Novel Inhibitors of Cytokine-induced IκBα Phosphorylation and Endothelial Cell Adhesion Molecule Expression Show Anti-inflammatory Effects in Vivo*

We have identified two compounds that inhibit the expression of endothelial-leukocyte adhesion molecules intercellular adhesion molecule-1, vascular cell adhesion molecule-1, and E-selectin. These compounds act by inhibiting tumor necrosis factor-α-induced phosphorylation of IκB-α, resulting in decreased nuclear factor-κB and decreased expression of adhesion molecules. The effects on both IκB-α phosphorylation and surface expression of E-selectin were irreversible and occurred at an IC50 of approximately 10 μm. These agents selectively and irreversibly inhibited the tumor necrosis factor-α-inducible phosphorylation of IκB-α without affecting the constitutive IκB-α phosphorylation. Although these compounds exhibited other activities, including stimulation of the stress-activated protein kinases, p38 and JNK-1, and activation of tyrosine phosphorylation of a 130–140-kDa protein, these effects are probably distinct from the effects on adhesion molecule expression since they were reversible. One compound was evaluated in vivo and shown to be a potent anti-inflammatory drug in two animal models of inflammation. The compound reduced edema formation in a dose-dependent manner in the rat carrageenan paw edema assay and reduced paw swelling in a rat adjuvant arthritis model. These studies suggest that inhibitors of cytokine-inducible IκBα phosphorylation exert anti-inflammatory activity in vivo.

The adhesion of circulating leukocytes to vascular endothelium is critical to inflammatory responses (reviewed in Refs. [1][2][3]. Interaction of the selectin family of adhesion proteins and lectin counter-receptors is the predominant mechanism mediating initial adhesion between leukocytes and the vessel wall. The expression of endothelial-leukocyte adhesion molecule-1 (E-selectin, CD62E), vascular cell adhesion molecule-1 (VCAM-1, 1 CD106), and intercellular adhesion molecule-1 (ICAM-1, CD54) on the surface of endothelial cells is elevated at sites of inflammation (2,4). Induction of these molecules by tumor necrosis factor-␣ (TNF␣) and other inflammatory cytokines is regulated at the level of gene transcription and requires binding of the transcription factor nuclear factor-B (NF-B) to the regulatory regions within the promoters of each of these genes (5)(6)(7)(8)(9)(10)(11)(12).
The NF-B/Rel transcription factor family plays an important role in cytokine-induced gene activation (13)(14)(15). The Rel family includes p50 (NFKB1), p52 (NFKB2), p65 (RelA), RelB, v-Rel, and c-Rel. In endothelial cells, the p50⅐p65 heterodimer is the predominant species that binds to B consensus sequences in the VCAM-1, ICAM-1, and E-selectin genes and activates gene transcription. NF-B is located in the cytoplasm of cells in an inactive form in association with the inhibitor IB-␣. In response to TNF␣ stimulation, IB-␣ is phosphorylated on 2 serine residues (Ser-32 and Ser-36), ubiquitinated, and degraded by a proteosome-dependent pathway allowing active NF-B to translocate to the nucleus where it can activate gene expression (16 -23). Many NF-B-dependent genes including the adhesion molecules and several cytokine genes are important mediators of inflammation (reviewed in Ref. 24). A diverse range of agents that block NF-B signaling has been shown to decrease expression of adhesion molecules (25)(26)(27)(28)(29)(30)(31). Recently, several clinically important anti-inflammatory agents including glucocorticoids, salicylates, and nitric oxide have been reported to inhibit NF-B-driven gene expression which may explain, at least in part, the anti-inflammatory actions of these drugs (24 -26, 31-34). Thus, novel agents that block NF-B/IB-␣ signaling have the potential to inhibit a wide range of inflammatory processes.
