The study was approved by the Institutional Review Board of Taichung Veterans General Hospital (IRB TCVGH NO: C09248), and informed written consent was obtained from patients. All animal experiments were conducted in accordance with the Guidelines for Animals Research of Agriculture Council, ROC and approved by the Ethical Committee for Animal Research of the National Taiwan University (IACUC: 20120439).

Mouse monoclonal antibody for α-tubulin, C23, NF-κB p65 and rabbit polyclonal antibody for IgG, IKKα/β, IκBα, NF-κB p50, NF-κB p65 and goat anti-mouse or anti-rabbit secondary antibody conjugated with horseradish peroxidase were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Mouse monoclonal antibody for phosphor-IκBα and rabbit monoclonal antibody for phosphor-IKKα/β were from Cell Signaling Technology (Danvers, MA, USA). Rabbit anti-5-LOX antibody was from Novus (Littleton, CO, USA). 5-LOX inhibitors, including MK-886, Nordihydroguaiaretic acid (NDGA); leukotriene B4 and leukotriene B4 receptor antagonist LY29311 were from Cayman Chemical Company (Ann Arbor, MI, USA). Collagenase and 4′,6-diamidino-2-phenylindole (DAPI) were from Sigma-Aldrich (St. Louis, MO, USA). Recombinant human TNF-α and enhanced chemiluminescent HRP substrate (ECL) were from Millipore (Bedford, MA, USA). We purchased RPMI-1640 medium, trypsin and anti-rabbit secondary antibody conjugated with Alexa Fluor 488 from Invitrogen (Carlsbad, CA, USA) and fetal bovine serum (FBS) from Biological Industries (Kibbutz Beit Haemek, Israel). Tri-zol was from MDBIO (Taipei, Taiwan). MMLV Reverse Transcriptase kit was from Promega (Madison, WI, USA). Taqman PCR Master Mix and qPCR probes were from Applied Biosystems/Invitrogen (Foster city, CA, USA).

Human synovial fibroblasts were isolated by collagenase treatment from synovial tissues obtained from patients with rheumatoid arthritis (RA) undergoing total knee replacement surgeries (Taichung Veterans General Hospital, Taichung, Taiwan) [17]. Patients with RA were fulfilled with diagnostic criteria of American College of Rheumatology (ACR). Fresh synovial tissues were minced and digested in a solution of collagenase, and DNase. Isolated synovial fibroblasts were filtered through 70 µm nylon filters. The cells were then grown on culture dishes in 95% air-5% CO₂ with RPMI-1640, which was supplemented with 10% heat-inactivated fetal calf serum, 100 U/ml penicillin, 100 µg/ml streptomycin, and 250 ng/ml fungizone (pH 7.6). Over 90% cultured cells were fibroblasts which were characterized by flow cytometry by using CD90 (Thy-1) antibody [18]. Synovial fibroblasts of passages four to nine were used in this study.

Total RNA were extracted from human synovial fibroblasts using a Tri-zol kit. The absorbance was measured in a spectrophotometer, Picodrop (Picodrop Ltd., Essex, UK) at 260 and 280 nm. RNA was used for RT-PCR by using two-step MMLV RT kit. Gene expression was detected by Real-Time PCR which was executed by using a SYBR Green PCR Master Mix (Applied Biosystems, Foster city, CA, USA) and an ABI StepOnePlus Real-time PCR system (Applied Biosystems). The cDNA was amplified with gene specific primers as shown below:

GAPDH: [19].

Forward primer: TGGGTGTGAACCATGAGAAG.

Reverse primer: GCTAAGCAGTTGGTGGTGC.

IL-6: [20].

Forward primer: TCACTGGTCTTTTGGAGTTTGA.

Reverse primer: AGAGCCCTCAGGCTGGACT.

MCP-1/CCL-2: [21].

Forward primer: ATGCAATCAATGCCCCAGTC.

Reverse primer: TGCAGATTCTTGGGTTGTGG.

Amplification was performed in the following cycling conditions: 50°C for 2 min and 95°C for 10 min and then 40 cycles at 95°C for 15 s followed by 60°C for 1 min. The optimal concentrations of primers and templates used in each reaction were established based on the standard curve created before reaction and corresponding to nearly 100% efficiency of the reaction. The reference household gene used to normalize the amount of mRNA was GAPDH. The fold change in gene expression relative to control was calculated by 2^−ΔΔCT.

After treatment with TNF-α or test substances, the levels of IL-6 or MCP-1 in the culture medium of human synovial fibroblasts and serum from mice were determined by enzyme-linked immunosorbent assay (ELISA). The conditioned medium or serum from mice was obtained and IL-6 or MCP-1 was detected according to the manufacturer’s protocol (R&D Systems, Minneapolis, MN, USA). All independent experiments were performed and the absorbance was determined using microplate reader (Bio-Tek, Winooski, VT, USA).

