Recombinant human IL-1β was obtained from Genzyme (Cambridge, MA, USA), and recombinant human TNF-α and recombinant human IL-17 were from R&D Systems (Minneapolis, MN, USA). Prostaglandin E₂ (PGE₂) was from Cayman Chemical Co. (Ann Arbor, MI, USA). SB203580, SP600125, PD98059, pyrrolidine dithiocarbamate (PDTC), MG-132 and SN-50 were from Calbiochem (La Jolla, CA, USA). DMEM, penicillin and streptomycin, FCS, and TRIzol^® reagent were from Invitrogen (Burlington, ON, Canada). All other chemicals were purchased from either Bio-Rad (Mississauga, ON, Canada) or Sigma-Aldrich Canada (Oakville, ON, Canada).

Human normal cartilage (from femoral chondyles) was obtained at necropsy, within 12 hours of death, from donors with no history of arthritic diseases (n = 18, age 61 ± 15 years (mean ± SD)). To ensure that only normal tissue was used, cartilage specimens were thoroughly examined both macroscopically and microscopically. Only those with neither alteration were processed further. Human OA cartilage was obtained from patients undergoing total knee replacement (n = 41, age 64 ± 14 years (mean ± SD)). All patients with OA were diagnosed on criteria developed by the American College of Rheumatology Diagnostic Subcommittee for OA 19. At the time of surgery, the patients had symptomatic disease requiring medical treatment in the form of non-steroidal anti-inflammatory drugs or selective COX-2 inhibitors. Patients who had received intra-articular injections of steroids were excluded. The Clinical Research Ethics Committee of Notre-Dame Hospital approved the study protocol and the use of human tissues.

Chondrocytes were released from cartilage by sequential enzymatic digestion as described previously 11. In brief, this consisted of 2 mg/ml pronase for 1 hour followed by 1 mg/ml collagenase for 6 hours (type IV; Sigma-Aldrich) at 37°C in DMEM and antibiotics (100 U/ml penicillin, 100 μg/ml streptomycin). The digested tissue was briefly centrifuged and the pellet was washed. The isolated chondrocytes were seeded at high density in tissue culture flasks and cultured in DMEM supplemented with 10% heat-inactivated FCS. At confluence, the chondrocytes were detached, seeded at high density, and allowed to grow in DMEM supplemented as above. The culture medium was changed every second day, and 24 hours before the experiment the cells were incubated in fresh medium containing 0.5% FCS. Only first-passaged chondrocytes were used.

Cartilage specimens were processed for immunohistochemistry as described previously 4. The specimens were fixed in 4% paraformaldehyde and embedded in paraffin. Sections (5 μm thick) of paraffin-embedded specimens were deparaffinized in toluene, then dehydrated in a graded ethanol series. The specimens were then preincubated with chondroitinase ABC (0.25 U/ml in PBS, pH 8.0) for 60 minutes at 37°C, followed by incubation with Triton X-100 (0.3%) for 30 minutes at 25°C. Slides were then washed in PBS followed by 2% hydrogen peroxide/methanol for 15 minutes. They were further incubated for 60 minutes with 2% normal serum (Vector Laboratories, Burlingame, CA, USA) and overlaid with primary antibody for 18 hours at 4°C in a humidified chamber. The antibody was a rabbit polyclonal anti-human PPARγ (Santa Cruz Biotechnology, Santa Cruz, CA, USA), used at 10 μg/ml. This antibody recognizes the epitope of the sequence mapping of amino acids 8 to 106 at the N terminus of PPARγ. Each slide was washed three times in PBS, pH 7.4, and stained with the use of the avidin-biotin complex method (Vectastain ABC kit; Vector Laboratories). The color was developed with 3,3'-diaminobenzidine (DAB) (Vector Laboratories) containing hydrogen peroxide. The slides were counterstained with eosin. The specificity of staining was evaluated by using antibody that had been preadsorbed (1 hour at 37°C) with a 20-fold molar excess of the protein fragment corresponding to amino acids 6 to 105 of human PPARγ (Santa Cruz), and by replacing the primary antibody with non-immune rabbit IgG (Chemicon, Temecula, CA, USA; used at the same concentration as the primary antibody). The evaluation of positive-staining chondrocytes was performed with our previously published method 4. For each specimen, six microscopic fields were examined under ×40 magnification. The total number of chondrocytes and the number of positive-staining chondrocytes were evaluated and results were expressed as the percentage of chondrocytes that stained positive (cell score).

