Human femoral condylar cartilages were obtained at total knee arthroplasty from 19 patients (n = 4 men, mean age 73.7 ± 6.4 years, range 66 to 80 years; n = 15 women, mean age 77.2 ± 11.3 years, range 50 to 90 years) with OA diagnosed in accordance with the criteria of the American College of Rheumatology 15. The study was approved by McGill University Ethics Review Board.

Femoral condylar articular cartilages were isolated and prepared for culture as described previously 316. All cartilages exhibited macroscopic articular surface differences from normal cartilages. To generate sufficient cartilage to perform each of these analyses on a patient, all of the available articular cartilage from each patient, regardless of the degree of degeneration (Mankin grades 4 to 12), was used, as described previously 31718. In brief, OA articular cartilages were washed three times with DMEM (Gibco BRL, Life Technologies, Grand Island, NY, USA) containing 20 mM HEPES (4-(2-hydroxyethyl)-1-piperazine ethanesulfonic acid) buffer pH 7.4 (Gibco BRL), 45 mM NaHCO₃, 100 units/ml penicillin, 100 μg/ml streptomycin and 150 μg/ml gentamycin sulphate. Full-depth cartilage slices from a single site, about 20 mm × 20 mm, were cut vertical to the articular surface and then into cubes of about 2 mm × 2 mm. Five to seven cubes were randomly obtained and wet weights of about 60 mg were distributed in each culture well (48-well Costar 3548 plate; Corning Inc., Corning, NY, USA). Samples were maintained before culture for 48 hours at 37°C in 1 ml per well of medium A in 95% air/5% CO₂.

In these and our previous studies of these human knee OA cartilages we used a standard sampling procedure 819. OA femoral condylar cartilages show variations in cartilage thickness and various degrees of degradation between weight-bearing and non-weight-bearing regions. These result in variation between samples, although we do our best to finely chop, mix and randomly distribute the tissue in our culture wells from a given joint and person. In spite of this variability we observed a significant difference in collagen cleavage activity between PGE₂-treated cartilage explants and the untreated controls.

The amount of OA cartilage, which is always macroscopically different from normal, is always limiting. We did not use Mankin grading because the degeneration is very variable within the joint and, as in the previous studies, we had to use all the cartilage available from each patient to be able to perform these analyses, otherwise it would not have been possible to conduct the experiments. This mixture of finely chopped cartilages therefore represents different degrees of degeneration existing within a given joint, from Mankin grade 4 to grade 12.

Media were changed after 48 hours (day 0) and thereafter were replaced every 4 days. Final concentrations of 1 pg/ml to 10 ng/ml PGE₂ (Sigma Chemical Co., St. Louis, Mo, USA) or 5 ng/ml TGF-β2 (R&D Systems, Minneapolis, MN, USA) were freshly added to medium A from day 0 at each medium change. The cartilage (triplicate cultures for each analytical point) was cultured for a total of 16 days and analyzed at day 16 for COL2A1 cleavage by collagenases and proteoglycan release. The conditioned media were collected every 4 days at each medium change from day 4 to day 16 and stored at -20°C until analyzed. For analyses of gene expression, separate cultures were maintained for up to 48 hours and analyzed as described below.

The OA cartilage explants from day 16 of culture were digested and extracted with α-chymotrypsin to solubilize denatured collagen including the carboxy-terminal neoepitope COL2–3/4C short (C1,2C) epitope generated by the cleavage of COL2A1 by collagenase. This was measured as described previously in α-chymotrypsin extracts and conditioned media by ELISA 1718. Total amount of cleavage neoepitope in cartilages and media was calculated by summation of the data from each medium change and cartilage analysis. Results were expressed as pmoles of epitope per mg wet weight of cartilage, based on a molecular mass of the standard peptide epitope of 608 Da.

