Withaferin A Strongly Elicits IκB Kinase β Hyperphosphorylation Concomitant with Potent Inhibition of Its Kinase Activity*

The transcription factor NFκB plays a critical role in normal and pathophysiological immune responses. Therefore, NFκB and the signaling pathways that regulate its activation have become a major focus of drug development programs. Withania somnifera (WS) is a medicinal plant that is widely used in Palestine for the treatment of various inflammatory disorders. In this study we show that the leave extract of WS, as well as its major constituent withaferin A (WA), potently inhibits NFκB activation by preventing the tumor necrosis factor-induced activation of IκB kinase β via a thioalkylation-sensitive redox mechanism, whereas other WS-derived steroidal lactones, such as withanolide A and 12-deoxywithastramonolide, are far less effective. To our knowledge, this is the first communication of IκB kinase β inhibition by a plant-derived inhibitor, coinciding with MEK1/ERK-dependent Ser-181 hyperphosphorylation. This prevents IκB phosphorylation and degradation, which subsequently blocks NFκB translocation, NFκB/DNA binding, and gene transcription. Taken together, our results indicate that pure WA or WA-enriched WS extracts can be considered as a novel class of NFκB inhibitors, which hold promise as novel anti-inflammatory agents for treatment of various inflammatory disorders and/or cancer.

The transcription factor NFB plays a critical role in normal and pathophysiological immune responses. Therefore, NFB and the signaling pathways that regulate its activation have become a major focus of drug development programs. Withania somnifera (WS) is a medicinal plant that is widely used in Palestine for the treatment of various inflammatory disorders. In this study we show that the leave extract of WS, as well as its major constituent withaferin A (WA), potently inhibits NFB activation by preventing the tumor necrosis factor-induced activation of IB kinase ␤ via a thioalkylation-sensitive redox mechanism, whereas other WS-derived steroidal lactones, such as withanolide A and 12-deoxywithastramonolide, are far less effective. To our knowledge, this is the first communication of IB kinase ␤ inhibition by a plant-derived inhibitor, coinciding with MEK1/ ERK-dependent Ser-181 hyperphosphorylation. This prevents IB phosphorylation and degradation, which subsequently blocks NFB translocation, NFB/DNA binding, and gene transcription. Taken together, our results indicate that pure WA or WA-enriched WS extracts can be considered as a novel class of NFB inhibitors, which hold promise as novel anti-inflammatory agents for treatment of various inflammatory disorders and/or cancer.
The search for anticancer drugs and anti-inflammatory agents from natural products represents an area of great interest worldwide (1). In the Palestinian traditional folk medicine, more than 700 plant species are known for their use as medicinal herbs or as botanical pesticides (2,3). As revealed by a recent ethnobotanical survey carried out in the West Bank, at least 63 plant species are still in use (4). Plants like Withania somnifera (WS) 4 are applied in different forms (decoctions, infusions, ointments, powder, syrup, etc.) (5,6) to treat burns, wounds, and/or dermatological disorders, to prevent infections, and to treat various diseases including gastrointestinal diseases, infertility, and cutaneous abscesses (7).
WS belongs to the Solanaceae family and is being used in many indigenous systems of medicine (8). It is a small, semiwoody shrub with ovate leaves and greenish orange fruit. It grows wild to about 1.5 m in height and can be found growing in Palestine as well as in some other Middle East countries, in Africa, and in India (where it is alternatively called ashwagandha in ayurvedic medicine). As a result of this wide growing range, there are considerable morphological and/or chemotypical variations in terms of local species, which may affect the composition of its metabolites and/or its activities (6, 9 -11). The biologically active constituents of WS are alkaloids (isopelletierine and anaferine), saponins (sitoindosides), and the C 28 steroidal lactones (withanolides) (12). Although some evidence on immunostimulatory and anti-inflammatory activities of WS has been reported before, the molecular mechanisms by which its constituents exert their immunomodulatory effects have not been studied and characterized sufficiently.
Because the NFB transcription factor is a key regulator of cellular processes involved in the immune response, differentiation, cellular proliferation, and apoptosis (13,14) and because the constitutive activation of NFB contributes to multiple pathophysiological conditions such as rheumatoid arthritis, asthma, inflammatory bowel disease (15), AIDS (16), and cancer (17), we investigated the effect of the WS leaf extract and its major constituent withaferin A (WA) on the NFB signaling pathway.
In mammalian cells, the NFB/Rel family consists of five members: RelA (p65), RelB, c-Rel, p105/p50 (NFB1), and * This work was supported in part by grants from Interuniversitaire Attractie-p100/p52 (NFB2). Each family member has a conserved Rel homology domain specifying DNA binding, protein dimerization, and nuclear localization. In most cells, NFB is composed of a heterodimer of p65 and p50 where the p65 protein is responsible for the transactivation potential. In unstimulated cells, NFB is sequestered predominantly in the cytoplasm in an inactive complex through interaction with IB inhibitor proteins. In response to stimulation by a variety of potent activators, such as tumor necrosis factor (TNF)-␣, interleukin (IL)-1, or lipopolysaccharide (18), IB␣ is rapidly phosphorylated at two conserved NH 2 -terminal serines (Ser-32 and Ser-36) and degraded through a ubiquitin-dependent proteolysis, resulting in the release of NFB, translocation into the nucleus, and induction of gene transcription. NFB regulates a wide variety of important target genes encoding cytokines (IL-6, TNF, and IL-1), chemokines (IL-8 and RANTES), adhesion molecules (E-selectin), and even its own inhibitors, IB␣ and IB␤ (19,20).
