Regulation by p38 MAPK - The Journal of Immunology

The Journal of Immunology

A Novel Mechanism for TNF-␣ Regulation by p38 MAPK: Involvement of NF-␬B with Implications for Therapy in Rheumatoid Arthritis1 Jamie Campbell, Cathleen J. Ciesielski,2 Abigail E. Hunt,3 Nicole J. Horwood, Jonathan T. Beech, Louise A. Hayes,4 Agnes Denys,5 Marc Feldmann, Fionula M. Brennan, and Brian M. J. Foxwell6 TNF-␣ is a key factor in a variety of inflammatory diseases. This study examines the role of p38 MAPK in the regulation of TNF-␣ in primary human cells relevant to inflammation, e.g., macrophages and rheumatoid synovial cells. Using a dominant negative variant (D168A) of p38 MAPK and a kinase inhibitor, SB203580, we confirm in primary human macrophages that p38 MAPK regulates TNF-␣ production using a posttranscriptional mechanism requiring the 3ⴕ untranslated region of the gene. However, in LPS-activated primary human macrophages we also detect a second previously unidentified mechanism, the p38 MAPK modulation of TNF-␣ transcription. This is mediated through p38 MAPK regulation of NF-␬B. Interestingly this mechanism was not observed in rheumatoid synovial cells. Importantly however, the dominant negative mutant of p38 MAPK, but not SB203580 was effective at inhibiting spontaneous TNF-␣ production in these ex vivo rheumatoid synovial cell cultures. These data indicate there are potential major differences in the role of p38 MAPK in inflammatory signaling that have a bearing on the use of this kinase as a target for therapy. These results indicate despite disappointing results with p38 MAPK inhibitors in the clinic, this kinase is a valid target in rheumatoid disease. The Journal of Immunology, 2004, 173: 6928 – 6937.

T

he success of biologicals, (i.e., Abs or soluble receptors) to block TNF-␣ function has demonstrated an essential role for this cytokine in the pathophysiology of rheumatoid arthritis and other autoinflammatory diseases (1, 2). However, the cost and inconvenience of using biologicals has lead to considerable effort being directed to finding orally available, small m.w. inhibitors of TNF-␣ expression. As the major TNF-␣-producing cells in inflammation are macrophages, there has been considerable interest in the signal transduction mechanisms that regulate TNF-␣ production, with the hope of identifying suitable “drug-prone” targets such as kinases (3, 4). However, despite considerable effort few signaling molecules have been validated as being involved in this process with NF-␬B probably being the best described (5– 8). Other signaling molecules implicated in regulating TNF-␣ expression are the family of MAPKs particularly p38 MAPK (9). Kennedy Institute of Rheumatology Division, Imperial College School of Medicine Hammersmith, London, United Kingdom Received for publication June 9, 2004. Accepted for publication September 17, 2004. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work was supported by the Arthritis Research Campaign and Wyeth. A.D. was a recipient of a European Union Marie Curie fellowship. 2 Current address: University of California San Francisco, Stein Clinical Research, 9500 Gilman Drive, La Jolla, CA 92093-0663. 3 Current address: University of California San Francisco, Cancer Center, 2340 Sutter Street, Box 0128, San Francisco, CA 94143-0128. 4 Current address: Institute of Cancer Research, Chester Beatty Laboratories, 237 Fulham Road, SW3 6JB London, U.K. 5 Current address: Laboratoire de Chimie Biologique, Unite´ Mixte de Recherche 8576 du, Centre National de Recherche Scientifique, Universite´ des Sciences et Technologies de Lille, F-59655 Villeneuve d’Ascq, France. 6

Address correspondence and reprint requests to Prof. Brian M. J. Foxwell, Kennedy Institute of Rheumatology, 1 Aspenlea Road, Hammersmith, London W6 8LH, U.K. Copyright © 2004 by The American Association of Immunologists, Inc.

Originally described as an LPS-activated kinase (10) the relevance of p38 MAPK to TNF-␣ production was demonstrated by the discovery of pyridinyl imidazole-based inhibitors (e.g., SB203580 (9)). Initial studies with these inhibitors indicated that p38 MAPK was exclusively involved in the translational control of TNF-␣ expression (9, 11), a view supported by more recent studies (12). Also, mice defective in the expression of MAPK-activated protein kinase-2, an immediate substrate of p38 MAPK, have a profound defect in translation of the TNF-␣ mRNA (13). In contrast other studies have shown that p38 MAPK controls TNF-␣ mRNA stability (14, 15) possibly by inhibiting the deadenylation process (16). It is unclear why such discrepancies exist between studies, but a contributing factor could be the transformed nature of the myeloid cell lines often used in these experiments (e.g., THP-1 (9), RAW264.7 (15)). Another possibility is that SB203580 may inhibit the related p54 JNK (17) as well as the phosphoinositide-dependent kinase 1 and 2, a part of the PI3K cascade (18), at higher concentrations (low micromolecular range). However, regardless of whether p38 MAPK regulates mRNA stability or translation, the mechanism involved appears to require the AU rich elements (ARE)7 found in 3⬘ untranslated regions (3⬘UTR) (19 – 22). p38 MAPK also regulates several transcription factors including activating transcription factor-2 (23), myocyte enhancer factor 2C (24), C/EBP homologous protein 1 (25), and also NF-␬B (26). A recent study has suggested that p38 MAPK regulates NF-␬B via MAPK and stress-activated protein kinase (SAPK)-1 that also requires p42/p44 Erk for activation (27). MAPK and SAPK-1 in turn phosphorylate serine 276 of the p65 NF-␬B subunit. Alternately, p38 MAPK has been proposed to regulate NF-␬B indirectly by regulating the access to DNA through histone H3 phosphorylation

7 Abbreviations used in this paper: ARE, AU rich element; SAPK, stress-activated protein kinase; M-CSF, macrophage CSF; PSI, proteasome inhibitor I; MOI, multiplicity of infection; UTR, untranslated regions.

