Vascular Cell Adhesion Molecule-1 Gene Expression during Human Smooth Muscle Cell Differentiation Is Independent of NF- k B Activation*

Vascular cell adhesion molecule-1 (VCAM-1) gene expression in cytokine-activated cells depends on two k B elements. Since VCAM-1 expression appears developmentally regulated and cytokine-inducible in smooth muscle cells (SMCs), we have studied the role of NF- k B in differentiated SMC VCAM-1 expression. Confluent SMCs were cultured either in a serum-free medium in order to induce differentiation, or in medium with serum, stimulated or not by tumor necrosis factor a (TNF- a ). The expression of smooth muscle myosin heavy chain, a SMC marker, and VCAM-1 was induced concomitantly in serum-free medium, whereas only VCAM-1 expression was induced by cytokine-treatment. We showed that the p50 and p65 subunits of NF- k B were localized in the cytoplasm of differentiating SMCs, whereas they were translocated into the nucleus of TNF- a -activated SMCs. Electrophoretic mobility shift assays with VCAM-1 gene k B elements failed to detect any induction of DNA-protein complex with nuclear extracts of differentiating SMCs in contrast to the cytokine-acti-vated SMC nuclear extracts. Furthermore, VCAM-1 mRNA induction was inhibited in TNF- a -stimulated SMCs, but not in differentiating SMCs, by pyrrolidine dithiocarbamate, an inhibitor of NF- k B protein activation. Taken together, these findings suggest that in contrast to TNF- a activation, NF- k B is not involved in VCAM-1 gene expression during SMC differentiation. were resolved by non-denaturing polyacrylamide electrophoresis in 5% acrylamide, employing 0.022 M Tris borate, 0.5 m M EDTA, pH 8, as buffer. Following electrophoresis, gels were dried and DNA-protein complexes were localized by autoradiography. Competition studies were performed by adding unlabeled double-stranded oligonucleotides to the binding reaction.

the immunoglobulin superfamily first identified as an adhesion protein expressed on cytokine-activated endothelial cells (1)(2)(3). Through its interaction with its integrin counter receptor, very late antigen-4 (VLA-4), it mediates cell-cell interactions, which have been shown to be involved in recruitment of mononuclear leukocytes to the vascular lesions observed in early atherosclerosis. VCAM-1 and VLA-4 are also implicated during skeletal muscle myogenesis and developing central nervous system on neuroepithelial cells in embryogenesis (4 -6).
In the vascular system, VCAM-1 is expressed in human SMCs during embryonic development and in atherosclerotic plaques, and it colocalizes with VLA-4 (7,8). VCAM-1-deficient mice embryos have been shown to have an abnormal placental development in which the allantois failed to fuse to the chorion, and they displayed abnormalities in their developing hearts (9,10). In vitro, VCAM-1 is induced in SMCs after cytokine stimulation (11,12). It is also coinduced with SMC marker proteins during SMC differentiation in vitro, and treatment with antibodies against VCAM-1 or its ligand VLA-4 has been found to interfere with the expression of mRNA of smooth muscle-specific markers during redifferentiation (8). Thus, VCAM-1 seems to be expressed and involved during smooth muscle differentiation. Investigation of mechanisms controlling its expression should provide some insight into specific gene regulation in vascular smooth muscle during differentiation.
Studies on the activity of the VCAM-1 gene promoter have shown that in endothelial cells, TNF-␣ induction of VCAM-1 is dependent on two adjacent B sites located at positions Ϫ77 and Ϫ63 relative to the single transcriptional start site (13,14). These B elements are also important for VCAM-1 induction in the embryonic carcinoma cell line P19 stimulated to differentiate along a neural pathway by treatment with retinoic acid (6). Thus, these previous works suggested that NF-B proteins could be involved in the VCAM-1 expression in SMCs. NF-B proteins belong to the Rel protein family, transcriptional regulatory proteins, including p65 (RelA), RelB, p52/100 (NF-B1), p50/p105 (NF-B2), v-Rel, and the product of the c-rel protooncogene (15)(16)(17)(18)(19). Interestingly, Bellas et al. (20) have described a novel NF-B member termed SMC-Rel, essential for proliferation of cultured bovine vascular SMCs. In most unstimulated cell types, NF-B is sequestered in the cytoplasm complexed with an inhibitory protein, I-B, in a non-DNAbinding form (21). Stimulation of cells with cytokines (22) and several other agents (23) results in the release of NF-B complex from I-B and its translocation into the nucleus, where it transcriptionally regulates the expression of a wide variety of genes through specific B elements (23). The activity of the NF-B/Rel family is controlled at different levels, e.g. by the different combination of NF-B/Rel dimers formed, and the sequence of B sites recognized by them.