IB kinase (IKK) is the protein kinase complex responsible for IB phosphorylation in response to proinflammatory stimuli, resulting in ubiquitination and degradation of the latter. IKK is a multisubunit complex that contains two catalytic subunits, IKK␣ and IKK␤, and a regulatory subunit, IKK␥/NEMO (NFB essential modulator) (20,21). Knock-out studies have revealed that IKK␤ is responsible for the proinflammatory cytokine-induced activation of NFB (22). The mechanisms that activate the IKK complex are not completely understood, but it is known that activation requires phosphorylation of Ser-177 and Ser-181 in the activation loop of IKK␤ (23). The physiological molecular mechanism for postinduction inactivation of IKK remains a mystery; regulated IKK inhibitors such as A20 (24) or other IKK inhibitors such as PP2C␤, PP2A, CYLD (cylindromatosis), hTid-1, and Hsp70 might be involved (25)(26)(27)(28)(29)(30). Alternatively, regulation of the conformational state of IKK␤ via hyperautophosphorylation of its COOH terminus could lead to IKK inactivation (23). Importantly, the disruption of the COOH-terminal sites in IKK␤ leads to persistently active IKK upon stimulation (23), suggesting inhibitory autophosphorylation as the major source of fast IKK down-regulation. Considerable work has been directed at identifying small molecules that inhibit IKK␤ as possible targets for the development of anti-inflammatory and antineoplastic drugs (31).
In this study, we highlight the potential of pure WA compound or WA-containing fractions of the WS leaf extract from Palestine as potent and specific inhibitors of the NFB pathway. We observed that WA inhibits NFB activation through inhibition of IKK␤ activity depending on thioalkylation-sensitive IKK␤ regulation concomitant with IKK␤ Ser-181 hyperphosphorylation. This results in stabilization of cytoplasmic IB␣ and hence the reduction of nuclear translocation, diminished NFB/DNA binding, and blockage of NFB-driven gene expression.

MATERIALS AND METHODS
Plant Collection and Preparation-The leaves of WS were collected during the months of January to April from the surrounding area of Ramallah, Palestine. The shade-dried and ground leaves of WS were extracted with a mixture of dichlo-romethane and methanol (1:1, v/v) for 24 h at room temperature. After filtration, the residue was covered with 100% methanol for 20 min and then drained into the same flask. The solvent was dried by rotary evaporation under reduced pressure and at a maximal temperature of 45°C. The resulting crude extracts were stored at Ϫ20°C until assayed. Before running the bioassay the extract was dissolved in Me 2 SO, diluted in Dulbecco's modified Eagle's medium, and added to the cell culture at the dose or concentration range indicated in figure legends.
Chromatography-The crude extract was fractionated via solid phase extraction using Bond Elut C 18 cartridges (Varian, St.-Kathelijne Waver, Belgium). After activation with methanol and conditioning with water, the extract (dissolved in methanol and diluted with 9 volumes of water) was applied to the cartridge. Fractions were eluted, respectively, with solvent of increasing strengths: water, water/methanol (3:1, v/v), water/ methanol (1:1), water/methanol (1:3), and finally methanol. Appropriate volumes were used depending on the scale of the fractionation. All extracts and fractions were analyzed using HPLC. Analyses were carried out using a Waters 2695 Alliance Separations Module equipped with a Waters 996 Photodiode Array detector. The column was a Varian Omnispher (C 18 , 250 ϫ 4.6 mm, 5 m) and was maintained at 35°C; the injection volume was 20 l. Gradient elution over 60 min was applied from 15% solvent B (a methanol/acetonitrile mixture with 0.025% formic acid) in solvent A (water with 0.025% formic acid) to 95% solvent B in solvent A. Another gradient from 45% solvent C (methanol with 0.025% formic acid) in solvent A to 85% solvent C in solvent A was used for the quantitative analysis of WA in the WS extract and in the fractions. Chromatograms at 220 nm were extracted from the three-dimensional data, and peaks were characterized based on their UV spectra and retention times and compared with authentic standards. Peak integrations were carried out using standard parameters. Isolation of WA and withanolide A was carried out using a Gilson 322 pump equipped with a Gilson UV-visible 156 detector and a Gilson 206 fraction collector. Solvents and gradients were similar to those used for analytical purposes. The column was a Varian Omnispher (C 18 , 350 ϫ 21.4 mm, 10 m), and fractions were collected either manually or based on peak intensity. The organic solvent remaining in the collected fractions was removed under reduced pressure using a rotavapor, and fractions were taken to dryness via lyophilization. Samples were analyzed with mass spectroscopy using flow injection on a Micromass Quattro Micro. The purified materials were also subjected to 1 H NMR and 13 C NMR using a Bruker 300 MHz spectrometer.
Cell Culture-Murine fibrosarcoma L929sA cells and human embryonic kidney 293T cells were maintained in Dulbecco's modified Eagle's medium supplemented with 5% newborn calf serum, 5% fetal calf serum, 100 units/ml penicillin, and 0.1 mg/ml streptomycin. Twenty-four hours before induction, cells were seeded in multiwell dishes such that they were confluent at the time of the experiment. IKK␣and IKK␤-deficient mouse embryonic fibroblasts (kindly provided by Dr. Emmanuel Dejardin) (32) and cervix cancer cells (HeLa) were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. MDA-MB-231 human breast cancer cells were cultured in L-15 medium supplemented with 10% fetal calf serum.
