Inhibition of Endothelial Vascular Cell Adhesion Molecule-1 Expression by Nitric Oxide Involves the Induction and Nuclear Translocation of IκBα*

The induction of vascular cell adhesion molecule-1 (VCAM-1) expression by tumor necrosis factor (TNF)-α requires the activation of nuclear factor-κB (NF-κB) via a process involving the phosphorylation and degradation of its cytoplasmic inhibitor, IκBα. We have shown that nitric oxide (NO) decreases VCAM-1 expression via inhibition of NF-κB activation. To determine how NO inhibits NF-κB, we studied the fate of IκBα following TNF-α stimulation in the presence of NO donorsS-nitrosoglutathione and sodium nitroprusside. Activation of NF-κB by TNF-α occurred within 15 min and coincided with rapid degradation of IκBα. Co-treatment with NO donors did not prevent IκBα phosphorylation or degradation. However, after 2 h of TNF-α stimulation, NO donors inhibited NF-κB activation and augmented IκBα resynthesis and nuclear translocation by 2.5- and 3-fold, respectively. This correlated with a 75% reduction in TNF-α-induced VCAM-1 expression. In a time-dependent manner, NO donors alone caused the nuclear translocation of IκBα. To confirm that NO donors have similar effects as endogenously derived NO, murine macrophage-like cells, RAW264.7, were co-cultured with endothelial cells. Induction of RAW264.7-derived NO inhibited lipopolysaccharide-induced endothelial VCAM-1 expression, which was reversed by the NO synthase inhibitorN ω-monomethyl-l-arginine. These findings indicate that NO inhibits NF-κB activation and VCAM-1 expression by increasing the expression and nuclear translocation of IκBα.


The induction of vascular cell adhesion molecule-1 (VCAM-1) expression by tumor necrosis factor (TNF)-␣ requires the activation of nuclear factor-B (NF-B) via a process involving the phosphorylation and degradation of its cytoplasmic inhibitor, IB␣. We have shown that nitric oxide (NO) decreases VCAM-1 expression via inhibition of NF-B activation. To determine how NO inhibits NF-B, we studied the fate of IB␣ following TNF-␣ stimulation in the presence of NO donors S-nitrosoglutathione and sodium nitroprusside. Activation of NF-B by TNF-␣ occurred within 15 min and coincided with rapid degradation of IB␣. Co-treatment with NO donors did not prevent IB␣ phosphorylation or degradation. However, after 2 h of TNF-␣ stimulation, NO donors inhibited NF-B activation and augmented IB␣ resynthesis and nuclear translocation by 2.5-and 3-fold, respectively. This correlated with a 75% reduction in TNF-␣-induced VCAM-1 expression. In a time-dependent manner, NO donors alone caused the nuclear translocation of IB␣.
To confirm that NO donors have similar effects as endogenously derived NO, murine macrophage-like cells, RAW264.7, were co-cultured with endothelial cells. Induction of RAW264.7-derived NO inhibited lipopolysaccharide-induced endothelial VCAM-1 expression, which was reversed by the NO synthase inhibitor N -monomethyl-L-arginine. These findings indicate that NO inhibits NF-B activation and VCAM-1 expression by increasing the expression and nuclear translocation of IB␣.
The adhesion of circulating leukocytes to the vessel wall is an initiating event in atherogenesis and vascular inflammation (1,2). Under certain conditions, the "activated" endothelium expresses cell surface adhesion molecules which mediate specific interactions between the endothelium and circulating leukocytes (3,4). Factors that affect the induction of endothelial cell adhesion molecules such as vascular cell adhesion molecule (VCAM-1) 1 therefore may be important in regulating vascular inflammatory processes. Recent studies suggest that the activation of the pleotropic transcription factor nuclear factor-B (NF-B) is required for the transcriptional induction of endothelial cell adhesion molecules (5).