The 5-LOX-shRNA conjugated on the vector of pLKO.1 with ampicillin-resistant region was purchased from National RNAi Core Facility (RNAi Core, Taipei, Taiwan). 3 µg 5-LOX-shRNA and 6 µl Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) were premixed with OPTI-MEM separately for 5 min and then mixed with each other for 25 min and then applied to human synovial fibroblasts. The empty vector of shRNA was used as negative control. After 36 hr, the cells were harvested or the culture medium was replaced with serum-free medium and test substances.

After treatment with test substances, synovial fibroblasts were then washed with cold PBS and lysed for 30 min at 4°C with lysis buffer as described previously [18]. For the separation of cytoplasmic extracts (CE) and nuclear extracts, cells were cultured onto 10 cm dish. After reaching confluence, cells were treated with test substances, cytoplasmic extracts and nuclear extracts were separated by NE-PER (Thermo Scientific-Pierce, Rockford, IL, USA). Equal protein (30 µg) was applied per lane, and electrophoresis was performed under denaturing conditions on a 8% or 12% SDS gel and transferred to nitrocellulose membranes (Invitrogen). The blots were blocked with 5% non-fat milk in TBS-T (0.5% Tween 20 in 20 mM Tris and 137 mM NaCl) for 1 h at room temperature and then probed with antibodies against 5-LOX, phosphor-IκBα, IκBα, phospho-IKKα/β, IKKα/β, NF-κB p50, NF-κB p65 (1∶1000) at 4°C overnight. After 3 washes by TBS-T, the blots were subsequently incubated with goat anti-rabbit or anti-mouse peroxidase-conjugated secondary antibody (1∶10000) for 1 h at room temperature. The blots were visualized by enhanced chemiluminescence using Amersham HyperfilmTM ECL (GE Healthcare, Pollards Wood, UK) or Biospectrum Imaging System (UVP, Upland, CA, USA). For normalization purposes, the same blot was also probed with mouse anti-α-tubulin antibody or mouse anti-C23 antibody (1∶1000).

For immunolabeling studies, synovial fibroblasts were seeded on glass overnight and then treated with test substances. After TNF-α (25 ng/ml) treatment for 30 min with or without inhibitors, 4% paraformaldehyde was used to fix the cells for 15 min. 0.1% triton was used to permeabilize the cells and 4% BSA was used for blocking non-specific binding for 1 hr. Synovial fibroblasts were stained with primary mouse monoclonal antibody against NFκB-p65 (1∶200) (Santa Cruz sc-8008) (Lin et al., 2011) overnight and then with Alexa Fluor 488-conjugated goat anti-mouse secondary antibody (Invitrogen, Carlsbad, California). The nucleus was counterstained by DAPI and the confocal images were obtained by using excitation wavelength 494 nm and emission wavelength 519 nm, respectively (for Alexa-488) (model SP5 TCS; Leica, Heidelberg, Germany) and the thickness of optical sections was set at 0.8 µm.

8 weeks-old male C57BL/6 mice and 15-LOX knockout mice were purchased from the Laboratory Animal Center of National Taiwan University. All mice were kept under standard temperature, humidity, and timed lighting conditions and provided mouse chow and water ad libitum.

Recombinant mouse TNF-α (40 µg/kg) in the presence or absence of 5-LOX inhibitor NDGA (10 mg/kg) was injected intravenously from femoral vein of male C57BL/6 mice. Mice were sacrificed 6 hr later and serum was collected to perform ELISA analysis.

For the paw swelling experiment, TNF-α (200 ng dissolved in 10 µl PBS) in the presence or absence of NDGA (10 µM) or MK-886 (10 µM) was injected intraplantarly into the right hind paw and the PBS-injected left hind paw was used as control. Paw thickness was measured with digimatic caliper (Mitutoyo, Japan). The change of paw thickness was compared to the paw thickness at 0 hr.

The data given were means ± S.E.M. The significance of difference between the experimental group and control was assessed by one-way analysis of variance (ANOVA) and 2-tailed Student’s t-test. The difference is significant if the p value is less than 0.05.