In this study, we identified two novel pharmacologic agents that inhibit the TNF␣-induced surface expression of ICAM-1, VCAM-1, and E-selectin in human endothelial cells. These compounds were examined for their effects on cytokine-induced NF-B/IB-␣ signaling. Both compounds decreased TNF␣-induced nuclear translocation of NF-B through inhibition of the TNF␣-induced phosphorylation of IB-␣. Compound 1 selectively inhibited the TNF␣-inducible phosphorylation of IB-␣ without affecting the constitutive IB-␣ phosphorylation. To determine whether these agents may inhibit other cellular phosphorylation events, we examined the effects of compound 1 on TNF␣-induced activity of the stress-activated protein kinases, p38 and JNK-1. This agent increased the activity of p38 kinase and JNK-1 but had little or no effect on the activity of the MAP kinase, ERK-1. The agent was also examined for effects on protein tyrosine phosphorylation since agents that block tyrosine phosphorylation have been reported to inhibit NF-B signaling and adhesion molecule expression (29). Treatment of endothelial cells with the test compound did not detectably inhibit protein tyrosine phosphorylation but rather resulted in an elevated level of a tyrosine-phosphorylated pro-tein of molecular mass 130 -140 kDa. We evaluated compound 2 in two animal models of inflammation. This agent reduced swelling in a dose-dependent manner in both the rat carrageenan paw edema assay and in a rat adjuvant arthritis model. Thus, we have identified a novel class of anti-inflammatory agents that act by selectively inhibiting the TNF␣-induced phosphorylation of IB␣ resulting in decreased expression of endothelial adhesion molecules.
Cell Culture, Cytokine Treatment, and Toxicity Assay-Human umbilical vein endothelial cells (HUVEC) were isolated and maintained in culture using previously described procedures (36). For experiments on cytokine induction, cells were exposed to recombinant human TNF␣ at a final concentration of 100 units/ml in complete media for the times indicated. The proteosomal inhibitor carbobenzoxyl-leucinyl-leucinylleucinal-H (MG115) was prepared as a 40 mM stock solution in Me 2 SO and added to complete medium to a final concentration of 40 M. Cell toxicity was assessed by morphology and by 3-(4,5-dimethyl thiazol-2yl)-2,5-diphenyl tetrazolium bromide (MTT) assay (37).
Interleukin-6 and Interleukin-8 Assays-The effects of compounds 1 and 2 on interleukin-6 (IL-6) and interleukin-8 (IL-8) production were evaluated on HUVEC that were grown to confluence on 96-well microtiter plates. The cells were preincubated with the drugs at concentrations of 0, 1, 5, 10, or 25 M and then incubated with TNF␣ (10 units/ml) and drug for 16 h. The culture supernatants were removed and assayed for IL-6 and IL-8 content using enzyme-linked immunoassay kits from R & D Systems (Minneapolis, MN).
Electrophoretic Mobility Shift Assay-Nuclear extracts were prepared from test or control HUVEC in the presence of 10 g/ml aprotinin, 10 g/ml leupeptin, 1.5 g/ml pepstatin A, 40 M ALLN (calpain inhibitor 1), 1 mM sodium orthovanadate, and 1 mM sodium fluoride as described previously (10). Oligonucleotides were gel-purified, annealed, and end-labeled with [␣-32 P]dCTP (50 Ci; specific activity of 3000 Ci/mmol, NEN Life Science Products) and the Klenow fragment of Escherichia coli DNA polymerase I. Binding reactions were performed in the presence of 10 mM Tris, pH 7.5, 1 mM dithiothreitol, 1 mM EDTA, 5% glycerol, and 1 g of poly(dI⅐dC) and electrophoresis was carried out as described previously (39). The following oligonucleotides were utilized: VCAM-B (vNF-WT), 5Ј CTGGGTTTCCCCTTGAAGGGATTTC-CCTC and the complementary strand. Protein DNA complexes were resolved on 4% polyacrylamide gels.
Western Blots-Following experimental treatment of HUVEC, cytosolic and nuclear protein extracts were prepared, subjected to electrophoresis on 10% SDS-polyacrylamide gels, and transferred to nitrocellulose in 25 mM Tris, 192 mM glycine, 5% methanol at 100 V for 1 h as described previously (10,27). Anti-IB␣ and anti-p38 antisera were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA) and used at dilutions of 1:1000. Rabbit antisera directed against the phosphorylated p38 (Tyr-182) were obtained from New England Biolabs (Beverly, MA) and used at a 1:1000 dilution. Mouse anti-phosphotyrosine antibodies were obtained from Upstate Biotechnology Inc. (Lake Placid, NY) and used at 1:1000 dilution. Immunoreactive proteins were detected by enhanced chemiluminescent protocol (Amersham Corp.) using 1:10,000 horseradish peroxidase-linked donkey anti-rabbit or sheep anti-mouse secondary antiserum. Blots were exposed to film for 1-15 min and then developed.
