Up-regulation of microsomal prostaglandin E synthase 1 in osteoarthritic human cartilage: Critical roles of the ERK-1/2 and p38 signaling pathways

ARTHRITIS & RHEUMATISMVol. 50, No. 9, September 2004, pp 2829–2838DOI 10.1002/art.20437© 2004, American College of Rheumatology

Up-Regulation of Microsomal Prostaglandin E Synthase 1 inOsteoarthritic Human Cartilage

Critical Roles of the ERK-1/2 and p38 Signaling Pathways

Kayo Masuko-Hongo,1 Francis Berenbaum,2 Lydie Humbert,1 Colette Salvat,1

Mary B. Goldring,3 and Sylvie Thirion1

Objective. Microsomal prostaglandin E synthase 1(mPGES-1) is the final enzyme of the cascade thatproduces prostaglandin E2 (PGE2), a key actor inarthritis. To study mPGES-1 synthesis in human carti-lage and its regulation by interleukin-1� (IL-1�), weused human cartilage and an immortalized humanchondrocyte cell line. Furthermore, we investigated thesignaling pathways involved in mPGES-1 expression.

Methods. We used real-time quantitative reversetranscription–polymerase chain reaction, Northernblotting, and Western blotting to measure mPGES-1messenger RNA (mRNA) and protein expression inhuman chondrocytes. PGE2 production was measuredby enzyme-linked immunosorbent assay.

Results. Cartilage specimens from osteoarthritis(OA) patients contained far greater amounts ofmPGES-1 and cyclooxygenase 2 (COX-2) mRNA thandid normal cartilage. Incubation with IL-1� markedly

increased mPGES-1 mRNA and protein in a dose-dependent and time-dependent manner, in parallel withan increase in PGE2 levels. Both PD98059, an ERKpathway inhibitor, and SB203580, a p38�/� MAPKinhibitor, abolished the increases in mPGES-1 mRNAand protein in response to IL-1�. The specific p38�MAPK inhibitor SC906 suppressed IL-1�–inducedCOX-2 expression but not IL-1�–induced mPGES-1expression, suggesting preferential involvement of p38�MAPK in IL-1�–induced mPGES-1 expression.

Conclusion. This study is the first to show thatmPGES-1 is stimulated in human chondrocytes by theproinflammatory cytokine IL-1� via activation of bothERK-1/2 and p38 MAPK in an isoform-specific manner.We postulate that mPGES-1 may be a novel target forOA therapy.

Prostaglandin E2 (PGE2) plays an important rolein cartilage metabolism. Its many effects on chondro-cytes (for review, see ref. 1) include enhanced produc-tion of matrix metalloproteinase 3 (2), modulation ofproteoglycan and collagen synthesis (2,3), and stimula-tion of chondrocyte apoptosis (4,5). The synovial fluid ofpatients with osteoarthritis (OA) and rheumatoid arthri-tis contains high concentrations of PGE2 (4), and carti-lage and chondrocytes from OA patients spontaneouslyrelease more PGE2 than does normal cartilage (2,6,7).These results suggest that PGE2 may be actively involvedin cartilage breakdown in OA patients.

PGE2 is synthesized by the isoenzymes cyclooxy-genase 1 (COX-1) and COX-2, both of which act onarachidonic acid to form the prostaglandin endoperox-ide H2 (PGH2). The prostanoid synthase prostaglandinE synthase (PGES) produces PGE2 from PGH2. Three

Supported by the Centre National de la Recherche Scienti-fique, Universite Pierre et Marie Curie, the Association de Recherchesur la Polyarthrite, and the Societe Francaise de Rhumatologie. Dr.Masuko-Hongo was recipient of a Rheumatology Traveling Fellowshipfrom the Japan Rheumatism Foundation. Dr. Goldring’s work wassupported by NIH grants R01-AG-22021 and AR-45378.

1Kyo Masuko-Hongo, MD, Lydie Humbert, Research Engi-neer, Colette Salvat, Research Engineer, Sylvie Thirion, PhD (currentaddress: UMR CNRS 6544, Universite de la Mediterranee, Marseilles,France): UMR CNRS 7079, Universite Pierre et Marie Curie (ParisVI), Paris, France; 2Francis Berenbaum, MD, PhD: UMR CNRS 7079and Rheumatology, UFR St. Antoine, and Universite Pierre et MarieCurie, Paris, France; 3Mary B. Goldring, PhD: Beth Israel DeaconessMedical Center, and New England Baptist Bone and Joint Institute,Harvard Institutes of Medicine, Boston, Massachusetts.

Address correspondence and reprint requests to FrancisBerenbaum, MD, PhD, UMR CNRS 7079, Universite Pierre et MarieCurie (Paris VI), 7 Quai St. Bernard, 75252 Paris Cedex 5, France.E-mail: [email protected].

Submitted for publication January 13, 2004; accepted inrevised form April 13, 2004.

2829

distinct types of PGES have been cloned: one cytosolictype (cPGES) and two membrane-associated types(glutathione-specific mPGES-1 and glutathione-nonspecific mPGES-2) (8). While cPGES is ubiquitous(9), mPGES-1 is not. This latter type of PGES is ahomotrimer that has been described as a regulatedenzyme (10), the production of which is stimulated byproinflammatory agents in several cells and tissues, suchas A549 and HeLa cancer cell lines (11), macrophages(12), vascular smooth muscle cells (13), rheumatoidsynovial cells (14,15), orbital fibroblasts (16), and ath-erosclerotic plaques (17). Several recent studies haveshown that mPGES-1 is critical for inflammatory re-sponses in vivo, such as in lipopolysaccharide-inducedpyresis (18,19), in mPGES-1–deficient mice (20,21), andin the rat model of adjuvant-induced arthritis (22).Although the role of PGE2 in cartilage is well docu-mented in inflammatory conditions, it is not knownwhether mPGES-1 is expressed in the cartilage of pa-tients with arthritis or whether its expression by chon-drocytes is regulated by proinflammatory agents. Wetherefore evaluated mPGES-1 expression in inflamma-tory conditions, i.e., in OA cartilage and underinterleukin-1� (IL-1�) stimulation.

