ARTHRITIS & RHEUMATISMVol. 46, No. 8, August 2002, pp 2079–2086DOI 10.1002/art.10443© 2002, American College of Rheumatology

Epigallocatechin-3-Gallate Inhibits Interleukin-1�–InducedExpression of Nitric Oxide Synthase and Production of

Nitric Oxide in Human Chondrocytes

Suppression of Nuclear Factor �B Activation by Degradation of the Inhibitor ofNuclear Factor �B

Rashmi Singh, Salahuddin Ahmed, Najmul Islam, Victor M. Goldberg, and Tariq M. Haqqi

Objective. The proinflammatory cytokineinterleukin-1� (IL-1�) induces the production of highlevels of nitric oxide (NO) in human chondrocytes.Green tea (Camellia sinensis) polyphenols are potentantiinflammatory agents and have been shown to inhibitNO production in tumor cell lines. In the present study,we examined the effect of epigallocatechin-3-gallate(EGCG), a green tea polyphenol, on IL-1�–inducedproduction of NO in primary human osteoarthritis(OA) chondrocytes.

Methods. Human chondrocytes were derived fromOA cartilage and were treated with EGCG (100 �M)and IL-1� (2 ng/ml) for different periods, and induciblenitric oxide synthase (iNOS) messenger RNA and pro-tein expression was determined by real-time quantita-tive reverse transcriptase–polymerase chain reactionand Western blotting, respectively. Production of NOwas determined as nitrite in culture supernatant. Acti-vation and translocation of nuclear factor �B (NF-�B),levels of inhibitor of nuclear factor �B (I�B), andNF-�B DNA binding activity were determined by West-ern blotting and a highly sensitive and specific enzyme-

linked immunosorbent assay. Activity of I�B kinase wasdetermined using in vitro kinase assay.

Results. Human chondrocytes cotreated withEGCG produced significantly less NO compared withchondrocytes stimulated with IL-1� alone (P < 0.005).The inhibition of NO production correlated with thesuppression of induction and expression of NF-�B–dependent gene iNOS. EGCG inhibited the activationand translocation of NF-�B to the nucleus by suppress-ing the degradation of its inhibitory protein I�B� in thecytoplasm.

Conclusion. Our results indicate that EGCG in-hibits the IL-1�–induced production of NO in humanchondrocytes by interfering with the activation ofNF-�B through a novel mechanism. Our data furthersuggest that EGCG may be a therapeutically effectiveinhibitor of IL-1�–induced inflammatory effects thatare dependent on NF-�B activation in human OAchondrocytes.

The proinflammatory cytokine interleukin-1�(IL-1�), produced in an arthritic joint by activatedsynovial cells and infiltrating macrophages, is consideredto be one of the most potent catabolic factors in jointdiseases (1). IL-1� induces the production of severalmediators of cartilage degradation, such as nitric oxide(NO) and matrix metalloproteinases (1–3), and inhibitsthe concentration of tissue inhibitor of metalloprotein-ases in arthritic joints (3). IL-1� also suppresses thebiosynthesis of type II collagen and aggrecan (4–6) andthe proliferation of chondrocytes (7–9), thus inhibitingthe repair process in the cartilage. Additional evidenceindicating the involvement of IL-1 in cartilage degrada-

Supported by NIH grants AR-44902, AR-20618 (NEOMAC),and AR-37726, and by funds from the Case Western Reserve Univer-sity Department of Orthopaedics.

Rashmi Singh, PhD, Salahuddin Ahmed, PhD, Najmul Islam,PhD, Victor M. Goldberg, MD, Tariq M. Haqqi, PhD: Case WesternReserve University, Cleveland, Ohio.

Drs. Singh and Ahmed contributed equally to this work.Address correspondence and reprint requests to Tariq M.

Haqqi, PhD, Department of Medicine, Division of Rheumatic Dis-eases, Case Western Reserve University, 2109 Adelbert Road, Cleve-land, OH 44106-4946. E-mail: txh5@po.cwru.edu.

Submitted for publication November 2, 2001; accepted inrevised form April 25, 2002.

2079

tion emerged from studies showing that intraarticularadministration of IL-1� into rabbit and mouse jointsresults in loss of proteoglycans from the cartilage(10,11). IL-1� exerts its inflammatory effects by activat-ing a diverse spectrum of signaling cascades in the cells(12,13) that lead to the induction of inducible nitricoxide synthase (iNOS) and the production of high levelsof the second messenger NO in arthritic joints (ref. 14and references therein).

Involvement of NO in the pathogenesis of arthri-tis is evident from studies showing that high levels ofnitrite/nitrate are present in serum and in the synovialfluid of arthritis patients, and the messenger RNA(mRNA) and protein of iNOS have been detected in thesynovial tissue of osteoarthritis (OA) patients (14,15).Generation of NO by chondrocytes is recognized as animportant factor in cartilage loss and degradation inarthritic joints (16–18). This is supported by experimen-tally derived evidence from animal models that demon-strated a significant relationship between NO produc-tion and apoptotic chondrocyte death in cartilage (17).Of importance to the present studies are the findingsthat inhibition of iNOS, the enzyme responsible for theproduction of high levels of NO in vivo, also protectedchondrocytes from apoptosis (19). In mammalian cells,regulation of iNOS expression is predominantly gov-erned by the ubiquitously expressed transcription factornuclear factor �B (NF-�B), which is required for theinducible expression of genes associated with inflamma-tory responses (for review, see ref. 20).

