Apigenin Blocks Lipopolysaccharide-Induced Lethality In Vivo and Proinflammatory Cytokines Expression by Inactivating NF-κB through the Suppression of p65 Phosphorylation

Sepsis affects 500,000 people in the USA alone and has a high mortality of up to 60%. The fatal syndrome was originally attributed to LPS, a component of the cell walls of Gram-negative bacteria. LPS-induced activation of monocytes and macrophages involves the production of proinflammatory cytokines including IL-1β, IL-8, and TNF (1, 2). LPS-induced signaling events include the activation of the IκB-kinase complex (IKK)⁴ (3). In the classical activation pathway, the phosphorylation and degradation of the NF-κB inhibitor, IκBα, results in the subsequent activation of NF-κB (4). Recently, multiple additional pathways that regulate NF-κB are becoming identified, but for many of them, the molecular mechanisms remained not fully characterized (5). Among them, a noncanonical regulation of NF-κB activation in the absence of IκBα degradation was described (6, 7). In this pathway, NF-κB transcriptional activity is independent of IκBα degradation and is regulated by phosphorylation of NF-κB. Phosphorylation at Ser536 on the p65 subunit is mediated by IKK during LPS-stimulation (8). Ser536 phosphorylation is responsible for the recruitment of coactivators such as p300, promoting the transcriptional activation of NF-κB and the subsequent production of inflammatory cytokines (9). Experimental models of sepsis showed increased levels of TNF in serum, preceding organ failure (10). However, anti-TNF-mAb therapy has so far failed in clinical trails. Hence, the identification of compounds with potential therapeutic activity represents an area of great relevance for sepsis.

Flavonoids are ubiquitous phenolic compounds broadly distributed in fruits and vegetables (11). Flavonoids have long been recognized as potential anti-inflammatory, anti-oxidant, antiviral, anti-microbial, and anti-allergic, providing important nutraceutical components to our diet (12, 13, 14, 15, 16). Apigenin was shown to inhibit with different potency the proliferation of several cancer cells (17, 18, 19, 20). Recently, we demonstrated that apigenin-induced-apoptosis in leukemia is mediated by the activation of protein kinase C δ (PKCδ) and the caspases (21). In comparison, very little is known about the mechanisms that mediate the anti-inflammatory activity of apigenin. Notably, apigenin inhibits the production of IL-1β and TNF in mouse cells lines more effectively than resveratrol, another plant phenolic compound (22). Previous studies also showed that apigenin regulates prostaglandin and NO production in mouse cell lines through the regulation of NF-κB (23). However, the molecular anti-inflammatory mechanisms of apigenin and its potential as an anti-inflammatory agent in vivo remain elusive.

In this study, we investigated the mechanisms of action of apigenin in LPS-stimulated primary human monocytes and in vivo in animal models of LPS-induced sepsis. We found that apigenin inhibited the release of the inflammatory cytokines IL-8, IL-1β, and TNF in LPS-stimulated primary human monocytes. Notably, apigenin was also effective when administered after LPS-stimulation. We showed, using RT-PCR, that apigenin specifically reduced the transcription of proinflammatory cytokines in human monocytes. Transient transfection experiments revealed that apigenin inactivated the transcriptional activity of NF-κB. Apigenin did not affect the LPS-induced degradation of IκBα in human primary monocytes. Moreover, EMSAs revealed that apigenin did not affect the DNA-binding activity of NF-κB in these cells. However, we found that apigenin inhibited IKK kinase activity and suppressed the LPS-induced phosphorylation of the p65 subunit of NF-κB in mouse macrophages. This effect was overcome by the overexpression of IKKβ. In addition, apigenin inhibits LPS-induced accumulation of TNF in vivo. More importantly, we demonstrated that the pretreatment with apigenin rescued C57BL/6J mice from LPS-induced mortality. Together, these studies show that apigenin, through a noncanonical pathway, inhibits LPS-stimulated NF-κB and our results provide evidence of the potent effect of apigenin in reducing mortality associated with LPS administration in a mouse model of LPS-induced sepsis.