This was determined in cartilage extracts and conditioned media as sulfated glycosaminoglycans (GAGs), which is primarily a measure of proteoglycan aggrecan content, using a modification of the colorimetric 1,9-dimethylmethylene blue dye-binding assay 20. Cumulative proteoglycan release (GAG in the medium) and its content in cartilages (GAG in the cartilage) were expressed as mirograms of GAG per milligram wet weight of cartilage. GAG release into medium was represented as a percentage of the total cumulative GAG in the cartilage plus cumulative GAG release into medium. This provides an accurate measure of release, accounting for the marked variation in cartilage GAG content in OA cartilage 1821.

PGE₂ concentrations were determined in the control undiluted conditioned media collected at each medium change, which were then pooled for analysis with a commercially available competitive ELISA kit (Cayman Chemical Company, Ann Arbor, MI, USA,) in accordance with the manufacturer's instructions. Results were expressed in picograms per milligram wet weight of cartilage. PGE₂ standards provided with the kit were diluted with DMEM.

Total RNA isolation for the detection of gene expression was based on methodology previously described 3. This was isolated from articular cartilage explants or isolated chondrocytes after up to 96 hours in culture as indicated in the Results section. Fresh cartilage tissue or cells in solution D (4 M guanidine isothiocyanate, 25 mM sodium citrate pH 7.0, 0.5% laurylsarcosine, 0.1 M 2-mercaptoethanol) were immediately frozen in liquid nitrogen and kept at -80°C until all samples had been collected. Samples were defrosted at 21–23°C and then vortex-mixed vigorously for 30 minutes. The debris was removed by centrifugation at 5,000 g for 10 minutes at 4°C. Proteins and nucleic acids in the supernatant were precipitated with 1 volume of propan-2-ol overnight at -20°C. The precipitate was removed at 10,000 g for 20 minutes at 4°C and resuspended in digestion buffer (10 mM Tris-HCl pH 8.0, 5 mM ethylenediaminetetraacetic acid, 1% SDS, with 2 mg/ml Proteinase K (Gibco BRL)) and incubated at 50°C until the pellet disappeared. After extraction with a mixture containing 1 volume of phenol, 0.2 volumes of chloroform and 0.1 volume of 2 M sodium acetate pH 4.0, the aqueous phase was recovered by centrifugation (10,000 g for 30 minutes at 4°C). An equal volume of 70% ethanol was added to each sample aqueous phase and loaded on an RNeasy spin column (Qiagen, Valencia, CA, USA). Further RNA purification was performed with an RNeasy kit (Qiagen) in accordance with the manufacturer's instructions.

The reverse transcriptase reaction was performed with total RNA isolated from articular cartilage explants and SuperScript TMII H^- Reverse Transcriptase as recommended by Gibco BRL-Invitrogen, (Burlington, ON, Canada) and as described previously 3.

Preformed primers and probes for the TaqMan assay (Applied Biosystems, Foster City, CA, USA) of human genes used in this study were MMP-13 Applied Biosystems, TaqMan Gene Expression Assays, accession no. Hs00233992_m1), COL10A1 (Hs00166657_m1), MMP-1(Hs00233958_m1), IL-1β (Hs00174097_m1) and TNF-α (Hs00174128_m1). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH), β-actin, 18s RNA and cyclophilin were examined as endogenous controls. GAPDH has been chosen for the final data presentation because it gives the lowest variation.

Quantification of gene expression levels of mRNA was performed with a 7500 Real-time PCR System (Applied Biosystems). After treatment with 1 μl of RNase H, 1 μl of reverse transcription product was subjected to real-time PCR in a 25 μl total reaction mixture containing 12.5 μl of TaqMan Universal PCR Master Mix (Applied Biosystems), 900 nM sense and antisense primers, 50 nM probe, and template cDNA. After a single step of 50°C for 2 minutes and initial activation at 95°C for 10 minutes, reaction mixtures were subjected to 40 amplification cycles (15 seconds at 95°C for denaturation, and 1 minute of annealing and extension at 60°C).

By using a sequence detection system the threshold cycle (C_t) was determined at which the exponential amplification of PCR products began. 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 by using SDS V 1.3 software provided by the manufacturer.

