NF-κB is a transcription factor that activates numerous genes that regulate several immune and inflammatory processes, including RA (Baeuerle and Henkel, 1994). NF-κB generally exists as a dimer in the cytosol bound to distinct inhibitory IκB subunits. NF-κB activation is primarily mediated by phosphorylation and rapid degradation of IκB, followed by translocation of NF-κB to the nucleus, where it activates transcription of specific genes (Karin, 1999). Examples of genes dependent on the activation of NF-κB include cytokines such as tumor necrosis factor α (TNFα), interleukin (IL)-6, IL-8, and IL-1β, chemokines such as MCP-1, RANTES, and matrix metalloproteinases (MMPs) such as MMP1, MMP3, and MMP13. NF-κB activity is tightly controlled by the IκB kinase complex, consisting of IκB kinase (IKK) α, IKKβ, and IKKγ. IKKβ is essential for the inflammatory cytokine-induced activation of NF-κB (Mercurio et al., 1997; Yamamoto and Gaynor, 2004). Because IKK-catalyzed phosphorylation of IκB proteins is an essential step in the signal-induced activation of NF-κB, targeting IKKβ represents an opportunity for developing novel therapeutics for inflammatory disease indications such as rheumatoid arthritis, asthma, and many others.

Rheumatoid arthritis is a chronic destructive disease of the joints, characterized by inflammation, synovial hyperplasia, and abnormal cellular and humoral immune responses. Several cell types have been implicated as major contributors in the pathophysiological process of RA (Gravallese, 2002), including fibroblast-like synoviocytes (FLS), macrophages, T cells, B cells, osteoclasts, chondrocytes, dendritic cells, and mast cells. Specific cytokines, chemokines, and tissue-destructive enzymes have been identified as key players in the pathologic process of RA. Multiple lines of evidence suggest that NF-κB plays a key role in regulating the inflammatory process in these cell types. Therefore, identification of selective IKKβ inhibitors as potential therapeutics has received considerable interest.

Here, we report that ML120B is a potent and selective IKKβ inhibitor. We have examined cellular mechanisms and functional effects of ML120B in RA-relevant cell systems. Our results demonstrate that ML120B specifically blocked the NF-κB signaling pathway. ML120B concentration-dependently blocked LPS-, TNFα-, or IL-1β-stimulated cytokine production in inflammatory cells, including human peripheral blood mononuclear cells (PBMC), human fibroblast-like synoviocytes (HFLS), and human mast cells (MC). ML120B also blocked IL-1β-stimulated collagenases (MMP1, MMP13), stromelysin (MMP3), and prostaglandin E₂ (PGE₂) production in the human chondrocyte cell line SW1353. Taken together, these results suggest a potential use of this class of compounds as therapeutic agents in the treatment of RA.

Materials and Methods

Reagents. Nonidet P-40, Triton X-100, Tween 20, BSA, LPS (Escherichia coli O26:B6), DMSO, toluidine blue, PMA, ionomycin, and PGN were purchased from Sigma-Aldrich (St. Louis, MO). αCD3 and αCD28 were obtained from BD Biosciences (San Jose, CA). rhTNFα, rhIL-1β, IL-4, and IL-6 were obtained from R&D Systems (Minneapolis, MN). Phospho-IκBα (Ser32/36) (12C2) monoclonal antibody was purchased from Cell Signaling Technologies (Beverly, MA) and labeled with W8044 europium chelate by PerkinElmer Life and Analytical Sciences (Boston, MA). Streptavidin Alexa 647 and Blasticidin were purchased from Invitrogen (Carlsbad, CA). GST-IκBα (5–55) was produced in E. coli as described previously (Chen et al., 1996; Lee et al., 1997, 1998) and labeled with biotin using PEO-Maleimide Activated Biotin from Pierce (Rockford, IL) as per the manufacturer's instructions. Titertube deep-well 96-well plates were purchased from Bio-Rad (Hercules, CA). PE-conjugated anti-human TLR antibodies were from eBioscience (San Diego, CA).

Compounds. β-Carboline compounds ML120B and PS1145 were synthesized at Millennium Pharmaceuticals, Inc. (Cambridge, MA). In all cell-based assays, compounds were preincubated with cells for 30 min or 1 h before stimulation. The compound was dissolved in DMSO. The final concentration of DMSO in all the assays was 0.5%.

Cell Lines. The 293 NF-κB-Luc and 293 AP-1-Luc reporter stable cell lines were generated at Millennium Pharmaceuticals, Inc. Briefly, the construct of NF-κB-TA-Luc or AP-1-Luc engineered with a Blasticidin selection gene was used to stably transfect the 293 cell line. Positive clones were selected using Blasticidin and evaluated by their responsiveness to TNFα or PMA/ionomycin stimulation. HFLS were obtained from Cell Applications, Inc. (San Diego, CA). Human chondrosarcoma cells SW1353 and HeLa cells were purchased from American Type Culture Collection (Manassas, VA).

Human Plasma. Normal human blood was obtained through Millennium Pharmaceuticals, Inc.'s blood donation program. Normal human plasma was pooled from multiple donors, and aliquots were stored at –80°C.