Immune Complex Kinase Assays-Extracts were prepared from control and TNF␣-treated HUVEC. Cells were solubilized with Triton lysis buffer (TLB, 20 mM Tris, pH 7.4, 1% Triton X-100, 10% glycerol, 137 mM NaCl, 2 mM EDTA, 25 mM ␤-glycerophosphate, 1 mM sodium orthovana-date, 2 mM pyrophosphate, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin). Extracts were centrifuged at 14,000 ϫ g for 15 min at 4°C. The JNK, p38, or ERK protein kinases were immunoprecipitated by incubation for 1 h at 4°C with specific rabbit polyclonal antibodies bound to protein-A Sepharose (Pharmacia Biotech Inc.). The rabbit polyclonal JNK-1 and p38 antibodies have been described (40). The immunoprecipitates were washed twice with TLB and twice with kinase buffer (20 mM Hepes, pH 7.4, 20 mM ␤-glycerophosphate, 20 mM MgCl 2 , 2 mM dithiothreitol, 0.1 mM sodium orthovanadate). The kinase assays were initiated by the addition of 1 g of substrate protein and 50 M [␥-32 P]ATP) (10 Ci/mmol) in a final volume of 25 l. The reactions were terminated after 15 min at 30°C by addition of Laemmli sample buffer. Control experiments demonstrated that the phosphorylation reaction was linear with time for at least 30 min under these conditions. The phosphorylation of the substrate proteins was examined by SDSpolyacrylamide gel electrophoresis followed by autoradiography.
In Gel Kinase Assay-In gel kinase assay for the proteins that phosphorylate IB-␣ was carried out according to the method of Hibi et al. (41) and as detailed below. Whole cell extracts were prepared from HUVEC treated with TNF␣ (100 units/ml) for 15 min in the presence or absence of compound 1 (20 M, pretreatment for 1 h) as indicated. Proteins were separated on a 10% SDS gel containing 0.5 mg/ml HIS-IB-␣. Gels were washed two times in 20% propanol, 50 mM Hepes, pH 7.6, for 30 min and two times in buffer A (50 mM Hepes, pH 7.6, 5 mM 2-mercaptoethanol) for 30 min, followed by a 1-h incubation with buffer A containing 6 M urea, 1 h each in 3, 1.5, and 0.75 M urea in buffer A and 0.05% Tween 20 and 1 h in buffer A with 0.05% Tween 20. The kinase assay was carried out for 1 h at 30°C in the presence of 50 M ATP, 5 Ci/ml [ 32 P]ATP, 20 mM Hepes, pH 7.6, 20 mM MgCl 2 , 20 mM ␤-glycerophosphate, 20 mM p-nitrophenyl phosphate, 1 mM sodium vanadate, 2 mM dithiothreitol. The gel was washed with 5% trichloroacetic acid and 1% sodium pyrophosphate, dried, and exposed to film. A separate gel with no HIS-IB-␣ was assayed as a control.
Carrageenan Paw Edema-Male Harlan Sprague Dawley rats 150 -175 g were used. A 1% suspension of carrageenan (Marine Colloids, Springfield, MA) in distilled water was administered to rats as 0.1 ml subplantar injection into the footpad of the right hind paw as described previously (42). One h prior to injection rats were treated intraperitoneally with vehicle (polyethylglycol 400 diluted 1:5 in 5% bovine serum albumin/H 2 O) or a fine suspension of compound 2 (1, 5, or 50 mg/kg) in vehicle. A positive control group was also included in which rats were pretreated with 20 mg/kg ibuprofen. Four hours after carrageenan administration, the volume of the injected paw was measured by means of a water displacement plethysmograph. Edema volumes were determined as the difference between the paw volumes of each rat at time 0 and 4 h. Each group contained five animals. Data were analyzed by a one-way analysis of variance and, if indicated, differences between groups analyzed by Bonferroni's modified t test. A p Ͻ 0.05 was considered significant.
