Promoter sequences of MMP-7, 9, 13, 14, ADAMTS-5, 4, and COL10A1 genes were obtained from CHIP Bioinformatics Tools 35. For each of these genes, we scanned the region from 1,500 base pairs upstream of the transcript start to 100 base pairs downstream of the coding-sequence start to find putative LEF-1 binding elements (5'-CTTTGWW-3'). In addition, LRP-5 promoter was tested for Smad binding sites (SBEs, 5'-GCCGnGCG-3').

Articular cartilage samples were obtained from femoral condyles and tibial plateaus of patients with primary osteoarthritis undergoing knee-replacement surgery at the Orthopaedics Department of University Hospital of Larissa. In total, 18 patients were included in this study (15 women/three men; mean age, 64.7 ± 9.7 years). All osteoarthritic specimens had Mankin scores of 10 to 14. Radiographs were obtained before surgery and graded according to the Kellgren-Lawrence system. All patients had K/L scores ≥ 2. The radiographs were assessed by two independent observers who were blinded to all data of the individuals. Normal articular cartilage was obtained from nine individuals (four women/five men; mean age, 44.6 ± 7.6 years) with 0 Mankin score, undergoing fracture-repair surgery, with no history of joint disease. Patients with rheumatoid arthritis and other autoimmune diseases, as well as chondrodysplasias, infection-induced OA, and posttraumatic OA, were excluded from the study. Written informed consent was obtained from all individuals in the study. The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki, as reflected in a priori approval by the Local Ethical Committee of the University Hospital of Larissa.

Articular cartilage was dissected and subjected to sequential digestion with 1 mg/ml proteinase mixture (Pronase; Roche Applied Science, Mannheim, Germany) and 1 mg/ml collagenase P (Roche Applied Science, Mannheim, Germany). Chondrocytes were counted and checked for viability by using trypan blue staining. More than 95% of the cells were viable after isolation. Isolated chondrocytes from individual specimens were separately cultured with Dulbecco Modified Eagle Medium/Ham F-12 (DMEM/F-12) (GIBCO BRL, Paisley, UK) plus 5% fetal bovine serum (FBS; Invitrogen, Life Technologies, Paisley, UK) at 37^οC under a humidified 5% CO₂ atmosphere until reaching confluence for 4 to 6 days.

Total cellular RNA was extracted from cultured chondrocytes by using Trizol reagent (Invitrogen, Life Technologies, Paisley, UK). Preservation of 28S and 18S ribosomal RNA (rRNA) species was used to assess RNA integrity. All the samples included the study had prominent 28S and 18S rRNA components. The yield was quantified spectrophotometrically. Transcription of 0.1 μg RNA to cDNA was performed by using SuperScript III reverse transcriptase (Invitrogen, Life Technologies, Paisley, UK) and random primers (Invitrogen, Life Technologies, Paisley, UK). Quantification of COL2A1, COL10A1, aggrecan, MMP-13, LRP-5, LRP-6, BMP-2, BMP-4, BMP-7, BMPR-IA, BMPR-IB, LEF-1, and TCF-4 mRNA expression was performed with real-time PCR (ABI 7300; Applied Biosystems, Foster City, CA, USA). Reactions were done in triplicate by using 2 μl of cDNA per reaction. All primers used are shown in Table 1. Real-time PCR validation was carried out by using the 2^-ΔΔCT method. Normalized gene-expression values for each gene based on cycle threshold (C_T) values for each of the genes and the housekeeping gene glyceraldehyde 3-phosphate dehydrogenase (GAPDH) were generated.