0022-1767/04/$02.00

The Journal of Immunology (28). In contrast, p38 MAPK has also been shown to inhibit NF-␬B activation (29, 30) although it is unclear how such conflicting mechanisms might coexist. Although NF-␬B can be required for TNF-␣ expression (7, 8), there are no data showing that p38 MAPK regulates TNF-␣ transcription. Given the importance of p38 MAPK to regulating TNF-␣ expression, it is disappointing that inhibitors of this kinase have not reached the clinic for the treatment of inflammatory disease such as rheumatoid arthritis. To investigate possible factors that may ultimately influence p38 MAPK as a target for treatment we have studied the role of this kinase in regulating TNF-␣ expression in systems highly relevant to disease, i.e., primary human macrophages, rheumatoid arthritis fibroblasts, and synovium. Because SB203580 has other cellular targets, we have also used a dominant negative kinase defective variant p38 D168A. We show for the first time that p38 MAPK does regulate TNF-␣ production in rheumatoid arthritis synovium, although SB203580 itself is a poor inhibitor in these circumstances. Moreover, inhibition of p38 MAPK results in a reduction of TNF-␣ mRNA in human macrophages. Furthermore, the kinase regulates TNF-␣ expression by a dual mechanism, in part by the 3⬘UTR as previously suggested but also via the 5⬘ promoter region, putatively through NF-␬B. However the regulation of NF-␬B by p38 MAPK is not universal, occurring in macrophages but not synovial fibroblasts or rheumatoid arthritis synovial cultures. These observations could have a major bearing on our approach to p38 MAPK as a target for inhibitors of TNF-␣ expression and anti-inflammatory drugs in general.

6929 is a modification of the pNF-␬B reporter vector (BD Biosciences/Clontech, Palo Alto, CA).

Viral infection of M-CSF-treated monocytes, rheumatoid arthritis synovial cells and fibroblast-like synoviocytes Macrophages susceptible to virus infection were generated by culturing freshly elutriated monocytes for 48 h with 100 ng/ml M-CSF; these cells were then infected, as previously described (7). Synovium from rheumatoid arthritis patients was infected as previously described (7). Human fibroblast-like synoviocytes were cultured and infected as previously described (35).

ELISA Following stimulation of human macrophages or culture of human tissue cells, supernatants were harvested. Concentrations of TNF-␣, IL-1R antagonist, and IL-10 (BD Pharmingen, San Diego, CA) were determined by ELISA, following the manufacturer’s instruction. Absorbance read and analyzed at 450 nM on a spectrophotometric ELISA plate reader (Labsystems Multiskan Biochromic) using the Ascent version 2.4.2 software program. Results are expressed as the mean concentration of triplicate cultures ⫾ SD.

Materials and Methods Cell culture The human embryonic kidney 293 cell lines were from BD Clontech (Palo Alto, CA). These cells were cultured in DMEM supplemented with 10% FCS (Sigma-Aldrich, Poole, U.K.). Primary human monocytes were isolated from single-donor buffy coats obtained from the North London Blood Transfusion Centre (Colindale, London, U.K.) by Ficoll-Hypaque centrifugation followed by centrifugal elutriation using a J6 elutriator (Beckman Coulter, High Wycombe, U.K.). Monocyte purity was assessed by flow cytometry and was routinely higher than 90%. Cells were cultured in RPMI 1640 (BioWhittaker, Wokingham, U.K.) supplemented with 5% FCS. For macrophage differentiation, monocytes were cultured with macrophage CSF (M-CSF) at 100 ng/ml (kindly provided by Dr. G. Larsen, Genetics Institute, Boston, MA) for 48 h before infection. Synovium from patients with rheumatoid arthritis undergoing joint surgery were dissociated by cutting into small pieces and were digested with collagenase and DNase (31). The total cell mixture was cultured at 37°C in RPMI 1640 medium with 25 mM HEPES and 2 mM L-glutamine, supplemented with 5% heat-inactivated FBS. Cells were stimulated with 10 ng/ml LPS (Sigma-Aldrich) purified through phenol/chloroform extraction. SB203580 (catalog no. 559389), PD98059 (catalog no. 513001), and proteasome inhibitor I (PSI; catalog no. 539160) were from Calbiochem (Nottingham, U.K.).

Production of recombinant adenoviruses Recombinant, replication-deficient adenoviral constructs were prepared using the AdEasy system, previously described (32). cDNAs encoding wildtype p38␣ MAPK and kinase-deficient p38␣ MAPK with a glutamate to alanine mutation at residue 168 (D168A), kindly provided by Dr. P. Young (SmithKline Beecham, King of Prussia, PA) were subcloned into AdTrackCMV transfer vector using KpnI and EcoRV. p38 MAPK-containing transfer vectors were inserted into the AdEasy viral backbone vector by bacterial recombination in the Escherichia coli strain BJ5183, as previously described (32). Recombined vectors were propagated in human embryonic kidney 293 cells and viral stocks prepared by ultracentrifugation through two caesium chloride gradients. Titers of viral stocks were determined by plaque assay and viral aliquots were stored at ⫺70°C. AdvGFP adenovirus was prepared according to this protocol using AdTrackGFP with no insert. pAdTrack-TNF-Luc-3⬘UTR and pAdTrack-TNF-Luc were generated, as previously described (33). The NF-␬B luciferase reporter adenovirus (Adv NF-␬B-Luc) contains four tandem copies of the ␬ enhancer element located upstream of the firefly luciferase gene (34). This adenovirus was provided by Dr. P. B. McCray, Jr. (University of Iowa, Iowa City, IA) and

FIGURE 1. Expression and effects of Advp38WT and Advp38D168Ainfected human macrophages stimulated by LPS. A, Macrophages were left uninfected or were infected with AdvGFP, Advp38WT or Advp38D168A at a MOI of 100. After 24 h, half the cells of each group were treated with LPS (10 ng/ml) for 10 min. The cells were lysed and cytosolic extracts were prepared and analyzed for p38 MAPK expression by immuno-Western blotting. A and B, Macrophages were left uninfected or were infected with AdvGFP, Advp38WT, or Advp38D168A at a MOI of 100. After 24 h cells were stimulated with LPS (10 ng/ml) or left unstimulated for 10 min. MAPK-activated protein kinase-2 activity was assayed by immunokinase assay using the heat shock protein hsp27 as a substrate (B). C, The data are represented by Phosphoimager analysis of the gel (B) and the same data normalized to percentage of radioactivity incorporated into hsp27 for each condition with AdvGFP control represented as 100%. Data shown are representative of three independent experiments.