Preparation of IKK Complex. The IKK complex was purified and activated as described previously (Chen et al., 1996; Lee et al., 1997, 1998) with minor modifications. HeLa cell S100 lysates were precipitated in 30% ammonium sulfate and resuspended in buffer A (50 mM Hepes, pH 7.5, 100 mM NaCl, 10 mM 2-glycerophosphate, 2 mM dithiothreitol, and 0.5 mM EDTA) and were passed over a Superose 6 gel filtration column (GE Healthcare Bio-Sciences AB, Uppsala, Sweden) pre-equilibrated in buffer A. Fractions containing IKK activity were pooled and activated with 100 nM recombinant MEKK1 at 37^°C for 45 min in buffer A supplemented with 10 mM MgCl₂, 250 μM ATP, and 2 μM Microcystin-LR (Sigma-Aldrich).

IKK Complex Enzyme Assay. Phosphorylation of GST-IκBα (5–55) was measured by incubating test samples in 3% DMSO with activated IKK complex as described above: 300 nM GST-IκBα (5–55) and varying ATP concentrations in 50 mM Hepes, pH 7.5, 10 mM 2-glycerophosphate, 10 mM MgCl₂, 5 mM dithiothreitol, and 0.1% BSA were incubated in a reaction volume of 30 μl for 1 h at room temperature. Reactions were terminated with the addition of 10 μlof 250 mM EDTA, followed by the addition of 40 μl of detection mixture containing 2 nM europium-labeled anti-phospho IκB and 50 nM streptavidin Alexa 647 in 50 mM Hepes, pH 7.5, 0.1% BSA, and 0.01% Tween 20. After 1 h, plates were read on a Wallac Victor plate reader (PerkinElmer Life and Analytical Sciences).

IKK Complex Reversibility Assay. IKK complex was incubated for 1 h with 10 μM ML120B. After incubation, samples were passed through a Zeba gel filtration column (Pierce) according to manufacturer's instructions. Samples were assayed as described above but at an ATP concentration of 1 mM.

IKKβ Kinase Assays. Recombinant IKKβ (Upstate Biotechnology, Lake Placid, NY) was assayed under identical conditions as the IKK complex assay with the following modifications. IKKβ (1 nM) was substituted for activated IKK complex, and the ATP concentration was adjusted to 1 μM.

Kinase Selectivity Assays. All kinase selectivity assays were performed against recombinant kinases at or below the K_m of ATP. Assays against IKKα were performed through Upstate Kinase Profiler (Upstate Biotechnology). Assays against PKCα, PKCδ, PKCθ, CSK, Src, and Lyn were performed in a time-resolved fluorescence format. All other assays were performed in Flashplate assays (PerkinElmer Life and Analytical Sciences).

Western Blot. HeLa cells or human PBMCs were preincubated with ML120B or DMSO for 1 h before the stimulation with TNFα (50 ng/ml) for 5 or 20 min. Cells were lysed using whole cell lysis buffer, and protein concentrations were determined using a bicinchoninic acid kit (Pierce). Twenty micrograms of cell lysate was loaded into each well of a 4 to 12% NuPAGE Bis-Tris gradient gel (Invitrogen). The blot was blocked in 5% milk/phosphate-buffered saline/Tween 20 and incubated with polyclonal anti-phosphorylated IκB antibody or polyclonal anti-IκB antibody (Cell Signaling) and, subsequently, anti-rabbit secondary antibody (Cell Signaling). ECL Plus (GE Healthcare, Little Chalfont, Buckinghamshire, UK) was used for detection.

HeLa Cells in NF-κB Nuclear Translocation Assay. HeLa cells were maintained in Dulbecco's modified Eagle's media with 10% FBS, 2 mM glutamine, and 1% penicillin/streptomycin solution. The effect of ML120B on the nuclear translocation of p65, in response to stimulus, was examined by using high-content screening technology from Universal Imaging Corporation (Downingtown, PA). Briefly, HeLa cells were seeded into a 96-well black, clear-bottom Packard ViewPlate at 5000 cells/per well the day before the experiment. ML120B was serially diluted 1:1 in DMSO with final concentrations ranging from 20 to 0.04 μM. After a 1-h preincubation with ML120B, cells were stimulated with TNFα (10 ng/ml) for 30 min, fixed with 4% paraformaldehyde solution, and permeabilized with 0.5% Triton X-100. After being washed with 0.5% phosphate-buffered saline, cells were blocked with 0.5% blocking reagent (Roche, Indianapolis, IN). Furthermore, cells were stained using “Cellomics Hit Kit” (Cellomics, Pittsburgh, PA) for the detection of NF-κB activation following the kit manual. Images were acquired with Discovery-1 system, and nuclear translocation data were analyzed by MetaMorph Software from Universal Imaging Corporation.

NF-κB DELFIA Assay. To quantitatively measure NF-κB binding activity upon activation, we developed an NF-κB DELFIA assay based on the Trans-AM NF-κB assay kit (Active Motif, Carlsbad, CA). Cells were plated in a 96-well plate the day before the experiment. After preincubation with serially diluted ML120B at 37°C for 1 h, cells were stimulated with TNFα (10 ng/ml) at 37°C for 30 min. Whole cell extract was prepared from the cell pellet. Twenty micrograms of whole cell extract was added in duplicate into a 96-well Trans-AM NF-κB plate containing the immobilized NF-κB consensus site (5′-GGGACTTTCC-3′). After a 1-h incubation with the primary antibody, wells were washed with DELFIA wash buffer (PerkinElmer Life and Analytical Sciences) and incubated with Eu-N1-labeled anti-mouse IgG (PerkinElmer Life and Analytical Sciences) for DELFIA readout. The NF-κB activation was quantified as europium counts of time-resolved fluorometry on a Wallac Victor plate reader.