Cytokines, Inducing Compounds, Antibodies, and Reagents-Recombinant murine TNF, produced in Escherichia coli and purified in our laboratory to at least 99% homogeneity, had a specific biological activity of 8.58 ϫ 10 7 IU/ml of protein as determined in a standard TNF cytolysis assay. Reference TNF (code 88/532) was obtained from the National Institute of Biological Standards and Control (Potters Bar, UK). U0126 was supplied by Promega (Madison, WI); SB203580 was purchased from Alexis (Lausanne, Switzerland); SP600125 was purchased from Calbiochem; cycloheximide was purchased from Sigma. Work solutions of the reagents were routinely prepared in culture medium. Control experiments verified that the final concentration of organic solvents did not interfere with any of the assays.
Biological IL-6 Assay-Secreted IL-6 was quantified according to its growth-stimulating effect on 7TD1 cells as described previously (33).
Transfection Procedure-Stable transfection of L929sA cells was performed by the calcium phosphate precipitation procedure according to standard protocols (33).
Reporter Gene Analysis-Luciferase and galactosidase reporter assays were carried out according to the manufacturer's instructions (Promega) and have been described previously (33). Normalization of luciferase activity was performed by measurement of ␤-galactosidase levels in a chemiluminescent reporter assay Galacto-Light kit (Tropix, Bedford, MA). Light emission was measured in a luminescence microplate counter (Packard Instrument Co.). Luciferase activity, expressed in arbitrary light units, was corrected for the protein concentration in the sample by normalization to the co-expressed ␤-galactosidase levels. ␤-Galactosidase protein levels were quantified with a chemiluminescent reporter assay Galacto-Light kit (Tropix).
Electrophoretic Mobility Shift Assay (EMSA)-L929sA, MDA-MB231, HeLa, and human embryonic kidney 293T cells were seeded in 6-well plates at 3 ϫ 10 5 cells/well. After appropriate induction, cells were washed with ice-cold phosphatebuffered saline (PBS), harvested with a rubber policeman, and precipitated in 1 ml of PBS by centrifugation for 10 min at 2600 rpm (4°C). Preparation of total, cytoplasmic, or nuclear cell extracts has been described previously (35,36). For EMSA, equal amounts of protein were incubated for 25 min with an NFB-specific 32 P-labeled oligonucleotide and binding mixture as described previously (36,37). Labeling of the oligonucleotides was performed with [␣-32 P]dCTP by using Klenow enzyme (Roche Applied Science). The NFB oligonucleotide comprises the sequence 5Ј-AGCTATGTGGGTTTTCCCAT-GAGC-3Ј in which the single IL-6 promoter-derived NFB motif is bold and underlined. Samples were loaded on a 6% polyacrylamide gel run in 0.5ϫ Tris borate-EDTA buffer (pH 8). The gel was dried after electrophoresis, and complexes formed were analyzed using PhosphorImager technology.
Western Blot Analysis-L929sA cells were seeded in 6-well plates at 3 ϫ 10 5 cells/well. Cells were pretreated with the extract/compound for 1 h followed by incubation with TNF for the indicated time periods. Cells were washed with ice-cold PBS and lysed in SDS sample buffer (62.5 mM Tris-HCl, 2% (w/v) SDS, 10% glycerol, 50 mM dithiothreitol (DTT), and 0.01% (w/v) bromphenol blue). To shear DNA and reduce sample viscosity, lysates were sonicated for 1 min in a water bath sonicator and then heated to 95°C for 5 min after which they were immediately cooled in ice and microcentrifuged for 5 min. The lysates were separated by 10% SDS-PAGE and electrotransferred onto a nitrocellulose membrane. Blots were probed using the appropriate antibodies, and the immunoreactive protein was detected using enhanced chemiluminescence reagents (Eastman Kodak Co.).
Zymosan-induced Inflamed Paw Model-Eight-week-old C57BL/6J mice were purchased from Iffa Credo. The experimental setup contained three groups with 10 animals per group. Group 1 received an intraperitoneal injection of 60 l of Me 2 SO (sterile) followed 30 min later by a subcutaneous injection of zymosan solution (15 mg/ml in phosphate-buffered saline, sterilized) in the right footpad; the left footpad was left uninjected. Group 2 was treated intraperitoneally with 60 l of 0.5 mg/mouse withaferin A (dissolved in Me 2 SO) followed 30 min later with zymosan injection. Group 3 was treated with 60 l of (150 g/mouse) dexamethasone (DEX) followed 30 min later with zymosan injection. 6, 12, 24, and 48 h after the zymosan treatment, the thickness of both footpads was measured by using a caliper, and the difference between zymosan-and noninjected footpads was compared for all three experimental groups.
Determination of Systemic IL-6 Levels in Blood-Blood samples were taken by sinus retro-orbital punction under isoflurane anesthesia from mice 6 h after zymosan injection, which followed pretreatment with Me 2 SO (60 l/mouse), withaferin A (0.5 mg/mouse), or DEX (150 g/mouse) injected intraperitoneally (six mice per group) for 30 min. IL-6 protein levels were tested by the biological 7TD1 assay as described elsewhere (33).