The activation of NF-B involves the degradation of its cytoplasmic inhibitor, IB (6,7). Presently, five distinct IB proteins have been shown to functionally retain NF-B in the cytoplasm and render it inactive (6). These IB proteins contain ankyrin repeat motifs that mask the nuclear localization sequence of NF-B subunits such as RelA (p65), c-Rel, and RelB (8,9). Of the different IB proteins, the best described is IB␣. Following cytokine stimulation, IB␣ is phosphorylated by a novel ubiquitinated serine kinase (10). Phosphorylation of IB␣ targets the IB␣ for ubiquitination and rapid degradation by 26 S proteasomes (10,11). The degradation of IB␣ then allows the unbound NF-B to translocate into the nucleus, where it can transactivate the enhancer elements of many proinflammatory genes (5).
Modulation of IB␣ function and expression has been shown to regulate NF-B activation. For example, the phosphorylation of IB␣ is a key regulatory step in the activation of NF-B (11,12). Indeed, recent studies indicate that salicylates and antioxidants inhibit NF-B and endothelial cell activation by preventing IB␣ phosphorylation and subsequent degradation (13)(14)(15)(16). Alternatively, the nuclear accumulation of IB␣ resulting from overexpression of IB␣ or following stimulation with tumor necrosis factor (TNF)-␣ has been shown to displace NF-B from its cognate DNA and terminate NF-B-mediated transcriptional activity (17,18). Finally, some of the anti-inflammatory effects of glucocorticoids may be mediated through their inhibitory effects on NF-B, since glucocorticoids are known to induce IB␣ expression (19,20).
We and others have shown that nitric oxide can inhibit NF-B and endothelial cell activation through non-cGMP-dependent mechanisms (21,22). Although NO donors appear to "stabilize" the NF-B-IB␣ heterotrimeric complex (23), the possibility that newly synthesized IB␣ could account for this stabilization has not been excluded. Furthermore, it is not known whether the induction of IB␣ expression by NO donors could actually lead to an increase in cytoplasmic or nuclear IB␣ protein levels. The purpose of this study, therefore, is to determine the mechanism(s) by which NO inhibits of NF-B activation and VCAM-1 expression in terms of its effects on IB␣ phosphorylation, expression, and nuclear accumulation.

EXPERIMENTAL PROCEDURES
Materials-Medium 199 was purchased from Life Technologies, Inc. Fetal calf serum was purchased from Atlanta Biologicals (Norcross, * This work was supported by Grants HL-52233 and HL-09650 from the National Institutes of Health. 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. GA). Recombinant TNF-␣ was purchased from Endogen (Cambridge, MA). Heparin sulfate, sodium nitrite, alkaline phospatase-conjugated secondary antibody, and p-nitrophenyl phosphate disodium were purchased from Sigma. Sodium nitroprusside (SNP) was purchased from Schwarz Pharma (Mannheim, Germany) and freshly prepared prior to each experiment in 5% dextrose. The NO donor, S-nitrosoglutathione (GSNO), was synthesized from reduced glutathione and sodium nitrite as described by Hart (24). The proteasome inhibitors, MG132 (Z-Leu-Leu-Leu-H) and ALL (N-acetyl-Leu-Leu-Met), were generously provided by Tucker Collins (Brigham & Women's Hospital, Boston, MA). Murine interferon-␥ (mIFN-␥) was purchased from Genzyme (Cambridge, MA). The anti-VCAM-1 antibody (mouse IgG) was obtained from Michael A. Gimbrone, Jr. (Brigham & Women's Hospital, Boston, MA). The rabbit polyclonal RelA and the affinity-purified rabbit IB␣/MAD-3 C-15 antibody were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Polyvinylidene fluoride transfer membranes (Immobilon P) were purchased from Millipore (Bedford, MA). The micro-BCA protein assay was obtained from Pierce. Low molecular weight protein standards were purchased from Bio-Rad.