NF-κB or AP-1 Luciferase Reporter Assay. 293/NF-κB-Luc and 293/AP-1-Luc cells were maintained in DMEM (Invitrogen) with 10% FBS (Hyclone, Logan, UT), 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM l-glutamine, and 5 μg/ml Blasticidin (Invitrogen). Cells were seeded at a density of 0.05 × 10⁶ cells per well in black poly-d-lysine 96-well plates (Discovery Labware Inc., Bedford, MA) the day before the assay. ML120B or PS1145 was serially diluted 1:1 in DMSO to produce concentrations ranging from 20 to 0.04 μM. Medium was changed to AIM-V (Invitrogen) for the assay. Cells were preincubated with IKKβ inhibitor (ML120B or PS1145) for 1 h at 5% CO₂ at 37°C. For the NF-κB assay, cells were stimulated with 0.5 ng/ml TNFα (R&D Systems) or 10 pg/ml PMA and 2 μM ionomycin for 3 h. For the AP-1 assay, cells were stimulated with 10 pg/ml PMA and 2 μM ionomycin for 3 h. Medium was removed, and cells were lysed in 1× passive lysis buffer (GE Healthcare). Lysate (20 μl) was used in accordance with luciferase assay kit (Promega, Madison, WI). The assay was run in triplicate, and data were analyzed using XLfit software (IDBS Ltd., Guildford, UK).

Transient Transfection Assays Using Luciferase Reporter Panel. The following PathDetect cis-reporter luciferase plasmids were purchased from Stratagene (La Jolla, CA): NF-κB, AP-1, cAMP response element, serum response element, interferon-stimulated response element, and nuclear factor of activated T cells. The IL-8-luciferase reporter was obtained from the A. Casola laboratory and contains the promoter region of IL-8 (–162 to +44) upstream of the luciferase gene (Casola et al., 2000) The ELAM-luciferase reporter was generated in-house and contains the promoter region of the human E-Selectin gene from –730 to +52 cloned into a modified TATA-less version of pMCS-Luc (Stratagene). Transient transfection and assay of ML120B and PS1145 on the luciferase reporter panel was carried out as follows: 293T cells were plated at a density of 17,000 cells per well in 100 μl of culture medium (DMEM, 10% FBS, 1% penicillin/streptomycin, and 1% l-glutamine) in Biocoat poly-d-lysine 96-well plates (Becton Dickinson, San Jose, CA). The following day, cells were transfected with a cocktail containing 100 ng of DNA consisting of 50 ng of luciferase reporter DNA, 5 ng of internal control pTK-Renilla (Promega), and 45 ng of empty vector using Fugene transfection reagent (Roche) following the manufacturer's instructions. After overnight incubation, fresh medium containing 0.1% FBS was added to the plates, and cells were incubated overnight. Cells were preincubated with IKKβ inhibitor compounds (ML120B or PS1145 prepared in DMSO) added at a dose range from 10 nM to 31.6 μM (3-fold serial dilution), and 30 min later, various stimulants for the reporters were added to induce the different pathways being monitored: 50 ng/ml PMA (Sigma-Aldrich), 1 μg/ml ionomycin (Sigma-Aldrich), 20 μM forskolin (Sigma-Aldrich), 1000 U/ml interferon-β (R&D Systems), and 50 ng/ml TNFα (R&D Systems). The cells were incubated for a further 6.5 h, after which cells were harvested and luciferase values were measured using the dual luciferase assay kit (Promega). Each individual assay was performed in quadruplicate per experiment, and the complete transfection procedure for the reporter panel was performed a minimum of two times. The data were analyzed using XLfit software.

Human PBMC Assay. Heparinized human whole blood was obtained from normal donors. PBMCs were separated on a Ficoll-Paque Plus (GE Healthcare) gradient. Compound was serially diluted 1:1 in DMSO with final concentrations ranging from 20 to 0.04 μM. Cells at a density of 4 × 10⁵ per well were seeded in a 96-well plate and incubated at 5% CO₂ at 37°C. Cells were preincubated with compound for 1 h and then stimulated with either LPS (100 ng/ml) or αCD3 (0.25 μg/ml) and αCD28 (0.25 μg/ml) for 5 h. Supernatant was collected for cytokine analysis.

Cytokine Analysis. Quantitative measurement of cytokines in culture supernatant or plasma was performed by Pierce Biotechnology, Inc. (Woburn, MA) using the SearchLight Human Cytokine Array, a multiplexed sandwich enzyme-linked immunosorbent assay (ELISA).

Cytotoxicity Lactate Dehydrogenase Release Assay. Compound cytotoxicity was monitored by the measurement of lactate dehydrogenase (LDH) activity released from the cytosol of damaged cells into the supernatant. The supernatant from human PBMC assays was tested for LDH release following the vendor's kit manual (Roche). Triton X-100-treated cells (2%) are used to determine the maximum releasable LDH activity in the cells. Cells treated with 0.5% DMSO were used to determine the basal level of LDH activity.

