Intercellular adhesion molecule-1 (ICAM-1), a transmembrane glycoprotein of 505 amino acids, is a member of the immunoglobulin supergene family and contains five extracellular immunoglobulin-like domains that function in cell-cell and cell-matrix adhesive interactions (van de Stolpe and van der Saag, 1996). In contrast to other cell adhesion molecules, ICAM-1 is expressed in both hematopoietic and nonhematopoietic cells and mediates adhesive interactions by binding to two integrins belonging to the β2 subfamily, i.e., CD11a/CD18 (LFA-1) and CD11b/CD18 (Mac-1). ICAM-1 adhesive interactions are critical for the transendothelial migration of leukocytes and the activation of T cells where ICAM-1 binding functions as a coactivation signal (Zuckerman et al., 1998). ICAM-1 is also associated with a variety of inflammatory diseases and conditions, including asthma, atherosclerosis, inflammatory bowl disease, acute respiratory distress syndrome, ischemia reperfusion injury, and autoimmune disease (Sampath et al., 1979; Welty et al., 1993; Colletti et al., 1998; Kacimi et al., 1998; Stanciu and Djukanovic, 1998). ICAM-1 is present low on the cell surface of a variety of cell types, and it is up-regulated in response to a number of inflammatory mediators, including retinoic acid, virus infection, oxidant stresses, and the proinflammatory cytokines IL-1β, TNF-α, and interferon-γ (Bassi et al., 1995; van de Stolpe and van der Saag, 1996; Chen et al., 2000, 2001; Chang et al., 2002). The level of ICAM-1 expression on the surface of any given cell type depends on the concentrations of pro- and anti-inflammatory mediators, on the availability of specific receptor-mediated signal transduction pathways, and their nuclear transcription factor targeting on the ICAM-1 promoter (Cornelius et al., 1993). Nuclear transcriptional factors important to the activation of ICAM-1 expression include AP-1, NF-κB, CCAAT/enhancer-binding protein, Ets, signal transducer and activator of transcription (STAT), or SP1 (van de Stolpe and van der Saag, 1996). The major intracellular signal transduction pathways involved in the regulation of ICAM-1 expression include PKC-dependent c-Src activation of the NF-κB or the STAT1 signaling pathways (Chen et al., 2000, 2001; Chang et al., 2002, 2004; Huang et al., 2003). The abundance of signaling pathways and transcription factors involved in ICAM-1 transcription reflects the complex cell-type- and stimulus-specific regulation of the ICAM-1 gene. The modulation of ICAM-1 expression is an important therapeutic target, as shown by the beneficial effect of anti-ICAM-1 antibodies and other pharmacological agent on the progression of inflammatory responses in several in vivo studies (Barton et al., 1989; Albelda et al., 1994).

Flavonoids are polyphenolic compounds that occur ubiquitously in food plants and vegetables. The polyphenolic flavonoids, which have the diphenylpropane (C6C3C6) skeleton, include monomeric flavanols, flavanones, flavones, and flavonols. Individual differences in flavonoids result from the variation in number and arrangement of the hydroxyl groups as well as from the nature and the extent of alkylation and/or glycosylation of these groups. Some flavonoids have been found to possess various clinically relevant properties such as anti-tumor, anti-platelet, anti-ischemic, and anti-inflammatory activities (Avila et al., 1994; Gerritsen et al., 1995), and these effects are believed to come from the antioxidant properties of the related flavonoids (Rice-Evans et al., 1996). Despite numerous interest in their pharmacological activities as anti-inflammatory agents, only a few studies have investigated the role of flavonoids in the regulation of cell adhesion processes. In this study, we chose three flavonols (kaempferol, quercetin, and myricetin) and six flavones (flavone, chrysin, apigenin, luteolin, baicalein, and baicalin) (Fig. 1) to examine the inhibitory effect on the TNF-α-induced ICAM-1 expression in A549 epithelial cells, in which the molecular mechanism of this event has been explored to involve the PKC-dependent c-Src pathway in inducing IKK and NF-κB activation (Chen et al., 2001; Huang et al., 2003). The structure-activity relationship of these naturally occurring flavonoids was studied by comparing their potency in inhibiting ICAM-1 expression and inhibiting the signaling molecules involved. Our results revealed the contributions of 4′-hydroxy structure in the B ring and 5,7-metadihydroxy arrangement in the A ring, which might help the design of analogs displaying anti-inflammatory effect, and provided evidence for the correlation between the anti-inflammatory properties of flavonoids and their efficiency in inhibiting signaling molecules.