We used primary cultures of human tissue andthe immortalized human chondrocyte cell line T/C-28a2,which was established by transfecting human juvenilecostal chondrocytes with SV40 (23). This cell line retainssome of the specific phenotype of chondrocytes inlong-term culture and is thus a suitable model forreproducible studies of chondrocyte metabolism (24–26). In chondrocytes, IL-1� and tumor necrosis factorrapidly activated the various members of the MAPKfamily, including ERK, JNK, and p38 MAPK (27–29).Therefore, we determined the influences of ERK-1/2and p38 MAPK on mPGES-1 expression. Our resultsshow that chondrocytes from OA cartilage synthesizemPGES-1, and that this synthesis is regulated by IL-1�acting via the ERK-1/2 and p38 MAPK signaling path-ways. The p38 MAPK signaling pathway may act in anisoform-specific manner.

MATERIALS AND METHODS

Materials. All reagents were purchased from Sigma-Aldrich (St. Quentin Fallavier, France), unless stated other-wise. Fetal calf serum (FCS) was obtained from Invitrogen(Cergy-Pontoise, France). Collagenase A, trypsin, hyaluroni-dase, and Complete protease inhibitor mixture were fromRoche Diagnostics (Meylan, France). Recombinant humanIL-1� was from PeproTech (Rocky Hill, NJ). Anti-mouseCOX-2 monoclonal antibody, anti-rabbit cPGES polyclonal

antibody, and anti-rabbit mPGES-1 polyclonal antibody werefrom SPI-BIO for Cayman (Massy, France). Anti-mouse p38monoclonal antibody was obtained from Tebu for Santa CruzBiotechnology (Le Perray-en-Yvelines, France), anti-rabbitphospho-ERK and anti-rabbit phospho-p38 antibodies werefrom New England Biolabs (Beverly, MA), and anti-rabbittotal ERK antibody was from Upstate Biotechnology (LakePlacid, NY). SB203580 and PD98059 were obtained fromCalbiochem (Meudon, France). The specific diarylpyrazoleclass p38� inhibitor SC906 (patent no. W000/31063) was a giftfrom Dr. J. P. Portanova (Pharmacia, St. Louis, MO). TheECL Western blot analysis system was purchased from Amer-sham Pharmacia Biotech (Orsay, France). The Immun-Blotpolyvinylidene difluoride (PVDF) membranes for Westernblotting and kaleidoscope prestained standards were obtainedfrom Bio-Rad (Ivry-sur-Seine, France).

Cell culture. We used an immortalized human chon-drocyte cell line (T/C-28a2) established by transfecting humanjuvenile costal chondrocytes with vectors encoding the simianvirus 40 large T antigen (23). The cells were maintained inDulbecco’s modified Eagle’s medium (DMEM) supplementedwith 10% FCS, 4 mM L-glutamine, 100 IU/ml of penicillin, and100 �g/ml of streptomycin on 10-cm plastic cell culture dishes(Corning, Corning, NY) at 37°C in an atmosphere of 5% CO2.Confluent cells were passaged routinely twice a week at a splitratio of 1:10.

Cartilage samples. Human knee cartilage specimenswere obtained from patients undergoing total knee arthro-plasty for OA at the St. Antoine Hospital. Informed consentwas obtained from each patient prior to surgery. The diagnosisof OA was based on clinical and radiographic evaluationsaccording to standard criteria (30). Our institutional ethicscommittee approved the study protocol.

Articular cartilage from the femoral condyles and tibialplateaus was cut into small pieces with a scalpel and digested at37°C with 0.05% hyaluronidase in phosphate buffered saline(PBS) for 30 minutes, with 0.2% trypsin for 45 minutes, andthen with 0.2% collagenase A in PBS for 120 minutes. Thechondrocytes were then washed with DMEM (4.5 gm/liter ofglucose) for 90 minutes. The resulting concentrated chondro-cyte suspensions were placed in 6- or 12-well plates (125,000cells/cm2) containing DMEM with 4.5 gm/liter of glucosesupplemented with 10% FCS, 100 IU/ml of penicillin, and 100�g/ml of streptomycin. The cells were allowed to grow for 5days and were then placed in serum-free DMEM containing0.3 % bovine serum albumin (BSA) for 24 hours. Cell viabilitybefore and after stimulation was assessed by trypan blueexclusion and was found to be similar under all the testconditions, indicating that the inhibitors were not cytotoxic(data not shown).

Northern blotting. Total RNA was extracted using theRNeasy kit (Qiagen, Hilden, Germany) according to themanufacturer’s instructions. Total RNA (15 �g), quantifiedspectrophotometrically and checked by agarose gel electro-phoresis, was denatured and electrophoresed on a 1% agarosegel containing formaldehyde. The resulting RNA was trans-ferred to a nylon membrane (Hybond-N; Amersham Interna-tional, Amersham, UK). The membrane was prehybridized for15 minutes in 1M sodium phosphate buffer containing 10%sodium dodecyl sulfate (SDS) and then hybridized overnight at65°C in the same solution with a human PGES probe (SPI-BIO

2830 MASUKO-HONGO ET AL

for Cayman) labeled with 32P-dCTP (1,000,000 counts perminute/ml). Membranes were washed with 0.1% SDS in 2�saline–sodium citrate at room temperature and exposed toKodak BioMax film (Kodak, Rochester, NY) at –80°C for 72hours and/or exposed to a Storage Phosphor Screen (Kodakfor Bio-Rad) for 3–5 days and scanned using a PhosphorIm-ager (Molecular Imager FX; Bio-Rad). The data thusobtained were quantified using Quantity One software(version 4.2.0; Bio-Rad). A 28S oligonucleotide probe la-beled with T4 polynucleotide kinase and �32P-ATP served asthe internal control.

