Protein A/G beads, anti-mouse and anti-rabbit IgG-conjugated horseradish peroxidase, rabbit polyclonal antibodies specific for thrombin, PAR1, PAR3, PAR4, c-Src, PKCδ, and Nrf2, and siRNA against PAR1, PAR3, and PAR4 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Rabbit polyclonal antibody specific for c-Src phosphorylated at Tyr⁴¹⁶ and PKCδ phosphorylated at Tyr³³¹ were purchased from Cell Signaling and Neuroscience (Danvers, MA, USA). Rabbit polyclonal antibody specific for Nrf2 phosphorylated at Ser⁴⁰ was purchased from Abcam Inc. (Cambridge, MA, USA). SFLLRN-NH₂ (a PAR1-agonist peptide), TFRGAP-NH₂ (a PAR3-agonist peptide), and GYPGQV-NH₂ (a PAR4-agonist peptide) were purchased from Bachem. Rottlerin, GF109203X, Ro320432, and PP2 were purchased from Calbiochem (Darmstadt, Germany). The c-Src dominant-negative mutant was a gift from Dr. S. Parsons (University of Virginia Health System, Charlottesville, VA, USA). The human HO-1 promoter construct was gift from Dr. Y.C. Liang (Taipei Medical University, Taipei, Taiwan). The pSV- β-galactosidase vector and luciferase assay kit were purchased from Promega (Madison, WI, USA). All other chemicals were purchased from Sigma-Aldrich (St. Louis, MO, USA).

We obtained approval from the local ethics committee, and subjects gave informed written consent. Human synovial fibroblasts were isolated by collagenase treatment of synovial tissue samples obtained from 32 patients with OA during knee-replacement surgeries and 18 samples of nonarthritic synovial tissues obtained at arthroscopy after trauma/joint derangement. The concentration of thrombin in the synovial fluid of selected patients was measured with an enzyme-linked immunosorbent assay (ELISA) according to the protocol provided by the manufacturer (Human thrombin ELISA kit; Abcam). OASFs were isolated, cultured, and characterized, as previously described 1718. Experiments were performed by using cells from passages 3 to 6.

Total RNA was extracted from synovial fibroblasts by using a TRIzol kit (MDBio Inc., Taipei, Taiwan). The reverse-transcription reaction was performed by using 2 μg of total RNA that was reverse transcribed into cDNA by using oligo (dT) primer 1920. The quantitative real-time PCR (qPCR) analysis was carried out by using Taqman one-step PCR Master Mix (Applied Biosystems, Foster City, CA, USA). cDNA templates (2 μl) were added per 25-μl reaction with sequence-specific primers and Taqman probes. Sequences for all target gene primers and probes were purchased commercially (β-actin was used as internal control) (Applied Biosystems). The qPCR assays were carried out in triplicate on a StepOnePlus sequence-detection system. The cycling conditions involved 10-minute polymerase activation at 95°C, followed by 40 cycles at 95°C for 15 seconds and 60°C for 60 seconds. The threshold was set above the nontemplate control background and within the linear phase of the target gene amplification to calculate the cycle number at which the transcript was detected (denoted CT).

Cellular lysates were prepared as described previously 2122. Proteins were resolved on SDS-PAGE and transferred to Immobilon polyvinyldifluoride (PVDF) membranes. The blots were blocked with 4% BSA for 1 hour at room temperature and then probed with rabbit anti-human antibodies against PKCδ, c-Src, or Nrf2 (1:1,000) for 1 hour at room temperature. After three washes, the blots were subsequently incubated with donkey anti-rabbit peroxidase-conjugated secondary antibody (1:3,000) for 1 hour at room temperature. The blots were visualized by enhanced chemiluminescence with Kodak X-OMAT LS film (Eastman Kodak, Rochester, NY, USA).