Green tea is a product of dried leaves of Camelliasinensis and has a long history of human consumptionwith no known serious side effects (21,22). Extensivestudies in many animal model systems have shown that apolyphenolic fraction isolated from green tea polyphe-nols (GTPs) is antiinflammatory and that the constituentpolyphenols are potent antioxidants (for review, see ref.22). We have previously shown that consumption ofGTPs was inhibitory to the development of type IIcollagen–induced arthritis in mice, a model of inflam-matory polyarthritis (23).

A majority of the biologic effects of green tea aremediated by one of its principal constituents,epigallocatechin-3-gallate (EGCG). Studies have shownthat EGCG inhibits the induction and expression ofiNOS in mouse macrophages stimulated with lipopoly-saccharide (LPS) (24). Since high levels of iNOS expres-sion and production of NO are also induced by IL-1� inhuman chondrocytes, in this study we addressed thequestion of a possible inhibitory effect of EGCG onIL-1�–induced production of NO in human chondro-

cytes. Our results showed that EGCG suppressed theIL-1�–induced iNOS mRNA and protein expression andproduction of NO in human OA chondrocytes and thatthese effects were concomitant with inhibited activationof the transcription factor NF-�B. We also showed thatinhibition of NF-�B was achieved by inhibiting thedegradation of the inhibitor of nuclear factor �B�(I�B�) in the cytoplasm. Our results thus identify aunique mechanism of action of a dietary constituent forexerting its antiinflammatory effects in human OA chon-drocytes and suggest that consumption of EGCG mayhave beneficial health effects in arthritis.

MATERIALS AND METHODS

Cytokines and EGCG. Highly purified green tea cate-chin EGCG was generously provided by Dr. Yukihiko Hara,Tokyo Food Techno (Tokyo, Japan), and was used withoutfurther purification. Recombinant human IL-1� was pur-chased from R&D Systems (Minneapolis, MN).

Preparation of articular chondrocytes. Macroscopi-cally “normal” cartilage samples were obtained from OApatients (7 samples) at the time of total joint replacement.Knee cartilage was procured according to the guidelines of theCooperative Human Tissue Network for the use of discardedhuman tissue and with the approval of the Institutional ReviewBoard of University Hospitals of Cleveland.

Chondrocytes were prepared from cartilage samples byenzymatic digestion essentially as previously described (25).Isolated chondrocytes were plated at a density of 1 � 106/ml in60-mm tissue culture dishes (Corning, Corning, NY) in Ham’sF-12 medium:Dulbecco’s modified Eagle’s medium (1:1), sup-plemented with L-glutamine, penicillin–streptomycin, Fungi-zone, and 10% fetal bovine serum (complete medium), andallowed to adhere to the tissue culture plates for 72 hours at37°C with 5% CO2 and 95% air.

Treatment of chondrocytes with IL-1� and EGCG.Human chondrocytes (1 � 106/ml) were plated in 6-well platesand serum-starved for 12 hours/overnight. The medium wasreplaced with fresh medium containing recombinant humanIL-1� (2 ng/ml), and chondrocytes were incubated for 24 hoursin a tissue culture incubator at 37°C and 5% CO2. Thisconcentration of IL-1� was chosen based on pilot experimentsperformed to determine the lowest concentration of IL-1� thatwill induce the maximum production of NO by OA chondro-cytes in vitro. To study the dose-dependence of the effect ofEGCG on IL-1�–induced iNOS expression, chondrocytes werepretreated for 30 minutes with varying concentrations ofEGCG (1–100 �M) before the addition of IL-1�. Controlsconsisted of chondrocytes incubated in the culture mediumwithout EGCG or IL-1�.

Nitrite determination. NO production was determinedby estimating the nitrite concentration in 24-hour culturesupernatant using a commercially available kit according to theinstructions of the manufacturer (R&D Systems). For eachexperiment, nitrite concentration was quantified in the samplesusing a standard curve prepared with known concentrations ofnitrite.