Materials and Methods

Materials. Mouse monoclonal anti-human ICAM-1 antibody and recombinant human TNF-α were purchased from R&D Systems (Minneapolis, MN). The rabbit polyclonal antibodies specific for IkBα, IKKβ, ERK2, p38, and JNK1 were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). T4 polynucleotide kinase and rabbit polyclonal antibodies specific for the phosphorylated forms of ERK1/2, p38, or JNK were obtained from New England Biolabs (Beverly, MA). Dulbecco's modified Eagle's medium (DMEM), fetal calf serum (FCS), penicillin, and streptomycin were purchased from Invitrogen (Carlsbad, CA). 2′,7′-Bis-(carboxyethyl)-5,6-carboxyfluorescein (BCECF) was obtained from Molecular Probes (Eugene, OR). Poly(dI/dC) was obtained from Amersham Biosciences Inc. (Piscataway, NJ). Reagents for SDS-PAGE were purchased from Bio-Rad (Hercules, CA). [γ-³²P]ATP (3000 Ci/mmol) was obtained from PerkinElmer Life and Analytical Sciences (Boston, MA). Horseradish peroxidase-labeled donkey anti-rabbit second antibody and the ECL-detecting reagent were purchased from Amersham Biosciences Inc. The ICAM-1 promoter constructs, pIC1352, pIC339, and pIC1352 (ΔAP-1) (pIC1352 with the –284/–279 region deleted) were gifts from Dr. P. T. van der Saag (Hubrecht Laboratory, Utrecht, The Netherlands). The NF-κB-Luc and AP-Luc expression plasmids were purchased from Stratagene (La Jolla, CA).

Cell Cultures. A549 cells, an alveolar epithelial cell carcinoma, were obtained from American Type Culture Collection (Manassas, VA), and cultured in DMEM supplemented with 10% FCS, 100 U/ml penicillin, and 100 μg/ml streptomycin in 96-well plates (ICAM-1 expression), in 6-well plates (transfection), in 6-cm dishes (IKK activity assay and RT-PCR), or in 10-cm dishes (NF-κB gel shift assay and MAPKs activation).

U937 cells, a human monocytic leukemia cell, were obtained from the Department of Microbiology (College of Medicine, National Taiwan University) and cultured in RPMI 1640 medium supplemented with 10% FCS, 100 U/ml penicillin, and 100 μg/ml streptomycin. Cells were split and fed every 3 to 4 days.

Quantification of ICAM-1 Expression. The level of cell surface ICAM-1 expression was determined by ELISA as previously described (Chen et al., 2000). Cells were treated with TNF-α for 4.5 h at 37°C following pretreatment with various flavonoids for 30 min. Each assay was performed in triplicate, and the basal absorbance (about 0.1 unit) was subtracted. None of the tested flavonoids affected the basal ICAM-1 expression.

Cell Adhesion Assay. A549 cells, grown in 96-well plates, were treated at 37°C with TNF-α for 4.5 h after pretreatment with flavonoids for 30 min, then washed twice with PBS. U937 cells were labeled for 30 min at 37°C with 10 ng/ml BCECF and washed twice with growth medium, then 2.5 × 10⁵ of the labeled cell were added to the A549 monolayer in a final volume of 100 μl and incubated in a CO₂ incubator for 1 h. Nonadherent cells were removed from the plate by gentle washing with PBS, and the number of adherent cells were determined by measuring the fluorescent intensity using a CytoFluor 2300 (Millipore Corporation, Bedford, MA).

RT-PCR. Total RNA was isolated from A549 cell using Trizol reagent (Invitrogen). Reverse transcription reaction was performed using 2 μg of total RNA and reverse-transcribed into cDNA using oligo(dT) primer, then amplified 30 cycles using two oligonucleotide primers derived from published ICAM-1, c-fos, c-jun, or β-actin sequence, including 5′-TGCGGCTGCTACCACAGTGATGAT-3′ and 5′-CCATCTACAGCTTTCGGCGCCCA-3′ (ICAM-1), 5′-CAGTCAGATCAAGGGAAGCCACAGACATCT-3′ and 5′-GAATAAGATGGCTGCAGCCAAATGCCGCA3′ (c-fos), 5′-GAACCCCTCCTGCTCATCTGTCACGTTCTT-3′ and 5′-GGAAACGACCTTCTATGACGATGCCCTCAA-3′ (c-jun), 5′-TGACGGGGTCACCCACACTGTGCCCATCTA-3′ and 5′-CTAGAAGCATTTGCGGGGACGATGGAGGG-3′ (β-actin). PCR is carried out at 94°C for 30 s, at 65°C for 30 s, and 1 min at 70°C for 34 cycles. The PCR products were subjected to 1% agarose gel electrophoresis. Quantitative data normalized to β-actin were obtained using a computing densitometer and ImageQuant software (Amersham Biosciences Inc.).