6930

p38 MAPK REGULATION OF TNF-␣ IN PRIMARY HUMAN MACROPHAGES

FIGURE 2. Effect of Advp38D168A and SB203580 on LPS-induced cytokine production by human macrophages and rheumatoid arthritis joint cells. Cells were uninfected or infected with AdvGFP (MOI of 200), Advp38WT (MOI of 200) or Advp38D168A at the indicated viral titers (A–D) or concentration of SB203580 (E and F). Cytokine production was analyzed in both macrophages (A, B, and E) and rheumatoid arthritis joint cells (C, D, and F). After 18 h LPS stimulation (10 ng/ml) supernatants were removed and assayed for TNF-␣ production in macrophages and rheumatoid arthritis joint cells (A, C, E, and F) or IL-10 and IL-1R antagonist (B and D), respectively. Levels of IL-10 and IL-1R antagonist in unstimulated cells were below the level of detection (data not shown) (B and D). Data are means of triplicate cultures ⫾ SD and are representative of three to six experiments performed using different donors. The broken line represents the 50% inhibition level (E and F).


6931

FIGURE 2.

Immunoprecipitation and MAPK-activated protein kinase-2 assays A total of 2 ⫻ 106 cells were lysed in 200 ␮l of lysis buffer, as previously described (17) and protein concentration of lysates were determined by Bradford assay. Lysates of equivalent protein content were incubated with 0.1 mg/ml anti-MAPK-activated protein kinase-2 Ab (Upstate Biotechnology, Lake Placid, NY) and 15 ␮l of protein G-Sepharose at 4°C overnight. Immunoprecipitates were washed twice with RIPA buffer and twice with kinase assay buffer and then resuspended in 20 ␮l of assay buffer containing recombinant hsp27 (Bioquote, York, U.K.) at a concentration of 0.1 mg/ml. Kinase reactions were initiated by the addition of a further 20 ␮l of assay buffer containing ATP at 20 ␮M and [␥-32P]ATP at 0.25 ␮Ci/ml. Reactions were terminated by the addition of 20 ␮l of 4⫻ gel sample buffer after 20 min. Samples were separated by SDS-PAGE (12.5%) and phosphorylated substrates were visualized by autoradiography and Phosphorimager analysis of dried gels.

Western blot analysis Western blotting for p38 MAPK, phospho-JNK/SAPK, and phospho-p42/ p44 ERK was performed, as previously described (36). Ab for p38 MAPK (SAK7) was from Prof. J. Saklatvala (Kennedy Institute, London, U.K.) and phosphospecific Abs were from New England Biolabs (Hitchin, Herts, U.K.). I␬B␣ and p65 phosphoserine 536 were obtained from Santa Cruz Biotechnology (Santa Cruz, CA) and Cell Signaling Technology (Beverly, MA), respectively.

Luciferase assays After LPS stimulation, cells were washed once in PBS and lysed with 100 ␮l of CAT lysis buffer (0.65%(v/v) of Nonidet P-40, 10 mM Tris-HCl pH

(Continued) 8, 0.1 mM EDTA pH 8, 150 mM NaCl). Cell lysate (50 ␮l) were transferred into the well of a luminometer cuvette strip containing 120 ␮l of luciferase assay buffer (25 mM Tris-phosphate pH 7.8, 8 mM MgCl2, 1 mM EDTA, 1%(v/v) Triton X-100, 1%(v/v) glycerol, 1 mM DTT, 0.5 mM ATP). Luciferase activity was measured with a LabSystem Luminometer by dispensing 30 ␮l of luciferin (Bright-Glo luciferase assay system; Promega, Madison, WI) per assay point. Cell lysates were assayed for protein concentration by bicinchoninic acid assay and luciferase activity adjusted accordingly.

Extraction of cytoplasmic and nuclear cellular fractions Following stimulation, cells were scraped into ice-cold PBS and lysed in hypotonic lysis buffer (0.125% Nonidet P-40, 5 mM HEPES (pH 7.9), 10 mM KCl, 1.5 mM MgCl2), and cells harvested by centrifugation at 1000 ⫻ g for 5 min. Supernatant, containing cytoplasmic fraction, was aspirated and nuclei resuspended in hypertonic lysis buffer (5 mM HEPES (pH 7.9), 25% glycerol, 500 mM NaCl, 1.5 mM MgCl2, 0.2 mM EDTA) for 2 h with constant agitation. Cellular debris was removed by centrifugation at 13,000 ⫻ g for 20 min.

NF-␬B EMSA NF-␬B DNA binding activities were determined by incubating 5 ␮g of nuclear protein of each extract with [␥-32P]ATP-labeled double-stranded NF-␬B consensus oligonucleotide (Promega), followed by resolution on a 5% (w/v) nondenaturing polyacrylamide gel. Gels were dried on to filter paper and retarded DNA; protein complexes were visualized by autoradiography.

6932


Northern blot analysis and hybridization Total RNA was isolated using the RNeasy kit (Qiagen, Hilden, Germany). A total of 2 ␮g of total RNA was resolved by electrophoresis on a 1.5% (w/v) agarose denaturing formaldehyde gel. The gel was transferred to Hybond N membrane by capillary blotting in 20⫻ SSC and RNA UV fixed to membrane. RNA was detected using a 495-bp TNF-␣-specific coding region cDNA probe (courtesy J. Dean, Kennedy Institute, London, U.K.) labeled with [␣-32P]CTP. Hybridization was conducted overnight in UltraHyb (Ambion, Austin, TX) at 42°C. The membrane was washed in a series of SSC/SDS buffers and visualized using Fuji FLA-2000 phosphor imager.

Taqman RT-PCR M-CSF-treated monocytes were plated at 5 ⫻ 105 cells/well in 24-well plates and infected as previously described. Total RNA was extracted using RNeasy kit (Qiagen) according to the manufacturer’s instructions. All semiquantitative RT-PCR was performed using an ABI PRISM 7700 Sequence Detection System, Taqman One Step RT-PCR reagent, and TNF-␣, IL-6, and GAPDH predeveloped assay reagents (Applied Biosystems, Foster City, CA and PerkinElmer, Wellesley, MA). ABI PRISM 7700 Sequence detector was programmed for the reverse transcription step of 30 min at 48°C followed by a 5 min deactivation step at 95°C. Subsequent PCR amplification consisted of 40 cycles of denaturation at 94°C for 15 s and annealing/extension at 60°C for 60 s. The cycle number at which the amplification plot crosses a fixed threshold above baseline is defined as threshold cycle (Ct). To control variation in mRNA concentration, all results were normalized to the housekeeping gene, GAPDH. Relative quantitation was performed using the comparative ⌬⌬Ct method according to the manufacturer’s instructions.