Human synovial fibroblasts were co-transfected with 0.8 μg HO-1-luciferase plasmid and 0.4 μg β-galactosidase expression vector. Fibroblasts were grown to 80% confluence in 12-well plates and were transfected the following day with Lipofectamine 2000 (LF2000; Invitrogen). DNA and LF2000 were premixed for 20 minutes and then applied to cells. After 24-hour transfection, cells were incubated with the indicated agents. After further 24-hour incubation, the media were removed, and cells were washed once with cold PBS. To prepare lysates, 100 μl of reporter lysis buffer (Promega, Madison, WI, USA) was added to each well, and cells were scraped from dishes. The supernatant was collected after centrifugation at 13,000 rpm for 2 minutes. Aliquots of cell lysates (20 μl) containing equal amounts of protein (20 to 30 μg) were placed into wells of an opaque black 96-well microplate. An equal volume of luciferase substrate was added to all samples, and luminescence was measured in a microplate luminometer. The value of luciferase activity was normalized to transfection efficiency monitored by the co-transfected β-galactosidase expression vector 23.

PKCδ activity was assessed with a PKCδ Kinase Activity Assay Kit according to manufacturer's instructions (Assay Designs, MI, USA). The PKCδ activity kit is based on a solid-phase ELISA that uses a specific synthetic peptide as a substrate for PKCδ and a polyclonal antibody that recognizes the phosphorylated form of the substrate.

Chromatin immunoprecipitation (ChIP) analysis was performed as described previously 24. DNA immunoprecipitated by anti-Nrf2 antibody was purified. The DNA was then extracted with phenol-chloroform. The purified DNA pellet was subjected to PCR. PCR products were then resolved by using 1.5% agarose gel electrophoresis and visualized with UV. The primers: 5'-CCATCAAACTTTAACTCGGTGA-3' and 5'- GACTTGGGAGATAGAAGGAACG-3' were used to amplify across the human HO-1 promoter region (-857 to -752) 24.

The values are reported as mean ± SEM. Statistical analysis between two samples was performed by using the Student t test. Statistical comparisons of more than two groups were performed by using one-way analysis of variance (ANOVA) with the Bonferroni post hoc test. In all cases, P < 0.05 was considered significant.

It has been reported that clotting factors and fibrinolytic products are increased in synovial fluid of patients with RA and OA 25. Therefore, we examined the levels of thrombin expression in samples from patients with OA and found that the expression of thrombin protein in human OASFs (Figure 1A, lines 4 to 6) was significantly higher than that in normal SFs (Figure 1A, lines 1 to 3). The OASFs medium showed a level of expression of thrombin that was significantly higher than that seen in the medium from normal SFs (Figure 1B). In addition, concentrations of thrombin in synovial fluid were significantly higher in patients with OA than in controls (Figure 1B). We applied thrombin directly to OASFs to examine the expression of HO-1. Treatment of OASFs with thrombin (0.1 to 3 U/ml) for 24 hours induced HO-1 expression in a concentration-dependent manner (Figure 1C, E), and this induction occurred in a time-dependent manner (Figure 1D, F). After thrombin (3 U/ml) treatment for 24 hours, the amount of HO-1 expression had increased in OASFs (Figure 1D, F). To confirm further this stimulation-specific mediation by thrombin, PPACK, a thrombin inhibitor, was used. Pretreatment of cells with PPACK effectively antagonized the potentiating effect of thrombin on HO-1 expression (Figure 2A). These data indicated that thrombin increases HO-1 expression in human OASFs.

Thrombin exerts its effects through interaction with specific PAR1, PAR 3, and PAR4 receptors 26. To investigate the role of PAR1, PAR3, and PAR4 subtype receptors in thrombin-mediated increase of HO-1 expression, cells were treated with PAR1-, 3-, and 4-specific agonist peptides and then examined for expression levels of HO-1. Of the agonist peptides tested, the PAR1-selective receptor agonist peptide, SFLLRN-NH₂ (100 μM) and TFRGAP-NH₂ (PAR3 agonist peptide; 100 μM) significantly increased the expression of HO-1 (Figure 2A). In contrast, GYPGQV-NH₂ (PAR4 agonist peptide; 100 μM) failed to upregulate HO-1 expression. Furthermore, thrombin, SFLLRN-NH₂, and TFRGAP-NH₂ only slightly increased HO-1 expression in normal synovial fibroblasts (Figure 2B), indicating that PAR agonist-induced HO-1 expression is more important during OA pathogenesis. To confirm that PAR1 and PAR3 subtype receptors are involved in the thrombin-mediated increase of HO-1 expression, specific inhibition of PAR-receptor expression was accomplished with siRNA (Figure 2C). It was found that PAR1 and PAR3 receptor-specific siRNA but not PAR4-receptor siRNA significantly blocked the thrombin-mediated increase of HO-1 expression (Figure 2D, E), indicating that interactions between thrombin and PAR1/PAR3 are important for HO-1 expression in human OASF cells.