Blood donors were obtained from the American Red Cross (Columbus, OH) or from normal volunteers, following the protocols approved by an appropriate institutional review committee. Human monocytes were purified by clumping, on average 70–80% pure as estimated by flow cytometry using an anti-CD14 marker (24) (BD Biosciences). Briefly, fresh human monocytes were diluted 1/1 with sterile saline solution and subsequently centrifuged through a Histopaque-1077 gradient column (Sigma-Aldrich) at 600 × g for 20 min at 4°C. The mononuclear layer was removed, washed, and spun twice in RPMI 1640 (Invitrogen). The cells were resuspended in RPMI 1640/10% FBS (HyClone) at a concentration of 5 × 10⁷ cell/ml. Cells were rotated at 70 rpm on a horizontal rotor for 1 h at 4°C to induce clumping and then sedimented by gravity for 20 min through FBS at 4°C. The sedimented cells were subsequently washed twice in RPMI 1640. This method as previously shown does not activate monocytes as they undergo apoptosis, an event that is halted during activation (25, 26, 27). For real time PCR experiments human monocytes were purified by a CD14 positive selection system using magnetic beads (Miltenyi Biotec), as previously described following the manufacturers’ recommendations (28). Purified monocytes were resuspended in RPMI 1640 (BioWhittaker) to a final concentration of 3–4 × 10⁶ cells/ml and cultured for indicated times at 37°C in 5% CO₂. RAW 264.7 macrophages were maintained in RPMI 1640 medium with l-glutamine supplemented with 5% FBS and 100 U/ml penicillin, and 100 μg/ml streptomycin (BioWhittaker). The expression plasmid for pGEX-5X-3 murine NF-κB p65 (aa 277–550) was a gift from Drs. B. C. Kone and Y. U. Zhiyuan (University of Texas-Houston Medical School, Houston, TX). GST-fusion proteins were expressed in bacteria and purified by affinity on GSH-Sepharose columns using standard procedures. The pCMV IKKβ and IKKα constructs and the vector control were gifts from Drs. R. B. Gaynor and Y.-T. Kwak (29). LPS (LPS from E. coli strain 0127:B8, Difco) was used at 10 ng/ml or 100 ng/ml. Apigenin (Sigma-Aldrich) was dissolved in DMSO and used at different concentrations as specified in the text.

For Western blot analyses, cells were lysed for 30 min on ice in lysis buffer (50 mM Tris, 10 mM EDTA 0.5% Nonidet P-40, 10 mM Na-glycerophosphate, 5 mM Na-pyrophosphate, 50 mM NaF, 1 mM orthovanadate, 1 mM DTT, 0.1 mM PMSF, 2 μg/ml of protease inhibitors: chymostatin, pepstatin, antipain, and leupeptin). Cell lysates were centrifuged (14,000 × g for 10 min at 4°C) and the supernatants were stored at −70°C for future analysis. Equal amounts of protein were loaded and separated by SDS-PAGE, transferred onto nitrocellulose membranes and probed with Abs of interest followed by HRP conjugated secondary Ab and visualized by ECL (Amersham Biosciences). Abs to β-tubulin and IkBα were obtained from Upstate. NF-κB phospho-Ser536-p65 from Cell Signaling and NF-κB p65 from Santa Cruz Biotechnology. The IKKα and IKKβ Abs used for Western blots were from Santa Cruz (sc-7218, clone H-744 and sc-8014, clone H-4 respectively).