Reverse Transcription-PCR-Reverse transcription-PCRs were carried out according to the instructions of the manufacturer (Promega). Briefly, an aliquot of 5 g of total RNA, 2 l of oligo(dT), and x ml of DEPC-bidi (diethyl pyrocarbonatetreated bidistilled water) (to adjust the final volume to 10 l) was placed in a microcentrifuge tube and incubated at 70°C for 10 min. The sample was briefly centrifuged and placed on ice for 10 min. A 20-l reaction mixture (containing 6 l of buffer 5ϫ, 2.5 l of 100 mM DTT, 2.5 l of 2.5 mM dNTP, 0.5 l of RNasin, 1 l if avian myeloblastosis virus reverse transcriptase, and 7.5 l of DEPC-bidi) was added to the annealed RNA sample. The reverse transcription was conducted using the following conditions: 5 min at 70°C followed by a decrease of 0.1°C/s to 42°C, 60 min at 42°C, 5 min at 90°C per cycle. The obtained cDNA fragments were then further analyzed with quantitative PCR. cDNA-specific primer sets are available upon request. The PCR products were analyzed by agarose gel electrophoresis. After staining the gel with ethidium bromide, the products were detected under UV light.
Northern Blot Analysis-After appropriate inductions, RNA isolations were obtained by using TRIzol TM reagent (Invitrogen). Briefly 10 ϫ 10 6 L929sA cells were grown to subconfluency in 14-cm Petri dishes. Cells were pretreated with WS for 1 h before TNF induction for 6 h. Total RNA was isolated with TRIzol; denaturation was achieved with a combination of Me 2 SO and deionized glyoxal (Sigma). A total of 23 g of RNA was separated on a 1% agarose gel in 20 mM phosphate buffer (pH 7). RNA was transferred to Hybond-Nϩ membranes (Amersham Biosciences) by standard capillary blotting and cross-linked by UV irradiation. Hybridization was done by sequentially probing with murine IL-6, human IB␣, RANTES, and murine glyceraldehyde-3-phosphate dehydrogenase cDNA fragments. The membrane was stripped (with boiling 0.5% SDS, left on the membrane until it is cold) before each new hybridization step. All fragments were labeled with [␣-32 P]dCTP by using a Random Primed labeling kit (Roche Applied Science).

WS Leaf Extract Inhibits Inflammatory Gene Expression-
IL-6 is a pleiotropic cytokine involved in acute phase and immune reactions and inflammatory responses. Previously, we have demonstrated potent IL-6 gene activation in response to TNF in the mouse fibrosarcoma cell line L929sA that can be reversed by a plethora of anti-inflammatory compounds, such as chemical kinase inhibitors, natural products, hormones, etc. (33)(34)(35)(36)(37)(38)(39)(40).
Taking into account the traditional use of WS in inflammatory affections, we explored the effect of WS leaf extract on TNF-induced IL-6 gene expression in L929sA mouse fibroblasts. As shown in Fig. 1A, the elevated levels of IL-6 protein detected after TNF treatment were strongly repressed in the presence of the extract in a concentration-dependent manner.
WS Inhibits NFB-driven Gene Expression-To verify whether the decrease in IL-6 protein levels by WS was due to transcriptional repression at the promoter level, L929sA cells stably transfected with various TNF-inducible reporter gene constructs, driven by the natural full-size IL-6 promoter (p1168hu.IL6P-lucϩ), the IL-8 promoter (p1481hu.IL8Plucϩ), and the E-selectin promoter (pE-selectin-lucϩ), respectively, were tested for WS effects in comparison with a constitutively expressed reporter gene construct (pPGKbGeobpA) controlled by the phosphoglycerokinase promoter (34). Enhanced luciferase expression levels were measured in response to TNF for the various reporter gene constructs, whereas pretreatment with WS was found to strongly inhibit reporter gene expression, leaving the constitutive housekeeping promoter phosphoglycerokinase expression unaffected under all conditions tested (Fig. 1B). Because we previously demonstrated an essential role for NFB in triggering IL-6 gene transcription in response to TNF, we also measured WS effects on a recombinant promoter with multiple NFB-responsive elements, i.e. p(IL-6B) 3 50hu.IL6P-lucϩ (Fig. 1C). The expression levels observed with the recombinant promoter or the various NFB-driven reporter gene constructs (IL-6, IL-8, E-selectin, and NFB) are in agreement with the drastic repression observed with the endogenous IL-6 protein levels and reconfirm the central role of NFB in TNF-induced IL-6 gene expression. In contrast, upon testing WS effects on another recombinant promoter construct, pAP1luc, with multiple AP1-responsive elements, it appeared that WS did not repress but slightly elevated AP1-driven gene expression, further illustrating the NFB-selective activities of WS (Fig. 1C). Furthermore and along the same line, mRNA analysis of various endogenous NFB target genes was evaluated in the presence or absence of the WS extract by Northern blot analysis and reverse transcription-PCR. As expected and shown in Fig. 1D, pretreatment of WS extract potently suppressed mRNA levels of basal and TNFinduced IL-6, RANTES, and IB␣, whereas the glyceraldehyde-3-phosphate dehydrogenase housekeeping transcription remained unaffected.
WS Inhibits NFB/DNA Binding-We further investigated whether the inhibition at the promoter level is due to the inhibition of NFB/DNA binding. As shown in Fig. 1E, TNF-induced NFB/DNA binding could be seen as early as 15 min and was maintained until 120 min. Pretreatment with WS caused a complete inhibition of NFB/DNA binding at all time points, whereas there was no effect on the levels of recombination signal sequence-binding protein J, binding onto the same DNA motif (36). Similar results were obtained in the human cell lines MDA-MB-231, HeLa, and human embryonic kidney 293T pointing to a cell-independent action mechanism.
WS Inhibits NFB Translocation-Because NFB activation requires nuclear translocation of the NFB p65 subunit prior to NFB/DNA binding, we next measured the level of p65 in the cytoplasm and nucleus under the various experimental conditions tested. As illustrated in Fig. 1F, and as expected, TNF induction resulted in an increase of p65 levels in the nucleus. Pretreatment of cells with WS extract abolished this effect, demonstrating that WS inhibits the release of p65 from its inhibitor molecule, IB␣, in line with the above mentioned observations that WS interferes with TNFinduced NFB/DNA binding.