Adjuvant Arthritis-Inbred, male Lewis rats 8 -10 weeks of age weighing 250 -275 g were obtained from Charles River (Wilmington, MA). Five animals per group were used, and the animals were allowed to feed ad libitum on laboratory rat chow and water. Heat-inactivated Mycobacterium butyricum (Difco) was suspended at 10 mg/ml in mineral oil (Purepac Lubinol, Purepac Pharmaceuticals, Elizabeth, NJ) and administered as 0.1-ml injection (1 mg/animal) at the base of the tail. Paw volumes were determined by a water displacement plethysmograph as described above. Volumes were determined on the indicated dates and values compared with initial time 0 measurements. Vehicle (0.5% methyl cellulose) or drug (compound 2 or dexamethasone at the indicated concentrations) was administered once a day as an intraperitoneal injection (200 l). Data from these studies are expressed as the mean difference in foot pad volume. At day 20 animals were sacrificed by CO 2 inhalation.

Identification of Novel Inhibitors of ICAM-1, VCAM-1, and E-selectin Surface
Expression-Two structurally related compounds, compound 1 and compound 2 ( Fig. 1), were identified as inhibitors of cytokine-induced surface expression of ICAM-1 as measured by fluorescence immunoassay as described by Gerritsen et al. (42). Drug effects on TNF␣-induced surface expression of E-selectin, VCAM-1, and ICAM-1 were determined by fluorescence immunoassay as described under "Experimental Procedures." Compound 1 and compound 2 inhibited the surface expression of all three adhesion molecules with IC 50 values in the range of 5-10 M (Fig. 2). To determine whether the effects of the test compounds were reversible, we compared E-selectin levels in cells stimulated with TNF␣ in the presence of compound 1 (10 M) with the levels obtained when the cells were pretreated with compound 1 followed by a 1 h "washout" period and then a 3-h stimulation with TNF␣ in the absence of test drug. Treatment of HUVEC with 10 M compound inhibited E-selectin expression by 57% and a similar level of inhibition was seen when the drug was "washed out" prior to TNF␣ treatment (Fig. 3). Thus, compound 1 irreversibly inhibits surface expression of E-selectin. Other effects of compound 1 were reversible (see below) suggesting that the inability to reverse the inhibition of TNF␣-induced E-selectin expression was not due to retention of the compound by the cell.
There was no detectable cytotoxicity as measured by MTT assay even after 16 h treatment of cells with this dose (10 M) of the test compound (IC 50 in the MTT assay ranged from 25-38 M). Thus, it is likely that the drug irreversibly modifies a cellular target.
We also determined whether the test compounds could inhibit E-selectin surface expression when added after initiation of TNF␣ treatment. Maximal inhibition occurred when the test compound was included from the start of the TNF␣ induction period, and no significant inhibition was observed when the test compound was added after 1 h (Fig. 3). These results are consistent with the drug acting rapidly and irreversibly within the 1st h of the TNF␣ induction.
Additionally, we determined whether the test compounds could inhibit cytokine production, as well as expression of leukocyte adhesion molecules. Compounds 1 and 2 also reduced TNF␣-induced IL-6 and IL-8 production in a dose-dependent manner. This inhibition was greater than 50% at 10 M and virtually complete at 25 M (data not shown).
Inhibition of Nuclear NF-B-The TNF␣-induced expression of adhesion molecules E-selectin, VCAM-1, and ICAM-1 requires the transcription factor NF-B (5-12). Therefore, we evaluated the test compounds for effects on nuclear translocation of NF-B. We carried out electrophoretic mobility shift assay to determine the levels of NF-B in nuclear extracts from HUVEC treated with TNF␣ in the presence of compound 1 or compound 2. As previously observed, TNF␣-induced nuclear translocation of NF-B occurs within 15 min in the absence of test compound (Fig. 4A, lane 2 and Fig. 4B, lane 2). At 20 M, both test compounds completely inhibited nuclear NF-B (Fig.  4A, lane 4, Fig. 4B, lane 3). A lower dose of compound 1 (10 M) also reduced nuclear NF-B (Fig. 4A, lane 3).