In this study, the role of NF-B in VCAM-1 expression dur-* This work was supported in part by the French Ministère de l'É ducation Nationale de la Recherche et de la Technologie (ACC-SV 9) and by the Fondation de France. Portions of this work were presented in abstract form at the 69th Scientific Sessions of the American Heart Association, New Orleans, LA, November 10 -13, 1996 (Lavie, J., Dandré, F., Lamazière, J.-M. D., and Bonnet, J. (1996) Circulation 94, I-167). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Cell Culture-SMCs were isolated from media of human thoracic aortas by enzyme digestion as described previously (26). SMCs were cultured in Ham's F-10 medium (Life Technologies, Inc.) supplemented with 5% fetal calf serum, 5% heat-inactivated human serum, 5 mM HEPES, 50 units/ml penicillin, and 50 mg/ml streptomycin, at 37°C in a 5% CO 2 , 95% air atmosphere. Studies were conducted on SMCs at passage two, seeded at a high seeding density (5 ϫ 10 4 cells/cm 2 ). In order to obtain the different phenotypes of SMCs (undifferentiated or differentiated), we used various culture conditions (8,27,28). At confluence, the cells were either maintained undifferentiated in medium supplemented with serum, or induced to differentiate in a defined serum-free medium containing a 1:1 mixture of Dulbecco's modified Eagle's medium (Life Technologies, Inc.) and Ham's F-10 plus insulin (10 Ϫ6 M, Sigma), transferrin (200 g/ml, Sigma), ascorbate (0.2 mM, Sigma), and sodium selenite (6.25 ng/ml, Sigma). These conditions designed to induce redifferentiation were usually maintained for 6 days. However, for time-course or inhibition experiments, as some smooth muscle markers and VCAM-1 mRNA were rapidly induced within day 1 in defined serum-free medium (8), cells were used 1, 6, or 24 h after the switch of medium.
In order to obtain a VCAM-1 cytokine-induced expression in the absence of any redifferentiation process, confluent SMCs maintained for 6 days in serum-supplemented medium, were treated with TNF-␣ (1000 units/ml) 24 h before analysis.
In order to inhibit NF-B activity, cells were treated with PDTC (Sigma) at a final concentration of 100 M. Initially the confluent cells were pretreated for 1 h with PDTC, and then maintained in culture for an additional 24 h in the presence of PDTC and either TNF-␣ in medium plus serum or in serum-free medium. Cells maintained in serum and treated for 25 h with PDTC were used as control. The effect of PDTC on SMCs viability was assessed by trypan blue exclusion; less than 5% cell mortality was observed. The well plates were rinsed with phosphate-buffered saline (PBS), and RNAs were extracted and analyzed by RT-PCR.
Flow Cytometry-For flow cytometric analysis, monodispersed suspensions of SMCs were prepared by brief incubations in buffer (6 mM glucose, 5.3 mM KCl, 125 mM NaCl, 18 mM Hepes, 1ϫ PBS, and 0.2 mM EDTA) at 37°C. The cells were then incubated with 50 l of monoclonal antibody against human VCAM-1 in 2% BSA/PBS for 30 min in ice. After washing, cells were treated with sheep anti-mouse Ig, fluoresceinlinked whole antibody (Amersham Pharmacia Biotech) for 30 min in ice, washed three times, and resuspended in 500 l of PBS. The samples were analyzed in a Coulter type XL flow cytometer.
RNA Analysis-Total cellular RNA was prepared using a single-step acid guanidium isothiocyanate/phenol/chloroform extraction method (29): first-strand cDNA synthesis by reverse transcription (30) using 1 g of total RNA and 20 g/ml oligodeoxythymidine and 200 units of Moloney murine leukemia virus RNase H-reverse transcriptase (Life Technologies, Inc.) with incubation for 1 h at 42°C. Specific primers for G3PDH, VCAM-1, and sm-MHC were used as described previously (8) to amplify regions of cDNA copies from total RNA.