WS Inhibits TNF-induced Phosphorylation and Degradation of IB␣-Because IB␣ degradation is required for translocation of NFB to the nucleus, we next determined whether inhibition of NFB/DNA binding by WS was due to the inhibition of IB␣ degradation or of new IB␣ synthesis. We found that 15 min after TNF induction, IB␣ was completely degraded, whereas it fully reappeared after 90 min. However, IB␣ degradation was completely blocked upon incubation with WS extract (Fig. 2A). The observation that inhibition of NFB/DNA binding by WS still persists in the presence of the protein synthesis inhibitor cycloheximide (Fig. 2B) further proves that WS interferes with the processing of IB rather than with IB protein synthesis.
As IB␣ needs to be phosphorylated by the IKK complex prior to its polyubiquitination and proteasomal degradation (14), we next investigated whether WS blocks either the TNF-induced IB␣ phosphorylation event or the degradation subsequent to its phosphorylation. To evaluate the level of IB␣ phosphorylation in Western blot analysis, we used the proteasomal inhibitors MG132 and PSI to block the degra-FIGURE 1. A, WS inhibits expression of the inflammatory cytokine IL-6. L929sA cells were pretreated for 1 h with the indicated concentrations of WS followed by 6-h treatment with 2000 IU/ml TNF. Corresponding levels of secreted IL-6 protein were quantified by the biological 7TD1 assay. B, WS inhibits NFB-driven reporter gene expression. L929sA cells were stably transfected with reporter gene plasmids containing the natural IL-6, IL-8, E-selectin (E-sel.), and phosphoglycerokinase (PGK) promoter luciferase constructs. L929sA transfectants were left untreated (untr) or were treated with 2000 IU/ml TNF for 6 h either alone or following a 1-h pretreatment with WS extract (63 g/ml). Lysates were prepared for quantification of reporter gene expression levels and normalization for protein concentration. The induction factor is defined as the amount of luciferase produced in TNF-treated cells after normalization for ␤-galactosidase expression compared with the non-induced state. C, WS inhibits NFB-but not AP1-driven reporter gene expression. Stable L929sA transfectants of recombinant reporter gene constructs, driven by either NFB-or AP1-responsive elements in front of a minimal promoter, were left untreated or were treated respectively with 2000 IU/ml TNF or 60 nM staurosporine (STS) (34) for 6 h either alone or following a 1-h pretreatment with WS extract (63 g/ml). Lysates were prepared for reporter gene analysis as described in B. D, WS represses endogenous NFB-regulated gene expression. L929sA cells with or without WS pretreatment (63 g/ml for 1 h) were stimulated with 2000 IU/ml TNF for 6 h. Total cytoplasmic RNA was isolated, and reverse transcription-PCR was performed with mRNA-specific primer sets for IL-6, IB␣, RANTES, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH), respectively. E, effect of WS on TNF-induced NFB/DNA binding. L929sA cells were left untreated or treated with 2000 IU/ml TNF alone or following a 1-h pretreatment with 63 g/ml WS for the indicated time points. Total cell lysates were incubated with a 32 P-labeled IL-6 B site-containing probe. Binding complexes formed were analyzed by EMSA. Loading of equal amounts of protein was verified by comparison with the binding activity of the repressor molecule recombination signal sequence-binding protein J (RBP-J) (36). Similar experiments were performed with human embryonic kidney (HEK) 293T, MDA-MB231, and HeLa cells. F, WS inhibits TNF-induced p65 nuclear translocation. L929 cells were left untreated or treated with 63 g/ml WS for 1 h and then stimulated with (2000 IU/ml) for the indicated time periods. Nuclear and cytoplasmic extracts were prepared, and the p65 protein levels were measured by Western blot analysis. dation of phosphorylated IB␣. In the presence of MG132 or PSI there was distinct phosphorylation upon TNF induction, whereas IB␣ phosphorylation was abolished in the presence of WS (Fig. 2C). Thus, WS inhibits the degradation of IB␣ most probably through the inhibition of its phosphorylation.
WS Inhibits TNF-induced IKK␤ Activity Coinciding with Its Hyperphosphorylation-In response to a multitude of factors, such as inflammatory cytokines, bacterial products, viruses, and irradiation, the IKK complex is activated, leading to IKK␤dependent phosphorylation of IB␣ (41)(42)(43). Because IB␣ phosphorylation and degradation are inhibited by WS, we assessed whether WS extract affects the TNF-induced IKK kinase activity. As shown in Fig. 3A, the IKK activity in the TNF-stimulated cells was totally inhibited when cells were pretreated with WS, providing conclusive evidence that the main inhibitory effect of WS is through the inhibition of the IKK complex.
To unravel whether WS directly blocks IKK␤ kinase activity or rather abrogates upstream signaling cascades toward IKK␤ Ser-181 phosphorylation in the activation loop, we per-formed phosphospecific Western blot analysis. Surprisingly (see Fig. 3B), pretreatment with WS induced hyperphosphorylation of IKK␤ at Ser-181. Because the antibody recognizes the phosphorylation of both IKK␣ Ser-180 and IKK␤ Ser-181, specificity of the Ser(P)-181 IKK␤ signal was confirmed in knock-out cells using IKK␣and IKK␤-null mouse embryonic fibroblast cells. As shown in Fig. 3C, cells that no longer express IKK␣ exhibited normal IKK␤ activation in response to TNF, and pretreatment with WS equally induced the sustained phosphorylation of IKK␤ as apparent in wild type cells. However, and as expected, the signal was completely lost in cells that no longer express IKK␤.