HFLS Assay. Cryopreserved HFLS cells were thawed and cultured in synoviocyte growth culture medium (Cell Applications, Inc., San Diego, CA). After two to three rounds of subculture, 1 × 10⁵ cells per well were plated in black poly-d-lysine 96-well plates (Discovery Labware, Inc.) 1 day before the experiment. ML120B was serially diluted 1:1 with DMSO and preincubated with the cells at 5% CO₂ at 37°C for 1 h. For cytokine studies, the cells were stimulated with 10 ng/ml human TNFα or 10 ng/ml human IL-1β and cultured for 16 h. The supernatant was harvested and analyzed for RANTES, MCP-1, IL-6, and IL-8 production (Pierce Biotechnology, Inc.). To measure the effects of ML120B on NF-κB activation, the cells were stimulated with either IL-1β (10 ng/ml) or TNFα (10 ng/ml) for 30 min. Whole cell extract was used to measure of NF-κB activation by the NF-κB DELFIA assay as described above, except that compound was serially diluted 1:1 in DMSO with final concentrations ranging from 50 to 0.1 μM.

Human Cord Blood-Derived Mast Cells Assay. Human PBMCs were separated by Ficoll (GE Healthcare) gradient from human cord blood (Cambrex Bio Science, Walkersville, Maryland) and then cultured in mast cell culture medium (RPMI 1640 containing 10% fetal bovine serum 2 mM l-glutamine, 0.1 mM nonessential amino acid, 100 U/ml penicillin, 100 mg/ml streptomycin, 2 μg/ml gentamycin, and 0.2 μM 2-mercaptoethanol) in the presence of 100 ng/ml human stem cell factor (R&D System), 10 ng/ml human IL-6, and 10 ng/ml human IL-10. Nonadherent cells were transferred to a new flask containing mast cell culture medium with fresh cytokines every week. To assess the maturity of the mast cells, starting from week 6, an aliquot of cells was stained with toluidine blue, and the expression of c-Kit (BD Bioscience, San Jose, CA) was analyzed by FACS. When >95% of the cells were confirmed to have differentiated into mature mast cells, the cells were processed for pharmacology studies. The cells were preincubated with serially diluted ML120B at 5% CO₂ at 37°C for 1 h and then stimulated with 10 ng/ml LPS or 100 μg/ml PGN overnight. The supernatant was collected and subjected to cytokine profiling analysis (Pierce Biotechnology, Inc.).

Expression of Toll-Like Receptors on Human Cord Blood-Derived Mast Cells. Mast cells (0.5 × 10⁵) were washed with ice-cold 1× phosphate-buffered saline with 1% BSA. Cells were stained with PE-conjugated anti-human TLR 2, TLR 4, TLR 3, or TLR 9 antibodies, respectively. PE-conjugated mouse IgG or PE-conjugated rat IgG2α was used as epitope controls. The expression of TLRs on mast cell surface was analyzed by FACS.

Adenovirus Transduction of IKKβ wt, IKKβ DN, and IκBSR in Human Mast Cells. The recombinant adenovirus constructs carrying the IKKβ wild type (IKKβ wt), IKKβ dominant-negative (IKKβ DN K44M), and IκB super-repressor (IκB SR) were kindly provided by Philip A. Barker (McGill University, Montreal, QC, Canada; Bhakar et al., 2002). Transduction efficiency was monitored by GFP expression. Mast cells were seeded in 24-well culture plate at 2 × 10⁵ per well. Adenovirus titer m.o.i. was optimized in a pilot experiment so that all the wells would have similar transduction efficiency. After a 24-h infection with corresponding adenoviruses, cells were stimulated with either purified LPS at 50 ng/ml (gift from Dr. Kate Fitzgerald, University of Massachusetts Medical School, Worcester, MA) or PGN at 100 μg/ml for 16 h. The supernatant was collected and subjected to cytokine profiling analysis (Pierce Biotechnology, Inc.).

Human Chondrocyte SW1353 Assay. The human chondrosarcoma cell line SW1353 was cultured in DMEM with 2 mM Glutamax (Invitrogen), 100 U/ml penicillin, 100 mg/ml streptomycin (Invitrogen), and 10% fetal bovine serum (Hyclone, Logan, UT). Cells per well (0.05 × 10⁶) were seeded into a 96-well poly-d-lysine plate (Discovery Labware, Inc.) 1 day before the experiment. The culture medium was changed to fresh AIM-V medium containing 2% fetal calf serum on the day of the experiment. Serially diluted ML120B in DMSO, ranging from 20 to 0.02 μM, was preincubated with cells for 1 h at 5% CO₂ at 37°C and then stimulated with 10 ng/ml human IL-1β (R&D System) for 24 h. The supernatant was harvested and used to measure MMP1, MMP3, MMP13, and PGE₂ production by ELISA (GE Healthcare).

Statistical Analyses. The IC₅₀ assays were run in triplicate, and XLfit software was used to fit dose response curves to the mean values. The IC₅₀ values and their 95% confidence intervals were calculated with equation 205 (Sigmoidal dose-response model). Differences between IC₅₀ estimates were assumed significant at p = 0.05 if the 95% confidence intervals of the estimates did not overlap. Where appropriate, differences in means were assessed using Student's t test in Microsoft Excel. All p values were two-tailed and were considered significant when p ≤ 0.01. Results are expressed as the mean ± S.D. of a representative experiment.