Real-time reverse transcription–polymerase chain re-action (RT-PCR) assays. Total RNA was extracted using theRNeasy kit, and 1 �g was reverse transcribed with Omniscript(Qiagen) in a final volume of 20 �l containing 50 ng of randomhexamers. The enzyme was then inactivated by heating, andthe mPGES-1 messenger RNA (mRNA) was quantitated byreal-time quantitative RT-PCR using the iCycler iQ Real TimePCR (Bio-Rad) and QuantiTect SYBR green PCR kits (Qia-gen). The PCR primer sequences were designed to amplify a101-bp fragment: mPGES-1 sense, 5�-GGA-ACG-ACA-TGG-AGA-CCA-TC-3� and mPGES-1 antisense, 5�-GGA-AGA-CCA-GGA-AGT-GCA-TC-3�. GAPDH mRNA was used tocheck and normalize the mPGES-1 RNAs and RT efficiency:GAPDH sense, 5�-CCA-TCA-CCA-TCT-TCC-A-3� andGAPDH antisense, 5�-CCT-TCT-CCA-TGG-TGG-T-3�.

The PCR reactions were performed in 25-�l finalvolumes using 0.06–0.25 �l of complementary DNA (cDNA),400 ng of specific primers, and 1� QuantiTect SYBR GreenPCR master mixture, including HotStar Taq DNA polymerase,QuantiTect SYBR Green PCR buffer, SYBR Green I, ROX,and 5 mM MgCl. Samples were denatured for 15 minutes at94°C, then amplified for 40 cycles as follows: denaturation at94°C for 1 minute, annealing at 54°C for 1 minute, andextension at 72°C for 90 seconds. Product formation wasdetected at 72°C in the fluorescein isothiocyanate channel. Thegeneration of specific PCR products was confirmed by melting-curve analysis. For each real-time RT-PCR run, cDNA wererun in quadruplicate in parallel with serial dilutions of a cDNAmixture tested for each primer pair to generate a linearstandard curve (mean cycle threshold [Ct], the cycle at whichfluorescence was considered significantly greater than thebackground level and within the linear range, plotted againstthe log of the template concentration), which was used toestimate relative quantities of mPGES-1 mRNA normalizedfor GAPDH in the samples.

PGE2 assay. The PGE2 concentrations in aliquots ofsupernatants from cultures of stimulated chondrocytes weremeasured using a PGE2 enzyme immunoassay kit (CaymanChemical, Ann Arbor, MI), as previously described (31). PGE2concentrations were assayed in duplicate and were read againsta standard curve.

Extraction of cell lysates. Confluent cells were incu-bated with or without IL-1�, washed with ice-cold PBS, andlysed in cold lysis buffer (20 mM Tris HCl, pH 7.6, 150 mMNaCl, 2 mM EDTA, 1% Triton, 10% glycerol, and Completeprotease inhibitor mixture). The lysed cells were disrupted bysonication and centrifuged at 13,000g for 10 minutes at 4°C.The supernatant was used as the whole cell lysate. Microsomaland cytosolic proteins were prepared by homogenization in abuffer containing 0.1M potassium phosphate, pH 7.4, 0.25M

sucrose, 2 mM EDTA, and Complete protease inhibitor mix-ture. The sonicated cells were centrifuged at 1,000g, then at10,000g, and the supernatant was removed and centrifuged athird time, at 170,000g, to separate the microsomal fraction(pellet) from the cytosolic fraction (supernatant). Proteinconcentrations were determined using the bicinchoninic acidassay kit (Perbio Science for Pierce, Bezons, France).

Immunoblot analysis. Cell lysates were separated by10% or 15% SDS–polyacrylamide gel electrophoresis andtransferred to PVDF membranes (31). The blots were soakedin Tween 20 in Tris buffered saline (TBST; 20 mM Tris HCl,pH 7.5, 100 mM NaCl, 0.1% Tween 20, and 5% BSA) for 2hours at room temperature. The blots were incubated withprimary rabbit polyclonal antibody to mPGES-1. The mem-brane was then stripped and reprobed using murine monoclo-nal antibody to �-actin as a primary antibody. Anti–phospho–p38 MAPK, anti–phospho-ERK, anti-ERK, and anti–p38

Figure 1. Expression of arachidonic acid cascade genes in osteoarthri-tis. Real-time reverse transcription–polymerase chain reaction (RT-PCR) assay demonstrating increased microsomal prostaglandin Esynthase 1 (mPGES-1) gene expression in interleukin-1� (IL-1�)–stimulated human chondrocytes. Standard curves for mPGES-1 andGAPDH were generated by serial dilution of a cDNA mixture. Theamount of mPGES-1 mRNA was normalized against the amount ofGAPDH mRNA measured in the same cDNA. All values are relativeto those in saline-treated control chondrocytes (C). Values are themean and SEM of triplicate experiments. �� P � 0.001 versuscontrol.

UP-REGULATION OF mPGES-1 BY IL-1� IN OA CARTILAGE 2831

MAPK antibodies were used to investigate MAPK activation.The blots were then incubated with horseradish peroxidase–conjugated secondary goat anti-rabbit or rabbit anti-mouseIgG antibody for 2 hours at room temperature. The mem-branes were washed repeatedly with TBST, and the signalswere detected using the enhanced chemiluminescence detec-tion system and exposed to Kodak BioMax MR-1 film.

Statistical analysis. Data are presented as the mean �SEM unless indicated otherwise. Data populations were testedfor normal distribution and equal standard deviations byStudent’s t-test. We used GraphPad InStat version 3.05 forWindows 95/NT (GraphPad Software, San Diego, CA). Pvalues less than 0.05 were considered significant.

RESULTS

Induction of mPGES-1 in human OA cartilageand human chondrocytes incubated with IL-1�. Induc-ible mPGES-1, which catalyzes PGE2 production, wasassayed in human chondrocytes to evaluate the regula-tion of PGE2 synthesis in cartilage. In previous studies,

mRNA for COX-2 and mPGES, and the COX-2 productPGE2, were found to be up-regulated in OA cartilage ascompared with normal cartilage (6,32). The concentra-tion of IL-1� was previously found to be markedlyelevated in the synovial fluid and cartilage of patientswith various arthropathies (33), including OA (34). Wetherefore used real-time RT-PCR to investigate theinfluence of IL-1� on mPGES-1 expression byprimary cultures of articular chondrocytes from OApatients. We found that IL-1� increased the amount ofmPGES-1 mRNA in human OA articular chondrocytes(Figure 1).