2080 SINGH ET AL

Western blot analysis. After the treatments, chondro-cyte lysate was prepared as described previously (25) and usedfor Western blot analysis. For studies involving NF-�B, nuclearand cytosolic fractions were prepared as previously described(26). Protein content of the lysates was determined using adetergent-compatible protein assay kit (Bio-Rad, Hercules,CA), and 25 �g of total protein/lane was resolved by denatur-ing sodium dodecyl sulfate–polyacrylamide gel electrophoresis(SDS-PAGE) and transferred to nitrocellulose membranes(Bio-Rad). Primary antibodies for NF-�B/p65, I�B�, phospho-I�B�, inhibitor of nuclear factor �B kinase (IKK), iNOS,endothelial cell nitric oxide synthase (eNOS), or neuronalnitric oxide synthase (nNOS) were either rabbit polyclonalantibodies or mouse monoclonal antibodies (Santa Cruz Bio-technology, Santa Cruz, CA). Membranes were blocked with5% nonfat dried milk or 2% bovine serum albumin in Trisbuffered saline (TBS) containing 0.05% Tween 20 (TBST).Primary antibodies were diluted in TBST and incubated withthe blots for �2 hours at room temperature, washed with freshTBST, and incubated with a 1:5,000 dilution of horseradishperoxidase (HRP)–conjugated anti-rabbit IgG or anti-mouseIgG (Southern Biotechnology, Birmingham, AL) in blockingbuffer for 1 hour at room temperature. After washing, theprotein bands were visualized with the ECL detection kit(Amersham, Arlington Heights, IL) and Kodak Bio-Max film.Blots shown are representative of 3 independent experimentsusing chondrocytes from independent cartilage samples.

NF-�B DNA binding activity assay. Cellular levels ofactive NF-�B/p65 in the nuclei of human chondrocytes treatedwith IL-1� and EGCG were determined with a highly sensitiveand specific colorimetric method reported recently (27), usinga commercially available reagent kit according to the instruc-tions of the manufacturer (Active Motif, Carlsbad, CA).Absorbance was read on an enzyme-linked immunosorbentassay reader at 450 nm with a reference wavelength of 655 nm.Wild-type and mutated NF-�B oligonucleotides supplied inthe kit were used as specificity controls and were added tosome wells prior to the addition of chondrocyte lysate.

In vitro kinase assays. Human chondrocytes weretreated with IL-1� (2 ng/ml, 20 minutes), and IKK complex wasimmunoprecipitated from cell lysates (250 �g total protein)using anti-mouse IKK� antibody (Santa Cruz Biotechnology)and protein A/G agarose. Immunoprecipitates were washed 3times with the lysis buffer (Cell Signaling Technology, Beverly,MA), twice with the kinase buffer (25 mM Tris [pH 7.5], 5 mM�-glycerophosphate, 2 mM dithiothreitol [DTT], 0.1 mMNa3VO4, 10 mM MgCl2, 100 �M ATP), and finally suspendedin 50 �l of kinase buffer. Assays were performed by adding 2�g of I�B� (Santa Cruz Biotechnology) and incubated at 30°Cfor 30 minutes. The samples were resolved by SDS-PAGE,transferred to nitrocellulose membranes, and the blot wasincubated with primary antibodies specific for phosphorylatedI�B� (Santa Cruz Biotechnology), washed, and then incubatedwith HRP-conjugated secondary antibodies and visualized byenhanced chemiluminescence (Amersham).

Quantitative reverse transcriptase–polymerase chainreaction (RT-PCR). Total cytoplasmic RNA was preparedfrom human chondrocytes using a commercially available kitaccording to the instructions of the manufacturer (Qiagen,Valencia, CA). We used real-time RT-PCR with internalfluorescent hybridization probes in an ABI Prism 7700 detec-

tion system (ABI/Perkin Elmer Biosystems, Foster City, CA),which allows the sensitive and specific quantification oftargeted mRNA transcripts. The target-specific RT primer(5�-CTCTggTCAAAC-3�), PCR primers (forward 5�-AgCggg-ATgACTTTCCAAgA-3�; reverse 5�-ATAATggACCCCAgg-CAAgATT-3�), and the TaqMan Probe for iNOS (6FAM-CCATAAggCCAAAgggATTTTAACTTgCAg-TAMRA)were designed using the “Primer Express” software (ABI/Perkin Elmer Biosystems). The probe was labeled with 5carbofluorescein (FAM) at the 5� end and with TAMRA at the3� end (ABI/Perkin Elmer Biosystems). The degradation of theprobe during PCR resulted in increased fluorescence of theprobe, which specifically annealed between the forward andreverse primer sites, and allowed the detection of the PCRproduct by monitoring the increase in fluorescence of thereporter dye.

To quantitate the expression of iNOS, single-strandedcomplementary DNA (cDNA) was synthesized using 100 ng oftotal RNA prepared from OA chondrocytes as describedabove. The RNA was mixed with 10 �M of RT primer,Moloney murine leukemia virus reverse transcriptase, RNaseinhibitor, 0.1M DTT, buffer, and dNTPs using a commerciallyavailable kit according to the instructions of the manufacturer(Invitrogen, Carlsbad, CA). The quantitative PCR reactionswere set up using 5 �l of cDNA and 20 �l TaqMan UniversalPCR Master mix (Perkin Elmer Biosystems), which containsreaction buffer, optimal amounts of AmpliTaq Gold DNApolymerase, AmpErase UNG (which protects against ampli-con carryover), and dNTPs with dUTP. Concentrations ofprimers and probe used were optimized in pilot studies toallow accurate quantitation of the target transcript. The PCRconditions were 1 cycle at 50°C for 2 minutes (for optimalAmpErase UNG enzyme activity), 1 cycle at 95°C for 10minutes (to activate AmpliTaq Gold DNA polymerase), fol-lowed by 40 cycles (95°C for 15 seconds, 60°C for 1 minute).The cycle threshold value (CT) for the samples analyzed wascompared with the CT value of the known amounts of standardiNOS cDNA constructed and amplified simultaneously. Toensure lack of DNA contamination in the RNA samples, atube of sample without RT was included as a control. Expres-sion of iNOS was corrected to the expression of ribosomal 18Sgene (R18) and the results were expressed as copies ofiNOS/1010 copies of R18 (equivalent to 1 � 106 humanchondrocytes).