Immunoprecipitation and in Vitro Kinase Activity Assay. Following a 10-min treatment with TNF-α or 30-min pretreatment with various flavonoids before the addition of TNF-α, cells were rapidly washed with PBS, then lysed with ice-cold lysis buffer, and the IKK proteins were immunoprecipitated. Total cell extract (50 μg) was incubated for 1 h at 4°C with 0.5 μg of anti-IKKβ, anti-ERK2, anti-p38, or anti-JNK1 antibody and collected using protein A-Sepharose Cl-4B beads (Sigma-Aldrich, St. Louis, MO). The beads were then washed three times with lysis buffer without Triton X-100 and incubated for 30 min at 30°Cin20 μl of kinase reaction mixture containing 20 mM HEPES, pH 7.4, 5 mM MgCl₂, 5 mM MnCl₂, 0.1 mM Na₃VO₄, 1 mM dithiothreitol, 1 μg of bacterially expressed GST-IκBα (1–100), 2 μg of MBP or GST-c-jun and 10 μM[γ-³²P]ATP. The reaction was stopped by the addition of Laemmli buffer and subjected to 10% SDS-PAGE, phosphorylated GST-IκBα (1–100), MBP or GST-c-jun being visualized by autoradiography.

Preparation of Nuclear Extracts and the Electrophoretic Mobility Shift Assay (EMSA). Control cells or cells pretreated with various flavonoids were treated with TNF-α for 1 h, then nuclear extracts were prepared as described previously (Chen et al., 2000). Oligonucleotides corresponding to the downstream NF-κB binding sequences (5′-AGCTTGGAAATTCCGGA-3′) or AP-1 sequences (5′-GACCGTGATTCAAGC-3′) in human ICAM-1 promoter were synthesized, annealed, and end-labeled with [γ-³²P]ATP using T4 polynucleotide kinase, and EMSA was performed as described previously (Chen et al., 2000). When supershift assays were performed, polyclonal antibodies specific for c-fos or c-jun were added to the nucleus extracts 30 min before the binding reaction.

Transient Transfection and Luciferase Assay. The human ICAM-1 firely luciferase (LUC) plasmids (pIC1352 or pIC1352ΔAP-1), or NF-κB or AP-1 luciferase reporter were transfected using Tfx-50 (Promega, Madison, WI) according to the manufacturer's recommendations. Briefly, reporter DNA (0.4 μg) and β-galactosidase DNA (0.2 μg) were mixed with 0.6 μl of Tfx-50 in 1 ml of serum-free DMEM. We used the plasmid PRK containing β-galactosidase gene driven by the constitutively active SV30 promoter to normalize the transfection efficiency. After a 10- to 15-min incubation at room temperature, the mixture was applied onto the cells. DMEM (1 ml) containing 10% FCS was added 1 h later, and the cells were grown in medium containing 5% FCS. The following day, cells were exposed to 10 ng/ml TNF-α or treated with various flavonoids for 30 min before challenge with TNF-α for 4.5 h, then cell extract was prepared, and luciferase (Promega) and β-galactosidase activity were measured. The luciferase activity of each well was normalized to β-galactosidase activity.

Preparation of Cell Extracts and Western Blot Analysis of Phosphorylated ERK1/2, Phosphorylated p38, Phosphorylated JNK, ERK2, p38, and JNK1. Following treatment with TNF-α or with various MAPK inhibitors before challenge with TNF-α for 10 min, the cells were rapidly washed with PBS, then lysed with ice-cold lysis buffer (50 mM Tris-HCl, pH 7.4, 1 mM EGTA, 150 mM NaCl, 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, 5 μg/ml leupeptin, 20 μg/ml aprotinin, 1 mM sodium fluoride, and 1 mM Na₃VO₄) as described previously (Chen et al., 2001). The lysates were subjected to SDS-PAGE using a 10% running gel. The proteins were transferred to nitrocellulose paper, and immunoblot analysis was performed as described previously (Chen et al., 2001).

Statistical Analyses. All data are expressed as mean ± S.E.M. Statistical analyses were done with Student's t test.

