Rabbit articular chondrocytes were released from cartilage slices of 2-week-old New Zealand white rabbits by enzymatic digestion. Briefly, cartilage slices were aseptically dissected and then dissociated enzymatically for 6 hr in 0.2% collagenase type II (381 U/mg solid, Sigma Louis, MO, U.S.A.) in phosphate-buffered saline, and then individual cells were obtained by collecting the supernatant after brief centrifugation. The cells were resuspended in Dulbecco's modified Eagle's medium (DMEM, Gibco-BRL, Gaithersburg, MD, U.S.A.) supplemented with 10% (v/v) fetal bovine-calf serum, 50 µg/mL streptomycin, and 50 units/mL penicillin, after which they were plated on culture dishes at a density of 5×10⁴ cells/cm². The medium was changed every 2 days after seeding, and cells reached confluence in approximately 5 days. Differentiation status of articular chondrocytes was determined by examining the accumulation of sulfated glycosaminoglycan with Alcian blue staining or expression of type II collagen (Chemicon, Temecula, CA, U.S.A.) by immunoblot analysis as described previously (23).

Whole cell lysates were prepared by extracting proteins using a buffer containing 50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Nonidet P-40, and 0.1% sodium dodecylsulfate, supplemented with protease inhibitors (10 µg/mL leupeptin, 10 µg/mL pepstatin A, 10 µg/mL aprotinin and 1 mM of 4-[2-aminoethyl] benzenesulfonyl fluoride) and phosphatase inhibitors (1 mM NaF and 1 mM Na₃VO₄). The proteins were size-fractionated by SDS-polyacrylamide gel electrophoresis and transferred to a nitrocellulose membrane. The nitrocellulose sheet was then blocked with 3% non-fat dry milk in Tris-buffered saline. COX-2 was detected using antibody purchased from Cayman Chemical (Ann Arbor, MI, U.S.A.). SOX-9, ERK-2, and PPAR-γ were detected using antibodies purchased from Santa Cruz Biotech (Santa Cruz, CA, U.S.A.). The bands were visualized using peroxidase-conjugated secondary antibodies and chemiluminescence.

Rabbit joint cartilage explants or arthritic cartilage were fixed in 4% paraformaldehyde in PBS for 24 hr at 4℃, washed with PBS, dehydrated with graded ethanol, embedded in paraffin, and sectioned at 4-µm thickness. The sections were stained by standard procedures using Alcian blue or antibody against type II collagen or COX-2 and visualized by developing with a kit purchased from DAKO (Carpinteria, CA, U.S.A.), following the procedure recommended by the manufacturer.

PGE₂ production was determined by measuring the levels of cellular and secreted PGE₂ using an assay kit (Amersham Pharmacia Biotech, NJ, U.K.). Briefly, chondrocytes were seeded in standard 96-well microtiter plates at 2×10⁴ cells/well. Following addition of the indicated pharmacological reagents, supernatant was used to quantify the amount of PGE₂, according to the manufacturer's protocol. PGE₂ levels were calculated against a standard curve of PGE₂ and normalized against the amount of genomic DNA.

The reporter construct (3xPPRE-TK-luc/pGL3) used in the transfections contained three copies of the peroxisome proliferator response element (PPRE) site from the aP2 enhancer (ARE7) inserted upstream of a minimal thymidine kinase (TK) promoter in the pGL3 basic luciferase vector and RXRα cDNA (a gift from Dr. JB, Kim at Seoul National University). Transfection of the reporter vector was performed as described previously (23). Chondrocytes cells were grown in 10% fetal calf serum/DMEM and co-transfected with various reporter plasmids (1 µg) and pCMV β-galactosidase plasmid (0.5 µg) (Clontech, Palo Alto, CA, U.S.A.) using Lipofectamine (Life Technologies, Inc., Grand island, NY, U.S.A.), following the procedure recommended by the manufacturer. The transfected cells were cultured in complete medium for 24 hr and used for further assay. Cells were treated with increasing concentrations (0-20 µM) or times (0-24 hr) of 15d-PGJ2. Luciferase and β-galactosidase activities were determined in all extracts using a Luciferase assay (Promega, Madison, WI, U.S.A.) and Galacto-light system (Tropix, Inc., Bedford, MA, U.S.A.). The activities were corrected for transfection efficiency with β-gal activity and expressed as fold-induction relative to the level in untreated cells. Each assay was performed in triplicate. Data are the mean +/- SD of three separate experiments performed with cells from different animals (^**pd 0.01 versus non treated cells).

Primary culture chondrocytes were treated with 15d-PGJ2. Total RNA was isolated using RNA STAT-60 (Tel-Test B, Inc., Friendswood, TX, U.S.A.) and reverse-transcribed with MoMLV-RT (Invitrogen, California, U.S.A.) as previously described (26). The following primers and conditions were used for PCR in rabbit articular chondrocytes: for type II collagen (370-bp product), 5'-GAC CCC ATG CAG TAC ATG CG-3' (sense) and 5'-AGC CGC CAT TGA TGG TCT CC-3' (antisense) with an annealing temperature of 55℃; for SOX-9 (386-bp product), 5'-GCG CGT GCA GCA CAA GAA GGA CCA CCC GGA TTA CAA GTA C-3' (sense) and 5'-CGA AGG TCT CGA TGT TGG AGA TGA CGT CGC TGC TCA GCT C-3' (antisense) with an annealing temperature of 62℃; for COX-2 (298 bp product), 5'-TCA GCC ACG CAG CAA ATC CT-3' (sense) and 5'-GTG ATC TGG ATG TCA GCA CG-3' (antisense) with an annealing temperature of 52℃ for glyceraldehydes 3-phosphate dehydrogenase (GAPDH; 299-bp products), 5'-TCA CCA TCT TCC AGG AGC GA-3' (sense) and 5'-CAC AAT GCC GAA GTG GTC GT-3' (antisense) with an annealing temperature of 50℃ The primers for rabbit type II collagen and rabbit SOX-9 were de- signed based on the sequence of human homologs, and sequencing of PCR product showed 89% and 93% homology with corresponding human genes.