Cytokines were measured by sandwich ELISA, in supernatants of 4 × 10⁶ human monocytes cultured for 16 h, as previously described (30). For IL-1β detection, monoclonal anti-human IL-1β Ab clone 8516 (R&D Systems) was used as coating Ab and a rabbit polyclonal that recognizes mature IL-1β (raised against entire 17-kDa mature IL-1β) for the sandwich. For IL-8, a mouse anti-human IL-8 Ab clone MAB208 (R&D Systems) was used as coating Ab, and an anti-human IL-8 polyclonal Ab (Endogen) was used for the sandwich. For TNF, an anti-human TNF mAb clone 2C8 (Advanced ImmunoChemical) was used for capture, and an anti-human TNF polyclonal Ab was used for the sandwich. All sandwich Abs were used at a dilution of 1:2000. For detection of TNF from RAW macrophages, supernatants of 3.0 × 10⁶ cells treated for 8 h with 100 ng/ml LPS alone or with varying concentrations of apigenin, were assayed using TNF specific ELISA Duo-set kit per manufacturer’s instructions (R&D Systems). The same kit was used to determine TNF in serum. HRP-conjugated goat anti-rabbit Ab (Bio-Rad) was used as a developing Ab at a dilution of 1:10,000 and colorimetrically quantified after addition of 3,3′,5,5′-tetramethylbenzidine (Moss) as substrate. Samples were read in a Dynatech MRX plate reader at 450 nm using the Revelation software (Dynatech Laboratories). rIL-1β, IL-8 and TNF used as standards were obtained from R&D Systems.

For real-time PCR, total mRNA was isolated from 4 × 10⁶ monocytes with Trizol (Invitrogen) and 1–2 μg of total RNA was reverse transcribed to cDNA by ThermoScript Rnase H reverse transcriptase (Invitrogen) and diluted to 100 μl. For quantitative PCR, a 20 μl reaction mixture contained 2 μl of cDNA (20–60 ng) template, 0.25 μM primers and 10 μl SYBR Green Master Mix (Applied BioSystems). The mixture was added to MicroAMp optical PCR tubes (Applied Biosystems) in a 96-well plate format. The real time PCR was performed in an ABI PRIZM 7700 using the software Sequence Detector version 1.7 (Applied Biosystems) as previously described (31). The following primers were used for IL-1β sense: 5′-CCAGTGAAATGATGGCTTATTAC-3′ and anti-sense: 5′-CTGTAGTGGTGGTGGTCGGAGATT-3′; for IL-8: sense: 5′-AGTTTTTGAAGAGGGCTGAGAAT-3′ and anti-sense: 5′ CAACAGACCCACACAATACATGA-3′. In the case of TNF sense: 5′- GCTTGTTCCTCAGCCTCTTCT-3′ and anti-sense: 5′-GGTTTGCTACAACATGGGCTA-3′; for IRAK2 sense: 5′-CTTGGAGTGGGCTTTCTGAG-3′ and anti-sense: 5′-TTCCTTCCCCATTCTCTGTG-3′; for IL1ra sense: 5′-CACTGACCTGAGCGAGAACA-3′ and anti-sense: 5′-CCTTCGTCAGGCATATTGGT-3′. Conditions for amplification were 95°C for 10 min, 40 cycles of 95°C for 1 min, followed by 60°C for 1 min, and 72°C for 1 min. The fold change of mRNA evaluation of relative copy number (RCN) and expression ratios of selected genes were normalized to the expression of two housekeeping genes: GAPDH (primers: sense: 5′-ACTTTGGTATCGTGGAAGGACT-3′ and anti-sense: 5′-GTAGAGGCAGGGATGATGTTCT-3′) and CAP (cyclic AMP-accessory protein, primers sense:5′-ATTCCCTGGATTGTGAAATAGTC-3′ and anti-sense: 5′-ATTAAAGTCACCGCCTTCTGTAG-3′), and calculated with the equation: RCN = E^−ΔCt × 100, where E = efficiency of PCR, and Ct = Ct target − Ct reference (average of two housekeeping genes). PCR efficiency was calculated by the equation: E = 10 ^(−1/slope). Amplification of genomic DNA was used to verify that the PCR conditions did not amplify any contaminating genomic DNA.