Cell Culture-Human saphenous vein endothelial cells of less than four passages were cultured in a growth medium containing M199, 5% fetal calf serum, 50 g/ml endothelial cell growth factor (Pel-Freez Biological, Rogers, AK), and 100 g/ml heparin sulfate. For cell surface immunoassays, endothelial cells were grown on gelatin-coated 96-well microplates (Greiner, Nü rtingen, Germany). All cell cultures were incubated at 37°C with 5% CO 2 atmosphere. Endothelial cells were identified by their typical morphological pattern (cobblestone morphology) and by immunostaining of representative plates for von Willebrandt factor antigen as described previously (25).
For the separated co-culture system, endothelial cells were grown on coverslips coated with 0.1% gelatin in 6-well culture dishes (Falcon, Franklin Lakes, NY). Murine macrophage-like cells, RAW264.7, were grown on inserts with 0.4-m pore size (Falcon). These inserts were placed above the underlying endothelial cells and shared the same culture medium of Dulbecco's modified Eagle's medium with 10% fetal calf serum and 50 g/ml endothelial cell growth factor.
Treatment Conditions-Endothelial cells were stimulated with TNF-␣ (1000 units/ml) in the presence and absence of the indicated concentrations of NO donors. The preincubation period was 30 min for GSNO, 10 min for SNP, and 60 min for N-acetylcysteine (30 mM) before TNF-␣ stimulation. Degradation of phosphorylated IB␣ was inhibited in the presence of a 26 S proteasome inhibitor, MG132 (10 M), which was added 30 min prior to TNF-␣ stimulation. Following incubation with the indicated agents, representative endothelial cell plates were tested for cellular viability by trypan blue exclusion. Only endothelial cell monolayers with more than 95% viable cells were used in experiments. For enzyme immunoassays, each experiment was performed in quadruplicate. The separated co-culture system was stimulated with LPS (10 ng/ml), mIFN-␥ (400 units/ml), alone or in combination, in the presence or absence of 3 mM N -monomethyl-L-arginine (LNMA) for 24 h.
Cell Surface Enzyme-linked Immunosorbent Assay-Enzyme immunoassays were performed on confluent human endothelial cell monolayers using a monoclonal VCAM-1 antibody. Nonbinding control antibodies (OX 6, against MHC class II antigen) were used in each experiment. The monolayers were washed with PBS. Following 4 h of stimulation with TNF-␣, endothelial cells were fixed on 96-well microtiter plates with 1% paraformaldehyde for 45 min and then incubated with the indicated monoclonal antibodies (1:100 dilution in PBS). Following 45 min of incubation with the primary antibody at 37°C, the alkaline phosphatase-conjugated secondary anti-mouse antibody was applied. Finally, the phosphatase substrate, p-nitrophenyl phosphate disodium, was added in an alkaline buffer solution, pH 10.3. Alkaline phosphatase activity was measured after 30 min at an absorbance wavelength of 405 nm with a Milenia microtiter plate reader (DPC MKA 220, DPC, Los Angeles, CA).
Immunohistochemistry-Endothelial cells grown on coverslips in the separated co-culture system were fixed with 4% formaldehyde at 4°C for 10 min and then incubated with VCAM-1 antibody for 1 h. After three washing steps in PBS, a biotinylated anti-mouse antibody (Bio Genex, San Ramon, CA) was added for 20 min. Following further washing steps, the cells were incubated with a streptavidin alkaline phosphatase-conjugated antibody and stained with fast red substrate (Dako, Carpinteria, CA). Immunostaining was visualized using a Nikon Optiphot-2 microscope.