Results

Inhibitory Effects of Flavonoids on TNF-α-Induced ICAM-1 Expression on A549 Cells and U937 Adhesion to A549 cells. Three flavonols (kaempferol, quercetin, and myricetin) and six flavones (flavone, chrysin, apigenin, luteolin, baicalein, and baicalin) (Fig. 1) were chosen to test their inhibitory effects on TNF-α-induced ICAM-1 expression on A549 cells. Since 30 μM was the highest concentration that had no toxicity to the cells, this concentration was examined first. As shown in Fig. 2A, five showed inhibitory effects at 30 μM. The extent of inhibition for kaempferol, quercetin, chrysin, apigenin, and luteolin was 74, 61, 72, 97, and 98%, respectively. Their IC₅₀ was 12.1, 14.4, 10.9, 0.82, and 0.80 μM, respectively (Fig. 2B). The action mechanism of kaempferol, chrysin, apigenin, and luteolin in inhibiting ICAM-1 expression was studied further. These four flavonoids also inhibited TNF-α-induced U937 adhesion to A549 cells in a dose-dependent manner (Fig. 2C).

Kaempferol, Chrysin, Apigenin, and Luteolin Inhibit TNF-α-Induced ICAM-1 mRNA Expression and ICAM-1 Promoter Activity. TNF-α has been shown to induce a dose- and time-dependent increase in ICAM-1 protein expression and U937 adhesion to A549 cells (Chen et al., 2001). It also induces the expression of ICAM-1 mRNA in a time-dependent manner; this effect is significant and maximal at 1 h, sustained at 3 h, and declined after 6 h (unpublished data). To determine the effects of these four flavonoids on TNF-α-induced ICAM-1 gene expression, cells were pretreated for 30 min with each of these compounds at a concentration of either 1 or 3 μM before the stimulation with TNF-α for 3 h. When ICAM-1 mRNA expression was analyzed by RT-PCR, kaempferol, cyrysin, apigenin, and luteolin inhibited TNF-α-induced ICAM-1 mRNA expression in a dose-dependent manner (Fig. 3A).

To elucidate the mechanism involved in flavonoid-mediated inhibition of ICAM-1 expression, transient transfections were performed using a human ICAM-1 promoter-luciferase construct, pIC339. As reported previously (Chen et al., 2001), treatment with TNF-α led to an increase in ICAM-1 promoter activity, and this effect was inhibited by kaempferol, chrysin, apigenin, and luteolin in a dose-dependent manner, whereas apigenin and luteolin were more potent than kaempferol and chrysin (Fig. 3B).

Kaempferol, Chrysin, Apigenin, and Luteolin Inhibit TNF-α-induced IKK Activation, IκBα Degradation and NF-κB-Specific DNA-Protein Complex Formation and NF-κB-Luc Activity. Because TNF-α-induced ICAM-1 expression was found to involve the IKKβ/NF-κB pathway (Chen et al., 2001; Huang et al., 2003), the effects of these compounds on the TNF-α-induced IKK activity were examined. TNF-α induced IKK activation at 10-min treatments (Fig. 4A, lane 2). Among the four flavonoids tested, both apigenin and luteolin at 1 μM and 3 μM inhibited the TNF-α-induced IKK activity in a dose-dependent manner (50% and 63% inhibitions for apigenin; 54% and 67% inhibitions for luteolin) (Fig. 4A, lanes 7–10). Kaempferol and chrysin only showed minor inhibitory effects at 3 μM (25% and 31% inhibitions for kaempferol and chrysin, respectively) (Fig. 4A, lanes 4 and 6).

Phosphorylation of IκBα is necessary for its degradation and the subsequent NF-κB activation. We next examined the effects of these four compounds on TNF-α-induced IκBα degradation. The degradation of IκBα was seen after a 10-min treatment with TNF-α (Fig. 4B, lane 2). Apigenin and luteolin at 3 μM partially reversed the TNF-α-induced IκBα degradation (Fig. 4B, lanes 8 and 10), whereas kaempferol at 1 μM and chrysin at 3 μM only showed minor reversal (Fig. 4B, lanes 7 and 9 and 4 and 6). Kampferol and chrysin at 1 μM had no effect on TNF-α-induced IκBα degradation (Fig. 4B, lanes 3 and 5).

To test whether the inhibitory effect of flavonoids on IKK activity led to NF-κB inhibition, the effects of these compounds on TNF-α-stimulated NF-κB-specific DNA-protein binding activity were examined. Apigenin and luteolin at 1 and 3 μM inhibited TNF-α-induced NF-κB-specific DNA-protein binding in a dose-dependent manner (Fig. 4C, lanes 7–10). Kaempferol and chrysin at 3 μM showed minor inhibitory effects (Fig. 4C, lanes 4 and 6).