Inhibition of TNF␣-inducible IB-␣ Phosphorylation and Irreversible Stabilization of IB-␣-The TNF␣-induced regulation of NF-B involves the phosphorylation, ubiquitination, and degradation of the cytoplasmic inhibitor, IB-␣ (16 -23). In endothelial cells, the phosphorylation and degradation of IB-␣ have been shown to occur within 15 min of TNF␣ treatment allowing NF-B to translocate to the nucleus where it can activate gene expression (8). To determine whether the test compounds may affect TNF␣-inducible phosphorylation and/or degradation of IB-␣, we examined the levels of IB-␣ in the cytoplasm of endothelial cells pretreated with increasing concentrations of test compounds and then stimulated with TNF␣ for 15 min. The results of Western blot analysis of endothelial cytoplasmic extracts using IB␣-specific antisera are shown in Fig. 5. A 37-kDa protein was detected in cytoplasmic extracts from unstimulated cells (Fig. 5, lane 1). Treatment of HUVEC with TNF␣ led to a rapid loss of IB-␣ from the cytoplasm (Fig.  5, lane 2). Both test compounds stabilized IB-␣ in a dose-dependent manner with an IC 50 value of approximately 10 M (Fig. 5, lanes 3-5). There was a clear correlation between the concentration of drug that stabilized IB-␣, the concentration that inhibited nuclear levels of NF-B, and the concentration that inhibited adhesion molecule expression. The levels of p38 were not significantly affected by the test compounds (Fig. 5) suggesting that these agents did not result in nonspecific effects on protein stability. In addition, these concentrations of test compounds did not affect protein synthesis as assessed by [ 3 H]leucine incorporation into 5% trichloroacetic acid-precipitable protein.
The effects of test compounds on levels of IB-␣ protein could be due to an inhibition of IB-␣ phosphorylation or a block of degradation. To determine whether these agents affect IB-␣ phosphorylation, we examined the IB-␣ levels in cells treated with both compound 1 and the proteosome inhibitor carbobenzoxyl-leucinyl-leucinyl-leucinal-H (MG115) (Fig. 5, lanes 6 and  7). As previously reported (19,22,27), the proteosome inhibitor blocks the degradation of IB-␣, allowing visualization of the phosphorylated form of IB-␣, detected as a slower migrating band (Fig. 5, lane 7, IB-␣-P). The faster migrating form present in unstimulated cells is basally phosphorylated IB␣ (18,19). The IB-␣ protein present in TNF␣-stimulated cells treated with compound 1 appears as a single faster migrating band corresponding to basally phosphorylated IB-␣ (Fig. 5,  lane 5). Stimulation of cells with TNF␣ in the presence of both compound 1 and the proteosome inhibitor resulted in stabilization of the basally phosphorylated form of IB-␣ (Fig. 5, lane  6). Little or no inducibly phosphorylated IB-␣ protein was detected (Fig. 5, lanes 6 and 7, IB-␣-P), suggesting that compound 1 inhibits TNF␣-inducible phosphorylation of IB-␣. Similar results were observed with compound 2 (data not shown). The IB-␣ which has not undergone inducible phosphorylation is not targeted for degradation (19,43). Thus, in contrast to the proteosome inhibitors that block degradation of the phosphorylated IB-␣, compounds 1 and 2 inhibit inducible phosphorylation of IB-␣. Compounds 1 and 2 may inhibit a TNF␣-inducible kinase and/or activate a cellular phosphatase activity.
Because the effects of these agents on adhesion molecule expression were found to be irreversible (Fig. 3), we tested whether the stabilization of IB-␣ was similarly irreversible.
The levels of cytoplasmic IB-␣ were examined in HUVEC pretreated with compound 1 for 1 h and then incubated with media alone for 1 h prior to treatment with TNF␣ in the absence of drug. The wash out of this compound had little or no effect on the levels of IB-␣ (Fig. 6, lanes 6 and 7). Thus, this drug appears to irreversibly stabilize IB-␣. Treatment of cells with the test compound 15 min after TNF␣ induction and immediately prior to preparation of extracts did not result in stabilization of IB-␣ (Fig. 6, lane 8). Clearly, the drug blocks TNF␣-induced phosphorylation and degradation of IB-␣ in intact cells and not during extract preparation. The effects on both IB-␣ phosphorylation and surface expression of E-selectin were rapid, irreversible, and occurred at an IC 50 of approximately 10 M. This suggests that the inhibition of IB-␣ phosphorylation and degradation causes the decrease in NF-B, resulting in decreased transcription and surface expression of the adhesion molecules.