(IC50 ⬎ 10 nM; Fig. 2, E and F). The possible reasons for this will be considered in Discussion. p38D168A MAPK decreases LPS-induced macrophage TNF-␣ mRNA production Previous studies with SB203580 have disagreed on whether inhibition of p38 MAPK resulted in suppression of TNF-␣ mRNA expression (9, 17); this issue was now revisited with the Advp38D168A. TNF-␣ mRNA expression was induced in human primary macrophages by LPS and this was inhibited to a similar degree (⬃30%) by 1 ␮M SB203580 and Advp38D168A (Fig. 3, A and B). The effect of SB203580 was similar to that seen by Dean et al. (17) in human monocytes. These data would indicate that p38 MAPK does have some role in regulating TNF-␣ expression before translation. However, the effect of blocking p38 MAPK was

Results p38 D168A is an effective inhibitor of TNF-␣ production either induced by LPS or released spontaneously in rheumatoid arthritis synovium Adenoviral vectors encoding p38␣ MAPK and the kinase-deficient variant D168A (Advp38WT and Advp38D168A, respectively) were generated. Infection of human macrophages or rheumatoid arthritis synovial cultures with these adenoviruses at a multiplicity of infection (MOI) of 100 resulted in ⬎90% infection (data not shown) as has been observed with other adenoviral vectors (7, 8). Infection of cells with either virus resulted in the over-expression of p38 MAPK as determined by immunoblotting (Fig. 1A), whereas a control virus (AdvGFP) had no effect. p38D168A acted as a dominant negative inhibiting the LPS-induced activation of the p38 MAPK substrate kinase, MAPK-activated protein kinase-2, in macrophages by ⬃90% as measured in immunokinase assays using hsp27 as a substrate (Fig. 1, B and C). The inhibitory effect of p38D168A appeared to be specific, as phospho-Western blot analysis of the related MAPKs, p42/p44/ERK and p54/JNK demonstrated no inhibition of LPS-induced activation of these kinases (data not shown). Infection of human macrophages with increasing MOI for Advp38D168A resulted in a dose-dependent inhibition of LPSinduced TNF-␣ production at 18 h poststimulation. This reached a plateau of ⬃80% at MOI of 100 (Fig. 2A). Similar results were achieved at 4 h post-LPS stimulation (data not shown). In contrast to the effect on TNF-␣ expression; p38D168A did not inhibit the production of the anti-inflammatory factors, IL-10 or IL-1R antagonist (Fig. 2B). Advp38D168A was also effective at blocking the spontaneous production of TNF-␣ (67% p ⬍ 0.002) by rheumatoid arthritis synovium (Fig. 2C). Again, p38D168A had no effect on IL-10 or IL-1R antagonist production (Fig. 2D), indicating that the effects seen are not due to any toxicity of inhibiting p38 MAPK or viral infection. In comparison, studies with SB203580 showed that the drug was much less effective on rheumatoid arthritis synovium compared with LPS-stimulated macrophages. In rheumatoid arthritis synovial cultures the IC50 of ⬎1000 nM was 100-fold higher than that found for LPS-activated macrophages

FIGURE 3. Influence of p38 MAPK on LPS induced TNF-␣ mRNA levels in human macrophages. Cells were infected with Advp38D168A or AdvGFP at MOI 200. After 24 h cells were preincubated with SB203580 (1 ␮M) or PSI (5 ␮M) for 30 min where indicated, then stimulated with LPS (10 ng/ml). After 30 min the cells were harvested and total RNA extracted. TNF-␣ mRNA was detected by Northern hybridization (A and B) or Taqman PCR (C). A, Phosphoimager data analysis (upper) of membrane and (B) densitometric analysis (lower) are represented. Changes in TNF-␣ mRNA were normalized with respect to G3PDH levels and are represented as percentage of control (control ⫽ uninfected LPS stimulated cells). Data are representative of two independent experiments from different donors.

The Journal of Immunology less than that seen with PSI (⬃80%), which is a proteasome inhibitor that is an effective blocker of NF-␬B activation. As the regulation of TNF-␣ mRNA has been a controversial area we confirmed our results by a second technique, semiquantitative Taqman RT-PCR (Fig. 3C). The results were nearly identical with those obtained by Northern blotting confirming that p38 MAPK does regulate TNF-␣ mRNA levels in macrophages. p38 MAPK D168A regulation of LPS-induced TNF-␣ reporter gene activity is mediated by both the TNF-␣ 5⬘ promoter and the 3⬘UTR region Studies in murine cells with SB203580 have indicated that p38 MAPK regulation of TNF-␣ was mediated only via the 3⬘UTR of the mRNA, which contains the ARE (15, 37). To determine whether TNF-␣ was regulated in human primary macrophages in a similar manner we used TNF-␣ promoter-driven luciferase reporter genes with (p5⬘3⬘UTR) or without (p5⬘ only) the 3⬘UTR encoded into adenovirus (reporter virus) (33, 38, 39). As expected, expression of p38D168A or SB203580 inhibited the LPS-stimulated activation of the p5⬘3⬘UTR reporter gene expression by 60 – 70% (Fig. 4). Surprisingly, both drug and dominant negative also caused a significant inhibition of LPS-induced reporter gene activity in the absence of the 3⬘UTR (Fig. 4). The level of inhibition

FIGURE 4. Investigation of regions of the TNF-␣ gene associated with p38 MAPK function. Schematic representation of Adv TNF-␣ luciferase reporter constructs (A). Cells were simultaneously infected with TNF-␣ reporter virus at a MOI of 50 and Advp38D168A or AdvGFP at MOI of 200 or reporter virus alone. After 24 h cells were preincubated with SB203580 (1 ␮M) for 30 min before stimulation with LPS (10 ng/ml) for 4 h. Cells were then lysed and assayed for luciferase activity (B). C, Cells were also infected with TNF-␣ reporter viruses at a MOI of 50 and Advp38D168A or AdvGFP at a MOI of 200. After 24 h cells were preincubated with PSI (5 ␮M) as above then stimulated with LPS (10 ng/ml) 4 h, lysed and assayed for luciferase activity. Data are expressed as percentage of AdvGFP ⫹PSI. NS, Not statistically significant; p ⬎ 0.2. Data are represented as percentage of activation of cells with LPS alone and are means of triplicate cultures ⫾ SD and representative of three to six experiments using different donors. Values for p represents Student’s two-tailed t test.