Mediation of IL-1� induction of mPGES-1 at thepretranslational level. Since only limited amounts ofhuman tissue were available, we used the immortalizedchondrocyte cell line T/C-28a2, which retains the char-acteristics of chondrocytes and provides enough cells forextensive studies (24,25,35). We used Northern blotting

Figure 2. Dose-dependent expression of mPGES-1 in response to IL-1�. T/C-28a2 humanchondrocytes were harvested and incubated for 24 hours with or without increasingconcentrations of IL-1�. The mPGES-1 mRNA was assessed by A, Northern blotting and B,real-time RT-PCR, and mPGES-1 proteins were assayed by C, immunoblotting. A, TotalRNA was subjected to Northern blot hybridization; the radioactive RNA/DNA hybrids werevisualized by exposing membranes to BioMax film. B, Real-time RT-PCR analysis wasperformed as described in Figure 1. Values are the mean and SEM of 5 independentexperiments. �� P � 0.01 versus control. C, Cell lysates were analyzed by sodium dodecylsulfate–polyacrylamide gel electrophoresis using 15%-gradient gels. Proteins were trans-ferred to a nylon membrane and blotted with anti–mPGES-1 antibody, then the blot wasstripped and reprobed with anti–�-actin antibody. Each blot is representative of 2–5independent experiments. D, Chondrocytes were maintained in medium alone or in mediumcontaining the indicated concentrations of IL-1� for 24 hours, and the amount ofprostaglandin E2 (PGE2) released into the supernatant (ng/ml of supernatant) wasmeasured. Values are the mean and SEM of 4 separate experiments. �� P � 0.01 versuscontrol. See Figure 1 for other definitions.


to measure mPGES-1 mRNA. T/C-28a2 human chon-drocytes contained small amounts of mPGES-1 mRNA(�2,000 bases) under basal conditions (Figure 2A).Stimulation with IL-1� significantly increased theamount of mPGES-1 mRNA (Figure 2A). The results ofquantitative real-time RT-PCR analysis were consistentwith those of Northern blotting (Figure 2B). We thenassessed mPGES-1 protein levels by immunoblottingusing a polyclonal antibody specific for mPGES-1. Theamount of mPGES-1 protein in T/C-28a2 human chon-drocytes increased in response to IL-1� in a dose-dependent manner (Figure 2C). The relationship be-tween mPGES-1 and PGE2 production by chondrocyteswas evaluated by measuring secreted PGE2. PGE2 se-cretion correlated with mPGES-1 concentrations afterIL-1� stimulation (Figure 2D), indicating that mPGES-1promoted PGE2 synthesis by chondrocytes.

We analyzed the rate of mPGES-1 expression byT/C-28a2 human chondrocytes stimulated with IL-1�.The amount of mPGES-1 increased during incubation,peaking after 18–24 hours and remaining high for up to48 hours (Figures 3A and B). The amounts of mPGES-1protein increased in parallel with the amounts of mRNAmPGES-1 (Figure 3C). Thus, PGE2 was produced dur-ing incubation (Figure 3D).

Since PGES exists as cytosolic and microsomalisoforms, we attempted to confirm that IL-1�–inducedPGES was found only in the microsomal fraction. To thisend, we separated T/C-28a2 human chondrocyte lysatesinto microsomal and cytosolic fractions, which we thenexamined by Western blotting. IL-1� stimulated theexpression of PGES in the microsomal fraction but notin the cytosolic fraction (Figure 4). Thus, mPGES-1 wasproduced by T/C-28a2 human chondrocytes, and itssynthesis was regulated by IL-1� in a dose-dependentand time-dependent manner.

Stimulation of p38� MAPK and mPGES-1 byIL-1� in human chondrocytes. The signaling pathwayresponsible for mPGES-1 synthesis by T/C-28a2 humanchondrocytes stimulated with IL-1� was identified byusing specific inhibitors of p38 MAPK and ERK-1/2MAPK. First, to determine whether IL-1� increasedERK, p38, and JNK MAPK activities in T/C-28a2 chon-drocytes, we used Western blotting to detect the phos-phorylated (active) and total forms of ERK-1/2, p38, andJNK MAPKs. T/C-28a2 cells contained both p38 andERK MAPK (Figure 5). Activation of p38 MAPKoccurred within 5 minutes of IL-1� treatment andpersisted for up to 24 hours (Figure 5). In contrast,ERK-1/2 activation was transient and returned to the

Figure 3. Rate of mPGES-1 expression in response to IL-1�. T/C-28a2 human chondro-cytes were incubated with 10 ng/ml of IL-1� for the indicated times, and mPGES-1 wasanalyzed by A, Northern blotting, B, real-time RT-PCR, and C, immunoblotting. Each blotshown in A is a single representative image of 2 independent experiments. The mean andSEM results of 4 independent experiments are shown in B (� � P � 0.05 versus control).Blots shown in C are representative of 3 experiments. D, Amount of prostaglandin E2

(PGE2) released into the supernatant in response to IL-1�. Values are the mean and SEMof 4 independent experiments. � � P � 0.05; �� P � 0.01; �� P � 0.001 versus control.See Figure 1 for other definitions.


control level within 1 hour. No JNK MAPK or activationby IL-1� was found in T/C-28a2 cells (data not shown).These results indicate that p38 and ERK-1/2 MAPKswere synthesized in T/C-28a2 human chondrocytes andwere activated by stimulation with IL-1�.

To assess the involvement of MAPK in IL-1�–induced stimulation of mPGES-1 synthesis by T/C-28a2human chondrocytes, we used several MAPK inhibitors.The ERK inhibitor PD98059 inhibited mPGES-1mRNA and protein synthesis (Figure 6). SB203580, aninhibitor of p38 MAPK through its �/�-isoforms, mark-edly inhibited mPGES-1 mRNA and protein synthesis(Figures 7A and B), indicating that p38�/� MAPKactivation was critical for mPGES-1 production by T/C-

28a2 human chondrocytes. In contrast, SC906, a specificinhibitor of the �-isoform of p38 MAPK, did not inhibitmPGES-1 production (Figures 7A and B), indicatingthat this isoform played no major role in mPGES-1synthesis. IL-1�–induced stimulation of PGE2 produc-tion was completely abolished by SB203580 and wasstrongly inhibited by SC906 (Figure 7C). Thus, while the�-isoform of p38 MAPK seemed to be specificallyFigure 4. Detection of inducible PGES in the microsomal fraction.

