probe. It can be seen that overexpression of Trx1 obviously reduced but depletion of endogenous Trx1 notably enhanced the ability of AP-1 to bind the double-stranded oligonucleotide containing an AP-1 binding site in ECs compared with that in wild-type cells, the neo cells, and the cells transfected with a scramble oligonucleotide (comparing AP-1-probe complex band in lanes 5, 11 with those in lanes 3, 4, and 9, 10). The EMSA results look as if to contradict the report by other investigators that Trx upregulates DNA-binding activity of AP-1 (24, 55). The reason for the observed apparent reduction of the DNA-binding activity in the nuclear extract of the Trx1-overexpressing cells and enhancement of the DNA-binding activity in the nuclear extract of the Trx1-knockdown cells might be due to the fact that overexpression or depletion of Trx1 not only changes Trx1 in nucleus but also significantly changes nuclear Ref-1, a key molecule in activation of DNA-binding activity of AP-1. The role of Ref-1 will be discussed later.

To further prove the inhibitory effect of Trx1 on AP-1 transcriptional activity, an AP-1 luciferase reporter containing seven tandem repeats of AP-1 consensus sequences in its promoter was transfected with various combination of Trx1 expression, c-Jun expression, and c-Fos expression vectors into BAECs for looking at whether Trx1 suppresses the transcriptional activity of AP-1. As shown in Fig. 5A, transfection of either c-Jun or c-Fos activated transcription of the AP-1 consensus sequences-driven transcription of luciferase gene. However, cotransfection of Trx1 substantially depressed the transcriptional activity of c-Jun and c-Fos. The results obtained with (AP-1)₇-luc reporter clearly demonstrate that Trx1 suppresses transcriptional activity of AP-1.

To demonstrate whether Trx1 downregulates MCP-1 expression through depressing transcriptional activity of AP-1, a MCP-1 full-length, promoter-driven luciferase reporter was transfected with various combination of Trx1 expression, c-Jun expression, and c-Fos expression vectors into BAECs. The reason for using BAECs is because they are easy for transfection. As Fig. 5B shows, transfection with either c-Jun expression or c-Fos expression vector significantly enhanced transcription activity of MCP-1 promoter (2.6-fold induction for c-Jun and 1.6-fold induction for c-Fos), indicating that c-Jun and c-Fos are indeed nuclear factors for transcription of MCP-1 gene. However, cotransfection of Trx1 expression vector obviously suppressed either c-Jun-enhanced or c-Fos-enhanced activity of MCP-1 promoter. Furthermore, transfection with both c-Jun expression and c-Fos expression vector greatly enhanced transcription activity of MCP-1 promoter (7-fold induction), and the enhanced activity was even more obviously suppressed by cotransfection with Trx1. The experiments with MCP-1 reporter further demonstrate that Trx1 downregulates MCP-1 expression at the transcriptional level and the downregulation is due to suppression of AP-1 activity.

DISCUSSION

It is only recently that few investigations have been devoted to look at effect of Trx on MCP-1 release. Ohashi et al. (34) reported that Trx1 in transgenic mice overexpressing Trx1 prevented development of chronic pancreatitis in mice, and overexpression of Trx1 reduced H₂O₂-induced MCP-1 production in isolated pancreatic acinar cells. Dai et al. (11) reported that homocysteine could induce expression of Trx1, and the induced Trx1 might be responsible for reduced MCP-1 secretion from monocytes. However, none of the investigations focused on the signaling mechanisms involved in the regulation of MCP-1 by Trx1 except for suppression of ROS production. In the above-mentioned reports, one dealt with the MCP-1 release in pancreatic acinar cells, the other was concerned with monocytes. No investigation on regulation of MCP-1 release by Trx in ECs has been reported. Therefore, we probably showed for the first time that Trx1 downregulates the oxLDL-stimulated MCP-1 secretion in human vascular ECs, and depression of nuclear translocation of AP-1 and Ref-1 is a critical mechanism for the downregulation.

In human ECs, the regulation of MCP-1 expression has been studied intensively (18, 53). A previous study showed that gene expression of MCP-1 was controlled in a stimulus-specific manner and involved differential activation of the redox-responsive transcription factors AP-1 and nuclear factor (NF)-κB (64). In this study, oxLDL was used to stimulate the endothelial-like EA.hy 926 cells as a cell model of atherogenesis because oxLDL is a key pathogenic mediator of atherogenesis. Other investigators have already demonstrated that oxLDL stimulation increased MCP-1 secretion/expression in ECs, and activation of NF-κB is not involved in oxLDL-mediated increase of MCP-1 secretion/expression (14). We also observed no nuclear translocation of NF-κB subunit p65 in the oxLDL-stimulated EA.hy 926 cells (data not shown). In contrast, we observed that c-Jun, the core member of AP-1 protein family, and Ref-1 translocated from cytosol to nuclei under oxLDL-stimulation in the endothelial-like cells.

Besides confirming that Trx1 downregulates MCP-1 release/expression by eliminating excrescent ROS generation, our results reveal two novel mechanisms by which Trx1 suppresses MCP-1 expression in oxLDL-stimulated ECs. First, Trx1 depresses nuclear translocation of AP-1 protein c-Jun (see Fig. 3, B and C), which must lead to less gene transcription of MCP-1. Second, Trx1 depresses nuclear translocation of Ref-1 (see Fig. 3D), which certainly should lead to less activation of AP-1. Ref-1 has been shown to mediate activation of AP-1 through a redox-sensitive mechanism in the cells under various stress, such as stimulation by phorbol myristate acetate (55), heat shock (13), treatment with hydrogen peroxide or DNA-damaging agents (56), and hypoxia (59). Since AP-1 is not a direct substrate of Trx (57), the real mechanism is that Trx1 activates AP-1 actually through physical interaction with Ref-1 and then redox modify Ref-1. It is Ref-1 to induce DNA-binding activity of AP-1 through a critical cysteine residue located in the basic DNA binding domain of c-Fos/c-Jun (56, 57) and a critical cysteine residue located at position 65 in the redox domain of Ref-1 (52). Therefore, a redox-sensitive signaling pathway leading from Trx1 to Ref-1 and then to the AP-1 complex participates in the upregulation of DNA-binding activity, and Ref-1 plays a key role in the induction of AP-1 DNA binding activity. In the present investigation, EMSA was performed with the nuclear extract from Trx1-overexpressing and Trx1-knockdown cells, and both EMSA experiments showed that Trx1 suppresses DNA-binding activity of AP-1. Nuclear content of Ref-1 was found significantly increased (twofold increase, see Fig. 3D) in Trx1-knockdown cells. Although nuclear Trx1 may be reduced to some extent in Trx1-knockdown cells, it may be still enough to associate and redox modify Ref-1, leading to an enhancement of DNA binding activity of AP-1. The observed apparent enhancement of DNA binding activity of AP-1 in Trx1-knockdown cells demonstrates the key role of Ref-1 rather than Trx1 itself in activation of AP-1 activity in binding DNA.

Since MCP-1 has a potent chemoattractant activity for monocytes/macrophages, and the accumulation of monocytes/macrophages in the intima of artery vessel is an important early event in atherosclerosis, MCP-1 has been considered as a critical chemokine involved in the pathogenesis of atherosclerosis. This study provides a novel evidence that Trx1 may play a role in preventing or decelerating atherosclerosis via downregulation of the expression and release of some pro-inflammatory factors from vascular ECs. Among them MCP-1 is a critical molecule.

GRANTS

This work was supported by the National Natural Science Foundation of China (Nos. 30330250 and 30770513).