Results

Identification of ML120B as a Potent and Specific Inhibitor of IKKβ. In an earlier report, we described β-carboline PS1145 as an IKKβ inhibitor with a certain level of specificity and potency (Fig. 1A) (Castro et al., 2003). Further chemical optimization has led to the identification of ML120B as a more potent and selective IKKβ inhibitor (Fig. 1B). ML120B was determined to have an IC₅₀ of 60 nM at 50 μM ATP in a time-resolved fluorescence assay measuring the phosphorylation of IκBα at various inhibitor concentrations in an in vitro kinase assay using IKK complex purified from HeLa cells and GST-IκBα as a substrate (Fig. 2A). Kinetic analysis shows ML120B to be a competitive inhibitor of ATP, as indicated by the double-reciprocal plots of velocity versus ATP concentration at fixed inhibitor concentrations (Fig. 2B). ML120B was also shown to be a reversible inhibitor of the IKK complex (Fig. 2C). To confirm that the activity of ML120B was indeed targeting IKKβ, an in vitro kinase assay was run using recombinant IKKβ. ML120B was determined to have an IC₅₀ of 45 ± 1 nM (95% confidence interval) against recombinant IKKβ (Fig. 2D).

To examine the selectivity of ML120B, inhibitory activities of MN120B were evaluated against a panel of 30 tyrosine and serine/threonine kinases, including IKKα and IKKϵ. As shown in Table 1, no inhibition below 50 μM was observed across this panel of kinases, thus demonstrating the selectivity of ML120B for IKKβ. ML120B was also screened at 10 μM against a NovaScreen (Hanover, MD) panel of 33 receptors, transporters, and channels. ML120B showed >100-fold selectivity over this panel (data not shown). To further evaluate the selectivity of ML120B, both ML120B and PS1145 were screened against a panel of eight luciferase reporters that represent different well characterized cellular pathways (Table 2). ML120B showed an inhibitory effect only on the NF-κB reporter or reporters that contained an NF-κB element in their promoter regions (IL-8 and ELAM). These data confirm that compared with the first generation β-carboline compound PS1145, ML120B is a selective IKKβ inhibitor specific to the NF-κB pathway.

ML120B Blocks NF-κB Signaling Pathway. Upon activation by cellular stimuli, IKK phosphorylated IκBα. Phosphorylated IκBα was rapidly degraded through ubiquitin-dependent pathways, which allowed released NF-κB factor to translocate to the nucleus and activate NF-κB-mediated gene expression. To dissect how ML120B exerts its effect on the NF-κB signaling pathway, we examined the effect of the IKKβ inhibitor on TNFα-induced IκB phosphorylation and IκBα degradation in HeLa cells. TNFα stimulation causes IκBα phosphorylation within 5 to 10 min and IκBα degradation within 10 to 30 min after stimulation. ML120B blocked TNFα-stimulated IκBα phosphorylation and degradation concentration-dependently when assessed by Western blot analysis (Fig. 3). To examine the effects of ML120B on NF-κB translocation and DNA binding activity, we used high-content image analysis to measure and quantify nuclear translocation of p65 in the absence or presence of ML120B on HeLa cells. As indicated in Fig. 4A, ML120B concentration-dependently blocked TNFα-induced nuclear translocation of p65 with an IC₅₀ of 2.4 μM (Fig. 4B). Furthermore, ML120B blocked TNFα-induced NF-κB-linked luciferase reporter activity in a concentration-dependent manner with an IC₅₀ of 1.1 μM (95% confidence interval: 0.9, 1.3) (Fig. 5A). However, ML120B showed minimal inhibition of PMA/ionomycin-induced AP-1-linked luciferase activity, as shown in Fig. 5A. In contrast, the first-generation β-carboline compound PS1145 inhibited AP-1 luciferase reporter activity with an IC₅₀ of 8.8 ± 2.9 μM (95% confidence interval). For comparison, PS1145 concentration-dependently inhibited NF-κB luciferase activity with an IC₅₀ of 1.7 μM (95% confidence interval: 5.9, 11.7) (Fig. 5B). Given these confidence intervals, PS1145 was significantly more potent against NF-κB than AP-1; however, ML120B is more selective and more potent than the first generation compound PS1145.