The results are expressed as the means±S.E. values calculated from the specified number of determinations. Analysis of variance (ANOVA) was performed to compare individual treatments with their respective control values. The Turkey's honestly significant difference method was employed when making multiple comparisons among means (24).

To examine the effects of 15d-PGJ2 on articular cartilage chondrocyte differentiation, rabbit articular chondrocytes in primary culture were treated with 10 µM 15d-PGJ2 for 48 hr. Under these experimental conditions 15d-PGJ2 displayed no cytotoxicity, as assessed by the conversion of a tetrazolium compound to its formazan product, as well as trypan blue exclusion (data not shown). 15d-PGJ2 inhibited type II collagen, a marker for differentiation of chondrocytes, and SOX-9 expression, a major transcription factor that regulates type II collagen expression, in a dose- and time-dependent manner as determined by immunoblot analysis (Fig. 1A). RT-PCR also showed that expression of type II collagen and SOX-9 was reduced (Fig. 1B). Similarly, 15d-PGJ2 treatment of primary culture cells blocked the accumulation of sulfated proteoglycan in a dose- and time-dependent manner (Fig. 1C, D). The effects of rosiglitazone, a PPAR-γ natural ligand, on dedifferentiation and COX-2 expression in articular chondrocytes showed the same results (data not shown). These results indicate that 15d-PGJ2 induces dedifferentiation of articular chondrocytes in primary culture cells.

The effect of 15d-PGJ2 on COX-2 expression and PGE₂ production was investigated using primary culture chondrocytes. In chondrocytes, 15d-PGJ2 increased protein levels of COX-2 in a time- and dose-dependent manner as determined by immunoblot analysis (Fig. 2A). COX-2 expression was increased 1 hr after 15d-PGJ2 treatment, and the levels peaked at 6 hr. 15d-PGJ2 treatment, however, did not affect PPAR-γ expression levels. 15d-PGJ2 also enhanced COX-2 expression as determined by RT-PCR (Fig. 2B). Consistent with the induction of COX-2 expression, 15d-PGJ2 stimulated PGE₂ production in a time- and dose-dependent manner (Fig. 2C, D). The production of PGE₂ was determined by using a PGE₂ assay kit. PGE₂ levels were calculated against a standard curve of PGE₂. These data indicate that 15d-PGJ2 not only causes dedifferentiation of articular chondrocytes but also stimulates COX-2 expression and PGE₂ production.

To examine the effects of 15d-PGJ2 on articular cartilage chondrocyte dedifferentiation and COX-2 expression, cartilage explant cultures were treated with 10 µM 15d-PGJ2 for 24 hr, and cartilage-specific matrix molecules were determined. 15d-PGJ2 caused a dramatic loss of type II collagen (Fig. 3, middle panel) and sulfated proteoglycans (Fig. 3, left panel) as determined by immunohistochemical staining and Alcian blue staining, respectively. However, 15d-PGJ2 stimulated COX-2 expression (Fig. 3, right panel) as determined by immunohistochemical staining. These data indicate 15d-PGJ2 also induces dedifferentiation and COX-2 expression in cartilage explants.

To determine whether the PPAR-γ expressed in the rabbit articular chondrocytes were transcriptionally active, a PPRE cloned upstream of luciferase (3xPPRE-TK-Luc/pGL3) was transiently transfected into articular chondrocyte cells, and the cells were then treated with 15d-PGJ2. As expected, expression of luciferase activity was significantly induced after 15d-PGJ2 treatment in a time- and dose-dependent manner (Fig. 4). The transfection efficiency was above 40% efficiency. These results indicate that 15d-PGJ2 increased PPAR-γ transcription activity in articular chondrocytes.

Because 15d-PGJ2 in articular chondrocytes causes both dedifferentiation and COX-2 expression/PGE₂ production, we next examined whether PPAR-γ activation is associated with dedifferentiation and/or COX-2 expression and PGE₂ production. 15d-PGJ2-induced inhibition of type II collagen and SOX-9 expression was prevented by the addition of 10 µM BADGE (bisphenol A diglycidyl ether, PPAR-γ antagonist) (Fig. 5A). The 15d-PGJ2-induced inhibition of sulfated proteoglycan on primary culture chondrocytes was completely recovered by BADGE pretreatment (Fig. 5B). BADGE did not affect 15d-PGJ2-stimulated COX-2 expression (Fig. 5A) or PGE₂ production (Fig. 5C). To elucidate the dedifferentiation mechanism, we examined PPAR-γ-response element (PPRE)-luciferase activity. As shown in Fig. 5D, 15d-PGJ2-enhanced PPAR-γ activity was blocked by BADGE pretreatment. On the basis of these results, we suggest that 15d-PGJ2 regulates dedifferentiation via a PPAR-γ-dependent pathway and COX-2 expression/PGE₂ production via a PPAR-γ-independent pathway.