T/C-28a2 human chondrocytes were unstimulated or were stimulatedwith 10 ng/ml of IL-1� for 24 hours. The cell extracts were thencentrifuged to separate the microsomal (Micros.) and cytosolic (Cy-tos.) fractions. The mPGES-1 and cPGES proteins in each samplewere detected by immunoblotting. The same samples were probed withanti–�-actin antibody as the internal control. Similar results wereobtained from 2 separate experiments. See Figure 1 for other defini-tions.

Figure 5. Interleukin-1� (IL-1�)–induced activation of p38 and ERK-1/2 MAPKs in T/C-28a2 human chondrocytes. Whole cell lysates fromT/C-28a2 chondrocytes were prepared at time 0 and at 5, 10, 20, and30 minutes and at 1, 6, and 24 hours after stimulation with 10 ng/ml ofIL-1�. The lysates were then analyzed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis using 10%-gradient gels. The pro-teins were transferred and blotted with anti–phospho–p38 MAPKantibody (P-p38) or anti–phospho-ERK antibody (P-p44 and P-p42).Blots were then stripped and reprobed with anti–native p38 (p38) andanti–native ERK (p44 and p42) antibodies. Each blot is representativeof 2–4 independent experiments.

Figure 6. Involvement of ERK-1/2 MAPK in the expression of micro-somal prostaglandin E synthase 1 (mPGES-1) in response tointerleukin-1� (IL-1�). T/C-28a2 chondrocytes were incubated with orwithout recombinant human IL-1� (10 ng/ml) for 24 hours in thepresence or absence of the ERK-1/2 inhibitor PD98059. The effect ofthe inhibitor on mPGES-1 expression was then analyzed by A, real-time reverse transcription–polymerase chain reaction, and B, immuno-blotting. Values in A are the mean and SEM of 6 experiments (� � P �0.05; �� P � 0.001 versus control). Blots in B are representative of3 independent experiments. C, Prostaglandin E2 (PGE2) released intothe supernatant in response to IL-1� and treatment with inhibitor.Values are the mean and SEM of 4 independent experiments. �� P � 0.001 versus control.


involved in IL-1�–induced stimulation of mPGES-1synthesis, the �-isoform apparently influenced anotherstep in the PGE2 synthesis pathway.

Expression of mPGES-1 by human articularchondrocytes incubated with IL-1�. To check our find-ings with T/C-28a2 cells, we measured mPGES-1 protein

in primary cultured articular chondrocytes from 3 pa-tients with OA. IL-1� significantly stimulated mPGES-1synthesis by these human articular chondrocytes,whereas PD98059 and SB203580 abolished this effect ofIL-1�; SC906 had no effect (Figure 8). These resultswere consistent with those obtained with the T/C-28a2cells. In primary cultured human chondrocytes, COX-2synthesis was blocked by both p38 MAPK inhibitors(SB203580 and SC906), whereas mPGES-1 inhibitionvaried across isoforms: the p38�/� inhibitor was effec-tive, but the p38� inhibitor was not. These resultsconfirmed that IL-1� regulated mPGES-1 expression viathe ERK and p38 MAPK signaling pathways, both inhuman articular chondrocytes and in immortalized hu-man chondrocytes.

DISCUSSION

We and other investigators have reported thatIL-1� stimulates the expression of secreted phospho-lipase A2 (36) and COX-2 (37) genes in articular chon-drocytes, resulting in increased PGE2 production. Here,we focused on PGES, the last enzyme in the PGE2biosynthesis pathway. Our results constitute the firstevidence of overexpression of mPGES-1, but notcPGES, in OA cartilage and in human articular chon-drocytes stimulated with IL-1�. Previous studies haveshown mPGES-1 overexpression in rheumatoid synovio-cytes stimulated with IL-1 or PGE2 (14,15). Moreover,recent studies in mice lacking mPGES-1 found markedlyreduced inflammatory responses in the collagen-inducedarthritis model (21). These studies, together with ourresults, strengthen the hypothesis that targetingmPGES-1 might hold promise as a novel antiinflamma-tory strategy in arthritis.

Although nonsteroidal antiinflammatory drugs,including selective COX-2 inhibitors, also abolish PGE2synthesis, new antiinflammatory treatments are needed.Moreover, studies in various experimental models haveshown disruption of homeostasis by COX-2 inhibition,the result being a prothrombotic state. COX-2 inhibitioncauses an imbalance between the COX-1–dependentproaggregant platelet mediator thromboxane A2 and theCOX-2–dependent vasodilator and endothelial media-tor prostacyclin (38). This prothrombotic state mayoffset the potential benefits of the decreased gastroin-testinal toxicity reported with selective COX-2 inhibitors(39). Thus, selective modulation of the prostanoid path-way achieved by targeting the PGE2-forming enzymedownstream of COX-2 might provide greater benefits.

IL-1�–induced expression of mPGES-1 mRNA

Figure 7. Involvement of p38 MAPK in microsomal prostaglandin Esynthase 1 (mPGES-1) expression in immortalized human chondro-cytes. T/C-28a2 chondrocytes were incubated with or without recom-binant human interleukin-1� (IL-1�; 10 ng/ml) for 24 hours in thepresence or absence of SB203580, an inhibitor of p38�/� MAPK, orSC906, a specific inhibitor of p38� MAPK. The effects of the inhibitors(inh) were analyzed by A, real-time reverse transcription–polymerasechain reaction and B, immunoblotting. Values in A are the mean andSEM of 8 experiments (�� P � 0.01; �� P � 0.001 versuscontrol). Blots in B are representative of 4 experiments. C, Prostaglan-din E2 (PGE2) released into the supernatant in response to IL-1� andtreatment with inhibitors. Values are the mean and SEM of 4experiments. �� P � 0.01; �� P � 0.001 versus control.


and protein was found to be correlated with increases inCOX-2 mRNA and protein in primary cultures ofhuman chondrocytes. These results are consistent withprevious studies in other cell types suggesting spatial andtemporal coupling of both enzymes, the result beingabundant production of PGE2 at sites of inflammation(19,40).