The inhibitory effects of the four flavonoids on NF-κB activation were further examined using a NF-κB-Luc construct. The results showed that apigenin and luteolin at 1 and 3 μM inhibited TNF-α-induced NF-κB luciferase activity in a dose-dependent manner (57 and 77% inhibitions for apigenin; 54 and 74% inhibitions for luteolin) (Fig. 4D). Kaempferol and chrysin displayed significant inhibition at 3 μM (49% for both) (Fig. 4D).

Role of AP-1 in TNF-α-Induced ICAM-1 Expression and the Inhibitory Effects of Kaempferol, Chrysin, Apigenin, and Luteolin. The above results showed that the extent of inhibition on TNF-α-induced NF-κB activation by the four flavonoids was not paralleled to their inhibitory effects on TNF-α-induced ICAM-1 expression, indicating that in addition to NF-κB, other transcriptional factors might be involved. The ICAM-1 promoter contains several AP-1 binding sites that may be important for ICAM-1 expression (Voraberger et al., 1991). To clarify whether AP-1 plays a role in the TNF-α-induced ICAM-1 expression in A549 cells, the human ICAM-1 promoter-luciferase constructs pIC 1352 (–1352/+1), which contains full-length human ICAM-1 promoter, pIC 1352 (ΔAP1), which does not contain the AP-1 site (–279/–284), and pIC339 (–339/+1), which contains the downstream NF-κB and AP-1 sites were transfected into cells to examine the TNF-α-induced promoter activity. As shown in Fig. 5, TNF-α led to a 2.5-fold increase with pIC1352 and a 2.7-fold increase with pIC339; however, TNF-α only induced a 1.6-fold increase in ICAM-1 promoter activity with pIC1352 (ΔAP1). These results indicated the requirement of AP-1 cis-acting element for mediating TNF-α-induced ICAM-1 expression on A549 cells.

Since AP-1 was found to involve the TNF-α-induced ICAM-1 promoter activity, AP-1 complex formation was examined by EMSA to determine whether AP-1 activity was stimulated by TNF-α. TNF-α stimulated AP-1 DNA-protein binding after 30 min treatment. The activation declined after 60 or 120 min (Fig. 6A). In the subsequent EMSA experiments, cells were treated with TNF-α for 30 min. To identify the specific subunits involved in the formation of AP-1 complex after TNF-α stimulation, supershift assays were performed in the presence of antibodies specific to either c-fos, c-jun, or p50 (NF-κB). Incubation of nuclear extracts with antibody specific for either c-fos or c-jun induced attenuation of AP-1 DNA-protein binding (Fig. 6B, lanes 3 and 4); however, no shift or attenuation occurred in the presence of antibody specific for p50, which is a component of another transcription factor NF-κB (Fig. 6B, lane 6). These results indicated the formation of c-fos and c-jun heterodimer in AP-1 complex. Excess cold AP-1 probe blocked the AP-1 DNA-protein binding (Fig. 6B, lane 5).

Since AP-1 was demonstrated to be responsible for mediating the TNF-α-induced ICAM-1 expression, and TNF-α indeed induced the AP-1 DNA-protein binding, effects of kaempferol, chrysin, apigenin, and luteolin on TNF-α-induced AP-1 DNA-protein binding and AP-1-luc activity were examined. As shown in Fig. 7A, apigenin and luteolin almost blocked TNF-α-induced AP-1 DNA-protein binding. The inhibitions at 1 μM were 84 and 88%, and those at 3 μM were both 100% for apigenin and luteolin (Fig. 7A, lanes 7–10). As for kaempferol and chrysin, the inhibitions at 1 μM were 44 and 42%, and those at 3 μM were 68 and 72%, respectively (Fig. 7A, lanes 3–6). These four flavonoids also showed significant inhibitions on TNF-α-induced AP-1 luciferase activity (Fig. 7B).

Kaempferol, Chrysin, Apigenin, and Luteolin Inhibit TNF-α-Induced c-fos and c-jun mRNA Expression and Mitogen-Activated Protein Kinases Activities. Since AP-1 complex was demonstrated to be a c-fos and c-jun heterodimer, the effect of the four flavonoids on c-fos and c-jun mRNA expression was examined by RT-PCR. As shown in Fig. 8A, apigenin and luteolin showed significant inhibition on TNF-α-induced c-fos mRNA expression (Fig. 8A, lanes 7–10), whereas kaempferol and chrysin had not much effect (Fig. 8A, lanes 3–6). As for c-jun mRNA, all the four flavonoids showed significant inhibition on the TNF-α-induced c-jun mRNA expression. However, apigenin and luteolin were more potent compared with kaempferol and chrysin (Fig. 8B, lanes 7–10 and 3–6, respectively).