WS Induces the Phosphorylation of IKK␤ through the MEK/ERK Pathway-Because MAP3K, Tpl2/ ERK, MEK/ERK/p90RSK, and TAK1 all have been described to act upstream of IKK (44 -48), we further investigated which pathway may be involved in WS-dependent IKK␤ hyperphosphorylation.
To address this question, L929sA cells were starved for 48 h in serumfree medium and treated with TNF alone or following a pretreatment with WS for 1 h before addition of TNF for different time periods. Pretreatment with WS was found to increase the basal and TNF-induced phosphorylation of p38, ERK, and JNK kinases (Fig. 4A). To further investigate potential cross-talk of MAPK pathways with sustained phosphorylation of IKK␤ in the presence of WS, L929sA cells were treated with TNF, WS, or their combination in the presence of SB20538, U0126, or SP600125, which are p38, MEK/ERK, and JNK inhibitors, respectively. In the presence of the MEK1 inhibitor the sustained phosphorylation of IKK␤, induced by WS, was strongly reversed (Fig. 4B). Similarly treatment with the MEK1 inhibitor also reversed WSinduced phosphorylation of ERK activity (Fig. 4B). In contrast, no significant inhibition of WS-induced IKK␤ hyperphosphorylation could be observed in the presence of the p38 and JNK inhibitors used (Fig. 4C). From these results we can conclude that WS-dependent IKK␤ hyperphospho-rylationstronglydependsonMEK1signaling.WhetherMEK1dependent IKK␤ hyperphosphorylation by WS is responsible for inhibition of IKK␤ kinase activity was further evaluated by IKK␤ kinase assays. As shown in Fig. 4D, treatment with U0126 did not reverse IKK␤ kinase inhibition by WS, although IKK␤ hyperphosphorylation was abrogated as  Fig. 1E was repeated, but L929sA cells were incubated with 100 g/ml cycloheximide (CHX) for 1 h before incubating with 63 g/ml WS for another 1 h and then stimulated with 2000 IU/ml TNF for the indicated time. Whole-cell lysates were analyzed for NFB/DNA binding by EMSA and for IB␣ levels by SDS-polyacrylamide gel electrophoresis and Western blotting, respectively. C, WS blocks TNF-stimulated phosphorylation of IB␣. L929sA cells were untreated or treated with a 10 M concentration of the protease inhibitors MG132 or PSI for 1 h, with or without a subsequent 1-h treatment with 63 g/ml WS, and then stimulated with 2000 IU/ml TNF for 30 min. Whole-cell extracts were prepared and analyzed by Western blot analysis using an anti-P-IB␣ antibody. RBP-J, recombination signal sequencebinding protein J.
shown above. Altogether this suggests that WS-induced IKK␤ hyperphosphorylation and kinase inhibition can be functionally uncoupled.
Withaferin A Is the Major Active Component of W. somnifera, Which Is Responsible for NFB Inhibition-Next we subfractionated the WS leaf extract by solid phase extraction using C 18 cartridges (supplemental Fig. 1A). Different fractions were evaluated for their potency of inhibiting NFB in a reporter gene assay (supplemental Fig. 1B). In the most active fraction, two peaks were identified. Based on retention times, UV spectra, and spiking with authentic standards (Chromadex), these peaks were characterized as the steroidal lactones WA (Fig. 5A) and withanolide A. The identities of these peaks were confirmed after isolation by semipreparative HPLC by mass spectrometry, 1 H NMR, and 13 C NMR. The purity as estimated by HPLC was Ն98%. Interestingly 12-deoxywithastramonolide (12DW), which has been demonstrated in WS, seems to be absent in the Palestinian chemotype of WS (9 -11).
Next dose-response curves for both the isolated WS compounds and 12DW in the NFB reporter gene assays were elab-orated. Because WA belongs to the class of steroidal lactones, we also included the glucocorticoid DEX and the sesquiterpene lactone helenalin (49) as reference compounds for NFB inhibition. As shown in Fig. 5B, it is apparent that WA with an IC 50 of 250 nM is the major and most active NFB-inhibiting compound in WS, whereas both WdA (50 M) and 12DW (90 M) required 200 -400-fold higher concentrations to elicit similar NFB repression. Furthermore NFB reporter gene assays confirmed that essentially all NFB-repressing activity is concentrated in the WA peak fraction as no repression could be observed in pre-or postpeak fractions as prepared by semipreparative HPLC (supplemental Fig. 1B). Interestingly in our assay system withaferin A revealed a stronger potency than the sesquiterpene lactone helenalin (IC 50 , 2.5 M) and a stronger efficacy than glucocorticoids. Although glucocorticoids revealed a lower IC 50 (80 nM) than WA, their maximal efficacy reached only 50% NFB inhibition as compared with 95% with WA. The strong differences in NFB inhibition between different WS constituents are remarkable and illustrate high specificity of the effects (Fig. 5B). With respect to structure function analysis, important roles have been attributed to the double bond (C-2ϭC-3) configuration and the C-26 lactone moiety for its biological activities (12,50), but all tested WS-derived reference compounds share these features. Thus, the presence of a C-5-C-6 epoxide and the absence of a C-20 hydroxyl seem critical for optimal activity.