Relative mRNA expression was determined with the ΔΔC_t method, as detailed by manufacturer guidelines (Applied Biosystems). A ΔC_t value was 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 cartilage) from the ΔC_t value of each treatment. Fold changes compared with the control were determined by raising 2 to the power -ΔΔC_t; that is, 2^-ΔΔCt. Each PCR was performed in duplicate on three separate occasions for each independent experiment. Three 'no template' controls were consistently negative for each reaction.

To avoid variation in efficiency between experiments, all samples of the same cartilage were simultaneously subjected to reverse transcription, and all samples of cDNA were simultaneously amplified by real-time PCR.

The relative quantification protocol for evaluating gene expression with real-time reverse transcription-mediated PCR is based on the comparison of gene expression level in the treated cartilage explants and untreated cartilage control samples, both of which are normalized to the GAPDH expression level. The expression level of each control is set to 1. We therefore show only one control bar for all the genes measured in the set.

The deviations in treated samples represent variations in gene expression in triplicates of cultured cartilage from each individual. Real-time PCR reaction was performed in duplicate for each cartilage sample.

Quantitative data are expressed as means ± SD. Assays were run in at least triplicate. A normality test showed that data were distributed in accordance with a Gaussian distribution curve. To analyze treatment effects, Mann–Whitney and paired t-test analyses were used. P < 0.05 was considered significant.

Our previous studies of cultured human OA articular cartilage explants have shown that conditioned media from cultures maintained with TGF-β2 for 16 days contained concentrations of PGE₂ in the range 4 to 125 pg/ml. We used this information to determine whether the addition of exogenous PGE₂ concentrations in this range would influence the cleavage of COL2A1 by collagenases in human OA explants. OA articular cartilage explants showed significant variability in the responsiveness to PGE₂. Collagen cleavage was downregulated by exogenous PGE₂ at concentrations as low as 0.1 pg/ml (data not shown) and 1 pg/ml in one of four patients (Figure 1b). However, in most examined cartilages and as shown by the group analyses, significant downregulation of collagen cleavage was observed at 10 pg/ml PGE₂ in these four individual patients (Figure 1e) and in a larger cohort of 13 other patients (Figure 1f). A mean inhibition of 39.7% (range 33.5 to 49.8%) by 10 pg/ml PGE₂ was observed and it was as effective as TGF-β2 (40.1%; range 26.9 to 49.4%). Interestingly, at higher concentrations (0.1 to 10 ng/ml) this inhibition was no longer seen. In one case, stimulation was observed at 10 ng/ml PGE₂ (Figure 1b).

We observed an inverse relationship between PGE₂ concentration and collagen cleavage activity in OA explant cultures that served as controls by plotting PGE₂ concentration versus collagen cleavage in the control OA articular cartilage explants (Figure 1g). An increase in PGE₂ levels in OA articular cartilage explants was accompanied by a linear (correlation coefficient r = -0.507; P = 0.044) decrease in collagen cleavage.

Downregulation of collagen cleavage was not accompanied by significant changes in GAG release in explants cultured in the presence of 10 pg/ml PGE₂ (Figure 2). However, TGF-β2 significantly upregulated the release of GAG in three out of four cartilages examined.

In comparison with controls, OA explants from five patients cultured in the presence of 10 pg/ml PGE₂ showed decreased expression of the genes related to chondrocyte hypertrophy, namely those encoding COL10A1 and MMP-13 (Figure 3). Expression of collagenase MMP-1 was significantly downregulated in four out of five patients. Cytokines were most strongly downregulated by prostaglandin and no expression of IL-1β (Figure 3a,c,d) or TNF-α (Figure 3a,b,c,d) was observed in explants cultured in its presence. However, expression of cyclin B2, caspase 3, TGF-β2, COX2 and PGES-1 was not significantly affected by 10 pg/ml PGE₂ (data not shown). The data for the whole group are shown in Figure 3f. The variability of the residual expression level of genes such as those encoding IL-1 and TNF may result from the differences in cytokine expression associated with various disease states or activities.