Since the transcriptional activity of AP-1 has been reported to be regulated by MAPKs (Karin, 1995), the roles of ERK1/2, p38, and JNK in TNF-α-induced c-fos and c-jun mRNA expressions were examined. Similar to our previous findings (Chen et al., 2001), PD98059 (50 μM) completely blocked TNF-α-induced ERK1/2 activation without having any effect on the p38 and JNK activations, and SB203580 (30 μM) caused marked inhibition of p38 activation without affecting ERK1/2 and JNK activations (Fig. 9A, lanes 3 and 4). SP600125 inhibited JNK1/2 activation in a dose-dependent manner without any effect on ERK1/2 and p38 activations (Fig. 9A, lanes 5–7). After cells were pretreated with either 50 μM PD98059, 30 μM SB203080, 10 μM SP600125, or 25 μM SP600125, the TNF-α-induced c-fos mRNA expression was almost completely inhibited by PD98059 and SB203580 (Fig. 9B, lanes 3 and 4), whereas SP600125 had no effect (Fig. 9B, lanes 5 and 6). As for the c-jun mRNA expression, only SP600125 showed inhibitory effect (Fig. 9B, lanes 5 and 6).

The three MAPKs were demonstrated to be involved in regulating c-fos and c-jun mRNA expressions; therefore, effects of kaempferol, chrysin, apigenin, and luteolin on their activities were examined. As shown in Fig. 10, luteolin showed the most potent inhibition on TNF-α-induced ERK, p38, and JNK activations (Fig. 10, A–C, lanes 9 and 10), and apigenin was the second (Fig. 10, A–C, lanes 7 and 8). Kaempferol and chrysin inhibited the JNK activity (Fig. 10C, lanes 3–6) while having no effect on ERK and p38 activities (Fig. 10, A and B, lanes 3–6).

Discussion

Naturally occurring flavonoids have been proposed to exert biological effects on cells through the inhibitions of different enzymes. They are considered as potential compounds for selectively blocking signal transduction pathways. Adhesion molecule ICAM-1 is known to play a central role in the regulation of cellular inflammatory responses (Albelda et al., 1994). In this study, we analyzed the effects of three flavonols and six flavones on TNF-α-induced ICAM-1 expression and found the inhibitory effects of five of them. We then studied the action mechanisms of kaempferol, chrysin, apigenin, and luteolin by examining their effects on the IKK/NF-κB pathway. We report here that in addition to the NF-κB pathway, the AP-1 seems to play a more significant role in these four flavonoid-induced down-regulation of ICAM-1 gene expression.

It is rational to study the ICAM-1 gene expression and the accompanying signaling pathways in the alveolar epithelial cells. Epithelial cells play an active role in inflammation by producing various cytokines that are involved in the late asthmatic response (Barnes, 1994). Of the potent alveolar macrophage-derived cytokines studied, TNF-α in particular has been implicated in the pathophysiology of neutrophilic infiltrating disorders, including acute lung injury from sepsis, silica-induced pulmonary fibrosis, allograft rejection, and acute respiratory tract infection (Beutler, 1995). Upon binding to its cell surface receptors, TNF-α activates the cytoplasmic form of NF-κB (Baldwin, 1996). In A549 alveolar epithelial cells, IKKβ/NF-κB pathway was required for TNF-α-induced ICAM-1 expression (Chen et al., 2001). Therefore, the inhibitory mechanisms of kaempferol, chrysin, apigenin, and luteolin were first examined for their effects on IKKβ/NF-κB pathway. The phosphorylation of IκB is a key regulatory step that dictates NF-κB activation. In nearly all instances, IKK controls IκB phosphorylation (Karin, 1999). IKK directly phosphorylates two serine residues at the N terminus of IκBα (Ser-32, Ser-36) and IκBβ (Ser-19, Ser-23). This phosphorylation event triggers the polyubiquitination of IκB and the subsequent degradation by the 26S proteasome. Among the parameters involved in IKKβ/NF-κB activation, the IKK activity, IκBα degradation, NF-κB-specific DNA-protein complex formation, and NF-κB-luc activity were examined. The results showed that apigenin and luteolin had more significant inhibitions than kaempferol and chrysin on these parameters; however, their extent of inhibitions was not paralleled with that on TNF-α-induced ICAM-1 expression. Therefore, other factors, in addition to IKKβ/NF-κB, might play a role in the flavonoid-induced inhibitions on ICAM-1 expression.