RT-PCR Product Analysis-The samples were analyzed on 1.5% agarose gels. In order to verify the specificity of generated PCR products, hybridizations were performed with specific probes for VCAM-1 and sm-MHC, or with an internal oligonucleotide for G3PDH after Southern transfer (31). Probes for human VCAM-1 were synthesized by RT-PCR and cloned in pBluescript vector (Stratagene) as described previously (12). For human sm-MHC probe, human sm-MHC cDNA was amplified by PCR for 30 cycles at 60°C for annealing using primers as described elsewhere (8). The PCR product obtained was purified, digested in a 392-bp fragment, and subcloned into pBluescript plasmid according to standard procedures. The specificity of the PCR product was ensured by the DNA sequence determined by the dideoxy chaintermination method with modified T7 DNA polymerase. The VCAM-1 and sm-MHC cDNA fragments were labeled by random priming with the kit Ready-To-Go (Biolabs). Hybridization was done at 42°C overnight in the presence of 10 6 cpm/ml in buffer 50% (v/v) formamide, 4ϫ SSC, 10ϫ Denhardt's solution (1ϫ Denhardt's solution is 0.02% polyvinyl pyrrolidone, 0.02% Ficoll, 0.02% BSA), 0.1% SDS, and sheared denatured salmon sperm DNA at 0.4 mg/ml. Filters were washed in 2ϫ SSC for 10 min at room temperature, followed by two washes in 0.1ϫ SSC for 30 min at 55°C. Filters were exposed to Kodak XAR-5 film with an intensifying screen.
Preparation of Nuclear Extracts-Preparation of nuclear extracts was based on the methodologies described by Li and Schreiber (32,33). The cultured cells (approximately 2 ϫ 10 7 cells) were washed with cold PBS and harvested after trypsin treatment. The cells were centrifuged at 4,000 rpm for 10 min, and the cell pellet was resuspended in 800 l of cold buffer A (10 mM HEPES, pH 8, 50 mM NaCl, 0.5 M sucrose, 1 mM EDTA, pH 8, 0.5 mM spermidine, 0.15 mM spermine). The cells were allowed to swell on ice for 15 min, then 25 l of a 10% solution of Nonidet P-40 was added to release the nuclei. Nuclei were recovered by centrifugation at 3,000 rpm for 10 min at 4°C. The nuclear pellet was resuspended in 500 l of buffer B (10 mM HEPES, pH 8, 50 mM NaCl, 25% glycerol, 0.1 mM EDTA, pH 8, 0.5 mM spermidine, 0.15 mM spermine) in the presence of protease inhibitors (pepstatin, leupeptin, aprotinin, antipain, and chymostatin at 5 g/ml, 1 mM Pefabloc, 2 mM benzamidine) and 1 mM dithiothreitol. The nuclear pellet was recovered by centrifugation, and nuclear proteins were extracted by incubation for 1 h at 4°C in 100 l of buffer C (the same as buffer B except 400 mM NaCl). The nuclear extract was centrifuged at 13,500 rpm for 40 min, and the supernatant was dialyzed three times against 250 ml of buffer D (20 mM HEPES, pH 8, 50 mM KCl, 20% glycerol, 0.2 mM EDTA, pH 8, 1 mM dithiothreitol, 2 mM benzamidine) for 1 h. Nuclear extracts were stored in aliquots at Ϫ80°C. Protein concentration was measured by the Bradford assay (Bio-Rad kit) (34).
Oligonucleotide Sequences-In order to prepare double-stranded oligonucleotides containing the upstream B elements of human VCAM-1 gene, synthetic complementary single-stranded oligonucleotides were annealed and then purified by polyacrylamide gel electrophoresis. Gelpurified double-stranded oligonucleotides were labeled on their 5Ј ends with [␥-32 P]ATP (5000 Ci/mmol) (Amersham Pharmacia Biotech) using T4 polynucleotide kinase (Life Technologies, Inc.) and used as probes in EMSA. Oligonucleotides were chosen according to Iademarco et al. (13). Oligonucleotide B corresponds to the two wild-type B sites (boldface) taken from the DNA sequence located at bp Ϫ77 to 63 of the VCAM-1 gene promoter: CTGCCCTGGGTTTCCCCTTGAAGGGATTTCCCT-CCGCCT. Oligonucleotide BM was identical to B wild type sites except for 4 nucleotide mutations (underlined) in each of the two NF-B-DNA binding sites according to Marui et al. (35): CTGCCCTGAG-TCACGCCTTGAAGAGACATCACTCCGCCT. An oligonucleotide (YY1) corresponding to a probe for another transcription factor was used in competition experiments to ensure the specific B DNA binding activity with NF-B proteins: TCTGTCTCCATTTTTTCTCT.