6933 (30 –50%) was similar to the effect seen on TNF-␣ mRNA (Fig. 3B). These data would suggest that p38 MAPK has some regulatory role on the TNF-␣ promoter but the significant increase in the inhibition seen in the presence of 3⬘UTR indicates that both elements are regulated by the kinase. p38 MAPK regulates LPS induced NF-␬B activation The mechanism by which p38 MAPK regulates TNF-␣ by the 3⬘UTR has been previously recognized and studied. In contrast, a regulatory mechanism involving the TNF-␣ promoter region has not been previously observed. Although p38 MAPK has been shown to putatively regulate several transcription factors, this effect has not been related to regulation of any actual genes. Of these transcription factors only NF-␬B (26, 27) is known to regulate TNF-␣ expression (4). Neither SB203580 nor expression of p38D168A was found to exert any effect in human primary macrophages on the LPS-induced degradation of I␬B␣ (Fig. 5A), or NF-␬B DNA binding activity as measured by EMSA (Fig. 5B). However, inhibition of p38 MAPK by either approach resulted in a 65% inhibition of NF-␬B-induced transcription as shown by an NF-␬B-luciferase reporter gene (Fig. 5C). Inhibition of NF-␬B via p38 MAPK was not quite as effective as over-expression of I␬B␣ (88% inhibition) or using an inhibitor of I␬B␣ degradation (PSI;

6934


FIGURE 5. Effect of p38 MAPK on NF-␬B activation in human macrophages. Cells were infected with AdvGFP, or Advp38D168A at a MOI of 200. After 24 h cells were preincubated with SB203580 for 30 min where indicated, then stimulated with LPS (10 ng/ml). After 30 min the cells were harvested and cytoplasmic and nuclear extracts prepared. Cytoplasmic extracts were analyzed for I␬B␣ (A) expression by Western blotting. ␤-actin was used to confirm equal protein loading. NF-␬B/DNA binding was analyzed by EMSA (B). Cells were simultaneously infected with NF-␬B reporter virus at a MOI 100 and Advp38D168A, AdvI␬B␣, AdvGFP at a MOI of 200 or reporter virus alone. After 24 h cells were preincubated with SB203580 (1 ␮M), PD98059 (10 ␮M), or PSI (5 ␮M) for 30 min before stimulation with LPS (10 ng/ml) for 4 h. Cells were then lysed and assayed for luciferase activity (C). Data are represented as percentage of activation of cells with LPS alone and are means of triplicate cultures ⫾ SD. Data are representative of six experiments using different donors. Values for p represent Student’s two-tailed t test.

95% inhibition) (Fig. 5C). These data indicate that the kinase was involved in modulating the transactivating function of NF-␬B. Next we addressed the question of whether the p38 MAPK regulation of NF-␬B would account for the effects seen on the 5⬘ promoter of the TNF-␣ gene. We therefore investigated whether inhibiting p38 MAPK would have any additional effect over PSI (NF-␬B inhibitor). As seen in Fig. 4D, p38D168A did not significantly inhibit 5⬘ promoter but did have an effect on construct containing the 3⬘UTR in the presence of PSI. Indicating that p38 MAPK regulates the 5⬘ promoter via NF-␬B. The role of p38 MAPK in regulating LPS-induced NF-␬B depends on cell type The regulation of NF-␬B by p38 MAPK has not been a widely reported phenomenon and in HeLa and Jurkat cells we found that SB203580 failed to inhibit TNF-induced NF-␬B-induced reporter gene activity (data not shown). Our studies on LPS were broadened to establish whether the role of p38 MAPK might be cell type- or stimuli-specific. LPS induction of the NF-␬B reporter gene in primary human rheumatoid arthritis synovial-derived fibroblasts was insensitive to SB203580 or p38D168A but still effectively inhibited by PSI, I␬B␣ or a dominant negative version of I␬B kinase 2 as expected (Fig. 6A). This was not only found for LPS stimulation but similar results were obtained from TNF-␣ and IL-1-stimulated cells (data not shown).

p38 MAPK does not have a role in spontaneous NF-␬B activation in rheumatoid arthritis synovium Because there are differences in the role of p38 MAPK in NF-␬B function depending on cell type and possible stimuli, we investigated whether the kinase has a role in NF-␬B activation in rheumatoid arthritis synovium where the stimulus is unknown. Infection of rheumatoid arthritis culture with NF-␬B reporter virus resulted in constitutive activation of NF-␬B as expected (Fig. 6B). However, this activation was not affected by SB203580 or p38D168A. The activation of the NF-␬B reporter gene was inhibited by PSI although less effectively than in other cells.

Discussion This study has investigated the regulation of TNF-␣ expression by p38 MAPK in human macrophages and rheumatoid arthritis synovial cells. Two approaches were used; the inhibitor SB203580 and an adenovirus encoding a kinase dead mutant of p38 MAPK that acts as dominant negative. This study confirmed that p38 MAPK regulates TNF-␣ in LPS-activated human macrophages via the TNF-␣ 3⬘UTR. In addition, we observed that p38 MAPK also regulated TNF-␣ transcription through NF-␬B. To our knowledge this is the first time that such mechanism has been observed. However, we also showed that the regulation of NF-␬B by p38 MAPK


6935

FIGURE 6. Role of p38 MAPK in LPS-induced NF-␬B activity in rheumatoid fibroblasts and synovium. Human fibroblast-like synoviocytes (A) or rheumatoid arthritis joint cells (B) were simultaneously infected with NF-␬B reporter virus at a MOI of 100 and Advp38D168A, AdvI␬B␣, AdvI␬B kinase 2DN, AdvGFP at MOI 200/400 or reporter virus alone. After 24 h cells were preincubated with SB203580 (1 ␮M), PD98059 (10 ␮M), or PSI (5 ␮M) for 30 min before stimulation with LPS (10 ng/ml) for 4 h. Cells were then lysed and assayed for luciferase activity. Data are represented as percentage of activation of cells with LPS alone and are means of triplicate cultures ⫾ SD and representative of six experiments and three experiments using different donors from human fibroblast-like synoviocytes (A) and rheumatoid joint cells (B), respectively. Values for p represents Student’s two-tailed t test.