Immunofluorescence-Endothelial cells were grown on gelatincoated coverslips in 6-well plates. Following incubation with TNF-␣ or NO donors for the indicated time intervals, cells were fixed and perme-abilized with acetone at Ϫ20°C for 2 min (RelA) or with 100% methanol at Ϫ20°C for 7 min (IB␣). Blocking was performed with 3% normal goat serum for 20 min. Cells were incubated with a rabbit polyclonal antibody directed against the NF-B subunit, RelA (p65), or with anti IB␣ for 1 h at room temperature. A biotinylated goat anti-rabbit antibody was used as secondary antibody. After 45 min of incubation with the secondary antibody, streptavidin-fluorescein isothiocyanate was added for 45 min. Immunofluorescence was visualized using an Olympus BX 60F microscope. Photographic images were taken from four random fields.
Nitrite Assay-The NO production was determined by assaying for nitrite accumulation. Briefly, following stimulation of the separated co-culture system with the indicated agents, a small aliquot of medium (50 l) was added to equal volumes (100 l) of 0.5% naphthylethylenediamine dihydrochloride and 1% sulfanilamide in 2.5% phosphoric acid (Griess reagents). Absorbance was measured at 543 nm, and nitrite concentration was determined using a standard curve of sodium nitrite concentrations from 1.6 to 200 nmol/ml. In some experiments, the medium was first treated with nitrate reductase (Sigma) to convert all of the nitrate to nitrite before applying the Griess reagents. Preliminary studies indicate that the nitrite to nitrate ratio in all treatment conditions was always greater than 5:1.
Western Blotting-After specific incubation periods with TNF-␣ and NO donors, endothelial cells were rinsed once with ice-cold PBS before addition of the lysis buffer (100 mM Tris-HCl, pH 6.8, 4% SDS, 20% glycerol, 1 mM sodium orthovanadate, 1 mM NaF, and 1 mM phenylmethylsulfonyl fluoride) to the dishes on an ice tray. The cell lysates were scraped, boiled, and centrifuged for 2 min at 14,000 ϫ g. Total cell lysates (100 g of protein) and low molecular weight markers were separated by SDS-polyacrylamide gel electrophoresis (12% running, 4% stacking) and electrophoretically transferred to polyvinylidene fluoride membranes (Immobilon P, 0.45 m pore size). Immunodetection was accomplished using the enhanced chemoluminescence kit (ECL Kit, Amersham Corp.). Densitometry was performed using the NIH Image 1.49 program (26).
Statistics-Readings from enzyme immunoassays are expressed as mean Ϯ S.D. Means were compared by Student's paired t-test. A confidence level of p Ͻ 0.05 was taken to represent a significant difference between two means. Multiple comparisons were done by analysis of variance.

Effect of NO Donor on VCAM-1 Expression-
The NO donor, GSNO, inhibited TNF-␣-induced VCAM-1 expression in a concentration-dependent manner (IC 50 ϭ ϳ30 M) (Fig. 1). Maximal 75% inhibition was achieved at a GSNO concentration of 200 M. The highest concentration of GSNO used in our experiments (1 mM) did not alter cell viability as assessed by cell number, cellular morphology, DNA content, and trypan blue exclusion.
Effect of Endogenous NO on VCAM-1 Expression-To confirm that the effects of the NO donors are comparable to the effects of endogenously derived NO, we developed a separated co-culture system where murine macrophage-like cells, RAW264.7, were grown on culture inserts above endothelial cell monolayers, but sharing the same culture medium (Fig. 2). In this separated co-culture system, no basal endothelial VCAM-1 expression or RAW264.7-derived NO production was observed ( Fig. 2B and Table I). Treatment with LPS (10 ng/ml) induced VCAM-1 expression in endothelial cells but not in RAW264.7 (Fig. 2D). Murine IFN-␥ was inactive in human endothelial cells (i.e. no MHC class II expression) and, by itself, could not induce Type II NO synthase (iNOS) expression in RAW264.7 cells (data not shown). The combination of LPS and mIFN-␥ induced iNOS expression and NO production in RAW264.7 but not in human endothelial cells.