Electrophoretic Mobility Shift Assay-EMSA were performed with 25 l of binding mixture containing 1 g of double-stranded poly(dI-dC), 40 fmol of 32 P-labeled annealed B oligonucleotides, 8 g of nuclear extract in 12 mM HEPES, pH 8, 50 mM KCl, 12% glycerol, 1 mM MgCl 2 , 0.6 mM dithiothreitol, 0.12 mM EDTA, 1.2 mM benzamidine. Reactions were incubated at room temperature for 30 min, and DNA-protein complexes were resolved by non-denaturing polyacrylamide electrophoresis in 5% acrylamide, employing 0.022 M Tris borate, 0.5 mM EDTA, pH 8, as buffer. Following electrophoresis, gels were dried and DNA-protein complexes were localized by autoradiography. Competition studies were performed by adding unlabeled double-stranded oligonucleotides to the binding reaction.

VCAM-1 Induction Model during SMC Differentiation-In
order to determine the role of the NF-B/Rel protein family in VCAM-1 expression in SMCs, we used an in vitro model of SMC phenotypic modulation. Concerning the modulation of differentiation process, we analyzed the protein and mRNA levels of sm-MHC, a marker of the fully differentiated phenotype of SMCs (36 -38). In 6-day post-confluent cells, maintained in medium with serum, a faint immunostaining for sm-MHC was observed, suggesting a dedifferentiated phenotype (Fig. 1, S). When these post-confluent cells in medium with serum were stimulated for 24 h with TNF-␣, the same low expression of sm-MHC was observed (Fig. 1, T), showing the lack of effect of TNF-␣ on SMC differentiation. On the other hand, when the confluent cells were cultured for 6 days in the defined medium without serum, a strong immunostaining for sm-MHC was noted, indicative of redifferentiation ( Fig. 1, D). Analysis of mRNA content by RT-PCR and Southern transfer revealed a low level of sm-MHC mRNA in cells cultured in medium with serum that was stimulated or not stimulated with TNF-␣ (Fig.  2). In SMCs cultured in defined serum-free medium, a high level of sm-MHC mRNA was detected (Fig. 2). In this in vitro model, VCAM-1 expression showed a different pattern to that of sm-MHC. Flow cytometry and RT-PCR analysis showed that SMCs maintained in medium with serum had a low level of VCAM-1 protein and mRNA (Table I and Fig. 2). However, administration for 24 h with TNF-␣ or cultivation in defined serum-free medium for 6 days induced an increase in protein and mRNA content of VCAM-1 (Table I and Fig. 2). These results showed that VCAM-1 expression was increased in two conditions, either by cytokine treatment in SMCs expressing low levels of sm-MHC, or by medium switch inducing redifferentiation characterized by a high level of sm-MHC expression. These results suggested that VCAM-1 was induced in two different cell populations, in TNF-␣-stimulated undifferentiated SMCs and in redifferentiated SMCs.
Localization of NF-B/Rel Proteins during VCAM-1 Induction on SMCs-In order to determine the cellular distribution of NF-B/Rel proteins in SMCs, the immunostaining of the p50 and p65 subunits of NF-B complex was studied in SMCs cultured under different conditions at 6 days (Fig. 3a). p65 and p50 were found in the nuclei of both undifferentiated SMCs cultured in serum and in TNF-␣-treated SMCs (Fig. 3a, S and  T, respectively). However, in contrast with the strong labeling of the cytokine-stimulated nucleus, a fainter nuclear staining and a persistent cytoplasmic labeling was observed in cells maintained only in medium with serum. In contrast, only a cytoplasmic labeling of the p50 and p65 subunits with no nuclear staining was observed in differentiating SMCs cultured in defined serum-free medium (Fig. 3a, D). This suggested that  obtained from post-confluent SMCs cultured in medium supplemented with serum for 6 days, activated (T) or not (S) with 1000 units/ml TNF-␣ for 24 h, and from post-confluent SMCs cultured in a defined serum-free medium for 6 days (D) were used for RT-PCR analysis. Ten l of DNA amplified product were analyzed in 1.5% agarose gel and transferred on nitrocellulose membrane. Hybridizations with specific end-labeled probes were used to ensure specificity of amplified products. DNA probes are described under "Materials and Methods." G3PDH, used as internal control, VCAM-1, and sm-MHC mRNA expression were analyzed in the three conditions. NF-B proteins were translocated to the nucleus when SMCs were cultured in medium supplemented with serum or stimulated by TNF-␣. This p50 and p65 protein translocation was not observed in redifferentiated SMCs.