is cell type-dependent as this was not seen in rheumatoid arthritis synovium or LPS-activated primary synovial fibroblasts. Inhibiting the activity of p38 MAPK resulted in 30 –50% reduction in TNF-␣ mRNA. This observation correlated with the reduction in TNF-␣ 5⬘ promoter reporter construct activity in the presence of p38 MAPK inhibitors. This study also showed in human primary macrophages that p38 MAPK regulates TNF-␣ expression through the 3⬘UTR, as previously demonstrated with murine cells (40). Kontoyiannis et al. (12) found that SB203580 is ineffective in mice deficient in the ARE regions of the TNF-␣ 3⬘UTR. This observation differs from our results in human macrophages, however as their study also showed no inhibitory effect of SB203580 on TNF-␣ mRNA expression in the wild-type mice controls there are clearly differences between this model and our cell system. In

addition, studies by others have shown an effect of this drug on mRNA expression (15, 17, 41). This disparity could be accounted for in part by differences between the regulation of TNF-␣ between murine and human systems (42). The regulation of LPS-induced NF-␬B by p38 MAPK provided a mechanism for the transcriptional control of TNF-␣ as this transcription factor is required for TNF-␣ expression (7). The data also suggest that NF-␬B accounted for all the inhibitory effect on transcription of blocking p38 MAPK. However there is the caveat that the NF-␬B inhibitor, PSI, used in our studies could share targets other than NF-␬B with the inhibitors of p38 MAPK. The observation that p38 MAPK regulates NF-␬B-induced transcription is in agreement with previous studies (26, 43). How p38 MAPK controls NF-␬B driven transcription is unclear but prior studies have

6936


suggested both direct mechanisms, via the phosphorylation of the p65 subunit (27) or indirect mechanisms, involving the phosphorylation of histone H3 to uncover NF-␬B sites (28). As blocking p38 MAPK produced a more potent inhibition of the simple NF-␬B consensus site reporter gene than that seen on 5⬘ TNF-␣ promoter or TNF-␣ mRNA, we need to examine the more highly complex nature of the promoter region. There are five potential NF-␬B sites in the TNF-␣ promoter (42) and the regulation of these sites, or their role in TNF-␣ expression, may not all require p38 MAPK activity. Alternately, p38 MAPK may have confounding effects on other elements in the 5⬘ promoter that ameliorate the inhibition of NF-␬B. Unraveling these questions will be the subject of further studies. Another important observation of this study is that a role for p38 MAPK in the activation of NF-␬B is not a universal event. We previously found no role for p38 MAPK in the TNF-induced activation of NF-␬B in the Jurkat T cell line (our unpublished observation). In primary human synovial fibroblasts, p38 MAPK was not required for NF-␬B activation even when induced by LPS. This observation has antecedents as we have found differences in signaling mechanisms leading to NF-␬B, with I␬B kinase 2 being required for LPS-induced NF-␬B in human primary fibroblasts and HUVECs but not macrophages (44). Whether the two observations are directly linked or just indications of differences in mechanism is unknown but the results again emphasize the diversity of signaling mechanisms that can exist between different cell systems and help to explain why the regulation of NF-␬B by p38 MAPK has not been a more widely observed event. It interesting to note that others have claimed that p38 MAPK is involved in the inhibition of NF-␬B activation (29, 30) indicating kinase may have multiple roles in NF-␬B function depending on cell type and stimulus. This study also shows that p38 MAPK would be an effective target for blocking TNF-␣ production in human rheumatoid arthritis synovium. However is interesting to note that SB203580 is not effective in this system, unlike p38 MAPK dominant negative. These latter observations in rheumatoid arthritis synovium and fibroblasts may have some bearing on the present problems p38 MAPK inhibitors have faced in clinical trials. It is unclear why the drug was ineffective but a possibility is that unlike the simple cell systems, SB203580 is rapidly metabolized by rheumatoid arthritis tissue. Another explanation could lie in the ability of the drug to inhibit the PI3K pathway via phosphoinositide-dependent kinase 1 and 2 (18). The PI3K pathway negatively regulates TNF-␣ production in rheumatoid arthritis tissue (45), thus SB203580 could have simultaneous opposing effects in this model. These data may explain why although many years have passed since the observation of the effect of pyridinyl imidazoles on p38 MAPK, no clinically effective anti-inflammatory drug has yet been developed from these molecules. Interestingly, blocking p38 MAPK did not inhibit the production of anti-inflammatory molecules, i.e., IL-10 or IL-1R antagonist in rheumatoid arthritis tissue. In summary, we have investigated the role of p38 MAPK in TNF-␣ expression by human macrophages induced by LPS and spontaneously in rheumatoid arthritis. We have shown that p38 MAPK is likely to regulate transcription of TNF-␣ mRNA via NF-␬B in addition to any effects on posttranscriptional mechanisms in primary human macrophages. The data also showed that the regulation of NF-␬B by p38 MAPK is cell-type dependent and this may have a consequence for the anti-inflammatory efficacy of inhibitors of p38 MAPK. However, even in the absence of regulation of NF-␬B by the kinase in human synovial tissue, inhibition was still effective in blocking TNF-␣. Overall our study, although demonstrating that p38 MAPK is an important target for TNF-␣

blockade, also reveals that there is still much to learn about how this kinase functions in different tissues, the results of which may modulate our approach to generating useful inhibitors.

Acknowledgments We acknowledge Drs. Cope, Dean, Brook, Clark, and Udalova for their helpful comments; Gena Mellett for typing the manuscript; and Amy Peters for collecting tissue and technical support. We also acknowledge Dr. Peter Youngs for the p38MAPK plasmids and Dr. P. B. McCray, Jr. for the NF-␬B luciferase adenoviral construct.