AP-1 is another transcription factor that plays an important role in the induction of the ICAM-1 gene (Voraberger et al., 1991; Roebuck et al., 1995). TNF-α indeed elicited increases in AP-1-specific DNA-protein complex formation and AP-1-luc activity. Its role in TNF-α-induced ICAM-1 expression on A549 cells was further demonstrated by the finding that ICAM-1 promoter activity was attenuated using pIC1352 (ΔAP1) with AP-1 site (–279/–284) deletion. Kaempferol, chrysin, apigenin, and luteolin inhibited TNF-α-mediated AP-1 activation. Fos and Jun family proteins function as dimeric transcription factors that bind to AP-1 regulatory elements in the promoter and enhanced regions of numerous mammalian genes (Chinenov and Kerppola, 2001). Supershift assays identified the heterodimer of c-fos and c-jun in the TNF-α-activated AP-1 complex in A549 cells. Our data demonstrated the inhibition of the induced c-jun mRNA expression by kaempferol and chrysin, whereas both c-fos and c-jun mRNA expressions were inhibited by apigenin and luteolin.

The activity of AP-1 is regulated by those that increase the abundance of AP-1 components c-jun and c-fos and by those that stimulate their activity (Karin, 1995; Gomez del Arco et al., 1997). There are two major MAPKs signaling pathways reported to regulate AP-1 activity, one through ERK1/2 and the other through JNK. Activation of ERK results in an increase in AP-1 activity via c-fos induction, whereas JNK activation leads to the c-jun phosphorylation (Karin, 1995; Whitmarsh and Davis, 1996). TNF-α has been found to activate ERK, p38, and JNK in A549 cells (Chen et al., 2001; Fig. 9A). In addition to the specific MEK inhibitor, PD98059, and the p38 inhibitor, SB203580 used previously (Chen et al., 2001), a specific JNK inhibitor SP600125 was also used to examine their individual role in regulating the TNF-α-induced c-fos and c-jun mRNA expressions. The TNF-α-induced c-fos mRNA expression was inhibited by PD98059 and SB203580, but not by SP600125. However, c-Jun mRNA expression was inhibited by SP600125, but not by PD 98059 and SB203580. These results indicated the involvement of ERK and p38 in c-fos mRNA expression and JNK in c-jun mRNA expression. By examining the inhibitory effects of flavonoids on MAPKs activity further, it was found that kaempferol and chrysin only inhibited JNK activity, whereas luteolin and apigenin were potent in inhibiting the activities of ERK, p38, and JNK. These results explained why kaempferol and chrysin only inhibited the TNF-α-induced c-jun mRNA expression but failed to inhibit the c-fos mRNA expression, whereas luteolin and apigenin were potent in inhibiting both c-fos and c-jun mRNA expressions. Therefore, the extent of luteolin- and apigenin-induced inhibition on AP-1 activity was greater than that elicited by kaempferol and chrysin. Down-regulation of ICAM-1 expression by these four flavonoids was shown to be more consistent with their inhibitions on AP-1 DNA-protein binding, AP-1-luc activity, c-fos and c-jun mRNA expressions, and MAPK activities. The inhibitory effects of kaempferol and chrysin on ICAM-1 expression act via inhibition of the JNK activity and the c-jun mRNA expression, and those of apigenin and luteolin act via inhibition of the ERK, p38, and JNK activities and the c-fos and c-jun mRNA expressions. All lead to the inhibition of AP-1 activity; however, inhibitory effects on the NF-κB activation pathway also contribute to the ICAM-1 inhibition induced by apigenin and luteolin. Our previous finding that MAPK pathways were not relevant for TNF-α-induced ICAM-1 expression on A549 cells seemed not correct (Chen et al., 2001). The major evidences supporting this conclusion was from the inability of PD98059 and SB203580 to inhibit TNF-α-induced ICAM-1 expression, and the inability of TNF-α to induce AP-1-specific DNA-protein binding in A549 cells using a commercially available probe. In the present study, we use oligonucleotides corresponding to AP-1 sequences in the human ICAM-1 promoter and find an increase in AP-1-specific DNA-protein binding upon TNF-α stimulation. Furthermore, ICAM-1 promoter activity is attenuated using AP-1 site (–279/–284) deletion mutant. Our unpublished data also show the inhibition of AP-1 but not NF-κB luciferase activity by PD and SB compounds. Therefore, MAPK pathways do not involve the TNF-α-induced NF-κB activation that is regulated by PKC/c-Src/IKK and NF-κB-inducing kinase/IKK (Huang et al., 2003), although they involve the AP-1 activation as demonstrated in this study.