RAW 264.7 macrophages at a density of 10⁶ cells/ml were transfected with 0.4 μg of NF-κB-SEAP vector or vector control-SEAP (BD Clontech) and pCMV-β-galactosidase, a gift from Dr. G. Leone (32), using Effectene (Qiagen). Twenty-four hours after transfection the cells were left untreated or treated for 2 h LPS alone (100 ng/ml final concentration), with LPS and apigenin (25 μM final concentration), or with apigenin alone. Supernatants were collected and assayed for NF-κB activity using the Great EscAPe SEAP chemiluminescence detection kit (BD Clontech) per manufacturer’s instructions and read on a Lumat LB 9507 (Berthold Technologies) and recorded as relative light units. Cell lysates were used to measure β-galactosidase activity using the β-Gal assay kit (Invitrogen), according to manufacturer’s instructions. Specific activity of β-galactosidase was used to normalize the luciferase activity obtained in transfected cells. All experiments were done in triplicate. EMSAs were performed as previously described (33, 34). For this purpose human primary monocytes were treated for 30 or 60 min with 10 ng/ml of LPS or with LPS and apigenin.

RAW 264.7 macrophages were lysed in buffer L (20 mM Tris pH 8.0, 0.5 M NaCl, 0.25% Triton-X-100, 1 mM EDTA, 1 mM EGTA, 10 mM PNPP, 300 μM Na₂VO₄, 10 mM NaF, 10 mM β-glycerophosphate, 1 mM benzamidine, 1 mM DTT, 1 mM PMSF, and 2 μg/ml each of chymostatin, leupeptin, antipain and pepstatin A). Lysates were diluted 1:1 in IP buffer (20 mM Tris pH 8.0, 250 mM NaCl, 0.05% Nonidet P-40, 3 mM EDTA, 3 mM EGTA containing DTT, PMSF, protease and phosphatase inhibitors as mentioned above). 250 μg of protein was immunoprecipitated with 1 μg of anti-IKKγ Abs (Santa Cruz, clone sc-8330, clone FL-419) and 20 μl of protein A Sepharose beads. Beads were washed 5 times with IP buffer followed by kinase buffer (KB: 150 mM HEPES (pH 7.6), 20 mM MgCl₂, DTT, PMSF, protease and phosphatase inhibitors as described above). For overexpression experiments, lysates from cells transfected with pCMV2-pFLAG IKKα, IKKβ or vector control and treated with LPS (100 ng/ml) for 60 min or with LPS and apigenin were used for Western analysis. Kinase reactions were performed in 45 μM of KB buffer in the presence of 500 μg of recombinant GST-p65:277–550 (35), 20 μM ATP, and 5 μCi of [γ-³²P]ATP for 30 min at 30°C. In experiments performed to evaluate the direct effect of apigenin in IKK, different concentrations of apigenin in KB buffer were added to the immunoprecipitates. Mixtures were incubated for 30 min prior the addition of GST-p65 and ATP. The reactions were stopped by addition of 5× SDS-PAGE Laemmli buffer and boiling for 5 min. Phosphorylated proteins were visualized by autoradiography. The same membranes were immunoblotted with the anti-IKKγ Abs (BD Biosciences; clone C73–1794).

Male C57BL/6J mice, 6-8 wk of age, were purchased from Harlan Teklad (Madison, WI) and used in accordance with approved regulations of the IACUC-OSU after 10 day-acclimatization. Mice were injected i.p. with apigenin (50 mg/kg of body weight) or diluents DMSO/PBS 3 h before administration of a lethal dose of LPS (37.5 mg/kg, Sigma, serotype 0127:B8, 1 × 10⁶ EU/mg) or PBS. Mice were monitored hourly. In experiments used to determine TNF, mice were treated with LPS and/or apigenin as outlined above, and serum was collected at 1 h postchallenge. Mice were anesthetized with 2–5% isofluorane (IsoFlo; Abbott Laboratories) and ∼200 μl of blood was drawn via retro-orbital puncture. Blood was clotted in Microtainer Gold blood collection tubes (BD Biosciences) for 1 h at 4°C, and serum was collected after centrifugation and stored at −80°C for future analysis. A Mantel-Haenszel log rank test was performed and a Kaplan-Meier survival curve was generated using Graph-Pad Prism (version 4.03 for Windows, GraphPad Software, www.graphpad.com).