Statistical analysis. All measurements were performedin triplicate and repeated 3 times using age- and sex-matchedsamples. Data were analyzed using Student’s t-test (SigmaPlotsoftware; SPSS, Chicago, IL). P values less than 0.05 wereconsidered significant.

RESULTS

Inhibition of IL-1�–induced NO production byEGCG in human chondrocytes. Primary chondrocytes(1 � 106/ml) were stimulated with IL-1� (2 ng/ml) orIL-1� � EGCG (100 �M) for 24 hours, and the concen-tration of nitrite (as an indicator of NO production) wasestimated in culture supernatant. Results demonstratedthat upon cotreatment with EGCG, human chondro-

GREEN TEA INHIBITS NO PRODUCTION IN HUMAN CHONDROCYTES 2081

cytes produced significantly less NO (P � 0.05) whencompared with the values obtained in cultures treatedwith IL-1� only (Figure 1). To confirm that the NOproduction in human chondrocytes was due to theactivity of iNOS and not of other NOS isoforms, humanchondrocytes were treated with IL-1� in the presence ofaminoguanidine (AG), a selective iNOS inhibitor, andthe accumulation of nitrite in culture supernatant wasanalyzed. As shown in Figure 1, in the presence of AG(1.0 mM), IL-1�–induced nitrite formation by humanchondrocytes was completely abolished, indicating thateNOS or nNOS isoforms were not activated by IL-1� toany significant degree in human chondrocytes.

Inhibition of IL-1�–induced iNOS protein ex-pression by EGCG in human chondrocytes. To deter-mine whether the suppression of NO production inhuman chondrocytes was due to the inhibition of iNOSexpression, human chondrocytes were treated with dif-ferent doses of EGCG and IL-1� for 24 hours asdescribed above, and the Western blots were probedwith antibodies specific for human iNOS. Our datarevealed that the expression of iNOS protein was up-regulated severalfold in human chondrocytes stimulatedwith IL-1� alone and that EGCG inhibited the IL-1�–induced increase in the expression of iNOS protein inhuman chondrocytes in a dose-dependent manner (Fig-ure 2A). The observed down-regulation of iNOS protein

expression (compare lane 2 with lanes 5 and 6, Figure2A) was consistent with the nitrite levels in the culturesupernatant of chondrocytes that were used to preparethe cell lysate. No effect of EGCG or IL-1� on theprotein levels of eNOS or nNOS in human chondrocyteswas found (unpublished data).

Inhibition of IL-1�–induced iNOS mRNA ex-pression by EGCG in human chondrocytes. To deter-mine that the inhibition of IL-1�–induced expression ofiNOS was due to the inhibition of iNOS induction andnot to the inhibition of translation, we analyzed theiNOS mRNA expression in human chondrocytes byreal-time quantitative RT-PCR. After 12 and 24 hours ofculture, iNOS mRNA expression in IL-1�–treated hu-man chondrocytes was severalfold higher in comparison

Figure 1. Inhibition of interleukin-1� (IL-1�)–induced nitric oxide(NO) production by epigallocatechin-3-gallate (EGCG) in humanchondrocytes. Human osteoarthritis chondrocytes were stimulatedwith IL-1� (2 ng/ml) and EGCG (100 �M) for 24 hours. Production ofNO was measured in culture supernatant using a commercially avail-able kit. Values shown are the mean and SD of 3 independentexperiments, each run in triplicate. � � P � 0.05 versus treatment withIL-1� alone. AG � aminoguanidine.

Figure 2. Inhibition of IL-1�–induced inducible nitric oxide synthase(iNOS) mRNA and protein expression by EGCG in human osteoar-thritis (OA) chondrocytes. A, OA chondrocytes were stimulated withIL-1� (2 ng/ml) and different doses of EGCG for 24 hours, andexpression of iNOS was determined by Western blotting (a represen-tative Western blot is shown). Lane 1, Control; lane 2, IL-1�; lane 3,IL-1� � 1 �M EGCG; lane 4, IL-1� � 10 �M EGCG; lane 5, IL-1� �50 �M EGCG; lane 6, IL-1� � 100 �M EGCG. Equal loading ofprotein was verified by probing the same blot for �-actin. B, OAchondrocytes were stimulated with IL-1� (2 ng/ml) and IL-1� plusEGCG (100 �M) for 12 and 24 hours, and expression of iNOS mRNAwas determined by real-time quantitative reverse transcriptase–polymerase chain reaction. Data shown are representative of 3 inde-pendent experiments. See Figure 1 for other definitions.