Quercetin, a flavonol with a structure similar to kaempferol, was also reported to act by attenuating the JNK pathway and the AP-1 activation to inhibit the TNF-α-orthe phorbol 12-myristate 13-acetate-induced ICAM-1 expression (Kobuchi et al., 1999). However, apigenin and luteolin were reported to act via down-regulation of NF-κB pathway in macrophages to inhibit the LPS-induced inducible nitric-oxide synthase and COX-2 expressions, and the releases of TNF-α and IL-6 (Liang et al., 1999; Xagorari et al., 2001). Quercetin was mediated through the down-regulation of NF-κB to inhibit LPS-induced TNF-α production (Wadsworth and Koop, 1999). (–)Epigallocatechin-3-gallate, a flavonol, was reported to act by blocking the NF-κB activation through IKK to inhibit the endotoxin-induced TNF-α production (Yang et al., 2001). In contrast to these reports, we investigated the role of AP-1 and NF-κB concomitantly and found the greater inhibitions of kaempferol, chrysin, apigenin, and luteolin on AP-1 activity, resulting in the attenuation of ICAM-1 expression. Two synthetic α-methylene-γ-butyrolactone derivatives, CYL-19s and CYL-26z have been found to down-regulate ICAM-1 expression and tumor cell invasion through the specific suppression of IKK activity and NF-κB activation (Huang et al., 2004).

We have screened a number of flavonoids and found that apigenin and luteolin are very effective in reducing the action of TNF-α on ICAM-1 expression. Similar results were also seen in TNF-α- or LPS-induced ICAM-1 expression and TNF-α-induced E-selectin expression on human umbilical vein endothelial cells (data not shown). Flavonoid aglycones consist of a benzene ring (A), fused with a six-membered ring (C) that carries a phenyl ring (B) at position 2 (Fig. 1). Our results showed that the presence of a double bond at position C₂-C₃ of the C ring with OXO function at position 4, along with the presence of OH groups at positions 3′ and 4′ of the B ring, are required for the optimal inhibition of TNF-α-induced ICAM-1 expression by luteolin. Apigenin, having similar structure to luteolin except only one OH group at position 4′ of the B ring, showed little difference in inhibiting ICAM-1 expression, indicating that one –OH group at position 4′ of the B ring is sufficient for their actions. OH group at position 4′ of the B ring is critical for the flavone types to exert their inhibitory actions, since the potency of apigenin and luteolin is much greater than that of chrysin, which lacks OH group at position 4′ of the B ring. Flavone, having a structure similar to chrysin except lacking –OH groups at positions 5 and 7 of the A ring, was ineffective. Comparing kaempferol and quercetin, with OH group at position 3 of the C ring, to apigenin and luteolin, they were a little bit less potent in inhibiting ICAM-1 expression, indicating that the –OH group at position 3 of the C ring slightly hindered the function of flavonoids. Addition of an OH group at position 5′ of the B ring (myricetin) abolished the biological activity. Our data are in line with the findings of Xagorari et al. (2001) who showed the inhibition of LPS-stimulated TNF-α and IL-6 release by luteolin, and those of Gerritsen et al. (1995) who showed the inhibition of TNF-α-induced IL-6 release in vascular endothelial cells by apigenin.

In summary, our results demonstrated that apigenin, luteolin, kaempferol, and chrysin inhibit ICAM-1 induction and monocyte adhesion on epithelial cells. Although apigenin and luteolin were more potent than kaempferol and chrysin in inhibiting ICAM-1 expression, they did not show much difference in inhibiting monocyte adhesion to epithelial cells. It seemed that to a certain extent, ICAM-1 inhibition was enough to induce reduction of monocyte adhesion. The potency variance among these four flavonoids in inhibiting ICAM-1, NF-κB, and AP-1 promoter activities was less than that in inhibiting ICAM-1 expression, either. This implied that flavonoids might affect not only ICAM-1 gene expression but also ICAM-1 protein presentation on the cell surface. The inhibitory effects of apigenin and luteolin are mediated by the sequential attenuation of the ERK1/2, p38, and JNK activities, the c-fos and c-jun mRNA expressions, and the AP-1 transcriptional activity. IKK/NF-κB pathway is also involved in their inhibitions; however, kaempferol and chrysin primarily mediate by the attenuation of the JNK activity, the c-jun mRNA expression, and the AP-1 activity. Therefore, the transcription factor AP-1 seems to play a more significant role than NF-κB in these inhibitions. The structure-activity relationships of flavonoids are also explored and found the significance of –OH groups at positions 5 and 7 of the A ring and at position 4′ of the B ring.