All data are expressed as mean ± SEM. For comparison of two groups, an unpaired t test was used. For comparisons that involved multiple variables and observations, ANOVA (JMP; SAS Institute) was used. Having passed statistical significance by ANOVA, Tukey-Kramer test for all pair comparisons or individual comparisons were made by using the contrast method. Statistical significance is stated in the figure legends.

To test whether apigenin was an effective anti-inflammatory compound in human primary monocytes, freshly isolated human monocytes were collected immediately after purification (fresh), treated with LPS (10 ng/ml) alone, with LPS and different doses of apigenin or left untreated for 16 h. Stimulation with LPS induced the release of IL-1β, undetectable in freshly isolated monocytes (fresh), or in monocytes left untreated (Fig. 1,A). The treatment with apigenin alone had no observable effect on IL-1β release (Fig. 1,A) or on the release of LDH (data not shown). Next, we tested the effect of apigenin in LPS-stimulated monocytes. We found that monocytes cultured with LPS and different doses of apigenin showed significant inhibition of IL-1β release at concentrations of apigenin 10 μM and higher (Fig. 1,A, black bars, ∗, p < 0.001 compared with LPS-treated cells). To determine whether apigenin was effective after the challenge with LPS, monocytes stimulated for 1 h with LPS were treated with different doses of apigenin (Fig. 1 A, gray bars, ∗, p < 0.001 compared with LPS-treated cells). We found that apigenin at 10 μM or higher doses significantly inhibited the release of IL-1β, even when administered postchallenge.

We next examined the effect of apigenin on the proinflammatory cytokines TNF and IL-8. We found that apigenin by itself did not induce the release of TNF or IL-8 (Fig. 1, B and C). However, a concentration of 10 μM or higher of apigenin significantly reduced the release of TNF by LPS-treated human monocytes (Fig. 1,B, black bars, ∗, p < 0.001). A similar significant effect was observed when apigenin was added 1 h after the treatment with LPS (Fig. 1,B, gray bars, ∗, p < 0.001). Moreover, concentrations of apigenin 10 μM or higher significantly reduced the release of IL-8 in LPS-treated monocytes (Fig. 1 C, black bars, ∗, p < 0.05). Although more modest than the effect on IL-1β, we also found a significant reduction of IL-8 release when apigenin was added 1 h after LPS (#, p < 0.5). Together, these results indicate that apigenin inhibits the release of the inflammatory cytokines in human primary monocytes when provided in conjunction with the inflammatory stimulus and it is also significantly effective when administered after LPS-stimulation.

To examine whether apigenin affected the release or the production of proinflammatory cytokines, we compared the mRNA level of inflammatory cytokines. RT-PCR was performed on mRNA isolated from human monocytes left nontreated (NT), treated with LPS alone (10 ng/ml), treated with apigenin alone (at a concentration of 10 μM) or with LPS and apigenin for 4 h. The treatment with LPS alone induced a 7,000-fold induction on the steady-state level of IL-1β, compared with nontreated monocytes. We found that the treatment with apigenin significantly reduced the accumulation of IL-1β mRNA on LPS-stimulated monocytes (Fig. 2). The treatment with apigenin alone had no effect on IL-1β mRNA accumulation (Fig. 2).