Chondrocytes were lysed by using RIPA buffer and a cocktail of protease and phosphatase inhibitors. Protein concentration was quantified by using the Bio-Rad Bradford protein assay (Bio-Rad Protein Assay; BioRad, Hercules, CA, USA) with bovine serum albumin as standard. Cell lysates from normal and OA chondrocytes were electrophoresed and separated on a 4% to 20% Tris-HCl gel (Bio-Rad, Hercules, CA, USA) and transferred to a Hybond-ECL nitrocellulose membrane (Amersan Biosciences, Piscataway, NJ, USA). The membrane was probed with anti-LRP-5 (sc-21389; Santa Cruz Biotechnology, Santa Cruz, CA, USA), BMP-2 (sc-57040; Santa Cruz Biotechnology, Santa Cruz, CA, USA), BMP-4 (AF757, R&D Systems, Minneapolis, MN, USA) BMPR-IA (sc-134285; Santa Cruz Biotechnology, Santa Cruz, CA, USA), LEF-1 antibody (sc-8591; Santa Cruz Biotechnology, Santa Cruz, CA, USA), and phospho-β-catenin (9561; Cell Signaling Technology, Boston, MA, USA), and signals were detected by using immunoglobulin (IgG) conjugated with horseradish peroxidase (1:10,000 dilution). The results were normalized by using anti-β-actin polyclonal antibody (Sigma-Aldrich, St. Louis, MO, USA). The nitrocellulose membranes were then exposed to photographic film, which was scanned, and the intensities of the protein bands were determined with computerized densitometry.

Cartilage specimens were fixed in 10% formalin overnight and decalcified in 13% EDTA (Sigma-Aldrich, St. Louis, MO, USA) for 3 weeks. Paraffin-embedded sections were deparaffinized in xylene, dehydrated through graded alcohols, and placed in 3% H₂O₂ to block endogenous peroxidase. Nonspecific staining was blocked with TBS containing 2% bovine serum albumin (BSA; Sigma-Aldrich, St. Louis, MO, USA) for 1 hour at room temperature, followed by incubation with primary antibody BMP-2 (c-57040; Santa Cruz Biotechnology, Santa Cruz, CA, USA; 1:20 dilution with TBS containing 2% BSA) overnight at room temperature. Normal serum of the same species as the primary antibody was used as a control for the primary antibody. After extensive washing with TBS, the sections were incubated with horseradish peroxidase-conjugated secondary antibody (Goat anti-Mouse; Invitrogen, Life Technologies, Paisley, UK; 1:250 dilution with TBS containing 2% BSA) for 1 hour at room temperature in a humid chamber. Finally, color reaction was performed by using the substrate reagent 3,3'-diaminobenzidine tetrahydrochloride (DAB; Thermo Scientific, Rockford, IL, USA). After being washed, the sections were incubated with DAB and coverslipped with mounting medium.

Normal and osteoarthritic chondrocytes were seeded on six-well plates at 3 × 10⁵ cells/well, and 3 days after seeding, cells were treated with 50 ng/ml of BMP-2 or BMP-4 for 12, 24, and 48 hours; each experiment was conducted in triplicate wells. RNA was extracted, and real-time PCR analysis for osteocalcin, LRP-5, and LRP-6 was performed, as described earlier. All primers used are shown in Table 1. Cell lysates were extracted, and LRP-5 (sc-21389; Santa Cruz Biotechnology, Santa Cruz, CA, USA), phospho-β-catenin (9561; Cell Signaling Technology, Boston, MA, USA), and total β-catenin protein levels (9582; Cell Signaling Technology, Boston, MA, USA) were evaluated by using Western blot analysis, as described earlier.

Normal and osteoarthritic chondrocytes were seeded on six-well plates at 3 × 10⁵ cells/well, and 3 days after seeding, cells were treated with 20 mM LiCl (Sigma-Aldrich, St. Louis, MO, USA) for 12, 24, and 48 hours; each experiment was conducted in triplicate wells. RNA was extracted, and real-time PCR analysis for MMP-7, 9, 13, 14, ADAMTS-4, 5, and collagen X was performed as described earlier. All primers used are shown in Table 1. Cell lysates were extracted, and phospho-β-catenin protein levels were evaluated by using Western blot analysis, as described earlier.

ChIP was performed by using a ChIP assay kit (Upstate USA, Inc., Charlottesville, VA, USA) on BMP-2-treated, LiCl-treated, and untreated normal chondrocytes, as previously described 21. The cell lysates after treatment with BMP-2 or LiCl were precleared by incubation with G-Sepharose beads and were incubated with monoclonal antibody Smad-1/5/8 (Cell signaling Technology, Boston, MA, USA) or polyclonal antibody LEF-1 overnight at 4°C, respectively. Antibody to human purified IgG was used as control (R&D Systems, Minneapolis, MN, USA). The immunoprecipitated DNAs were used for PCR amplification. All primers were designed according to the nucleotide sequence of the gene promoters, and each PCR fragment covered 250 to 400 bp of the promoter. Table 2 shows only the primer sets that amplify the promoter region containing putative sites, as observed after ChIP assay. The PCR products were fractionated on 3% agarose gels and were stained with ethidium bromide.