Effect on Constitutive Phosphorylation of IB-␣-The IB-␣ protein is regulated by cytokine-inducible phosphorylation on Ser-32 and Ser-36 (16 -23). In addition, the C-terminal sequences of IB-␣ contain a consensus sequence for casein kinase II that may be important for basal phosphorylation of the IB-␣ protein (19, 44 -46). To determine whether compound 1 selectively inhibited TNF␣-inducible phosphorylation of IB-␣ or might also inhibit the activity of a constitutive IB-␣ kinase that basally phosphorylates IB-␣, we carried out an in gel kinase assay for proteins that phosphorylate IB-␣ (Fig. 7). Two proteins of molecular mass of approximately 36 to 41 kDa were observed in whole cell extracts. The activities of these kinases were unaffected by 15 min treatment with TNF␣ or by compound 1 (20 M). The molecular weights of these kinases correspond to those expected for the catalytic subunits of casein kinase II; however, Bennett et al. (47) have recently described IB-␣ kinases of similar molecular weight in endothelial cells that are distinct from casein kinase II. Our results suggest that compound 1 had no effect on the activity of these IB-␣ kinases. Thus, compound 1 selectively inhibits the TNF␣-inducible phosphorylation of IB-␣ without affecting the constitutive IB-␣ phosphorylation.
Activation of Stress-activated Protein Kinases, p38 and JNK-1, with No Effect on ERK-1-The treatment of endothelial cells with TNF␣ induces multiple signaling events that might be affected by the test compounds. It has been shown that TNF␣ treatment stimulates the stress-activated protein kinase cascade (40, 41, 48 -53). The JNK-1 and p38 kinases are acti- vated by dual phosphorylation of threonine and tyrosine. To determine whether compound 1 selectively inhibits IB-␣ phosphorylation without affecting other TNF␣-induced phosphorylation events, we examined the effects of compound 1 on MAP kinase activity. We carried out Western blot analysis with antisera specific for the phosphorylated form of p38 to determine whether this agent affected the phosphorylation of the p38 kinase. Results are shown in the bottom panel of Fig. 5.
There was a small increase in phosphorylation of p38 with TNF␣ alone and a marked increase in phosphorylation of p38 in HUVEC-treated with TNF␣ in the presence of 20 M compound 1 (Fig. 5, lanes 2 and 5). In addition, there was a slight but detectable increase in p38 phosphorylation in HUVEC treated with compound 1 alone (data not shown). The level of p38 phosphorylation in cells treated with both compound 1 and TNF␣ was higher than that seen in with TNF␣ alone or compound 1 alone (Fig. 5, lanes 2 and 5). The total levels of p38 in these cells did not change. These results suggest that compound 1 stimulates TNF␣-induced phosphorylation of p38 and that compound 1 does not globally inhibit all TNF␣-induced phosphorylation. In contrast to the effects on IB-␣ and the effects on adhesion molecule expression, the observed activation of p38 phosphorylation was reversible with 1 h treatment in the absence of drug (data not shown). Thus, the increase in p38 phosphorylation did not correlate with the inhibition of adhesion molecule expression.
The effect of this compound on MAP kinase signaling was also measured by immunoprecipitation kinase assays for activity of p38, JNK-1, or ERK-1. The results are shown in Fig. 8. The level of ERK-1 activity detected in cells treated with TNF␣ in the presence of compound 1 (20 M) was similar to the ERK-1 activity in cells treated with TNF␣ alone suggesting that com-pound 1 does not affect ERK-1 kinase activity (Fig. 8, lanes 2  and 3). The JNK-1 and p38 kinase activities were induced with 15-min TNF␣ treatment (Fig. 8, lane 2). The level of JNK-1 activity was somewhat enhanced in cells treated with TNF␣ in the presence of compound 1 (Fig. 8, lanes 2 and 3). Notably, there was a significant increase in the activity of the p38 kinase activity in cells treated with TNF␣ in the presence of 20 M compound 1 (Fig. 8, lanes 2 and 3). This suggests that compound 1 stimulates the TNF␣-induced p38 kinase and JNK-1 kinase activities with no detectable effect on ERK-1 kinase activity. This compound may activate signaling events upstream of p38 and JNK-1 that are distinct from the pathway activating ERK-1 (54,55). Alternatively, the compound may inhibit a dual specificity phosphatase such as M3/6 which selectively regulates p38 and JNK-1 but not ERK-1 (56). Taken together, our Western blot and immunoprecipitation kinase assays suggest that the compound does not act as a global inhibitor of TNF␣-induced phosphorylation events but selectively inhibits phosphorylation of IB-␣.