PKC has been shown to play an important role in cellular functions modulated by several stimuli, including thrombin 2728. To determine whether PKC isoforms were involved in thrombin-triggered HO-1 expression, OASFs were pretreated for 30 minutes with either GF109203X, a pan-PKC inhibitor, or rottlerin, a selective PKCδ inhibitor 29, and then incubated with thrombin for 24 hours. As shown in Figure 3A and 3C, pretreatment with GF109203X and rottlerin but not PKCα inhibitor (Ro320432, 10 μM) reduced thrombin-induced HO-1 expression, which suggests that PKCδ plays a specific role in thrombin-induced HO-1 expression in OASFs. Transfection of cells with PKCδ siRNA also reduced thrombin-induced HO-1 expression (Figure 3B). We directly measured the PKCδ phosphorylation response to thrombin. Stimulation of OASFs with thrombin led to a significant increase in phosphorylation of PKCδ (Figure 3D). The PKCδ activity in OASFs was increased by thrombin treatment in a dose-dependent manner (Figure 3E). Pretreatment of cells with PPACK or transfection of cells with PAR1 and PAR3 siRNA also reduced thrombin-mediated PKCδ kinase activity (Figure 3F). Based on these results, thrombin appears to act through a PAR1/PAR3 and PKCδ-dependent signaling pathway to enhance HO-1 expression in human synovial fibroblasts.

PKCδ-dependent c-Src activation is involved in the regulation of COX-2 expression 30, and we investigated the role of Src in mediating thrombin-induced HO-1 expression with the specific Src inhibitor PP2. As shown in Figure 4A and 4B, thrombin-induced HO-1 expression was markedly attenuated by pretreatment of cells for 30 minutes with PP2 or by transfection of cells for 24 hours with c-Src mutant. The major phosphorylation site of c-Src at the Tyr⁴¹⁶ residue results in activation of c-Src autophosphorylation 31. To examine directly the crucial role of c-Src in HO-1 expression, we measured the level of c-Src phosphorylation at Tyr⁴¹⁶ in response to thrombin. As shown in Figure 4C, treatment of OASFs with thrombin resulted in a time-dependent phosphorylation of c-Src at Tyr⁴¹⁶. Pretreatment of cells with PPACK and rottlerin markedly inhibited the thrombin-induced c-Src kinase activity (Figure 4D). Based on these results, thrombin appears to act through a signaling pathway involving the PAR1/PAR3 receptor, PKCδ, and c-Src to enhance HO-1 expression in human synovial fibroblasts.

Activation of Nrf2 has been reported to play an important role in HO-1 expression 32. To determine the role of Nrf2 in thrombin-mediated HO-1 expression, OASFs were transfected with Nrf2 siRNA. Transfection of cells with Nrf2 siRNA suppressed the thrombin-induced HO-1 expression (Figure 5A). Stimulation of cells with thrombin also induced Nrf2 phosphorylation in a time-dependent manner (Figure 5B). Pretreatment of cells with GF109203X, rottlerin, and PP2 attenuated thrombin-induced Nrf2 activation (Figure 5C).

We next investigated whether Nrf2 binds to the ARE element on the HO-1 promoter after thrombin stimulation. The in vivo recruitment of Nrf2 to the HO-1 promoter (-857 to -752) was assessed with chromatin immunoprecipitation assays. In vivo binding of Nrf2 to the ARE element of the HO-1 promoter occurred after thrombin stimulation (Figure 6A). The binding of Nrf2 to the ARE element by thrombin was attenuated by GF109203X, rottlerin, and PP2 (Figure 6A).

To study further the pathways involved in the action of thrombin-induced HO-1 expression, transient transfection was performed by using the human HO-1 promoter-luciferase construct. Treatment of cells with thrombin led to dose-dependent increases in HO-1 promoter activity (Figure 6B). PPACK, GF109203X, rottlerin, and PP2; PAR1, PAR3, and PKCδ siRNA; and c-Src mutant all antagonized the stimulation of HO-1 promoter activity by thrombin (Figure 6C, D). These results indicated that thrombin-induced HO-1 expression is mediated through the PAR1/PAR3, PKCδ, c-Src, and Nrf2 pathways in OASFs.