Total RNA from homogenized cartilage or stimulated chondrocytes was isolated by using TRIzol^® reagent (Invitrogen) in accordance with the manufacturer's instructions. To remove contaminating DNA, isolated RNA was treated with RNase-free DNase I (Ambion, Austin, TX, USA). The RNA was quantified with the RiboGreen RNA quantitation kit (Molecular Probes, Eugene, OR, USA), dissolved in diethylpyrocarbonate-treated water and stored at -80°C until use. Total RNA (1 μg) was reverse-transcribed with Moloney murine leukemia virus reverse transcriptase (Fermentas, Burlington, ON, Canada) as detailed in the manufacturer's guidelines. One-fiftieth of the reverse transcriptase reaction was analyzed by real-time quantitative PCR as described below. The following primers were used: PPARγ1 sense, 5'-AAAGAAGCCAACACTAAACC-3'; PPARγ2 sense, 5'-GCGATTCCTTCACTGATAC-3'; common PPARγ1 and PPARγ2 antisense, 5'-CTTCCATTACGGAGAGATCC-3'; glyceraldehyde-3-phosphate dehydrogenase (GAPDH) sense, 5'-CAGAACATCATCCCTGCCTCT-3'; and GAPDH antisense, 5'-GCTTGACAAAGTGGTCGTTGAG-3'.

Quantitative PCR analysis was performed in a total volume of 50 μl containing template DNA, 200 nM sense and antisense primers, 25 μl of SYBR^® Green master mix (Qiagen, Mississauga, ON, Canada) and uracil-N-glycosylase (UNG, 0.5 U; Epicentre Technologies, Madison, WI, USA). After incubation at 50°C for 2 minutes (UNG reaction), and at 95°C for 10 minutes (UNG inactivation and activation of the AmpliTaq Gold enzyme), the mixtures were subjected to 40 amplification cycles (15 s at 95°C for denaturation, and 1 minute for annealing and extension at 60°C). Incorporation of SYBR Green dye into PCR products was monitored in real time with a GeneAmp 5700 Sequence detection system (Applied Biosystems, Foster City, CA, USA) allowing determination of the threshold cycle (C_t) at which exponential amplification of PCR products begins. After PCR, dissociation curves were generated with one peak, indicating the specificity of the amplification. A threshold cycle (C_t value) was obtained from each amplification curve with the software provided by the manufacturer (Applied Biosystems).

Relative amounts of mRNA in normal and OA cartilage were determined with the use of the standard curve method. Serial dilutions of internal standards (plasmids containing cDNA of target genes) were included in each PCR run, and standard curves for the target gene and for GAPDH were generated by linear regression with a plot of log(C_t) against log(cDNA relative dilution). C_t values were then converted to the number of molecules. Relative mRNA expression in cultured chondrocytes was determined with the ΔΔC_t method, as detailed in the manufacturer's guidelines (Applied Biosystems). A ΔC_t value was first calculated by subtracting the C_t value for the housekeeping gene GAPDH from the C_t value for each sample. A ΔΔC_t value was then calculated by subtracting the ΔC_t value of the control (unstimulated cells) from the ΔC_t value of each treatment. Fold changes compared with the control were then determined by raising 2 to the -ΔΔC_t power. Each PCR reaction generated only the expected specific amplicon as shown by the melting-temperature profiles of the final product and by gel electrophoresis of test PCR reactions. Each PCR was performed in triplicate on two separate occasions for each independent experiment.

The luciferase reporter construct pGL3-PPARγ1p3000, containing a 3,000-base-pair fragment of the human PPARγ1 gene promoter, was kindly provided by Dr Johan Auwerx (Institut de Génétique et de Biologie Moléculaire et Cellulaire, Illkirch, France) 9. β-Galactosidase reporter vector under the control of SV40 promoter (pSV40-β-galactosidase) was from Promega (Madison, WI, USA). Transient transfection experiments were performed with FuGene-6 (1 μg of DNA to 4 μl of FuGene 6; Roche Applied Science, Laval, QC, Canada) in accordance with the manufacturer's recommended protocol. In brief, chondrocytes were seeded and grown to 50 to 60% confluence. The cells were transfected with 1 μg of the reporter construct and 0.5 μg of the internal control pSV40-β-galactosidase. Six hours later, the medium was replaced with DMEM containing 1% FCS. The next day, the cells were treated for 18 hours with or without IL-1. After harvesting, luciferase activity was determined and normalized to β-galactosidase activity 16.