IL-1� has been reported to activate the 3 mainMAPKs, namely, ERK-1/2, p38, and JNK MAPKs, inhuman articular chondrocytes (27). Here, we haveshown that ERK-1/2 and p38 MAPK activation alsodepends on IL-1� in T/C-28a2 chondrocytes. IL-1�activates ERK-1/2 rapidly and transiently but activatesp38 MAPK in a more sustained manner. Moreover,mPGES-1 gene stimulation by IL-1� in articular chon-drocytes depends on both ERK-1/2 and p38 MAPKs. Incontrast, in previous studies of immortalized T/AC62human articular chondrocytes, we found that IL-1–induced COX-2 gene expression was exclusively medi-ated by p38 MAPK (29). From these studies, it wouldappear that IL-1� may use overlapping, but distinct,signaling pathways to induce COX-2 and mPGES-1expression in chondrocytes (Figure 9).

Four isoforms of p38 MAPK (�, �, �, �) havebeen identified to date (for review, see ref. 41). They are

Figure 8. Selective inhibition of the expression of interleukin-1� (IL-1�)–stimulated microsomalprostaglandin E synthase 1 (mPGES-1) and cyclooxygenase 2 (COX-2) in primary cultures ofhuman articular chondrocytes. Articular chondrocytes from osteoarthritic (OA) joints were treatedfor 24 hours with IL-1� (10 ng/ml) in the presence or absence of SB203580, an inhibitor of p38�MAPK, SC906, a specific inhibitor of p38� MAPK, or PD98059, an inhibitor of the ERK-1/2pathway. A, Effects of the inhibitors on the increased levels of mPGES-1 and COX-2 proteins afterIL-1� treatment. Blots are representative of 1 OA patient; similar results were obtained withchondrocytes from 3 OA patients. B, Prostaglandin E2 (PGE2) released into the supernatant inresponse to IL-1� and treatment with inhibitors. Values are the mean and SEM of 3 independentexperiments, each performed in triplicate. � � P � 0.05; �� P � 0.001 versus control.

Figure 9. Proposed signal transduction pathway that leads to theinduction of microsomal prostaglandin E synthase 1 (mPGES-1) byinterleukin-1� (IL-1�) in human articular chondrocytes. Nanomolarconcentrations of IL-1� bind to the IL-1� receptor (IL-1R), causingphosphorylation of p38 and ERK-1/2 MAPKs. The resulting expres-sion of cyclooxygenase 2 (COX-2) and mPGES-1 genes may bedifferentially regulated by MAPK-mediated signaling. PG � prosta-glandin.


activated by distinct MAPK kinases (42) and may havedifferent substrate specificities and distributions (43).The pyridinyl imidazole MAPK inhibitor SB203580 in-hibits p38� and p38�, but not p38� or p38� (41). Thus,blocking of mPGES-1 gene expression by SB203580(Figures 7 and 8) implies that the �- and/or �-isoformsare responsible for PGES gene expression in humanchondrocytes. The novel p38 MAPK inhibitor SC906belongs to the diarylpyrazole class and specificallyblocks activation of the �-isoform of p38 MAPK (patentno. W000/31063). This SC906 did not inhibit mPGES-1mRNA or protein expression in T/C-28a2 cells or inprimary cultured human chondrocytes, suggesting thatIL-1� stimulated mPGES-1 gene expression in chondro-cytes via a pathway that was not specific for p38�MAPK.

Taken together, our results suggest that p38�, butnot p38�, may be important in IL-1–induced mPGES-1expression in human chondrocytes. This finding suggestspotential therapeutic targets for combating PGE2-related inflammation in cartilage, most notably, ERK-1/2 and p38 MAPKs. MAPK activation may be involvedin inflammatory and degenerative arthropathies (44,45),and several therapeutic trials have indicated that p38MAPK inhibitors can halt joint disease (46–48). How-ever, the use of MAPK inhibitors in clinical practice islimited by the ability of these compounds to blockoverall MAPK activity. Because MAPK activity is ubiq-uitous, MAPK inhibitors may have unwanted effects onnormal tissues. Thus, inhibitors of specific MAPK path-ways are needed. Targeting p38� MAPK may inhibit theproinflammatory mediator PGE2 without interferingwith the prostacyclin pathway. However, this requiresconfirmation in endothelial cells.

In summary, we have demonstrated that themPGES-1 gene is regulated by IL-1� via ERK-1/2 andputative �-isoform signaling of the p38 MAPK pathwayin human chondrocytes. Further research into the regu-lation of inflammation-associated prostaglandins shouldhelp to unravel the mechanisms underlying cartilagedestruction in inflammatory disorders, thus generatingnew approaches to the development of preventive strat-egies.

ACKNOWLEDGMENT

The authors thank Prof. Alain Sautet (Hopital Saint-Antoine, Paris, France) for providing the human cartilagespecimens.

REFERENCES

1. Pelletier JP, Martel-Pelletier J, Abramson SB. Osteoarthritis, aninflammatory disease: potential implication for the selection ofnew therapeutic targets. Arthritis Rheum 2001;44:1237–47.

2. Abramson SB. The role of COX-2 produced by cartilage inarthritis. Osteoarthritis Cartilage 1999;7:380–1.

3. Goldring MB, Suen LF, Yamin R, Lai WF. Regulation of collagengene expression by prostaglandins and interleukin-1� in culturedchondrocytes and fibroblasts. Am J Ther 1996;3:9–16.

4. Miwa M, Saura R, Hirata S, Hayashi Y, Mizuno K, Itoh H.Induction of apoptosis in bovine articular chondrocytes by prosta-glandin E2 through cAMP-dependent pathway. OsteoarthritisCartilage 2000;8:17–24.

5. Notoya K, Jovanovic DV, Reboul P, Martel-Pelletier J, Mineau F,Pelletier JP. The induction of cell death in human osteoarthritischondrocytes by nitric oxide is related to the production ofprostaglandin E2 via the induction of cyclooxygenase-2. J Immunol2000;165:3402–10.