Effects on Tyrosine Phosphorylation-Protein tyrosine kinase inhibitors have been reported to block NF-B/IB-␣ signaling and inhibit adhesion molecule expression (29). To determine whether the selected agents may act as protein tyrosine kinase inhibitors, we assayed the tyrosine-phosphorylated proteins in whole cell extracts from endothelial cells treated with TNF␣ or TNF␣ and compound 1. Extracts were analyzed by Western blot with anti-phosphotyrosine antisera as described under "Experimental Procedures." Multiple bands at a variety of molecular weights were reactive with the anti-phosphotyrosine antisera. There was not a general reduction in the pattern or intensity of the bands in compound 1-treated cells. Indeed, cells treated with TNF␣ and compound 1 showed a dramatic increase in a tyrosine-phosphorylated protein with a molecular mass of 130 -140 kDa (Fig. 9). There was some increase in this tyrosine-phosphorylated protein observed with compound 1 alone, and further analysis of nuclear and cytoplasmic extracts was done to determine that this protein was predominantly present in the cytoplasm (data not shown). The effect of the drug on tyrosine phosphorylation was reversible and thus probably distinct from the irreversible effects on IB-␣ phosphorylation and adhesion molecule expression. Our data suggest that compound 1 does not act as a global protein tyrosine kinase inhibitor and may in fact activate some specific tyrosine phosphorylation.
Anti-inflammatory Actions of Compound 2-Compound 2 was evaluated in two in vivo models of inflammation. As shown in Fig. 10, compound 2 demonstrated a dose-dependent reduction in swelling in the rat carrageenan paw model. Compound 2 was also evaluated in established rat adjuvant arthritis (Fig.  11). In the vehicle-treated control group, the mean volume of both hind paws increased by 0.39 Ϯ 0.15 ml. Compound 2, given intraperitoneally at 20 mg/kg, but not at 5 mg/kg, significantly reduced the mean paw edema of the rats, to levels similar to those observed with the positive control, dexamethasone, at 1 mg/kg intraperitoneally. Thus, this compound acted as an anti-inflammatory agent in both the rat carrageenan paw and the rat adjuvant arthritis model. DISCUSSION We have identified two structurally related compounds that inhibit the expression of ICAM-1, VCAM-1, and E-selectin in human endothelial cells. These compounds act by selectively inhibiting TNF␣-induced phosphorylation of IB-␣, resulting in decreased nuclear NF-B and decreased expression of adhesion molecules. These compounds selectively inhibited the TNF␣-inducible phosphorylation of IB-␣ without affecting the constitutive IB-␣ phosphorylation. Although these agents were shown to exhibit reversible effects on other cellular phosphorylation events including activation of the stress-activated protein kinases, p38 and JNK-1, and activation of protein ty-rosine phosphorylation, it is likely that these effects are distinct from the effects on adhesion molecule expression which were irreversible. One of these agents, compound 2, was tested in vivo in two animal models of inflammation. The compound reduced swelling in both the rat carrageenan paw edema assay and in a rat adjuvant arthritis model. These studies suggest that novel pharmacologic agents that inhibit cytokine-inducible phosphorylation of IB␣ can act as anti-inflammatory agents.
The precise molecular target for these agents is not yet clear. While these drugs were shown to inhibit IB-␣ phosphorylation, this may be the result of direct inhibition of a TNF␣inducible IB-␣ kinase or due to inhibition of a signaling event upstream of the IB-␣ kinase. Alternatively, the regulation of IB-␣ phosphorylation involves cellular phosphatase activities that may be activated by these drugs (19). Once the TNF␣inducible IB-␣ kinase(s) and regulatory phosphatase(s) are identified, it will be interesting to determine if these molecules are the direct target of compound 1 or 2. It has been observed that upstream activators of the MAP kinase pathway can induce NF-B/IB-␣ signaling, suggesting that the MAP kinase and NF-B cascades share some common intermediates (57)(58)(59)(60)(61). Since TNF␣ signaling of MAP kinases was not inhibited by compound 1, the target for this drug is likely to be downstream of the events that are common to both NF-B and MAP kinase signaling pathways. Recent reports have suggested that the TNF␣ signaling of NF-B may occur by a ceramide-dependent mechanism, whereas the TNF␣ signaling of p38 and JNK-1 kinases in endothelial cells may be ceramide-independent (62). Therefore, ceramide-dependent protein kinases (63) and/or ceramide-dependent protein phosphatases (64) may be potential targets of the drug action.