When the separated co-culture system was stimulated with the combination of LPS and mIFN-␥, NO production was increased significantly (0.6 Ϯ 0.1 M to 27.3 Ϯ 4.5 M, p Ͻ 0.01), while endothelial VCAM-1 expression was decreased by 76% compared with stimulation with LPS alone (Fig. 2E and Table  I). However, addition of the iNOS inhibitor, LNMA (3 mM), to the separated co-culture system that had been stimulated with LPS and IFN-␥ inhibited NO production by 91% and augmented endothelial VCAM-1 expression by 5-fold compared with the level of VCAM-1 expression stimulated with LPS and IFN-␥ (Fig. 2F). Interestingly, treatment with LNMA alone resulted in mild endothelial VCAM-1 expression (12% of LPSinduced expression), suggesting that basal endothelial NO production can functionally and tonically inhibit VCAM-1 expression (Fig. 2C).
Effect of NO on Cellular Localization of RelA-Stimulation of endothelial adhesion molecule expression requires the activation of NF-B (5). Since the activation of NF-B involves the nuclear translocation of NF-B subunits, we followed the intracellular localization of NF-B subunit, RelA, by immunofluorescence following TNF-␣ stimulation in the presence or absence of GSNO. In unstimulated endothelial cells, RelA is predominantly localized to the cytoplasm with little, if any, present in the nucleus (Fig. 3A). Within 15 min following TNF-␣ stimulation, there is observable nuclear accumulation of RelA, which was sustained at 2 h (Fig. 3, B and D).
Treatment with GSNO (200 M) did not prevent the nuclear translocation of RelA after 15 min of TNF-␣ stimulation (Fig.  3C). After 2 h of TNF-␣ and GSNO, however, RelA was localized to both the cytoplasm and nucleus, suggesting partial inhibition of NF-B activation and possible reverse nuclear to cytoplasmic translocation of NF-B by GSNO (Fig. 3E). Treatment with GSNO alone did not activate NF-B at the 2 h time point (Fig. 3F). Specificity was determined by the absence of RelA immunofluorescence when nonimmune serum was used instead of RelA antibody. These results indicate that the inhibitory effects of NO donors on NF-B activation had a delayed onset (i.e. Ͼ 15 min) and suggest that NO does not prevent the initial degradation of IB␣ following TNF-␣ stimulation.
Effect of NO on IB␣ Phosphorylation-Since the activation of NF-B involves the phosphorylation and subsequent degradation of its cytoplasmic inhibitor, IB␣, we followed the fate of IB␣ after stimulation with TNF-␣ (1000 units/ml) in the presence and absence of GSNO (200 M). Stimulation with TNF-␣ for 15 min caused an almost complete disappearance of IB␣, which was not prevented by co-treatment with GSNO (Fig. 4). Since phosphorylated IB␣ has a very short half-life and is difficult to visualize by immunoblotting, we treated endothelial cells with a relatively specific IB␣ 26S proteasome inhibitor, MG132 (10 M), to "protect" phosphorylated IB␣. In the presence of MG132, stimulation with TNF-␣ for 15 min produced two bands corresponding to phosphorylated and nonphosphorylated IB␣.
Treatment with GSNO (200 M) did not prevent TNF-␣stimulated IB␣ phosphorylation when compared with that of TNF-␣ alone (Fig. 4). Incubation with a higher concentration of GSNO (500 M) or with a second NO donor, SNP (500 M), also had no effect on TNF-␣-stimulated IB␣ phosphorylation (data not shown). Treatment with N-acetylcysteine (30 mM) stabi-lized IB␣ and prevented its degradation following TNF-␣ stimulation. In contrast to MG132, a nonspecific proteasome inhibitor, ALL (10 M), did not prevent IB␣ degradation following IB␣ phosphorylation.