6. Amin AR, Attur M, Patel RN, Thakker GD, Marshall PJ, RediskeJ, et al. Superinduction of cyclooxygenase-2 activity in humanosteoarthritis-affected cartilage: influence of nitric oxide. J ClinInvest 1997;99:1231–7.

7. Jacques C, Sautet A, Moldovan M, Thomas B, Humbert L,Berenbaum F. Cyclooxygenase activity in chondrocytes from os-teoarthritic and healthy cartilage. Rev Rhum Engl Ed 1999;66:701–4.

8. Tanikawa N, Ohmiya Y, Ohkubo H, Hashimoto K, Kangawa K,Kojima M, et al. Identification and characterization of a novel typeof membrane-associated prostaglandin E synthase. Biochem Bio-phys Res Commun 2002;291:884–9.

9. Tanioka T, Nakatani Y, Semmyo N, Murakami M, Kudo I.Molecular identification of cytosolic prostaglandin E2 synthasethat is functionally coupled with cyclooxygenase-1 in immediateprostaglandin E2 biosynthesis. J Biol Chem 2000;275:32775–82.

10. Thoren S, Weinander R, Saha S, Jegerschold C, Pettersson PL,Samuelsson B, et al. Human microsomal prostaglandin E syn-thase-1: purification, functional characterization and projectionstructure determination. J Biol Chem. 2003;278:22199–209.

11. Jakobsson PJ, Thoren S, Morgenstern R, Samuelsson B. Identifi-cation of human prostaglandin E synthase: a microsomal, gluta-thione-dependent, inducible enzyme, constituting a potentialnovel drug target. Proc Natl Acad Sci U S A 1999;96:7220–5.

12. Dieter P, Scheibe R, Jakobsson PJ, Watanabe K, Kolada A,Kamionka S. Functional coupling of cyclooxygenase 1 and 2 todiscrete prostanoid synthases in liver macrophages. BiochemBiophys Res Commun 2000;276:488–92.

13. Soler M, Camacho M, Escudero JR, Iniguez MA, Vila L. Humanvascular smooth muscle cells but not endothelial cells expressprostaglandin E synthase. Circ Res 2000;87:504–7.

14. Stichtenoth DO, Thoren S, Bian H, Peters-Golden M, JakobssonPJ, Crofford LJ. Microsomal prostaglandin E synthase is regulatedby proinflammatory cytokines and glucocorticoids in primaryrheumatoid synovial cells. J Immunol 2001;167:469–74.

15. Kojima F, Naraba H, Sasaki Y, Beppu M, Aoki H, Kawai S.Prostaglandin E2 is an enhancer of interleukin-1�–induced expres-sion of membrane-associated prostaglandin E synthase in rheuma-toid synovial fibroblasts. Arthritis Rheum 2003;48:2819–28.

16. Han R, Tsui S, Smith TJ. Up-regulation of prostaglandin E2synthesis by interleukin-1� in human orbital fibroblasts involvescoordinate induction of prostaglandin-endoperoxide H synthase-2and glutathione-dependent prostaglandin E2 synthase expression.J Biol Chem 2002;277:16355–64.

17. Cipollone F, Prontera C, Pini B, Marini M, Fazia M, De Cesare D,et al. Overexpression of functionally coupled cyclooxygenase-2 andprostaglandin E synthase in symptomatic atherosclerotic plaques


as a basis of prostaglandin E2-dependent plaque instability. Cir-culation 2001;104:921–7.

18. Mancini JA, Blood K, Guay J, Gordon R, Claveau D, Chan CC, etal. Cloning, expression, and up-regulation of inducible rat prosta-glandin E synthase during lipopolysaccharide-induced pyresis andadjuvant-induced arthritis. J Biol Chem 2001;276:4469–75.

19. Yamagata K, Matsumura K, Inoue W, Shiraki T, Suzuki K,Yasuda S, et al. Coexpression of microsomal-type prostaglandin Esynthase with cyclooxygenase-2 in brain endothelial cells of ratsduring endotoxin-induced fever. J Neurosci 2001;21:2669–77.

20. Uematsu S, Matsumoto M, Takeda K, Akira S. Lipopolysaccha-ride-dependent prostaglandin E2 production is regulated by theglutathione-dependent prostaglandin E2 synthase gene induced bythe Toll-like receptor 4/MyD88/NF-IL6 pathway. J Immunol 2002;168:5811–6.

21. Trebino C, Stock J, Gibbons C, Naiman B, Wachtmann T, UmlandJEA. Impaired inflammatory and pain responses in mice lackingan inducible prostaglandin E synthase. Proc Natl Acad Sci U S A2003;100:9044–9.

22. Claveau D, Sirinyan M, Guay J, Gordon R, Chan CC, Bureau Y,et al. Microsomal prostaglandin E synthase-1 is a major terminalsynthase that is selectively up-regulated during cyclooxygenase-2-dependent prostaglandin E2 production in the rat adjuvant-in-duced arthritis model. J Immunol 2003;170:4738–44.

23. Goldring MB, Birkhead JR, Suen LF, Yamin R, Mizuno S,Glowacki J, et al. Interleukin-1�-modulated gene expression inimmortalized human chondrocytes. J Clin Invest 1994;94:2307–16.

24. Goldring MB, Berenbaum F. Human chondrocyte culture modelsfor studying cyclooxygenase expression and prostaglandin regula-tion of collagen gene expression. Osteoarthritis Cartilage 1999;7:386–8.

25. Kokenyesi R, Tan L, Robbins JR, Goldring MB. Proteoglycanproduction by immortalized human chondrocyte cell lines culturedunder conditions that promote expression of the differentiatedphenotype. Arch Biochem Biophys 2000;383:79–90.

26. Loeser RF, Sadiev S, Tan L, Goldring MB. Integrin expression byprimary and immortalized human chondrocytes: evidence of adifferential role for �1�1 and �2�1 integrins in mediating chon-drocyte adhesion to types II and VI collagen. OsteoarthritisCartilage 2000;8:96–105.

27. Geng Y, Valbracht J, Lotz M. Selective activation of the mitogen-activated protein kinase subgroups c-Jun NH2 terminal kinase andp38 by IL-1 and TNF in human articular chondrocytes. J ClinInvest 1996;98:2425–30.