2082 SINGH ET AL

with the levels present in untreated human chondro-cytes, which exhibited a constitutive low level of expres-sion (Figure 2B). Remarkably, and consistent with theabove results, EGCG (100 �M) was found to inhibit theIL-1�–induced increase in the expression level of iNOSmRNA in human chondrocytes at both time pointsanalyzed (Figure 2B).

Inhibition of IL-1�–induced activation of NF-�B/p65 transcription factor by EGCG in human chondro-cytes. Stimulation by IL-1 leads to the activation of acascade of adaptor proteins and kinases, resulting in thephosphorylation of I�B by I�B kinases (IKK�/�). Thisresults in the activation of the transcription factorNF-�B (for review, see ref. 28). Since the expression ofthe iNOS gene is regulated by NF-�B, we determinedwhether EGCG inhibited the IL-1�–induced activationof NF-�B in human chondrocytes. The DNA bindingactivity present in nuclear extracts of human chondro-cytes showed a rapid increase upon stimulation withIL-1�, reaching a peak between 5 and 30 minutesposttreatment (Figure 3A). However, even at 60 minutesposttreatment, IL-1�–treated chondrocytes had signifi-cantly (P � 0.005) higher levels of active NF-�B thanuntreated controls. Although the activation of NF-�Bwas not totally blocked in chondrocytes cotreated withEGCG, the levels of DNA-bound NF-�B/p65 weresignificantly reduced (P � 0.005) in comparison with thelevels detected in human chondrocytes treated withIL-1� alone (Figure 3A). That there was a reduction inthe levels of NF-�B in the nuclei of chondrocytescotreated with EGCG was also supported by Westernblot results that showed reduced levels of NF-�B/p65 inthe nuclei of these cells in comparison with the nuclei ofchondrocytes treated with IL-1� alone (results notshown).

Taken together, these results indicated thatEGCG was interfering with an event essential for IL-1�–induced activation and translocation of NF-�B to thenucleus in human chondrocytes. Since activation andtranslocation of NF-�B to the nucleus are dependent onthe phosphorylation of I�B, we next determined theeffect of EGCG on the activity of IKK in humanchondrocytes using the in vitro kinase assay. Our resultsshowed that EGCG had no effect on the I�B� phos-phorylating activity of the IKK complex isolated fromhuman chondrocytes cotreated with EGCG (Figure 3B).These results indicated that the observed reduction inthe nuclear levels of NF-�B was not due to an inhibitionof IKK activity in EGCG-treated human chondrocytes.

The next event downstream of the phosphoryla-tion of I�B� is ubiquitination and subsequent degrada-

tion by the 26S proteasome. Therefore, we determinedwhether EGCG had any effect on the levels of cytosolicI�B� in human chondrocytes. As can be seen in Figure4, the levels of I�B� in human chondrocytes cotreatedwith EGCG were similar to the levels present in con-trols, indicating that I�B� was not being degraded in

Figure 3. IL-1�–induced increase in the levels and DNA bindingactivity of nuclear factor �B (NF-�B)/p65 in the nucleus of humanosteoarthritis (OA) chondrocytes inhibited by EGCG. A, OA chon-drocytes were stimulated with IL-1� (2 ng/ml) and EGCG for varyingtimes, and activation of NF-�B/p65 was determined by a highly specificenzyme-linked immunosorbent assay. Specificity of NF-�B binding inthis assay was determined by preincubating nuclear protein extractswith a 50-fold excess of kit-supplied wild-type NF-�B or with themutant NF-�B oligonucleotides. Values shown are the mean and SDof 3 independent experiments, each performed in duplicate. � � P �0.005 versus control. B, Effect of EGCG on activity of inhibitor ofnuclear factor �B kinase (IKK). Human OA chondrocytes werestimulated with IL-1� (2 ng/ml) and EGCG (100 �M) for 15 and 30minutes, and the glutathione S transferase (GST)–inhibitor of nuclearfactor �B� (I�B�) (GST-I�B�) phosphorylating activity was deter-mined by an in vitro kinase assay. Based on the intensity of phosphor-ylated I�B� bands detected with an antibody specific for phospho-I�B�, no significant difference in the GST-I�B� phosphorylatingactivity of the IKK complex in chondrocytes treated with IL-1� alone(lanes 2 and 3) or in chondrocytes cotreated with EGCG (lanes 4 and5) was detected (lane 1 is control). Similar results were obtained in 2additional independent experiments. Band intensities shown wereobtained by scanning and analyzing the blot using Un-Scan-It gelsoftware (Silk Scientific, Orem, UT). See Figure 1 for other defini-tions.

GREEN TEA INHIBITS NO PRODUCTION IN HUMAN CHONDROCYTES 2083

these cells. On the other hand, I�B� was barely detect-able in human chondrocytes stimulated with IL-1�alone, indicating that it was being degraded in these cells(compare lanes marked 30 minutes in Figure 4). Takentogether, these results indicated that EGCG inhibitedthe IL-1�–induced activation of NF-�B by inhibiting thedegradation of its inhibitor I�B�, possibly by inhibitingthe proteasome pathway in human chondrocytes.