In endothelial cells, cytokines increase superoxide anion production (30,65), and reactive oxygen intermediates may act as an important regulator of NF-B (reviewed in Ref. 66). A number of antioxidants have been reported to inhibit cytokineinduced IB-␣ phosphorylation, nuclear translocation of NF-B, and NF-B-dependent transcription of VCAM-1 (24, 30, 67-70). However, it has been suggested that the expression of ICAM-1 and E-selectin may be less affected by antioxidants A 1% suspension of carrageenan in distilled water was administered to rats as 0.1-ml subplantar injection into the footpad of the right hind paw. One hour prior to injection, rats were treated intraperitoneally with vehicle (polyethylglycol 400 diluted 1:5 in 5% bovine serum albumin/H 2 O) or a fine suspension of compound 2 (1, 5, or 50 mg/kg) in vehicle. A positive control group was also included in which rats were pretreated with 20 mg/kg ibuprofen. Four hours after carrageenan administration, the volume of the injected paw was measured by means of a water displacement plethysmograph. Edema volumes were determined as the difference between the paw volumes of each rat at time 0 and 4 h. Data were analyzed by a one-way analysis of variance and, if indicated, differences between groups analyzed by Bonferroni's modified t test. A p Ͻ 0.05 was considered significant. *, significantly different from control.
FIG. 11. Rat adjuvant arthritis. Inbred, male Lewis rats 8 -10 weeks of age weighing 250 -275 g were used. Heat-inactivated M. butyricum was administered as 0.1-ml injection (1 mg/animal) at the base of the tail as indicated under "Experimental Procedures." Paw volumes were determined on the indicated dates and values compared with initial time 0 measurements. Vehicle (0.5% methyl cellulose) or drug (compound 2 or dexamethasone at the indicated concentrations) was administered once/day as an intraperitoneal injection (200 l). Data from these studies are expressed as the mean difference in foot pad volume. (30,70,71). The potential of the novel compounds described in this study to act as antioxidants by inhibiting the generation of reactive oxygen intermediates or by scavenging free radicals has not been evaluated.
Protein tyrosine kinase inhibitors have been shown to block phosphorylation of IB-␣ and adhesion molecule expression. Recent reports suggest that reactive oxygen intermediates activate NF-B by a tyrosine kinase-dependent mechanism (29,72). It is possible that the compounds described in this study could act to inhibit a specific protein tyrosine kinase that is upstream of the IB-␣ kinase; however, it does not appear that these compounds act as globally active tyrosine kinase inhibitors since we observed an increase in phosphotyrosine activity as measured by Western blot with anti-phosphotyrosine antibodies. In addition, these compounds activated tyrosine phosphorylation of p38 as measured by Western blot with phosphospecific antisera and stimulated p38 and JNK-1 activities that are up-regulated by tyrosine phosphorylation (40,48,49). The mechanism by which test compounds may stimulate tyrosine phosphorylation has not been determined; however, it is possible that these compounds may inhibit some protein tyrosine phosphatase activity. Protein tyrosine phosphatase inhibitors have been reported to inhibit NF-B signaling (73,74); however, the mechanism for this inhibition is unclear. Serine protease inhibitors have also been reported to inhibit NF-B signaling and decrease adhesion molecule expression (28). We have not tested our novel compounds for specific effects on serine protease activity.
The anti-inflammatory effects of the test compounds in two animal models are striking and are consistent with the action of other pharmacologic agents that inhibit adhesion molecule expression and leukocyte recruitment (75). Since these novel compounds inhibit NF-B signaling, they would be expected to affect B-dependent expression of many other genes including IL-1, IL-6, tissue factor, and TNF␣ in lymphoid cells, monocytes, and endothelial cells (reviewed in Ref. 24). Thus, the observed anti-inflammatory action probably reflects not just inhibition of adhesion molecules but also effects on many other important mediators of inflammation in a variety of cell types. Understanding the mechanism by which these agents disrupt the NF-B/IB regulatory pathway will be useful in identifying novel anti-inflammatory agents that are both highly specific and effective. Such drugs may be useful as therapeutic agents in disorders involving up-regulation of endothelial adhesion molecules including ischemia, reperfusion injury, asthma, transplantation, inflammatory bowel disease, rheumatoid arthritis, and atherosclerosis.