Upon quantifying the amount of withaferin A present in the Palestinian WS leaf extract and NFB repression levels obtained with pure withaferin A, it was observed that almost 100% of the NFB-inhibiting activity in the WS extract originated from withaferin A (Fig. 5C). Finally in analogy with effects of WS extract on IKK␤ regulation, we measured the effects of single WS constituents on IKK␤ phosphorylation and IB␣ degradation. Interestingly strong IKK␤ Ser-181 hyperphosphorylation by WA again coincided with lack of IB␣ degradation, whereas similar concentrations of WdA and 12DW showed much weaker IKK␤ hyperphosphorylation and failed to significantly inhibit IB degradation (Fig. 5D). This demonstrates that IKK␤ hyperphosphorylation and inhibition of IKK␤-dependent IB␣ degradation are similarly mediated by the single constituent WA in the WS extract.
Reducing Agents Reverse WA-mediated IKK␤ Hyperphosphorylation and Suppression of TNF-induced NFB Activation-Of special note, it has been suggested that withaferin A may be involved in Michael addition thioalkylation reactions through either its epoxide or its lactone ring (51,52). Interestingly lactones have been described to directly suppress IKK␤ kinase activity upon attack of Cys-179 in the kinase domain activation loop or of Cys-662/716, which affects IKK␣/␤ complex formation (31,(53)(54)(55)(56) independently of IKK␤ phosphorylation regulation. Furthermore considering the observed cross-talk of MEK/ERK signaling with IKK␤ hyperphosphorylation, resorcylic acid lactones were described to target mitogen-activated protein kinase pathways at four levels: mitogen receptors, mitogen-activated protein kinase kinases MEK1/2, and ERK1/2, and certain substrates downstream of ERK (57). Also dual specificity tyrosine phosphatases were found to be susceptible to thioalkylation due to a conserved "XHCXXGXSRS" motif in the  Total cell extracts were assayed by Western blot analysis using anti-P-p38, anti-P-JNK (p46/p54), and anti-P-ERK (p42/p44) antibodies. B, MEK1/ERK inhibitor (U0126) suppresses WS-induced hyperphosphorylation of ERK and IKK␤. L929sA cells were left untreated or were treated with 10 M U0126 for 1 h followed by treatment with 63 g/ml WS for 1 h and then stimulated with 2000 IU/ml TNF for the indicated times. Total cell lysates were assayed by Western blot analysis using anti-P-IKK␣/␤ (Ser-180/Ser-181) or anti-P-ERK (p42/p44) antibodies. C, L929sA cells were left untreated (untr) or were pretreated with U0126, SB203580, or SP600125 inhibitors for 1 h followed by WS treatment (63 g/ml) for 1 h with or without subsequent cell stimulation with TNF (2000 IU/ml) for 5, 10, 20, or 30 min. Total cell lysates were assayed by Western blot analysis using anti-P-IKK␣/␤ (Ser-180/Ser-181), and specific signal intensities were subsequently quantified by free ImageJ software (National Institutes of Health). The P-IKK signal intensities obtained with the various treatments during 5-30-min TNF treatment are represented as dots in a column scatter plot. Statistical significant differences between the various treatments were calculated through the analysis of variance. Pairwise comparison of selected pairs of columns revealed the following significances: no significance (ns; p Ͼ 0.05), p Ͻ 0.05 (*), p Ͻ 0.01 (**), and p Ͻ 0.001 (***). D, L929sA cells were left untreated or were treated with 10 M U0126 for 1 h followed with 63 g/ml WS for 1 h and then stimulated with 2000 IU/ml TNF for 20 min after which the IKK complex was immunoprecipitated with anti-IKK␥ antibody. Subsequently IKK␤ kinase activity present in the immunoprecipitated IKK complex was determined by in vitro kinase assay using GST-IB␣ as a substrate and an excess of ATP. The final yield of phosphorylated IB␣ in the reaction mixture was revealed by Western blot analysis with a P-IB␣ antibody.
catalytic domain, which may result in sustained MEK/ERK signaling (58 -63). Finally NFB p65 DNA binding itself was demonstrated to be sensitive to Cys-38 alkylation too (31,49,64). Therefore, we further investigated whether WA mediates part of its NFB-inhibitory effects by alkylation of thiol-sensitive redox pathways. Therefore, L929sA cells were co-treated with WA and DTT for 1 h, and then TNF-induced NFB activation was examined. We found that DTT had no effect on TNF-induced NFB activation but that it significantly reversed the inhibitory effects of WA on NFB activation (Fig. 6A). Furthermore DTT also prevented WA-dependent IKK␤ hyperphosphorylation (Fig. 6B). This suggests that WA may target various cysteine residues of multiple kinases/phosphatases (25,(65)(66)(67)(68)(69)(70)(71)(72) that affect the phosphorylation status of p38, MEK/ERK, JNK, and IKK␤, ultimately resulting in inhibition of IKK␤ kinase activity upon attack of critical cysteine residues involved in kinase activation and/or complex formation.