DISCUSSION

A characteristic feature of arthritic joints is thepersistence of proinflammatory cytokines produced bythe inflamed synovium as well as by chondrocytes in theaffected joints (for review, see refs. 29 and 30). It is wellknown that human chondrocytes are highly responsive toIL-1 and the most striking effect of IL-1 on chondrocytesis to induce the production of NO, which is an importantsource of highly reactive free oxygen radicals, within thechondrocytes and synoviocytes by inducing the expres-sion of iNOS (14,15). Although arthritis is present inevery population and OA is the most common jointdisorder, treatment is still limited to a few classes ofdrugs, primarily nonsteroidal antiinflammatory drugsand injectable corticosteroids (31–34). However, whileproviding relief from pain, none of these agents has beenshown to inhibit cartilage breakdown or to inhibit dis-ease progress; they also have varying degrees of gastro-intestinal toxicity (32,35).

These issues have shifted the attention towardthe development or identification of new compoundsthat can impinge on the activity of cartilage-degradingfactors in an arthritic joint and are better tolerated by

the gastrointestinal tract than currently available anti-rheumatic medicines. A class of compounds that isgaining attention is the “dietary supplements” and,recently, glucosamine, a commonly used dietary supple-ment, has been shown to inhibit cartilage breakdownand disease progress in patients with OA (35). Otherstudies have shown that glucosamine prevents the IL-1�–induced activation of human chondrocytes throughthe inhibition of inflammatory processes (32). Glu-cosamine is also a potent inhibitor of iNOS inductionand production of NO (36), and this may explain itsobserved cartilage protective effects in OA (35). In otherstudies, inhibition of iNOS has been shown to protectagainst both inflammation and cartilage matrix loss inexperimental models of arthritis (19,37). Taken to-gether, these data suggest that agents capable of inhib-iting proinflammatory cytokine–induced iNOS activityand NO production may be of potential benefit in boththe degenerative and inflammatory joint diseases.

In the present study, we demonstrated that thegreen tea polyphenol EGCG inhibited IL-1�–inducedproduction of NO in human chondrocytes. Our resultsfurther show that the inhibitory effect on NO productionwas mediated by inhibiting the expression of iNOSmRNA, although we cannot rule out an inhibitory effectof EGCG on the activity of iNOS as well (24). Theinhibition of NO production by EGCG appears to be aspecific effect, because epigallocatechin or sodium gal-late alone or in combination was not effective in sup-pressing the IL-1�–induced expression of iNOS or pro-duction of NO in human chondrocytes (unpublisheddata), indicating that the EGCG structure was importantin mediating this effect. Although EGCG did not com-pletely abolish the production of NO in chondrocytes,low levels of residual NO detected were not toxic tohuman chondrocytes, since no apoptosis was seen inthese cultures (unpublished data).

Several studies have shown that the expression ofiNOS is dependent on the activation of the ubiquitouslyexpressed transcription factor NF-�B (24,38). Our dataalso show that there are decreased levels of NF-�B/p65in the nuclei of chondrocytes cotreated with EGCG,suggesting that the inhibitory effect of EGCG on theinduction and expression of iNOS is partly a result ofreduced levels of activated NF-�B/p65 in the nucleus ofchondrocytes. Decreased levels and DNA binding ofNF-�B/p65 to its consensus site in human chondrocytescotreated with EGCG may be of particular significancebecause this subunit potently transactivates target genes,while the NF-�B/p50 homodimers exert relatively lowtransactivation of target genes (39). We also found that

Figure 4. Inhibition of nuclear factor �B� (I�B�) degradation inhuman osteoarthritis (OA) chondrocytes by EGCG. OA chondrocyteswere treated with IL-1� alone or with IL-1� plus EGCG, and the levelsof cytosolic I�B� were determined by Western blotting. Each lane wasloaded with 25 �g of total protein, and protein bands reacting with ahuman I�B�-specific antibody were visualized by enhanced chemilu-minescence. Band intensities shown were obtained by scanning andanalyzing using Un-Scan-It gel software (Silk Scientific). See Figure 1for other definitions.

2084 SINGH ET AL

in chondrocytes costimulated with EGCG, cytosolic lev-els of I�B� are similar to the I�B� levels present inunstimulated chondrocytes, indicating that I�B� is notbeing degraded in chondrocytes costimulated withEGCG. Previously, Lin and Lin (24) also noted that inmurine macrophages, EGCG inhibited the LPS-induceddegradation of I�B�, but whether this was due toinhibition of IKK complex activity was not studied. Onthe other hand, Yang et al (40) reported that in intesti-nal epithelium cell line IEC-6, EGCG blocked NF-�Bactivation by inhibiting the IKK complex activity, caus-ing I�B� to accumulate in the nucleus.