Similarly, we found that the treatment with apigenin inhibited by ∼9-fold the accumulation of IL-8 and TNF mRNAs from LPS-stimulated monocytes (Fig. 2). In contrast, the level of IL-1ra (IL-1-receptor antagonist) and IRAK2 (not shown) mRNAs was similar in LPS-stimulated monocytes and monocytes treated with LPS and apigenin (Fig. 2). These experiments indicate that apigenin affects the transcript levels of several inflammatory cytokines.

Because the activation of NF-κB is responsible for the production of proinflammatory cytokines upon LPS-stimulation, we next determined the effect of apigenin on NF-κB transcriptional activity. Raw 264.7 mouse macrophages were stimulated with LPS alone or with LPS in the presence of different doses of apigenin for 8 h. We found that apigenin at a final concentration of 25 μM significantly inhibited the induction of TNF induced by LPS in this model system (Fig. 3,A). Moreover, apigenin alone did not promote production of TNF (Fig. 3,A). Next, macrophages transfected with an NF-κB luciferase reporter construct or a vector control (see Materials and Methods) were treated with LPS, with apigenin or with LPS and apigenin for 2 h. We found that the LPS-induced-transcriptional activity of NF-κB was significantly reduced by the treatment with apigenin (Fig. 3 B, black bars, ∗, p < 0.001 LPS-treated cells compared with LPS and apigenin-treated cells).

To examine the molecular mechanisms by which apigenin inhibits the NF-κB transcriptional activity, we investigated the effect of apigenin on IκBα turnover in response to inflammatory mediators. For this purpose, primary human monocytes were treated with LPS alone (10 ng/ml) or LPS and apigenin (10 μM) for different lengths of time. We repeatedly found that the degradation of IκBα induced by LPS did not change when the cells were treated with apigenin (Fig. 4,A, compare LPS and LPS + apigenin). Next, we investigated whether apigenin affected the NF-κB DNA binding activity. For this purpose, nuclear lysates from human primary monocytes treated as described above were analyzed by EMSA, as we previously described (34). In line with the lack of IκBα regulation, apigenin treatment did not affect the DNA-binding activity of NF-κB in LPS-stimulated human monocytes (Fig. 4,B). Recent studies have shown that activation of NF-κB can occur through the phosphorylation of Ser536 on the p65 subunit independent of ΙκΒ turnover (6, 7). To investigate the possibility that apigenin regulates NF-κB by a mechanism involving p65 phosphorylation, mouse macrophages were treated with LPS alone (100 ng/ml) or LPS and apigenin (25 μM final concentration) for different lengths of time. Immunoblots showed phosphorylation of Ser536 after 30 min of LPS-stimulation. In contrast, apigenin completely suppressed the phosphorylation of Ser536 induced by LPS (Fig. 4,C). Because the phosphorylation of Ser536 can be mediated through IKK, we next analyzed the effect of apigenin on IKK activity in LPS-stimulated macrophages. IKK immunoprecipitated from cells treated with LPS and apigenin or LPS alone were subjected to kinase assays using GST-p65 as a substrate. We found that apigenin significantly inhibited the activity of IKK induced by LPS (Fig. 4 D, compare LPS and LPS + Apigenin).

Next, we attempted to determine whether the effect of apigenin on LPS-induced p65 phosphorylation occurred through IKKα or IKKβ. Mouse macrophages over-expressing IKKα or IKKβ were treated with LPS or LPS and apigenin (25 μM) for 60 min. Immunoblots showed that LPS-stimulated phosphorylation of Ser536 on the p65 subunit of NF-κB was reduced by the treatment with apigenin, regardless of whether the cells over-expressed IKKα or were transfected with the control vector (Fig. 5,A), suggesting that IKKα is not likely to be involved in the response to apigenin. In contrast, we found that over-expression of IKKβ overcame apigenin-mediated suppression of phosphorylation induced by apigenin on Ser536 on the p65 subunit of NF-κB (Fig. 5 A).