28. Mengshol JA, Vincenti MP, Coon CI, Barchowsky A, BrinckerhoffCE. Interleukin-1 induction of collagenase 3 (matrix metallopro-teinase 13) gene expression in chondrocytes requires p38, c-JunN-terminal kinase, and nuclear factor �B: differential regulation ofcollagenase 1 and collagenase 3. Arthritis Rheum 2000;43:801–11.

29. Thomas B, Thirion S, Humbert L, Tan L, Goldring MB, BereziatG, et al. Differentiation regulates interleukin-1�1�-induced cyclo-oxygenase-2 in human articular chondrocytes: role of p38 mitogen-activated protein kinase. Biochem J 2002;362:367–73.

30. Altman R, Asch E, Bloch D, Bole G, Borenstein D, Brandt K, etal. Development of criteria for the classification and reporting ofosteoarthritis: classification of osteoarthritis of the knee. ArthritisRheum 1986;29:1039–49.

31. Berenbaum F, Humbert L, Bereziat G, Thirion S. Concomitantrecruitment of ERK1/2 and p38 MAPK signalling pathway isrequired for activation of cytoplasmic phospholipase A2 via ATPin articular chondrocytes. J Biol Chem 2003;278:13680–7.

32. Ashok A, Attur M, Dave M, Leung M, Serhan C, Abramson S.Functional and pharmacogenomic analysis of lipid mediators inhuman osteoarthritis-affected cartilage [abstract]. ArthritisRheum 2002;46 Suppl 9:S221.

33. Bertazzolo N, Punzi L, Stefani MP, Cesaro G, Pianon M, Finco B,Todesco S. Interrelationships between interleukin (IL)-1, IL-6 andIL-8 in synovial fluid of various arthropathies. Agents Actions1994;41:90–2.

34. Attur M, Patel I, Patel R, Abramson S, Amin A. Autocrineproduction of IL-1� by human osteoarthritis-affected cartilage anddifferential regulation of endogenous nitric oxide, IL-6, prosta-glandin E2, and IL-8. Proc Assoc Am Physicians 1998;110:65–72.

35. Robbins JR, Thomas B, Tan L, Choy B, Arbiser JL, Berenbaum F,et al. Immortalized human adult articular chondrocytes maintaincartilage-specific phenotype and responses to interleukin-1�. Ar-thritis Rheum 2000;43:2189–201.

36. Massaad C, Paradon M, Jacques C, Salvat C, Bereziat G, Beren-baum F, et al. Induction of secreted type IIA phospholipase A2gene transcription by interleukin-1�: role of C/EBP factors. J BiolChem 2000;275:22686–94.

37. Thomas B, Berenbaum F, Humbert L, Bian H, Bereziat G,Crofford L, et al. Critical role of C/EBP� and C/EBP� factors inthe stimulation of the cyclooxygenase-2 gene transcription byinterleukin-1� in articular chondrocytes. Eur J Biochem 2000;267:6798–809.

38. Cheng Y, Austin SC, Rocca B, Koller BH, Coffman TM, GrosserT, et al. Role of prostacyclin in the cardiovascular response tothromboxane A2. Science 2002;296:539–41.

39. Mukherjee D, Topol EJ. COX-2: where are we in 2003? Cardio-vascular risk and COX-2 inhibitors. Arthritis Res Ther 2003;5:8–11.

40. Murakami M, Naraba H, Tanioka T, Semmyo N, Nakatani Y,Kojima F, et al. Regulation of prostaglandin E2 biosynthesis byinducible membrane-associated prostaglandin E2 synthase thatacts in concert with cyclooxygenase-2. J Biol Chem 2000;275:32783–92.

41. Lee JC, Kumar S, Griswold DE, Underwood DC, Votta BJ,Adams JL. Inhibition of p38 MAP kinase as a therapeutic strategy.Immunopharmacology 2000;47:185–201.

42. Enslen H, Raingeaud J, Davis RJ. Selective activation of p38mitogen-activated protein (MAP) kinase isoforms by the MAPkinase kinases MKK3 and MKK6. J Biol Chem 1998;273:1741–8.

43. Hale KK, Trollinger D, Rihanek M, Manthey CL. Differentialexpression and activation of p38 mitogen-activated protein kinase�, �, �, and � in inflammatory cell lineages. J Immunol 1999;162:4246–52.

44. Pelletier JP, Fernandes JC, Jovanovic DV, Reboul P, Martel-Pelletier J. Chondrocyte death in experimental osteoarthritis ismediated by MEK 1/2 and p38 pathways: role of cyclooxygenase-2and inducible nitric oxide synthase. J Rheumatol 2001;28:2509–19.

45. Schett G, Tohidast-Akrad M, Smolen JS, Schmid BJ, Steiner CW,Bitzan P, et al. Activation, differential localization, and regulationof the stress-activated protein kinases, extracellular signal–regulated kinase, c-Jun N-terminal kinase, and p38 mitogen-activated protein kinase, in synovial tissue and cells in rheumatoidarthritis. Arthritis Rheum 2000;43:2501–12.

46. Badger AM, Cook MN, Lark MW, Newman-Tarr TM, Swift BA,Nelson AH, et al. SB 203580 inhibits p38 mitogen-activatedprotein kinase, nitric oxide production, and inducible nitric oxidesynthase in bovine cartilage-derived chondrocytes. J Immunol1998;161:467–73.

47. Badger AM, Griswold DE, Kapadia R, Blake S, Swift BA,Hoffman SJ, et al. Disease-modifying activity of SB 242235, aselective inhibitor of p38 mitogen-activated protein kinase, in ratadjuvant-induced arthritis. Arthritis Rheum 2000;43:175–83.

48. Kumar S, Boehm J, Lee JC. p38 MAP Kinases: key signallingmolecules as therapeutic targets for inflammatory diseases. NatRev Drug Disc 2003;2:717–26.


Up-regulation of microsomal prostaglandin E synthase 1 in osteoarthritic human cartilage: Critical roles of the ERK-1/2 and p38 signaling pathways

Documents