ML120B Inhibits NF-κB-Mediated Gene Activation in RA-Relevant Inflammatory Cells. Activated macrophages are known to contribute to synovial inflammation in RA, and IKKβ has been shown to be a key regulator for macrophage function through activation of the NF-κB signaling pathway (Guha and Mackman, 2001). It is noteworthy that anti-TNFα therapy such as etanercept and infliximab has proven to be effective therapeutic agents for RA. We therefore investigated the effect of ML120B on TNFα release from human PBMCs. Human PBMCs were isolated from normal donors and exposed to LPS in the absence or presence of inhibitor. ML120B concentration-dependently inhibited LPS-stimulated TNFα production with an IC₅₀ of 3.3 μM (Fig. 6A). Consistent with the mechanism of IKKβ inhibition, ML120B also concentration-dependently blocked TNFα-induced IκBα phosphorylation in human PBMCs (Fig. 6C). T cell activation leads to IKKβ-dependent NF-κB up-regulation, which contributes to inflammatory responses in RA. To evaluate the effect of ML120B on T cell signaling pathways, we evaluated αCD3/αCD28-induced IL-2 production with ML120B treatment in human PBMCs. As indicated in Fig. 6B, ML120B concentration-dependently blocks αCD3/αCD28-induced IL-2 production with an IC₅₀ of 3.4 μM. The cytotoxicity of ML120B was monitored by measuring LDH release in the supernatant of human PBMC. We found that ML120B did not cause LDH leakage at concentrations up to 20 μM (Fig. 6D), suggesting that the effect of ML120B is not due to cellular toxicity of the inhibitor. The LDH release in ML120B-treated cells is not statistically different compared with DMSO-treated cells (p > 0.01).

In addition to macrophage and T cells, HFLS also play an important role in the pathogenesis of RA. Cytokine-induced NF-κB activation in HFLS is IKKβ-dependent (Aupperle et al., 2001; Hammaker et al., 2003). To evaluate the effect of an IKKβ inhibitor on HFLS, we tested the inhibitory effect of ML120B on IL-1β- and TNFα-induced cytokine and chemokine production. ML120B inhibits TNFα- or IL-1β-induced IL-6 and IL-8 production with IC₅₀ values of 5.7 to 14.5 μM. Interestingly, we found that ML120B concentration-dependently inhibits either TNFα- or IL-1β-induced production of chemokines (RANTES, MCP-1) and MMPs (MMP3, MMP13, and MMP1) with greater potency compared with IL-6 and IL-8 production (Fig. 7, A and B). The IC₅₀ values using either IL-1β or TNFα as stimulator are summarized in Table 3. To confirm that the inhibition was NF-κB-dependent, we measured NF-κB activity at a 30-min time point with IL-1 or TNFα stimulation in the presence or absence of ML120B. Consistent with the reduction of cytokine/chemokine production, ML120B showed a concentration-dependent inhibition of p65 binding activity measured by NF-κB DELFIA (Fig. 7C) with an IC₅₀ of 3.4 μM after IL-1β stimulation and 5.7 μM after TNFα stimulation. These new findings indicate that the chemokine system may play a more direct role in the destructive phase of RA than is currently suspected, and IKKβ inhibition may lead to a greater effect in protecting joint disruption and blocking cell infiltration to the joints. We observed differential inhibitory effects when HFLS was stimulated by TNFα or IL-1β (Table 3 and Fig. 7). Donor variation is very large, and the mechanisms for differences in response to IL-1β and TNFα stimulations remain to be elucidated. Recent studies indicated that different inflammatory stimuli induce distinct IKK profiles (Cheong et al., 2005; Covert et al., 2005; Werner et al., 2005). Therefore, different genes may be alternately regulated by the differential usage of κB site by different promoters.

The potential role of MCs in the pathophysiological process of RA has been demonstrated previously (Woolley and Tetlow, 2000; Woolley, 2003). To further investigate the function of IKKβ in TLR signaling in human MC, we investigated the effect of ML120B on mast cell function with human cord blood PBMC-derived mast cells, characterized by positive toludine blue staining and positive c-Kit expression by FACS (data not shown). In addition, we evaluated the expression pattern of TLRs in human cord blood-derived mast cells by FACS analysis. Human MCs express both TLR2 (receptor for PGN) and TLR4 (receptor for LPS) but have minimal expression of TLR3 and TLR9 (Fig. 8). Purified LPS (gift from Dr. Kate Fitzgerald) was used to avoid potential cross-signaling through other TLRs from contamination. We used adenoviral-mediated overexpression of a dominant-negative version of IKKβ (IKKβ DN) or IκB SR to elucidate the functional role of IKKβ and NF-κB in TLR signaling in human MCs. Both IKKβ DN and IκB SR can prevent LPS- or PGN-induced NF-κB-dependent cytokine production (IL-1β, TNFα, and IL-6) (Fig. 9, A and B). Consistent with our findings with the adenoviral-mediated overexpression system, ML120B showed concentration-dependent inhibition of PGN- or LPS-induced IL-1β, TNFα, and IL-6 production with IC₅₀ values ranging from 2.2 to 5.6 μM (Fig. 9, C and D). p values from t test were used to analyze the statistical differences.

Finally, chondrocyte activation is a known contributor to joint inflammation and destruction in RA, and it has been shown that IL-1-mediated chondrocyte activation leads to production of MMPs and other mediators (Fernandes et al., 2002; Abramson, 2004). We used ML120B to evaluate the role of IKKβ in IL-1β-mediated signaling in the human chondrocyte cell line SW1353. ML120B concentration-dependently inhibited IL-1β-induced MMP13 and MMP1 production with the IC₅₀ of 0.9 and 1.2 μM, respectively (Fig. 10A). In addition, IL-1β-stimulated PGE₂ production was also blocked by ML120B with an IC₅₀ of 1.7 μM (Fig. 10B).