Normal and osteoarthritic chondrocytes were seeded in six-well plates in DMEM/F-12 containing 5% FBS. After overnight incubation, cells were transfected with specific siRNA against LRP-5 (Ambion, Life Technologies, Paisley, UK), LEF-1 (Invitrogen, Life Technologies, Paisley, UK), and scrambled siRNA (used as a control; Invitrogen, Life Technologies, Paisley, UK) by using LipofectAMINE 2000 (Invitrogen, Life Technologies, Paisley, UK), as described by the manufacturer. No cell toxicity was detected from the transfection agent. Four hours after siRNA transfection, the medium was changed, and cells were treated with 50 ng/ml BMP-2 (cultures with siRNA against LRP-5) or 20 mM LiCl (cultures with siRNA against LEF-1) for 48 hours. In each experiment, the results from three of the six-well plates were averaged and considered as n = 1. No significant variance was observed among the individual wells in each averaged group. RNA was extracted, and real-time PCR analysis for MMP-7, 9, 13, 14, ADAMTS-4, 5, and collagen X was performed, as described earlier. Moreover, cell lysates were extracted, and phospho- as well as total β-catenin protein levels were evaluated in BMP-2-treated chondrocytes after LRP-5 siRNA transfection by using Western blot analysis, as described earlier.

Statistical significance was determined by using the Student t test with a confidence level of 95% (P < 0.05).

To evaluate differences in chondrocyte differentiation-related genes, we evaluated COL2A1, COL10A1, aggrecan, and MMP-13 mRNA expression levels in normal and osteoarthritic chondrocytes. We found that COL2A1 and aggrecan were significantly upregulated in normal chondrocytes (P < 0.05; Figure 1), whereas COL10A1 and MMP-13 were significantly upregulated in osteoarthritic chondrocytes (P < 0.05; Figure 1).

To investigate the involvement of BMP-2 and Wnt/β-catenin signaling pathway in osteoarthritis, we evaluated BMP-2, BMP-4, BMP-7, BMPR-IA, BMPR-IB, LRP-5, LRP-6, phospho-β-catenin, LEF-1, and TCF-4, the major transcription factors of the Wnt canonic signaling pathway expression levels in normal and osteoarthritic chondrocytes. We found that osteoarthritic chondrocytes had significantly higher LRP-5, BMP-2, BMP-4, BMPR-IA, LEF-1 mRNA, and protein expression compared with normal (P < 0.05; Figure 2a, b, c). Moreover, BMP-2 was found to be upregulated in osteoarthritic compared with normal cartilage after immunohistochemistry (Figure 2d). Western blot analysis showed a significant reduction of phospho-β-catenin protein levels in osteoarthritic chondrocytes compared with normal (P < 0.05; Figure 1e and 1f). No significant difference was found in LRP-6, BMP-7, BMPR-IB, and TCF-4 mRNA levels between normal and osteoarthritic chondrocytes (Figure 2a).

BMP-2 and Wnt/β-catenin signaling pathways have been reported to play a significant role in the control of chondrocyte differentiation in the growth plate. We examined the effect of BMP-2 on Wnt/β-catenin signaling-pathway activation after treatment with BMP-2 in normal and osteoarthritic chondrocytes. We found that BMP-2 treatment at 12, 24, and 48 hours increased the active form of nuclear β-catenin protein levels (Figure 3b), decreased phospho-β-catenin protein levels (Figure 3b), and significantly upregulated LRP-5 mRNA and protein expression, but not LRP-6 (P < 0.05; Figure 3a and 3c). However, no difference was noted in the active form of nuclear β-catenin protein levels after treatment with BMP-4 (Figure 3d). Further to investigate the intracellular signaling pathway involved in BMP-2-induced LRP-5 expression, we subjected BMP-2-treated chondrocytes to chromatin immunoprecipitation by using an antibody against Smad-1/5/8 and tested whether Smads bind to LRP-5 promoter via Smad-binding elements and subsequently LRP-5 expression. We observed that LRP-5 promoter contains conserved a Smad-binding site in -280 to -270 from the ATG initiation codon, and Smad1/5/8 binding was enhanced after treatment with BMP-2. Smads binding to osteocalcin promoter served as positive control (Figure 3e).