In this regard, our results are distinct from thoseof previous studies because they show that EGCG hasno inhibitory effect on the I�B� phosphorylating activityof IKK, indicating that the inhibition of activation andtranslocation of NF-�B/p65 to the nucleus may resultfrom inhibition of the proteasome pathway, causing thestabilization and accumulation of I�B� in the cytoplasmof human chondrocytes. Thus, our results point outanother inhibitory mechanism of NF-�B activation andinduction of iNOS by EGCG, namely by inhibiting theproteasome pathway and preventing I�B� degradationin human chondrocytes. These results are supported byprevious studies showing that inhibitors of the protea-some pathway interfere with the induction of iNOS byblocking the activation of NF-�B (41). Furthermore,these results, along with the absence of any knownserious toxic effects of EGCG, suggest that EGCG maybe a therapeutically effective agent, in combination withcurrent treatment modalities, for inhibiting IL-1�–induced cartilage degradation in arthritis.

ACKNOWLEDGMENTS

The authors thank Timothy Kermode and MathewLalonde for their expert technical assistance.

REFERENCES

1. Van der Kraan PM, van den Berg WB. Anabolic and destructivemediators in osteoarthritis. Curr Opin Clin Nutr Metab Care2000;3:205–11.

2. Van de Loo FAJ, Joosten LA, van Lent PLEM, Arntz OJ, van denBerg WB. Role of interleukin-1, tumor necrosis factor �, andinterleukin-6 in cartilage proteoglycan metabolism and destruc-tion: effect of in situ blocking in murine antigen- and zymosan-induced arthritis. Arthritis Rheum 1995;38:164–72.

3. Wood DD, Ihrie EJ, Dinarello CA, Cohen PL. Isolation of aninterleukin-1–like factor from human joint effusions. ArthritisRheum 1983;26:975–83.

4. 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.

5. Gouze JN, Bordji K, Gulberti S, Terlain B, Netter P, Magdalou J,

et al. Interleukin-1� down-regulates the expression of glucurono-syltransferase I, a key enzyme priming glycosaminoglycan biosyn-thesis: influence of glucosamine on interleukin-1�–mediated ef-fects in rat chondrocytes. Arthritis Rheum 2001;44:351–60.

6. Nietfeld JJ, Wilbrink B, Den Otter W, Huber J, Huber-Bruning O.The effect of human interleukin 1 on proteoglycan metabolism inhuman and porcine cartilage explants. J Rheumatol 1990;17:818–26.

7. Guerne PA, Sublet A, Lotz M. Growth factor responsiveness ofhuman articular chondrocytes: distinct profiles in primary chon-drocytes, subcultured chondrocytes, and fibroblasts. J Cell Physiol1994;158:476–84.

8. Blanco FJ, Lotz M. IL-1-induced nitric oxide inhibits chondrocyteproliferation via PGE2. Exp Cell Res 1995;218:319–25.

9. Hauselmann HJ, Oppliger L, Michel BA, Stefanovic-Racic M,Evans CH. Nitric oxide and proteoglycan biosynthesis by humanarticular chondrocytes in alginate culture. FEBS Lett 1994;352:361–4.

10. Pettipher ER, Higgs GA, Henderson B. Interleukin 1 inducesleukocyte infiltration and cartilage proteoglycan degradation inthe synovial joint. Proc Natl Acad Sci U S A 1986;83:8749–53.

11. Van de Loo AA, van den Berg WB. Effects of murine recombinantinterleukin 1 on synovial joints in mice: measurement of patellarcartilage metabolism and joint inflammation. Ann Rheum Dis1990;49:238–45.

12. Geng Y, Valbracht J, Lotz M. Selective activation of 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.

13. Eder J. Tumor necrosis factor � and interleukin 1 signalling: doMAPKK kinases connect it all? Trends Pharmacol Sci 1997;18:319–22.

14. Amin AR, Abramson SB. The role of nitric oxide in articularcartilage breakdown in osteoarthritis. Curr Opin Rheumatol 1998;10:263–8.

15. Grabowski PS, Wright PK, van’t Hof RJ, Helfrich MH, OhshimaH, Ralston SH. Immunolocalization of inducible nitric oxidesynthase in synovium and cartilage in rheumatoid arthritis andosteoarthritis. Br J Rheumatol 1997;36:651–5.

16. Farrell AJ, Blake DR, Palmer RM, Moncada S. Increased con-centrations of nitrite in synovial fluid and serum samples suggestincreased nitric oxide synthesis in rheumatic diseases. Ann RheumDis 1992;51:1219–22.

17. Hashimoto S, Takahashi K, Amiel D, Coutts RD, Lotz M.Chondrocyte apoptosis and nitric oxide production during exper-imentally induced osteoarthritis. Arthritis Rheum 1998;41:1266–74.

18. 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.

19. Pelletier JP, Jovanovic DV, Lascau-Coman V, Fernandes JC,Manning PT, Connor JR, et al. Selective inhibition of induciblenitric oxide synthase reduces progression of experimental osteo-arthritis in vivo: possible link with the reduction in chondrocyteapoptosis and caspase 3 level. Arthritis Rheum 2000;43:1290–9.