Chondrocytes were lysed in ice-cold lysis buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 2 mM EDTA, 1 mM PMSF, 10 μg/ml each of aprotinin, leupeptin, and pepstatin, 1% Nonidet P40, 1 mM Na₃VO₄, 1 mM NaF). Lysates were sonicated on ice and centrifuged at 12,000 r.p.m. for 15 minutes. The protein concentration of the supernatant was determined with the bicinchoninic acid method (Pierce, Rockford, IL, USA). Total cell lysate (20 μg) was subjected to SDS-PAGE and electrotransferred to a nitrocellulose membrane (Bio-Rad). After blocking in 20 mM Tris-HCl, pH 7.5, containing 150 mM NaCl, 0.1% Tween 20, and 5% (w/v) non-fat dry milk, blots were incubated overnight at 4°C with the primary antibody and washed with a Tris buffer (Tris-buffered saline, pH 7.5, containing 0.1% Tween 20). The blots were then incubated with horseradish peroxidase-conjugated secondary antibody (Pierce), washed again, incubated with SuperSignal Ultra Chemiluminescent reagent (Pierce), and, finally, exposed to Kodak X-Omat film (Eastman Kodak Ltd, Rochester, NY, USA).

Data are expressed as means ± SEM unless stated otherwise. Statistical significance was assessed by the two-tailed Student's t test; p < 0.05 was considered significant.

To examine the expression and localization of PPAR-γ protein in cartilage, we performed an immunohistochemical analysis. We found that chondrocytes in both normal and OA cartilage express PPARγ protein. The immunostaining for PPARγ was essentially located in the superficial zones, and was lower in OA cartilage than in normal cartilage. Statistical evaluation of the cell score for PPARγ indicated significant differences between normal cartilage (22 ± 2.5% (mean ± SEM)) and cartilage from mild to moderate OA (11 ± 3%; Figure 1a,b). Similarly, PPARγ expression was significantly reduced in severe OA cartilage (10 ± 2%, data not shown). By contrast, in intact OA cartilage, the positive staining seemed lower, but the differences were not significant (data not shown). The specificity of the staining was confirmed by using antibodies that had been preadsorbed (1 hour, 37°C) with a 20-fold molar excess of the protein fragment corresponding to amino acids 6 to 105 of human PPARγ (Figure 1c) or non-immune serum (Figure 1c). PPARα and PPARβ were also expressed in normal, mild to moderate, and severe OA cartilage, but no significant differences were observed between the cartilage groups (Additional file 1).

PPARγ has two isoforms, PPARγ1 and PPARγ2, which are generated by alternative promoters and differential splicing 9. To examine which PPARγ transcripts were expressed in cartilage, we determined absolute mRNA concentrations of PPARγ1 and PPARγ2 by quantitative real-time PCR. As shown in Figure 2, PPARγ1 abundance represents about 90% of the total PPARγ mRNA. Thus, human cartilage expresses high levels of γ1 mRNA, the isoform that is generally expressed in various tissues, and low levels of the γ2 isoform, which is more selectively expressed in adipose tissue 10. The level of PPARγ1 expression in OA cartilage was 2.4-fold lower than in normal cartilage (p < 0.005). However, no significant differences in mRNA levels of PPARγ2 were seen between normal and OA cartilage (Figure 2). These observations demonstrate a selective downregulation of PPARγ1 in OA cartilage. In preliminary experiments we showed that the amplification efficiency of PPARγ1, PPARγ2, and GAPDH were approximately equal, ranging between 1.95 and 2.

The reduced expression of PPARγ1 in OA cartilage suggests that humoral factors produced in the OA joint downregulate PPARγ1 expression. We therefore evaluated the effect of IL-1, one of the most prominent mediators in OA, on PPARγ1 expression in cultured chondrocytes. OA chondrocytes were treated with 100 pg/ml IL-1 for 0, 3, 6, 12, and 24 hours; the levels of PPARγ1 protein were then analyzed by Western blotting. In preliminary experiments we found that, as in cartilage, cultured chondrocytes express predominantly the PPARγ1 isoform but not the adipocyte-specific PPARγ2 isoform. As shown in Figure 3a, PPARγ1 protein expression was not significantly affected after 3 hours of stimulation with IL-1. The level of PPARγ1 protein then started to decline gradually at 6 hours and remained low until at least 24 hours. Subsequently, we examined the effect of various concentrations of IL-1 on PPARγ1 protein expression. As shown in Figure 3b, the expression of PPARγ1 was downregulated by IL-1 in a concentration-dependent manner; significant decreases were observed at a concentration as low as 10 pg/ml. Maximal decreases were obtained at an IL-1 concentration of 100 pg/ml (Figure 3b). No modulation of PPARα and PPARβ expression was seen (Additional file 2).