In order to exclude an early involvement of NF-B system in VCAM-1 expression during SMC redifferentiation process, we analyzed the time course of p50 Rel protein translocation in our culture model (Fig. 3b). As the result obtained after 6 days, we observed a weak p50 subunit staining into nucleus of SMCs maintained in serum. After TNF-␣ stimulation, a slight p50 nuclear staining was evident at 1 h (Fig. 3b, T). A maximal positive nucleus labeling was seen after 6 h of TNF-␣ stimulation and was maintained for 24 h (Fig. 3b, T). In redifferentiated SMCs, maintained in serum free-medium, we did not observe any nucleus staining but only a cytoplasmic labeling ( Fig. 3b, D). These results seemed to confirm the non-involvement of NF-B proteins in redifferentiated SMC VCAM-1 expression.
Characterization of Binding Activity of NF-B/Rel Proteins during VCAM-1 Induction on SMCs-EMSA was employed in an attempt to find out whether nuclear proteins interact specifically with B elements of the VCAM-1 gene promoter. Nuclear extracts of SMCs harvested in the three culture conditions described above were incubated with a double-stranded end-labeled oligonucleotide probe extending from position Ϫ88 to Ϫ56 of the human VCAM-1 promoter. As shown in Fig. 4, the nuclear extract of undifferentiated SMCs cultured in medium with serum (S) formed one DNA-protein complex (B1) with the probe. The complex was competed with a 10 -50-fold molar excess of unlabeled B oligonucleotide, whereas addition of a 50-fold molar excess of an unrelated YY1 oligonucleotide, or a 10 -50-fold molar excess of the mutated BM oligonucleotide did not compete, indicating the specificity of this binding. Nuclear extracts from undifferentiated TNF-␣-stimulated SMCs (T) formed two specific DNA-protein complexes (B1 and B2 complexes) with the labeled B probe. Competition experiments demonstrated the specificity of the two DNA-protein complexes. On the other hand, EMSA with nuclear proteins from differentiated SMCs (D) showed no specific DNA-protein complex. These results showed that the presence of serum was associated with the presence of the B1 complex, TNF-␣ stimulation inducing the appearance of the B2 complex. These complexes seemed to disappear in culture in defined serum-free medium, which induced redifferentiation.

Role of NF-B/Rel Proteins in VCAM-1 Induction on SMCs-
In an attempt to clarify the role of NF-B/Rel proteins in the two conditions of VCAM-1 expression, after cytokine stimulation and during SMC differentiation, we examined whether the antioxidant PDTC, a blocker of NF-B/Rel protein activation (20,23,35) interfered with VCAM-1 expression on SMCs. Since it has been shown that VCAM-1 mRNA in confluent SMCs is induced within 24 h after switching to a defined serum-free medium (8), the experiments with PDTC were performed at the onset of VCAM-1 mRNA induction during cytokine stimulation or SMC differentiation. As shown in Fig. 5, as expected, TNF-␣ treatment alone (lane 3) and switch to a defined serum-free medium (lane 5) induced a strong increase in VCAM-1 mRNA expression relative to that observed by FIG. 3. NF-B protein distribution in cultured SMCs during redifferentiation process. Cells were stained with rabbit polyclonal anti-p50 or anti-p65 antibody. The secondary antibody, a biotin-conjugated anti-rabbit donkey IgG, was linked with streptavidin-Texas Red complex. a, a p50 and p65 subunit localization in culture model was studied. Confluent SMCs were cultured for 6 days in serum-supplemented medium (S and T) or in serum-free medium (D). In contrast with the strong nuclear labeling observed in TNF-␣-stimulated SMCs (T), only the cytoplasm was seen to be labeled in defined serum-free medium (D). A faint nuclear staining is noted in SMCs maintained in serum (S). b, a time course of p50 subunit was performed. Confluent SMCs were cultured for 1, 6, or 24 h in serum-supplemented medium without (S) or with TNF-␣ (T), or in serum-free medium (D). Antiserum to p50 was used to immunostain SMCs. After TNF-␣ stimulation (T), a positive nuclear labeling was evident after 6 h of stimulation and maintained for 24 h. In redifferentiating SMCs in serum-free medium (D), only cytoplasmic staining was observed at any time. Original magnification, ϫ20.