Next, we evaluated whether apigenin directly affects the activity of IKK proteins. The IKK complex was therefore immunoprecipitated from lysates of mouse macrophages treated with LPS for 90 min. Immunoprecipitates were then treated with various concentrations of apigenin and subjected to an IKK kinase assay. We found that apigenin did not directly affect the activity of IKK, suggesting that apigenin inhibits IKK activity through an indirect mechanism (Fig. 5 B).

Because apigenin efficiently inhibited LPS-stimulation of proinflammatory cytokines ex vivo, we next investigated whether apigenin protects animals from LPS lethal toxicity in a mouse model of E. coli LPS-induced mortality. C57BL/6J mice were injected with apigenin (50 mg/kg) or vehicle control and, 3 h later, they were challenged i.p. with LPS. Pretreatment of mice with apigenin before the LPS-challenge significantly attenuated the increase in serum TNF levels elicited after 1 h of LPS-challenge (Fig. 6 A, 1436.7 ± 149.7 pg/ml for the LPS group vs 476.2 ± 74.3 pg/ml for the apigenin plus LPS group, n = 3, 12 animals per treatment ∗, p < 0.01).

We next investigated whether apigenin can increase survival in mice treated with a lethal dose of LPS. Animals were injected i.p. with apigenin (or control diluent) before injecting a lethal dose of LPS (as described above). Mice receiving apigenin alone exhibited no lethality. In contrast, all the mice receiving LPS died 28 to 42 h postchallenge. Pretreatment with apigenin increased survival to 100% at 70 h (Fig. 6 B). Only one mouse died after 5 days while the rest remained alive after 7 days.

The recognition of flavonoids as potential anti-inflammatory agents is emerging but the molecular mechanisms responsible for their action and their efficacy in vivo remain unclear (14). LPS stimulation elicits a cascade leading to the activation of NF-κB and the production of proinflammatory cytokines. Inhibition of the LPS-stimulated signal transduction cascade has been proposed as a promising target for the treatment of sepsis (36). The present study demonstrates that apigenin inhibits LPS-stimulated production of proinflammatory cytokines in human primary monocytes and macrophages through the inactivation of NF-κB by suppressing the LPS-induced phosphorylation of the p65 subunit. In addition, we show in vivo that apigenin inhibits LPS-stimulated TNF and induces survival in a mouse model of lethal LPS-induced sepsis.