In inflammatory arthritis, synovial fibroblasts contribute to synovium inflammation, angiogenesis, matrix degradation by its abilities to produce inflammatory cytokines, matrix degradation enzymes and angiogenic factor [22]. In addition, TNF-α is the potent inflammatory cytokine and is responsible for synovial fibroblasts-involved cartilage degradation. First of all, we examined whether TNF-α can enhance the production of pro-inflammatory cytokine, IL-6. Human synovial fibroblasts were incubated with TNF-α (10 ng/ml) in serum-free medium for several time intervals. The mRNA and released protein level of IL-6 were evaluated by quantitative PCR and ELISA, respectively. Figures 1A and 1B showed that treatment with TNF-α enhanced the mRNA and protein levels of IL-6 in a time-dependent manner. Treatment of TNF-α for 6 hr increased mRNA expression of IL-6 to 28.2±3.9 fold of control (Figure 1A). In addition, treatment of TNF-α also increased IL-6 protein release to 9.9±2.3 fold and 40.0±7.0 fold of control at 6 or 12 hr, respectively (Figure 1B).

To investigate the role of 5-LOX in rheumatoid arthritis, human synovial fibroblasts were pretreated with 5-LOX inhibitors NDGA (5 and 10 µM) and MK-886 (5 µM) for 1 hr and then treated with TNF-α (10 ng/ml) for next 6 hr. mRNA extracted from cell lysates was evaluated by quantitative PCR and the conditioned medium collected was measured for IL-6 protein release by ELISA assay. Real-time PCR analysis showed that TNF-α significantly increased IL-6 mRNA levels and pretreatment with 5-LOX inhibitors NDGA (5 and 10 µM) and MK-886 (5 µM) could inhibit the upregulatory effect of TNF-α by 54.7±6.0%, 91.4±4.3% and 57.3±12.6%, respectively (Figure 1C). In addition, treatment with TNF-α increased the secreted IL-6 in the conditioned medium and pretreatment of 5-LOX inhibitors NDGA (5 and 10 µM) and MK-886 (5 µM) decreased the TNF-α-induced IL-6 protein levels by 83.0±1.1%, 96.3±0.4% and 48.1±2.8%, respectively (Figure 1D).

One of potential therapy in RA is targeting chemokines and a randomized clinical trial with an anti-CCL2/MCP-1 monoclonal antibody in RA patients has been reported previously [23]. MCP-1 plays a key role in leukocyte migration and recruitment of mononuclear phagocytes during inflammation in the joint and previous study indicated that there is overproduction of CCL-2/MCP-1 in RA patient’s synovium [24]. To further elucidate the anti-inflammatory effects of 5-LOX inhibitors in the rheumatoid arthritis, human synovial fibroblasts were pretreated with 5-LOX inhibitors NDGA (10 µM) and MK-886 (5 µM) for 1 hr and then treated with TNF-α (10 ng/ml) for another 6 hr. mRNA levels extracted from cell lysates was evaluated by Real-time PCR and the supernatant was collected for measuring MCP-1 protein levels by ELISA assay. Real-time PCR analysis showed that TNF-α significantly increased MCP-1 mRNA expression to 4.5±0.5 fold of vehicle control (Figure 2A). Pretreatment of 5-LOX inhibitors NDGA (10 µM) and MK-886 (5 µM) decreased TNF-α-induced MCP-1 mRNA levels by 70.6±11.0%, and 77.9±3.2%, respectively (Figure 2A). The released MCP-1 was analysed by ELISA analysis and showed that treatment with TNF-α increased the secreted MCP-1 in the conditioned medium to 4.0±0.1 fold of vehicle control (Figure 2B). Pretreatment of 5-LOX inhibitors NDGA (10 µM) and MK-886 (5 µM) antagonized TNF-α-induced MCP-1 protein levels by 89.6±9.1%, and 63.8±12.7%, respectively (Figure 2B).

Leukotriene B4 (LTB4) is one of 5-Lipoxygenase downstream metabolites and plays an important role in pathogenesis of RA [25]. Recent study has suggested that LTB4 is an essential and non-redundant role in both acute and chronic inflammation of rheumatoid arthritis [10]. In addition, exogenous LTB4 could increase the expression of cytokines like TNF-α and IL-1β in human synovial fibroblasts [26]. We thus investigated whether LTB4 is involved incytokines release in human synovial fibroblasts. Human synovial fibroblasts were pretreated with LTB4 receptor antagonist LY29311 (1 µM) for 1 hr and then treated with TNF-α (10 ng/ml) for another 6 hr. mRNA were extracted for Real-time PCR analysis, and the results show that TNF-α-induced up-regulation of IL-6 and MCP-1 were reduced by LY29311 (Figure 3A, 3B). In addition, treatment of TNF-α for 4 hr increased the secretion of LTB4 from 3.9±0.9 to 16.1±2.9 (pg/ml) (Figure 3C). Treatmentwith LTB4 at different concentrations in serum-free RPMI medium for 6 hr and the quantity of IL-6 in conditioned medium was measured. It was found that treatment with LTB4 (10, 100 nM) increased secreted IL-6 protein levels to 2.5±0.5-fold and 3.5±0.3 fold of control, respectively (Figure 3D).