Discussion

IKKβ has been identified as the primary kinase responsible for NF-κB activation in cytokine-stimulated cells, and thus, identification of selective IKKβ inhibitors as potential therapeutics for chronic inflammatory diseases has received considerable interest (Yamamoto and Gaynor, 2001). In this report, we describe the characterization of a potent and selective IKKβ inhibitor, ML120B. We demonstrate that ML120B is highly selective and specifically inhibits IκBα phosphorylation, NF-κB nuclear translocation, and transcription in a concentration-dependent manner. To evaluate the pharmacological effects of ML120B, we tested several cell types relevant to RA including HFLS, mast cells, and chondrocytes for inhibition of cytokine, chemokine, and metalloprotease expression. In all cases, we found that ML120B displays inhibitory effects in a concentration-dependent manner.

A chemical screening effort using the endogenous IKK complex allowed us to identify β-carboline derivatives as novel IKKβ inhibitors (Castro et al., 2003). Lead optimization led to the identification of ML120B, a potent ATP-competitive inhibitor of IKKβ. ML120B selectively inhibits only IKKβ kinase activity when screened against a panel of 30 other kinases. In addition, ML120B does not inhibit the kinase activity of the related proteins IKKα or IKKϵ. Furthermore, in a panel of cellular pathway selectivity reporter assays, ML120B only showed an inhibitory effect on the NF-κB reporter or reporters that contain an NF-κB element in their promoter region (IL-8 and ELAM). Compared with the first-generation β-carboline compound PS1145, ML120B has been shown to be a selective inhibitor specific to the NF-κB pathway. Thus, ML120B is a useful tool to evaluate the effects of IKKβ inhibition on the activation of the NF-κB pathway in a variety of cell-based systems relevant to RA.

To determine whether IKKβ activation can be blocked by ML120B in cells, we tested the inhibitory effect of ML120B on IκBα phosphorylation, on nuclear translocation of NF-κB, and on transcriptional regulation. ML120B inhibited TNFα-induced IκB phosphorylation and degradation in HeLa cells and inhibited nuclear translocation of NF-κB with an IC₅₀ of 2.4 μM. ML120B also concentration-dependently blocked TNFα-induced NF-κB DNA binding activity.

The cellular activity and selectivity of ML120B was further evaluated at the transcription level using NF-κB and AP-1 luciferase reporter assays. ML120B selectively inhibits NF-κB activity but does not inhibit AP-1 activity. It has been shown that cytokine-activated IKKβ is essential for NF-κB activation (Mercurio et al., 1997). LPS-induced cytokine production in human monocytes and T cell receptor-mediated IL-2 production in T cells are both IKKβ-dependent (O'Connell et al., 1998). As expected, ML120B inhibited both LPS-stimulated TNFα production and αCD3/αCD28-costimulated IL-2 production in human PBMCs with similar IC₅₀ values around 3 μM. Overall, the data demonstrated that ML120B is a selective IKKβ inhibitor that blocks IκBα phosphorylation, nuclear translocation, and NF-κB-mediated transcription.

RA is a chronic autoimmune disease characterized by synovitis and bone destruction (Gravallese, 2002). Multiple lines of evidence suggest that several cell types play critical roles in the pathophysiological process of RA. We evaluated HFLS, mast cells, and chondrocytes for ML120B-mediated inhibition of NF-κB-dependent responses. HFLS play a critical role in the process of cartilage and bone erosion in RA, presumably via the synthesis and secretion of inflammatory mediators, including IL-6 and IL-8 (Georganas et al., 2000). In vitro IKKβ blockade with a dominant-negative adenoviral construct inhibited IL-1- or TNFα-stimulated IL-6, IL-8, and intercellular adhesion molecule-1 production (Aupperle et al., 2001). In our study, ML120B inhibited TNFα- or IL-1β-induced IL-6 or IL-8 production in HFLS with IC₅₀ values ranging from 5.7 to 14.5 μM (Table 3). Several lines of evidence indicate that NF-κB and MAP kinase/AP-1 pathways are both involved in IL-6 and IL-8 expression in HFLS (Neff et al., 2001) and that cross-talk exists between IL-1 and IL-6 signaling (Deon et al., 2001). Therefore, blocking NF-κB with a selective IKKβ inhibitor may only partially inhibit IL-6 or IL-8 production induced by TNFα or IL-1β in HFLS.

Several studies have examined the pathogenic role of chemokine and chemokine receptor interactions in RA (Nanki et al., 2001). We report here for the first time that in HFLS, both RANTES and MCP-1 production induced by TNFα or IL-1β are IKKβ/NF-κB-dependent. The selective IKKβ inhibitor ML120B concentration-dependently blocked TNFα- or IL-1β-induced RANTES and MCP-1 production with IC₅₀ values ranging from 0.7 to 1.8 μM (Table 3). We found ML120B is more potent in blocking the RANTES and MCP-1 production than its effect on blocking IL-6 and IL-8 production (Table 3). Pharmacia Corporation's IKKβ inhibitor SC514 inhibits IL-1β-stimulated IL-6 and IL-8 production in HFLS with an IC₅₀ of 20 μM (Kishore et al., 2003). Compared with the first generation β-carboline IKKβ inhibitor PS1145 and other classes of IKKβ inhibitors, ML120B is a highly selective, potent inhibitor. In a separate report, we demonstrated that ML120B was efficacious in an experimental model of RA. The oral administration of ML120B dose-dependently inhibited paw swelling and offered significant protection from arthritis-induced body weight loss, cartilage, and bone erosion (L. Schopf, A. Savinainen, K. Anderson, J. Kujawa, M. DuPont, M. Silva, E. Siebert, S. Chandra, J. Morgan, P. Gangurde, et al., unpublished data). Considering the pivotal role of HFLS in the pathophysiology of RA, the inhibitory effect of ML120B on RANTES, MCP-1, IL-6, and IL-8 production may directly contribute to the in vivo efficacy of the compound.