References 1. Feldmann, M., and R. N. Maini. 2003. Lasker Clinical Medical Research Award: TNF defined as a therapeutic target for rheumatoid arthritis and other autoimmune diseases. Nat. Med. 9:1245. 2. Feldmann, M., and R. N. Maini. 2001. Anti-TNF␣ therapy of rheumatoid arthritis: what have we learned? Annu. Rev. Immunol. 19:163. 3. Hammaker, D., S. Sweeney, and G. S. Firestein. 2003. Signal transduction networks in rheumatoid arthritis. Ann. Rheum. Dis. (62 Suppl.)2:ii86. 4. Foxwell, B., E. Andreakos, F. Brennan, M. Feldmann, C. Smith, and M. Conron. 2003. Prospects for the development of small molecular weight compounds to replace anti-tumour necrosis factor biological agents. Ann. Rheum. Dis. (62 Suppl.)2:ii90. 5. Baldwin, A. S., Jr. 1996. The NF-␬B and I␬B proteins: new discoveries and insights. Annu. Rev. Immunol. 14:649. 6. Barnes, P. J., and M. Karin. 1997. Nuclear factor-␬B: a pivotal transcription factor in chronic inflammatory diseases. N. Engl. J. Med. 336:1066. 7. Foxwell, B., K. Browne, J. Bondeson, C. Clarke, C. de Martin, F. Brennan, and M. Feldmann. 1998. Efficient adenoviral infection with I␬B␣ reveals that macrophage tumor necrosis factor a production in rheumatoid arthritis is NF-␬B dependent. Proc. Natl. Acad. Sci. USA 95:8211. 8. Bondeson, J., B. Foxwell, F. Brennan, and M. Feldmann. 1999. Defining therapeutic targets by using adenovirus: blocking NF-␬B inhibits both inflammatory and destructive mechanisms in rheumatoid synovium but spares antiinflammatory mediators. Proc. Natl. Acad. Sci. USA 96:5668. 9. Lee, J. C., J. T. Laydon, P. C. McDonnell, T. F. Gallagher, S. Kumar, D. Green, D. McNulty, M. J. Blumenthal, J. R. Heys, S. W. Landvatter, et al. 1994. A protein kinase involved in the regulation of inflammatory cytokine biosynthesis. Nature 372:739. 10. Han, J., J. D. Lee, L. Bibbs, and R. J. Ulevitch. 1994. A MAP kinase targeted by endotoxin and hyperosmolarity in mammalian cells. Science 265:808. 11. Young, P., P. McDonnell, D. Dunnington, A. Hand, J. Laydon, and J. Lee. 1993. Pyridinyl imidazoles inhibit IL-1 and TNF production at the protein level. Agents Actions 39:C67. 12. Kontoyiannis, D., M. Pasparakis, T. T. Pizarro, F. Cominelli, and G. Kollias. 1999. Impaired on/off regulation of TNF biosynthesis in mice lacking TNF AUrich elements: implications for joint and gut-associated immunopathologies. Immunity 10:387. 13. Kotlyarov, A., A. Neininger, C. Schubert, R. Eckert, C. Birchmeier, H. D. Volk, and M. Gaestel. 1999. MAPKAP kinase 2 is essential for LPS-induced TNF-␣ biosynthesis. Nat. Cell Biol. 1:94. 14. Wang, C. Y., D. C. Guttridge, M. W. Mayo, and A. S. Baldwin, Jr. 1999. NF-␬B induces expression of the Bcl-2 homologue A1/Bfl-1 to preferentially suppress chemotherapy-induced apoptosis. Mol. Cell. Biol. 19:5923. 15. Brook, M., G. Sully, A. R. Clark, and J. Saklatvala. 2000. Regulation of tumour necrosis factor ␣ mRNA stability by the mitogen-activated protein kinase p38 signalling cascade. FEBS Lett. 483:57. 16. Dean, J. L., S. J. Sarsfield, E. Tsounakou, and J. Saklatvala. 2003. p38 Mitogenactivated protein kinase stabilizes mRNAs that contain cyclooxygenase-2 and tumor necrosis factor AU-rich elements by inhibiting deadenylation. J. Biol. Chem. 278:39470. 17. Dean, J. L., M. Brook, A. R. Clark, and J. Saklatvala. 1999. p38 Mitogen-activated protein kinase regulates cyclooxygenase-2 mRNA stability and transcription in lipopolysaccharide-treated human monocytes. J. Biol. Chem. 274:264. 18. Lali, F. V., A. E. Hunt, S. J. Turner, and B. M. Foxwell. 2000. The pyridinyl imidazole inhibitor SB203580 blocks phosphoinositide-dependent protein kinase activity, protein kinase B phosphorylation, and retinoblastoma hyperphosphorylation in interleukin-2-stimulated T cells independently of p38 mitogen-activated protein kinase. J. Biol. Chem. 275:7395. 19. Miyazawa, K., A. Mori, H. Miyata, M. Akahane, Y. Ajisawa, and H. Okudaira. 1998. Regulation of interleukin-1␤-induced interleukin-6 gene expression in human fibroblast-like synoviocytes by p38 mitogen-activated protein kinase. J. Biol. Chem. 273:24832. 20. Xu, N., C. Y. Chen, and A. B. Shyu. 1997. Modulation of the fate of cytoplasmic mRNA by AU-rich elements: key sequence features controlling mRNA deadenylation and decay. Mol. Cell. Biol. 17:4611. 21. Lasa, M., K. R. Mahtani, A. Finch, G. Brewer, J. Saklatvala, and A. R. Clark. 2000. Regulation of cyclooxygenase 2 mRNA stability by the mitogen-activated protein kinase p38 signaling cascade. Mol. Cell. Biol. 20:4265. 22. Winzen, R., M. Kracht, B. Ritter, A. Wilhelm, C. Y. Chen, A. B. Shyu, M. Muller, M. Gaestel, K. Resch, and H. Holtmann. 1999. The p38 MAP kinase pathway signals for cytokine-induced mRNA stabilization via MAP kinase-activated protein kinase 2 and an AU-rich region-targeted mechanism. EMBO J. 18:4969.