WA Displays Anti-inflammatory Properties in Vivo-Because NFB is a critical player in the inflammatory signaling pathway and because WA inhibits NFB activation, we further investigated the possible anti-inflammatory effects of WA in vivo using the zymosan-induced inflamed paw model. Mice are injected subcutaneously with zymosan in the footpad, and  (Hel). B, cells, stably transfected with p(IL-6B)350hu.IL6P-lucϩ, were left untreated or were pretreated with various concentrations of WA, WdA, 12DW, helenalin (HEL), or DEX for 1 h and then stimulated with 2000 IU/ml TNF for 6 h. Lysates were prepared for quantification of reporter gene expression levels and normalization for protein concentration. C, L929sA cells, stably transfected with p(IL-6B) 3 50hu.IL6P-lucϩ, were left untreated or were pretreated for 1 h either with crude extract of WS or with equimolar amounts of WS compounds (Cpds) as can be measured in the Palestinian WS leaf extract followed by stimulation with 2000 IU/ml TNF for 6 h. Lysates were prepared for quantification of reporter gene expression levels and normalization for protein concentration. D, WA, but not WdA or 12DW, elicits hyperphosphorylation of IKK␤. L929sA cells were left untreated or were pretreated for 1 h with a 2.8 g/ml concentration of either WA, WdA, or 12DW and then stimulated with 2000 IU/ml TNF for the indicated time points. Whole-cell extracts were prepared, and Western blot analysis was performed with anti-P-IKK␣/␤ and anti-IB␣. swelling is determined at several time points after the injection. Pretreatment with WA (intraperitoneal) was performed to evaluate anti-inflammatory effects, whereas DEX (intraperitoneal) was chosen as a positive control. Results in Fig. 7A show the mean of the differences between the zymosan-treated and untreated footpad for all the three groups. Both DEX and WA pretreatment showed clear anti-inflammatory activities. At the first time intervals DEX and WA were more equally active, whereas they become less equally active, and a significant difference can be observed at later time points (48 -72 h). These results show that WA displays an anti-inflammatory activity that is more persistent than that of DEX. Moreover IL-6 protein levels were measured at an early stage, and as can be observed in Fig. 7B, pretreatment with DEX or with WA significantly reduced the levels of IL-6 compared with Me 2 SO, which is in line with our results obtained in vitro.

DISCUSSION
The transcription factor NFB regulates the expression of cytokines, chemokines, adhesion factors, and inducible proinflammatory receptors (73). The abnormal activation of NFB has been established for a series of inflammatory diseases and cancer (74,75). Thus, NFB is an ideal target for anticancer and anti-inflammatory drug development. In fact, several anti-inflammatory agents such as aspirin, sulfasalazine, and steroids have been suggested to act at least partially by inhibiting NFB activation (73,76).
Use of W. somnifera as a traditional remedy for several illnesses is widespread in many countries. Today this plant is used as a constituent in more than 100 herbal preparations (77). However, the molecular mechanism of its immunomodulatory action is poorly understood. In this report we demonstrate that pure WA or WA-containing extract of WS completely suppressed NFB activation induced by inflammatory agents irrespective of the cell type, but not AP1 activity. Indeed we showed that pure WA and WS extract block IKK␤ kinase activity, which in turn abrogates IB␣ phosphorylation and degradation, p65 translocation, NFB/DNA binding, and subsequent NFB-driven gene expression, which finally results in the down-regulation of various NFB-regulated gene products induced by TNF.
Although Mohan et al. (78) previously reported that WA might inhibit NFB activity by impaired IB degradation and ubiquitination, we clearly show that WS and WA inhibit NFB activation by directly suppressing IKK␤ kinase activity, which blocks IB␣ phosphorylation prior to its ubiquitination and degradation. Our results show that, unlike the previously described IKK␤ inhibitors (79 -81), WA and WA-containing preparations of WS inhibit the activity of IKK␤ kinase via a thioalkylation-sensitive redox mechanism concomitant with inducing Ser-181 phosphorylation in a MEK1/ERK1-dependent way. This is surprising as IKK␤ mutants (S176E/S181E) that mimic constitutive IKK␤ phosphorylation are constitutively active (14,43); however, in the case of WA, we cannot exclude cross-talk of Ser-181 with other less well defined COOH-terminal IKK␤ phosphorylations, which have been described to decrease general IKK activity (23). Furthermore various critical cysteine residues have already been identified in IKK␤ that interfere with kinase activity and complex conformation and may overrule phosphorylation-dependent control of the kinase activity (31). Surprisingly we were unable to completely block WA-dependent NFB inhibition by either IKK␤ C179A or C662A/C716A mutants (data not shown), which suggests that WA may target multiple IKK␤ cysteine residues to mediate its full inhibitory activity. Furthermore independently of IKK␤ regulation, WA/WS may elicit (hyper)phosphorylation of other kinases too that depend on cysteine-sensitive regulation of alternative phosphatase targets (66 -72).
In summary, these experiments demonstrate that pure WA or WA-containing preparations of WS potently inhibit TNFinduced NFB activation via inhibition of IKK␤ kinase activity in the nanomolar range. We further demonstrate that WA FIGURE 7. A, mice were injected subcutaneously with zymosan in the footpad, and swelling was determined at 6, 12, 24, and 48 h after the injection. Mice were preinjected intraperitoneally with WA, DEX, or Me 2 SO (DMSO) 30 min before zymosan injection. B, effect of WA or DEX on blood IL-6 protein levels. Blood IL-6 protein concentration was determined at 6 h after zymosan injection from the same mice as used in Fig. 6A. The dashed line indicates background IL-6 levels of untreated mice. The statistical significance between the treated groups and the control group was calculated through the analysis of variance. p values less than 0.05 were considered as indicative of significance. ***, p ϭ 0.0004; **, p ϭ 0.001. clearly displays anti-inflammatory characteristics not only in vitro but also in vivo, i.e. in an acute inflammatory mouse model. Taken together, the results of this study extend our understanding of the molecular mechanisms underlying the anti-inflammatory and the antitumor activity of WS plant extracts that are used in traditional medicine and thus provide scientific support for the use of WA preparations of WS as a folk remedy for the treatment of inflammation. Finally WA is a promising lead compound for the design of potent IKK␤ inhibitors for anti-inflammatory, antitumoral, and/or chemopreventive applications.