Active arthritis is associated with increased cell infiltration, as well as expression of proinflammatory cytokines and MMPs (Smeets et al., 2003). ML120B concentration-dependently blocks TNFα- or IL-1β-induced MMP13, MMP3, and MMP1 production. Results from two to three donors are summarized in Table 3. These data suggest that IKKβ inhibition may play a dominant role in protecting joint disruption and blockade of inflammatory cell migration in RA joints.

The mast cell is a tissue-based inflammatory cell of bone marrow origin. Mast cells respond to signals of innate and acquired immunity with the immediate and delayed release of inflammatory mediators (Taylor and Metcalfe, 2001). The mast cell is now considered to play a pivotal role not only in allergic reactions but also in a number of other inflammatory disorders (Stevens and Austen, 1989; Boyce, 2003, 2004). A potential role for mast cells in RA has also been highlighted recently (Benoist and Mathis, 2002; Lee et al., 2002; Woolley, 2003), and it has been suggested that mast cells provide a critical cellular link between soluble factors and the synovial eruption through activation of cytokine production upon activation. In addition, cytokine production by mast cells has been linked to NF-κB-mediated events (Boyce, 2004). We therefore evaluated ML120B for inhibition of mast cell cytokine production in response to inflammatory initiators. The TLR family is the essential recognition and signaling component of mammalian host defense (Akira and Sato, 2003). TLRs are responsible for the production of inflammatory cytokines in bone marrow-derived mast cells (Stassen et al., 2001). Human cord blood PBMCs can be differentiated in vitro into mast cells (Hsieh et al., 2001). We studied the expression of TLR in this cell type by FACS. Human cord blood-derived mast cells express both TLR 2 and TLR 4 but have minimal expression of TLR 3 and TLR 9. We used adenoviral overexpression of a dominant-negative form of IKKβ (IKKβ DN) and IκB SR to study the function of IKKβ/NF-κB in TLR signaling. Either LPS- or PGN-induced IL-1β, TNFα, and IL-6 production was abolished by IKKβ DN or IκB SR. In line with this finding, the selective IKKβ inhibitor ML120B had similar effects on human MCs. ML120B concentration-dependently inhibited LPS- or PGN-induced IL-1β, TNFα, and IL-6 production with IC₅₀ values of 2.2 to 5.6 μM. Our results indicate that inhibition of IKKβ may have a beneficial effect on attenuating local inflammation through the direct blockade of mast cell activation. Interestingly, in the same experiment we found that IL-4 and IL-13 production in MC is not affected by ML120B (data not shown), indicating that these two cytokines may use an NF-κB-independent regulatory mechanism in human MCs.

Activated chondrocytes are also thought to play an important role in joint destruction in RA. MMPs, a family of zinc-dependent enzymes, play a prominent role in matrix degradation. Collagenases (MMP1, MMP13) and stromelysins (MMP3) are increased in RA synovium/cartilage (Close, 2001). IL-1 stimulates the release of degrading enzymes by fibroblast-like synoviocytes at the cartilage-pannus interface. At the same time, IL-1 may also activate chondrocytes to release these enzymes, which contributes to the cartilage destruction at sites distant to the pannus. ML120B blocked IL-1β-induced MMP13 and MMP1 production in SW1353 cells with IC₅₀ values of 0.9 and 1.2 μM, respectively, suggesting that chondrocytes may be a suitable cell target for IKKβ inhibition in RA. Recent evidence suggests that cyclo-oxygenase (COX-2), a NF-κB-mediated gene product, is a mediator of angiogenesis, and COX-2 activity is known to be up-regulated in the rheumatoid arthritis synovium (Woods et al., 2003). Activation of COX-2 leads to the production of the proinflammatory mediator PGE_2. We investigated the effects of ML120B in IL-1β-stimulated PGE₂ production in human SW1353 cells and found that ML120B inhibited IL-1β-induced PGE₂ with an IC₅₀ of 1.7 μM. These data suggest a potent inhibitory effect of ML120B on chondrocyte activation and cartilage degradation.

RA is a chronic progressive inflammatory disease of the joint, involving several cell types thought to play key roles in the abnormal immune response, synovial hyperplasia, and joint destruction observed in the course of the disease. These cellular processes are controlled through production of several molecular mediators, including cytokines and chemokines, and it is likely that NF-κB is a key transcription factor regulating this inflammatory response. Based on the critical role that IKKβ plays in cytokine-mediated NF-κB activation, targeting IKKβ represents a novel approach for the treatment of RA. We have developed ML120B, a potent and selective inhibitor of IKKβ that is specific to the NF-κB pathway. ML120B has shown consistent inhibitory effects toward key cell types and inflammatory mediators that are involved in the pathology and progression of RA. The results of these studies suggest a potential use of this class of IKKβ inhibitors as new therapeutic agents in the treatment of RA.