The Journal of Immunology 23. Chen, K. D., L. Y. Chen, H. L. Huang, C. H. Lieu, Y. N. Chang, M. D. Chang, and Y. K. Lai. 1998. Involvement of p38 mitogen-activated protein kinase signaling pathway in the rapid induction of the 78-kDa glucose-regulated protein in 9L rat brain tumor cells. J. Biol. Chem. 273:749. 24. Han, J., Y. Jiang, Z. Li, V. V. Kravchenko, and R. J. Ulevitch. 1997. Activation of the transcription factor MEF2C by the MAP kinase p38 in inflammation. Nature 386:296. 25. Wang, X. Z., and D. Ron. 1996. Stress-induced phosphorylation and activation of the transcription factor CHOP (GADD153) by p38 MAP kinase. Science 272:1347. 26. Raingeaud, J., A. J. Whitmarsh, T. Barrett, B. Derijard, and R. J. Davis. 1996. MKK3- and MKK6-regulated gene expression is mediated by the p38 mitogenactivated protein kinase signal transduction pathway. Mol. Cell. Biol. 16:1247. 27. Vermeulen, L., G. De Wilde, P. Van Damme, W. Vanden Berghe, and G. Haegeman. 2003. Transcriptional activation of the NF-␬B p65 subunit by mitogen- and stress-activated protein kinase-1 (MSK1). EMBO J. 22:1313. 28. Saccani, S., S. Pantano, and G. Natoli. 2002. p38-Dependent marking of inflammatory genes for increased NF-␬B recruitment. Nat. Immunol. 3:69. 29. Bowie, A. G., and L. A. O’Neill. 2000. Vitamin C inhibits NF-␬B activation by TNF via the activation of p38 mitogen-activated protein kinase. J. Immunol. 165:7180. 30. Alpert, D., P. Schwenger, J. Han, and J. Vilcek. 1999. Cell stress and MKK6bmediated p38 MAP kinase activation inhibit tumor necrosis factor-induced I␬B phosphorylation and NF-␬B activation. J. Biol. Chem. 274:22176. 31. Buchan, G., K. Barrett, M. Turner, D. Chantry, R. N. Maini, and M. Feldmann. 1988. Interleukin-1 and tumour necrosis factor mRNA expression in rheumatoid arthritis: prolonged production of IL-1␣. Clin. Exp. Immunol. 73:449. 32. He, T. C., S. Zhou, L. T. da Costa, J. Yu, K. W. Kinzler, and B. Vogelstein. 1998. A simplified system for generating recombinant adenoviruses. Proc. Natl. Acad. Sci. USA 95:2509. 33. Udalova, I. A., A. Richardson, A. Denys, C. Smith, H. Ackerman, B. Foxwell, and D. Kwiatkowski. 2000. Functional consequences of a polymorphism affecting NF-␬B p50 –p50 binding to the TNF promoter region. Mol. Cell. Biol. 20:9113. 34. Sanlioglu, S., C. M. Williams, L. Samavati, N. S. Butler, G. Wang, P. B. McCray, Jr., T. C. Ritchie, G. W. Hunninghake, E. Zandi, and J. F. Engelhardt. 2001. Lipopolysaccharide induces Rac1-dependent reactive oxygen species formation and coordinates tumor necrosis factor-␣ secretion through IKK regulation of NF-␬B. J. Biol. Chem. 276:30188.

6937 35. Bondeson, J., F. Brennan, B. Foxwell, and M. Feldmann. 2000. Effective adenoviral transfer of I␬B␣ into human fibroblasts and chondrosarcoma cells reveals that the induction of matrix metalloproteinases and proinflammatory cytokines is nuclear factor-␬B dependent. J. Rheumatol. 27:2078. 36. Hunt, A. E., F. V. Lali, J. D. Lord, B. H. Nelson, T. Miyazaki, K. J. Tracey, and B. M. Foxwell. 1999. Role of interleukin (IL)-2 receptor ␤-chain subdomains and Shc in p38 mitogen-activated protein (MAP) kinase and p54 MAP kinase (stressactivated protein kinase/c-Jun N-terminal kinase) activation: IL-2-driven proliferation is independent of p38 and p54 MAP kinase activation. J. Biol. Chem. 274:7591. 37. Baldassare, J. J., Y. Bi, and C. J. Bellone. 1999. The role of p38 mitogen-activated protein kinase in IL-1␤ transcription. J. Immunol. 162:5367. 38. Denys, A., I. A. Udalova, C. Smith, L. M. Williams, C. J. Ciesielski, J. Campbell, C. Andrews, D. Kwaitkowski, and B. M. Foxwell. 2002. Evidence for a dual mechanism for IL-10 suppression of TNF-␣ production that does not involve inhibition of p38 mitogen-activated protein kinase or NF-␬B in primary human macrophages. J. Immunol. 168:4837. 39. Horwood, N. J., T. Mahon, J. P. McDaid, J. Campbell, H. Mano, F. M. Brennan, D. Webster, and B. M. Foxwell. 2003. Bruton’s tyrosine kinase is required for lipopolysaccharide-induced tumor necrosis factor ␣ production. J. Exp. Med. 197:1603. 40. Kumar, S., J. Boehm, and J. C. Lee. 2003. p38 MAP kinases: key signalling molecules as therapeutic targets for inflammatory diseases. Nat. Rev. Drug. Discov. 2:717. 41. Nick, J. A., N. J. Avdi, S. K. Young, L. A. Lehman, P. P. McDonald, S. C. Frasch, M. A. Billstrom, P. M. Henson, G. L. Johnson, and G. S. Worthen. 1999. Selective activation and functional significance of p38␣ mitogen-activated protein kinase in lipopolysaccharide-stimulated neutrophils. J. Clin. Invest. 103:851. 42. Kuprash, D. V., I. A. Udalova, R. L. Turetskaya, D. Kwiatkowski, N. R. Rice, and S. A. Nedospasov. 1999. Similarities and differences between human and murine TNF promoters in their response to lipopolysaccharide. J. Immunol. 162:4045. 43. Carter, A. B., K. L. Knudtson, M. M. Monick, and G. W. Hunninghake. 1999. The p38 mitogen-activated protein kinase is required for NF-␬B-dependent gene expression: the role of TATA-binding protein (TBP). J. Biol. Chem. 274:30858. 44. Andreakos, E., C. Smith, S. Kiriakidis, C. Monaco, R. de Martin, F. M. Brennan, E. Paleolog, M. Feldmann, and B. M. Foxwell. 2003. Heterogeneous requirement of I␬B kinase 2 for inflammatory cytokine and matrix metalloproteinase production in rheumatoid arthritis: implications for therapy. Arthritis. Rheum. 48:1901. 45. Brennan, F. M. 2001. A follow-up to “Anti-cytokine therapy in chronic destructive arthritis” by Wim B van den Berg. Arthritis Res. 3:211.