20. Ghosh S, May MJ, Kopp EB. NF-�B and Rel proteins: evolution-arily conserved mediators of immune responses. Annu Rev Immu-nol 1998;16:225–60.

21. Pisters KM, Newman RA, Coldman B, Shin DM, Khuri FR, HongWK, et al. Phase I trial of oral green tea extract in adult patientswith solid tumors. J Clin Oncol 2001;19:1830–8.

22. Katiyar SK, Mukhtar H. Tea consumption and cancer. World RevNutr Diet 1996;79:154–84.

23. Haqqi TM, Anthony DD, Gupta S, Ahmad N, Lee MS, KumarGK, et al. Prevention of collagen-induced arthritis in mice by a

GREEN TEA INHIBITS NO PRODUCTION IN HUMAN CHONDROCYTES 2085

polyphenolic fraction from green tea. Proc Natl Acad Sci U S A1999;96:4524–9.

24. Lin YL, Lin JK. (�)-Epigallocatechin-3-gallate blocks the induc-tion of nitric oxide synthase by down-regulating lipopolysaccha-ride-induced activity of transcription factor nuclear factor-�B. MolPharmacol 1997;52:465–72.

25. Islam S, Kermode T, Sultana D, Moskowitz RW, Mukhtar H,Malemud CJ, et al. Expression profile of protein tyrosine kinasegenes in human osteoarthritis chondrocytes. Osteoarthritis Carti-lage 2001;9:684–93.

26. Li WQ, Dehnade F, Zafarullah M. Oncostatin M-induced matrixmetalloproteinase and tissue inhibitor of metalloproteinase-3genes expression in chondrocytes requires Janus kinase/STATsignaling pathway. J Immunol 2001;166:3491–8.

27. Renard P, Ernest I, Houbion A, Art M, Le Calvez H, Raes M, etal. Development of a sensitive multi-well colorimetric assay foractive NF-�B. Nucleic Acids Res 2001;29:E21.

28. Karin M. The beginning of the end: I�B kinase (IKK) and NF-�Bactivation. J Biol Chem 1999;274:27339–42.

29. Clancy RM, Amin AR, Abramson SB. The role of nitric oxide ininflammation and immunity. Arthritis Rheum 1998;41:1141–51.

30. Goldring MB. Osteoarthritis and cartilage: the role of cytokines.Curr Rheumatol Rep 2000;2:459–65.

31. Darlington LG, Stone TW. Antioxidants and fatty acids in theamelioration of rheumatoid arthritis and related disorders. Br JNutr 2001;85:251–69.

32. Shikhman AR, Kuhn K, Alaaeddine N, Lotz M. N-acetylglu-cosamine prevents IL-1�-mediated activation of human chondro-cytes. J Immunol 2001;166:5155–60.

33. Jacobs JW, Rasker JJ, Bijlsma JW. Alternative medicine in rheuma-tology: threat or challenge? Clin Exp Rheumatol 2001;19:117–9.

34. Sowers M, Lachance L. Vitamins and arthritis: the roles ofvitamins A, C, D, and E. Rheum Dis Clin North Am 1999;25:315–32.

35. Reginster JY, Deroisy R, Rovati LC, Lee RL, Lejeune E, BruyereO, et al. Long-term effects of glucosamine sulphate on osteoar-thritis progression: a randomised, placebo-controlled clinical trial.Lancet 2001;357:251–6.

36. Meininger CJ, Kelly KA, Li H, Haynes TE, Wu G. Glucosamineinhibits inducible nitric oxide synthesis. Biochem Biophys ResCommun 2000;279:234–9.

37. Van de Loo FAJ, Arntz OJ, Van Enckevort FHJ, van Lent PLEM,van den Berg WB. Reduced cartilage proteoglycan loss duringzymosan-induced gonarthritis in NOS2-deficient mice and inanti–interleukin-1–treated wild-type mice with unabated joint in-flammation. Arthritis Rheum 1998;41:634–46.

38. Xie QW, Kashiwabara Y, Nathan C. Role of transcription factorNF-�B/Rel in induction of nitric oxide synthase. J Biol Chem1994;269:4705–8.

39. Cao Z, Tanaka M, Regnier C, Rothe M, Yamit-hezi A, WoroniczJD, et al. NF-�B activation by tumor necrosis factor and interleu-kin-1. Cold Spring Harb Symp Quant Biol 1999;64:473–83.

40. Yang F, Oz HS, Barve S, de Villiers WJ, McClain CJ, Varilek GW.The green tea polyphenol (�)-epigallocatechin-3-gallate blocksnuclear factor-�B activation by inhibiting I�B kinase activity in theintestinal epithelial cell line IEC-6. Mol Pharmacol 2001;60:528–33.

41. Griscavage JM, Wilk S, Ignarro LJ. Inhibitors of the proteasomepathway interfere with induction of nitric oxide synthase in macro-phages by blocking activation of transcription factor NF-�B. ProcNatl Acad Sci U S A 1996;93:3308–12.

2086 SINGH ET AL