In addition to IL-1, the pro-inflammatory mediators TNF-α, IL-17, and PGE₂ also contribute to the pathogenesis of OA 123. We therefore examined their effects on PPARγ1 protein expression. Cultured OA chondrocytes were incubated for 24 hours with IL-1 (100 pg/ml), TNF-α (1 and 10 ng/ml), IL-17 (10 and 100 ng/ml), and PGE₂ (0.1 and 1 μM), and the expression levels of PPARγ1 were determined by Western blotting. As shown in Figure 4, and like IL-1, TNF-α, IL-17, and PGE₂ also downregulated PPARγ1 protein expression. Similar results were obtained with normal chondrocytes (n = 3; data not shown).

To elucidate the mechanism responsible for the changes in amounts of PPARγ1 protein, we measured the steady-state level of PPARγ1 mRNA by quantitative real-time PCR. Expression of the gene encoding GAPDH was used for normalization. The relative expression level of PPARγ1 mRNA was plotted as a percentage decrease compared with untreated control cells (Figure 5a). Consistent with its effects on protein expression (Figure 3b), IL-1 downregulates PPARγ1 mRNA expression in a dose-dependent manner in OA chondrocytes. The effect of IL-1 on PPARγ1 mRNA expression was maximal (about 85% decrease) at 100 pg/ml. A dose-dependent effect of IL-1 on PPARγ1 mRNA expression was also observed in normal chondrocytes (n = 3; data not shown).

To characterize the effect of IL-1 on PPARγ1 expression further, we performed transient transfection experiments with the reporter construct pGL3-PPARγ1p3000, containing about 3,000 base pairs of regulatory sequence of the gene encoding human PPARγ1 9. As shown in Figure 5b, IL-1 suppressed PPARγ1 promoter activity in a dose-dependent manner. The effect of IL-1 on PPARγ1 promoter activity was optimal at 100 pg/ml (about 65% decrease). Taken together, these data strongly suggest that IL-1 suppressed PPARγ1 expression at the transcriptional level.

IL-1 is known to induce its effects in chondrocytes through activation of a plethora of signaling pathways, including the mitogen-activated protein kinases (MAPKs) c-Jun N-terminal kinase (JNK), p38, and extracellular signal-regulated kinase (ERK) 20. To assess the contribution of these pathways in the IL-1-mediated downregulation of PPARγ1, OA chondrocytes were pretreated for 30 minutes with selective inhibitors for the above pathways, and then stimulated or not with IL-1 for 18 hours. Total cell lysates were analyzed for PPARγ1 protein expression by Western blotting. As shown in Figure 6a, IL-1 reduced PPARγ1 expression remarkably, confirming the results seen previously (Figure 3). Pretreatment with SB203580, a specific p38 MAPK inhibitor, as well as pretreatment with SP600125, a selective inhibitor of JNK, dose-dependently abolished IL-1-induced downregulation of PPARγ1 expression. Conversely, PD98059, a selective inhibitor of ERK, had no effect on IL-1-induced downregulation of PPARγ expression, even at a high concentration (20 μM). None of the MAPK inhibitors had an effect on PPARγ expression in the absence of IL-1. These results suggest that the MAPKs JNK and p38, but not ERK, are involved in the suppression of PPARγ1 expression by IL-1.

Because NF-κB mediates many of the effects of IL-1 in a variety of cell types including chondrocytes, we examined the role of this transcription factor in the repression of PPARγ1. We used three different pharmacological inhibitors of the NF-κB pathway: the antioxidant PDTC, a proteasome inhibitor MG-132, and an inhibitor of NF-κB translocation (SN-50). Cells were pretreated with increasing concentrations of each inhibitor for 30 minutes and then subsequently treated with 100 pg of IL-1 for 18 hours.

As shown in Figure 6b, treatment with IL-1 decreased PPARγ1 expression, but this IL-1 effect was dose-dependently abolished in the presence of each of the three NF-κB inhibitors (PDTC, MG-132, and SN-50). None of the NF-κB inhibitors had an effect on basal PPARγ1 expression. These results imply that NF-κB activation participates in the IL-1-mediated downregulation of PPARγ1 expression.