DISCUSSION
VCAM-1 is induced or up-regulated by proinflammatory cytokines such as TNF-␣ and interleukin-1␤ on many cellular components of the arterial wall, including endothelial cells, SMCs, and fibroblasts (3, 39 -41). More recently, VCAM-1 has been reported to be inducible by TNF-␣ stimulation in a timeand dose-dependent manner on cultured SMCs (12). The pathophysiological significance in vivo of VCAM-1 expression on arterial SMC is unknown, although it may amplify inflammatory processes in pathological conditions such as atherosclerosis. However, as suggested by Rosen's work on myocytes (4) and more recently by our own work in SMCs (8), homotypic cell-cell interactions mediated by VCAM-1/VLA-4 may participate in the regulation of cell differentiation. In a previous study, we demonstrated in an in vitro model of redifferentiation, SMCs reexpressed VCAM-1 and smooth muscle-specific markers during induction of differentiation (8). These observations suggested that VCAM-1 is expressed on the surface SMCs either after cytokine stimulation or during the process of differentiation. These two kinds of VCAM-1 up-regulation on SMCs were reproduced in the present experiments (cf. Fig. 1). However, the lack of induction of sm-MHC expression after TNF-␣ stimulation suggested two different mechanisms for VCAM-1 induction on SMCs. Previous studies have shown that B elements are involved in VCAM-1 gene induction by cytokines in endothelial cells, and for VCAM-1 expression during P19 cell differentiation into a neural pathway (6,13). In an attempt to find out whether the VCAM-1 induction during SMC differentiation differed from the cytokine induction pathway, we analyzed the involvement of the NF-B system in both processes.
Members of the NF-B family are typically present in the cytoplasm bound to the inhibitory I-B proteins (21). Activation of NF-B involves the phosphorylation and subsequently the degradation of inhibitory proteins (42), allowing NF-B to translocate to the nucleus. The members of this family, which include p50, p52, p65, c-Rel, v-Rel, RelB, exist as homo-and heterodimers that bind to B sites in the enhancer regions of a large number of target genes. As in many instances NF-B binds as a heterodimer of a p50 in combination with p65 to stimulate gene expression (43), we analyzed the cellular distribution of p50 and p65 proteins in the different conditions of VCAM-1 expression. An intense staining of the nuclear compartment was observed for these two Rel proteins after TNF-␣ stimulation, the time course of NF-B translocation into the nucleus showing an early nuclear staining for p50. In culture with serum, a weaker staining of the nucleus along with a labeling of the cytoplasm was observed. However, when the cells were cultivated for in defined serum-free medium allowing strong VCAM-1 expression and redifferentiation process, no immunostaining of the p50 and p65 subunits of the NF-B/ Rel family was observed in the nuclear compartment, whereas the cytoplasm was strongly labeled. The negative nuclear staining during the time-course experiments of NF-B translocation into the nucleus excludes the hypothesis that NF-B could be involved in VCAM-1 gene expression early in the differentiation stages. Cytokines such as TNF-␣ have been found to induce NF-B activation in many cell types (22,23). Moreover, growth factors that are contained in serum such as PDGF-BB, bFGF, EGF, and IGF-1 induce the translocation of p50 and p65 into the nuclei of SMCs, whereas growth inhibitors such as TGF-␤ and interferon-␥ do not induce nuclear localization of these subunits of the NF-B protein family (44). These findings are consistent with an involvement of the NF-B complex in the TNF-␣-induced VCAM-1 expression we observed. The absence of p50 and p65 subunit nuclear translocation during the redifferentiation induced by the switch to a defined serum-free medium indicated that VCAM-1 expression was independent of NF-B complex activation in this case.