To provide evidence that BMP-2 stimulates catabolic enzymes and hypertrophic markers through Wnt/β-catenin signaling in osteoarthritic chondrocytes, we blocked LRP-5 mRNA expression by using siRNA against LRP-5 in BMP-2-treated normal and osteoarthritic chondrocytes and evaluated different MMPs (MMP-7, 9, 13, and 14), ADAMTSs (ADAMTS-4 and 5), and collagen X expression levels. siRNA against LRP-5 effectively inhibited LRP-5 expression, as we observed a downregulation of LRP-5 mRNA and protein expression in si-RNA-transfected chondrocytes compared with the untransfected and the scrambled siRNA-transfected chondrocytes 24 hours after siRNA transfection (P < 0.05; Figure 4a). Moreover, we found that LRP-5 silencing in BMP-2-treated normal and osteoarthritic chondrocytes resulted in downregulation of β-catenin protein levels (Figure 4b) and MMPs (MMP-9, 13, and 14), ADAMTS-5, and collagen X mRNA levels (Figure 4d). In addition, phospho-β-catenin protein levels were upregulated after LRP-5 siRNA transfection in BMP-2-treated normal and osteoarthritic chondrocytes (Figure 4c), suggesting that LRP-5 upregulation by BMP-2 contributed to Wnt/β-catenin signaling activation and chondrocyte catabolic activity and hypertrophy.

To obtain direct evidence that the Wnt/β-catenin signaling pathway can trigger extracellular matrix degradation and hypertrophic chondrocyte differentiation in osteoarthritis, we activated the pathway in vitro by using LiCl in normal and osteoarthritic chondrocytes and evaluated the expression levels of basic catabolic (MMP-7, 9, 13, 14, ADAMTS-5, and 4) and hypertrophic markers (collagen X). We found that normal and osteoarthritic chondrocyte treatment with LiCl reduced phospho-β-catenin levels (Figure 5), suggesting the stabilization of β-catenin and the activation of Wnt canonic signaling. We then investigated the expression of different MMPs, ADAMTSs, and collagen X in treated and untreated-LiCl normal and osteoarthritic chondrocytes. We found that MMP-13 and MMP-9 mRNA levels were significantly upregulated after LiCl treatment in normal and osteoarthritic chondrocytes (P < 0.05), whereas MMP-14, ADAMTS-5, and collagen X expression was significantly upregulated in normal chondrocytes (P < 0.05) and showed a trend to increase in osteoarthritic chondrocytes (Figure 6a). siRNA against LEF-1 decreased MMP-13, 9, and 14 mRNA levels in LiCl-treated normal and osteoarthritic chondrocytes (P < 0.05), whereas ADAMTS-5 and collagen X mRNA expression was significantly downregulated in LiCl-treated osteoarthritic chondrocytes (P < 0.05) and showed a trend to decrease in LiCl-treated normal chondrocytes (Figure 6a). However, MMP-7 and ADAMTS-4 mRNA levels remained at the same levels after LiCl treatment and LEF-1 silencing (Figure 6a). To investigate whether LEF-1 binds to MMP-9, 13, 14, ADAMTS-5, and COL10A1 promoters through conserved LEF-1 binding sites and upregulates their expression, we performed a ChIP assay in chondrocytes after treatment with LiCl for 24 and 48 hours. ChIP assay revealed LEF-1 binding sites on MMP-9 (-737 to -733 from the ATG initiation codon), MMP-13 (from -1142 to -1138), MMP-14 (from -575 to -571), ADAMTS-5 (from -755 to -751), and COL10A1 (from -784 to -781) promoters, and the binding was stronger in treated-LiCl chondrocytes, as LiCl stabilizes β-catenin active form and increases nuclear β-catenin protein levels, enhancing thus the complex between β-catenin and LEF-1 and the subsequent binding on gene promoters, suggesting that these genes are Wnt targets in adult articular cartilage (Figure 6b). No LEF-1 binding sites were observed in MMP-7 and ADAMTS-4 promoter in the region from 1,500 base pairs upstream of the transcript start to 100 base pairs downstream of the coding-sequence start (Figure 6b). LEF-1 binding to AXIN-2 promoter served as positive control.