Effect of NO on IB␣ Expression-To investigate the effects of NO on IB␣ expression, endothelial cells were incubated for longer time periods with GSNO. Stimulation with TNF-␣ (1000 units/ml) resulted in almost complete IB␣ degradation by 15 min, followed by the reappearance or resynthesis of IB␣ at 2 h (Fig. 5A). Co-treatment with GSNO (200 M) did not prevent TNF-␣-induced IB␣ degradation at 15 min but did augment TNF-␣-induced IB␣ expression by 2.5-fold after 2 h compared to that of TNF-␣ alone. Treatment with GSNO alone produced a time-dependent 1.5-and 2.5-fold increase in IB␣ levels above basal levels after 1 and 2 h, respectively (Fig. 5B).
Effect of NO on IB␣ Nuclear Translocation-To determine whether NO can affect the cellular localization of IB␣, we performed immunofluorescence studies of IB␣ following GSNO treatment. In unstimulated endothelial cells, IB␣ is predominantly localized to the cytoplasm with little, if any, localized to the nucleus (Fig. 6A). In a time-dependent manner, treatment with GSNO (200 M) caused a progressive nuclear accumulation of IB␣ (Fig. 6, B-D). Incubation with nonimmune serum yielded no appreciable immunofluorescence (data not shown). The nuclear translocation of IB␣ by NO, therefore, occurred in the absence of NF-B activation, since GSNO alone does not activate NF-B (Fig. 3F).
To determine whether the nuclear accumulation of IB␣ induced by NO can terminate NF-B signaling in the nucleus by binding to RelA, we studied the appearance of RelA-associated IB␣ in the nucleus. Immunoblotting of nuclear extracts that have been immunoprecipitated with agarose-conjugated anti-RelA antibody demonstrated that RelA-associated IB␣ begins to accumulate in the nuclear extracts of endothelial cells following 2 h of TNF-␣ (1000 units/ml) stimulation (Fig. 7). This time frame is consistent with the appearance of IB␣ in the nucleus following GSNO treatment (Fig. 6, C and D).
Compared withTNF-␣ alone, co-treatment with GSNO (200 M) caused a further increase in nuclear RelA-associated IB␣ levels resulting in a 3-fold increase after 6 h (Fig. 7). Possible cytoplasmic contamination of the nuclear extracts was excluded by the absence of G-protein ␣ i2 subunit in the nuclear extracts. In addition, IB␣, which is present in the cytoplasm of The separated co-culture system consisting of RAW264.7 and endothelial cells was stimulated with the indicated conditions for 24 h as described. Nitrite assay was performed on the co-culture medium, and surface enzyme-linked immunosorbent assay for VCAM-1 was performed on endothelial cell monolayers. Values are means Ϯ S.E. Experiments were performed three times in quadruplicate. *p Ͻ 0.05 compared with control (no stimulation). **p Ͻ 0.01 compared with LPS and mIFN-␥. unstimulated endothelial cells, was not observed in the corresponding nuclear extracts. DISCUSSION The mechanism by which NO inhibits endothelial VCAM-1 expression occurs, in part, via the inhibition of NF-B (21,22). Since the activation of NF-B is dependent upon its association with its inhibitor, IB␣, which not only binds and prevents the nuclear translocation of NF-B (6, 7) but also may displace nuclear NF-B from its cognate DNA (16), the level of IB␣ in both the cytoplasm and nucleus may be important in determin-ing the time of onset, duration, and magnitude of NF-B activation. Regulation of NF-B activation by IB␣ therefore may ultimately determine the level of VCAM-1 expression in response to cytokines. We find that NO inhibits NF-B activation and VCAM-1 expression by increasing cytoplasmic and nuclear levels of IB␣.