To directly address whether VCAM-1 expression in SMCs in different states of differentiation was associated with NF-B activation and nuclear translocation, the binding activity of the B sequence from the human VCAM-1 promoter for the different nuclear extracts was tested by EMSA. Specific DNA-protein complexes were observed on EMSA with nuclear extracts of TNF-␣-treated and serum-treated SMCs. Only one DNAprotein complex, B1, was observed with serum-treated cells, and one faster migrating, additional complex, B2, with TNF-␣treated SMCs. Similar patterns have been reported by other workers: two NF-B DNA-protein complexes in phorbol esteractivated Jurkat cells; the more rapidly migrating complex (B2) was composed of a p50 subunit and partial degradation products of p50, while the more slowly migrating complex (B1) contained not only these smaller proteins but also p65 and c-Rel (45). These two complexes were also observed in the same migration positions with nuclear extracts from rat aortic SMCs stimulated by various growth factors (44). In our experiments, complex B2 appeared only in human TNF-␣-activated SMCs and may be associated with VCAM-1 induction. Complex B1, described after serum stimulation (44), did not appear to be in sufficient amount for VCAM-1 induction. In contrast, EMSA of nuclear extracts from differentiated SMCs, cultured for 6 days in defined serum-free medium, did not evidence any specific DNA-protein complex. Interestingly, Lehtinen et al. (46) reported that, after 48 h in low serum differentiation medium, the binding activity of NF-B was very low in all myogenic cell lines tested (mouse C2C12 myocytes and rat L6 myocytes) and coincided with a most abondant nuclear staining of myogenin, a protein involved in terminal differentiation of skeletal muscle cells. In rat aortic smooth muscle cells (between third and sixth passage) starved with 0.1% fetal bovine serum for 3 days, only a faint B1 band was observed on EMSA by Obata et al. (44). In our experiments, the absence of the B1 band could have been due to the human origin of the cells or to the more differentiated state of the SMCs (8). The disappearance of a specific protein interacting with VCAM-1 gene promoter B sites in nuclear extracts of serum-free medium cultured SMCs expressing VCAM-1, indicated that the NF-B system was not in- FIG. 5. Antioxidant PDTC treatment prevents VCAM-1 mRNA induction in undifferentiated TNF-␣-stimulated SMCs but not in redifferentiated SMCs. After pretreatment for 1 h with 100 M PDTC, SMCs were exposed to TNF-␣ (1000 units/ml) for 24 h or cultured during 24 h in defined serum-free medium and in the continuous presence of 100 M PDTC. One g of total RNA was used for the RT-PCR analysis. Amplified products were electrophoresed, blotted, and hybridized with an ␣-32 P-labeled VCAM-1 probe. Labeled DNA products were visualized by autoradiography. Autoradiogram of PCR amplification of retrotranscribed G3PDH mRNA is shown as control.
volved in its regulation during the process of differentiation.
In order to verify this idea, we analyzed the level of VCAM-1 mRNA expression after inhibition of NF-B system activation by an antioxidant, PDTC. It has been shown that the activity of NF-B as transcriptional factor is regulated by changes in the redox state of the cell (47). In human vascular endothelial cells, TNF-␣-and lipopolysaccharide-mediated transcriptional activation of the human VCAM-1 promoter has been found to be mediated by NF-B DNA elements, and the associated NF-B DNA-binding protein is blocked by PDTC (35). PDTC has been shown to inhibit NF-B stimulation by all inducers known including H 2 O 2 (24), by preventing dissociation of the NF-B/ I-B complex (48). The antioxidative effect of thiol compound such as PDTC might hinge on a metal-chelating action as chelation of free iron, and copper is thought to be an important protective mechanism against oxidants. On the other hand, dithiocarbamates can act directly as free radical scavengers. In our experiments, the ability of the antioxidant PDTC to prevent the TNF-␣-induced increase in VCAM-1 mRNA in SMCs supports an NF-B-dependent signaling pathway for the cytokine-induced VCAM-1 expression. In contrast, the persistence of VCAM-1 mRNA expression after PDTC treatment of redifferentiated SMCs points to an NF-B-independent pathway in this population of cells.
Taken together, these results indicate that VCAM-1 expression on SMCs is regulated by two independent pathways, an NF-B-dependent pathway after cytokine activation and an NF-B-independent pathway during the redifferentiation process.