Several flavonoids were shown to modulate the production of NF-κB in mouse macrophage cell lines (22, 37). Our findings in human primary monocytes show that apigenin blocks the release and production of IL-1β, TNF, and IL-8 in LPS-stimulated cells at apigenin concentrations of 10 μM and higher (Fig. 1 and 2). These results appear in agreement with previous reports using mouse macrophages cell lines (38, 39). Remarkably, we find that apigenin is also very effective when administered after LPS-stimulation. Apigenin administration after LPS reduced TNF and IL-1β levels in a similar manner as when apigenin was administered post-LPS challenge. The reduction of IL-8 was less dramatic, but also remained significant (Fig. 1). Moreover, we showed that apigenin inhibited the production of proinflammatory cytokines (Fig. 2). Our results suggest that apigenin acts with some level of specificity on proinflammatory cytokines because the expression of other genes such as IL-1ra and IRAK2 was not affected (Fig. 2). Our findings demonstrate that apigenin’s inhibition of the transcriptional activity of NF-κB in macrophages is consistent with previous observations (Fig. 3,B and Refs. 23 and 38). Moreover, the inhibitory effect of apigenin on NF-κB transcriptional activity in macrophages is at least four times higher than the activity reported on epithelial cells (38, 39). In addition, we found that apigenin reduced IKK activity induced by LPS. We demonstrate that apigenin modulated LPS-induced phosphorylation of NF-κB p65 subunit at Ser536 (Fig. 4). In contrast with previous studies which used mouse macrophages and epithelial cell lines (38, 39), our findings clearly and repeatedly show that apigenin regulates NF-κB activity independently of the LPS-triggered degradation of IκBα (Fig. 4,A). In this context, recent studies demonstrated that NF-κB activity is regulated by phosphorylation of the p65 subunit in the absence of IκBα degradation (7, 40, 41, 42). In this nonclassical mechanism, changes in the phosphorylation of NF-κB affect the interaction of p65 with other components of the basal transcriptional machinery without affecting NF-κB DNA-binding activity (42). In agreement with this model, we found that the binding of LPS-stimulated NF-κB to DNA was not affected by apigenin (Fig. 4,B). However, the LPS-induced phosphorylation of p65 on Ser536 was dramatically reduced by apigenin (Fig. 4,C). It is clear that our results support a noncanonical pathway, however why apigenin decreased IKK activity without affecting the turnover of IκBα is not yet known. Although it is becoming well accepted that the activity of NF-κB can be regulated through many nonclassical pathways in addition to the canonical, the details of their signal transduction cascades remain not fully characterized (5). In light of our results, it is possible to speculate that the IKK activity that remained after the exposure to apigenin (Fig. 4,D) was sufficient to regulate IκBα proteolysis, but not adequate to maintain p65 Ser536 phosphorylation. An alternative explanation of these observations is that a factor downstream of IKK required for the phosphorylation of p65, may itself also be blocked by apigenin, thus reducing p65, but not affecting IκBα. It was previously shown that LPS-induced phosphorylation of p65 on Ser536 was regulated by IKKβ (8). In agreement with these results, we found that over-expression of IKKβ can overcome the inhibitory effect of apigenin in the LPS-induced phosphorylation on Ser536 of the p65 subunit (Fig. 5,A). Several plant natural products have been previously shown to inhibit the IKK activity by either directly or indirectly regulating the IKK proteins (43, 44). However, our findings show that the effect of apigenin is not direct on IKK proteins (Fig. 5 B), suggesting that apigenin regulation occurs indirectly of IKKβ, possibly involving an associated member or an upstream regulator of the IKK complex.

Most studies that investigate the potential use of flavonoids as anti-inflammatory compounds rely on in vitro models using cell lines. As these cells lines have aberrant mutations, results obtained in these systems, although helpful, may not necessarily represent the mechanisms used by apigenin in normal cells of the innate immune system. One of the most significant findings of the present study is that apigenin protects mice from LPS-induced lethal toxicity in vivo (Fig. 6 B). This effect was accompanied by the decreased of LPS-stimulated production of TNF. The increased survival of mice pretreated with apigenin is in agreement with our in vitro and ex vivo experiments using human primary monocytes. In addition, apigenin was shown to block the activity of enzymes involved in inflammation, including lipo- and cyclooxygenase (38). Moreover, apigenin inhibits neutrophil and lymphocyte adhesion to endothelial cells by regulating ICAM expression (45). In vivo, during sepsis, up-regulation of proinflammatory cytokines initiates and contributes to organ dysfunction syndrome. Inhibition of the expression of these molecules may ameliorate inflammation.

In summary, we have demonstrated that apigenin can protect mice from LPS-induced lethality. Moreover, apigenin inhibits TNF production in vivo and ex vivo in primary human monocytes. Our findings provide novel insights into the molecular mechanisms by which apigenin regulates inflammation and show that apigenin functions by regulating NF-κB activity through the suppression of LPS-induced phosphorylation of p65.

We thank Drs. B. C. Kone and Y. U. Zhiyuan of University of Texas-Houston Medical School for providing the GST-NFκB clones. We thank Dr. G. Leone of The Ohio State University for the pCMV-β-galactosidase construct. We thank the prompt assistance of Drs. R. B. Gaynor and Y. T. Kwak in providing the pCMV2-pFLAG IKKα and IKKβ constructs. We thank the reviewers of this manuscript for their valuable suggestions.

The authors have no financial conflict of interest.