The NO donor, GSNO, has been shown to slowly release NO under physiological pH conditions (27). Furthermore, the findings of the separated co-culture system suggest that the effects of the NO donors is most likely due to biologically active NO and not some other metabolite. However, we cannot exclude possible interactions of NO with other reactive oxygen intermediates such as superoxide anion to form another NO derivative such as peroxynitrite (28). Neither the NO donors used in our experiments nor endogenously derived NO from RAW264.7 cells cause any cellular toxicity as determined by cell count, DNA content, cellular morphology, and trypan blue exclusion. Thus it is unlikely that changes in NF-B activation or VCAM-1 expression were due to direct or indirect toxic effects of NO. Furthermore, we have previously shown that two constitutively expressed endothelial cell surface proteins, E1/1 and MHC class I, were not affected by similar treatment with these NO donors (21).
We find that the induction of IB␣ by NO or TNF-␣ resulted in a delayed increase in IB␣ protein levels (i.e. Ն2 h). Higher levels of IB␣ protein in the cytoplasm could replace the initially degraded IB␣ and help retain and "stabilize" the NF-B/IB␣ complex in the cytoplasm. The binding of newly synthesized IB␣ to cytoplasmic NF-B therefore may prevent further nuclear translocation of NF-B subunits and limit VCAM-1 gene transcription. Treatment with NO donors, however, did not inhibit IB␣ phosphorylation or degradation by 26 S proteasomes following TNF-␣ stimulation. This is in contrast to our previous study, which suggested that NO may stabilize IB␣ (23). In the present study, a more detailed analysis using shorter stimulation periods indicates that stabilization of latent IB␣ is not responsible for inhibition of NF-B. Retention of NF-B in the cytoplasm occurs later only after IB␣ resynthesis. An increase in IB␣ protein level in the nucleus may also serve to terminate NF-B-mediated VCAM-1 gene transcription by displacing NF-B from its putative cis-acting element(s) (16). This is consistent with our finding that following treatment with NO, and to a lesser extent TNF-␣, IB␣ not only accumulates in the nucleus but also is bound to nuclear translocated RelA. Binding of RelA to IB␣ may also serve to expel RelA from the nucleus into the cytoplasm. Thus, the observed increase in nuclear IB␣ levels and possible reverse translocation of NF-B may be additional mechanisms by which NO can rapidly terminate NF-B-mediated gene transcription. It is not known, however, whether this increase in nuclear IB␣ levels is due to enhanced nuclear translocation of IB␣ or occurs Nuclear extracts (100 g) were immunoprecipitated with agarose-conjugated RelA antibody followed by immunoblotting with IB␣-specific antibody. As controls, the corresponding nuclear and cytoplasmic extracts were immunoblotted with G i␣2 -specific antibody. Studies were performed three times with similar results.
indirectly as a result of stoichiometric increases in cytoplasmic IB␣ levels.
The activation of NF-B following TNF-␣ stimulation leads to the transcriptional induction of IB␣ via B cis-acting elements in the IB␣ promoter (29 -31). The induction of IB␣ by NF-B, therefore, serves as an autoregulatory mechanism for terminating NF-B-mediated gene transcription (15,31). In our studies, the induction of IB␣ by NO donors occurred independent of NF-B activation in unstimulated human endothelial cells. The induction of IB␣ by NO, however, is somewhat similar to the effect of glucocorticoids, which have also been shown to induce IB␣ gene transcription (19,20). Although the NO-responsive cis-acting element(s) has yet to be identified, analysis of the Ϫ2.1-kilobase porcine, Ϫ1.6-kilobase murine, and Ϫ385-base pair human IB␣ promoter does not reveal site corresponding to putative glucocorticoid response elements (GRE) (29 -31). Thus, it is interesting to speculate whether NO and glucocorticoids share a similar transcriptional signaling pathway leading to the transactivation of the IB␣ gene.
In summary, NO inhibits VCAM-1 expression in cultured human vascular endothelial cells via a novel mechanism involving the induction and nuclear translocation of IB␣. These effects of NO appear to be distinct from that of antioxidants and salicylates, which prevent IB␣ phosphorylation and degradation (13)(14)(15)(16). Further investigation into how NO donors induce IB␣ gene transcription and modulate NF-B activity may provide greater insights into the role of NO in the vascular wall.