Several mechanisms could account for the regulation of VCAM-1 expression during the redifferentiation process, either nonspecific or specific. Among the nonspecific mechanisms of VCAM-1 regulation during differentiation, an autocrine/paracrine activation comes to mind. However, the lack of involvement of NF-B proteins in the transcriptional regulation suggests that VCAM-1 expression is independent of an autocrine/ paracrine stimulation by growth factors such as PDGF-BB, bFGF, EGF, and IGF-1 or cytokines as TNF-␣ and interleukin-1␤ synthesized by arterial wall cells. All these secreted factors are known to activate the NF-B system. Another possibility is the involvement of pleiotropic transcription factors, different from the NF-B system. Two GATA DNA elements are present in opposite transcriptional orientations in the VCAM-1 gene promoter at positions between Ϫ255 and Ϫ235 relative to the transcription start site. GATA zinc finger proteins, which are involved in TNF-␣ induction of VCAM-1 in endothelial cells, belong to a transcription factor family composed of at least eight members. GATA-1, -2, and -3 are involved in differentiation of hematopoietic cells. GATA-4 has a tissue distribution limited to heart and endodermally derived tissue. Furthermore, inhibition of GATA-4 protein expression blocks cardiac muscle differentiation in vitro. Recently, GATA-6 human transcripts have been detected in lung and liver and in cultures of human and rat vascular SMCs (49). In SMCs, GATA-6 transcripts are down-regulated when quiescent cultures are stimulated to proliferate in response to mitogen activation. As withdrawal from the cell cycle is a prerequisite for differentiation, it is possible that the GATA-6 protein participates in regulation of the differentiative state of SMCs and thus to induction of the VCAM-1 gene expression. However, a search for GATA DNA-binding proteins in redifferentiated SMC nuclear extracts using the GATA probe on EMSA did not uncover any specific DNA-protein complexes (data not shown).
VCAM-1 expression could be also regulated by a musclespecific differentiation pathway. During striated muscle differentiation, VCAM-1 is expressed in a developmentally specific patter. The cell-cell interactions underlying myogenesis have been shown to involve an interaction between VCAM-1 and VLA-4 (4). Study of VCAM-1 promoter activity in C2C12 mouse myoblasts has identified a position-specific enhancer located between bp Ϫ17 and Ϫ5, consensus for an interferon regulatory factor (IRF) binding element that overrides the effect of the other promoter elements, resulting in VCAM-1 constitutive expression (50,51). The transcription factor IRF-2 that interacts with this VCAM-1 gene promoter in muscle cells is expressed concomitantly with VCAM-1 in differentiating skeletal muscle in mouse embryos. Similarly, VCAM-1 expression has been described in human aorta during ontogenesis (8), suggesting that VCAM-1 could be regulated by this IRF binding site. However, in human differentiated SMCs, we did not find any specific protein-DNA complex by EMSA with this element (data not shown). These negative results suggested that IRF binding site is not involved in SMC expression of VCAM-1.
These results led us to suspect a smooth muscle-specific transcriptional regulation of VCAM-1 gene in differentiated SMCs. Several studies on smooth muscle gene promoters (SM22␣, sm-MHC, sm-␣ actin) have found evidence for roles of various transcription factors such as MEF2 (52,53) and CArGbox binding proteins (53,54). Several potential CArG-boxes and MEF2 binding sites are present in the VCAM-1 gene promoter. The isoforms of the MEF2 transcription factor belonging to the MADS box family transcription factors are expressed in vascular SMCs and in vascular tissue (55) and during Drosophila ontogenesis, MEF2 is required for later aspects of differentiation of the three major types of musculature (56,57). MEF2 binding sites are found in the control of the majority of musclespecific genes (58). Interestingly, a MEF2-like binding site is present in the VCAM-1 gene promoter located from Ϫ1992 to Ϫ1982 (5Ј-TTTTAATAAA-3Ј). It resembles the MEF2-like sequence of the rabbit sm-MHC gene promoter found at Ϫ1540 to Ϫ1530 (5Ј-TATTAATATAA-3Ј) involved in gene expression in rat aortic SMCs (52). It is of note that this MEF2-like binding site specifically interacts with MEF2B (59). These findings suggest that many potential smooth muscle-specific cis-regulator elements in the VCAM-1 gene promoter could participate in the regulation of VCAM-1 during redifferentiation of smooth muscle cells. This might constitute a smooth muscle-specific pathway. Further experiments will be required to gain further understanding of the mechanisms underlying the intriguing pattern of VCAM-1 expression in smooth muscle tissue.