To further examine the role of 5-LOX in TNF-α-induced IL-6 and MCP-1 expression, 5-LOX knockdown was performed by using 5-LOX shRNA and the protein release in the conditioned medium in response to TNF-α administration was measured. Total protein extracts from RASFs transfected with empty vector or 5-LOX shRNA were analyzed by immunoblotting. Figure 4A showed that cells transfected with 5-LOX shRNA No.3 and No.7 significantly decreased 5-LOX expression. Human synovial fibroblasts transfected with empty vector or 5-LOX shRNA clone No.3 and No.7 were then used to treat with TNF-α for 6 hr. IL-6 and MCP-1 protein release in the conditioned medium was evaluated by ELISA analysis. It was found that knockdown of 5-LOX with shRNA No.3 and No.7 decreased TNF-α-induced IL-6 protein release by 16.8±4.7% and 23.8±4.9%, respectively (Figure 4B). In addition, knockdown of 5-LOX with shRNA No.3 and No.7 reduced TNF-α-induced MCP-1 protein release by 36.6±5.2% and 32.6±7.9% following 6 hr treatment with TNF-α (Figure 4C).

NF-κB is implicated in the transcriptional regulation of inflammatory response by TNF-α. Bondeson et al. [27] have demonstrated that the activation of NF-κB plays an important role in the inflammatory action of TNF-α in synovial fibroblasts. To examine whether NF-κB pathway is involved in the antagonism of 5-LOX inhibitors on TNF-α-induced cytokine expression in human synovial fibroblasts, IKKα/β activation, IκBα phosphorylation, and IκBα degradation were evaluated by immunoblotting after stimulation by TNF-α. Since the canonical NF-κB pathway is induced by TNF-α and this activation leads to the recruitment and activation of an IKK complex comprising IKK alpha and/or IKK beta catalytic subunits. The IKK complex then phosphorylated IκBα at ser32 and ser36 to produce ubiquitination of IκBα at lysine residues and subsequent degradation by the 26 s proteasome. NF-κB complex (p50 and p65), which is associated with IκB to retain in the cytosol under resting state, is released and translocates to the nucleus [28]. Therefore, we first investigated the time-course of IκBα phosphorylation in human synovial fibroblasts. Cells were treated by TNF-α (10 ng/ml) for 5, 10, 15, 30, and 60 min and total proteins were extracted for Western blotting. The results showed that treatment with TNF-α (10 ng/ml) enhanced the phosphorylation of IκBα time-dependently in human synovial fibroblasts, and the phosphorylation of IκBα reached a peak at 10 min (Figure 5A). Human synovial fibroblasts were then pre-incubated with 5-LOX inhibitors NDGA (10 µM) or MK-886 (5 µM) for 1 hr and then exposed to TNF-α for another 10 min. It was found that pretreatment with NDGA or MK-886 could antagonize TNF-α-induced IKKα/β activation, IκBα phosphorylation and IκBα degradation (Figure 5B). Pretreatment of NDGA or MK-886 also inhibited the TNF-α-induced nuclear translocation of NF-κB subunits of p50 and p65 (Figure 5C). Immunofluorescence staining was also used to evaluate whether 5-LOX antagonized TNF-α-induced NF-κB nuclear translocation. Human synovial fibroblasts were treated with TNF-α (25 ng/ml) for 30 min and p65 translocated into nucleus (Figure 6). Pretreatment with 5-LOX inhibitors, including NDGA (10 µM) and MK-886 (5 µM) for 1 hr antagonized TNF-α-induced nuclear translocation of p65 (Figure 6). These results indicated that 5-LOX inhibitors decreased TNF-α-induced cytokine/chemokine expression via the abrogation of NF-κB signalling.

We have elucidated the role of 5-LOX in TNF-α-induced cytokines and chemokines production in vitro. We then examined the effect of 5-LOX inhibitor in animal model. Male C57BL/6 mice were injected with TNF-α (40 µg/kg, via femoral vein) with or without 5-LOX inhibitor NDGA (10 mg/kg). After 6 hr, mice were sacrificed and serum was collected to perform ELISA analysis. It was found that mice injected with TNF-α significantly increased MCP-1 levels in serum to 4.4±0.6 fold of vehicle-treated mice. However, co-treatment TNF-α with 5-LOX inhibitor NDGA (10 mg/kg) significantly reduced TNF-α-induced MCP-1 protein release in serum by 67.9±4.3% (Figure 7A). Moreover, we examined the effects of 5-LOX inhibitors on TNF-α-induced acute paw edema in mice model. As shown in figure 7B, intraplantar injection of TNF-α-increased the paw thickness and the thickness reached the peak at 4 hr. Co-administration with 5-LOX inhibitors, NDGA (10 µM) or MK-886 (10 µM) could markedly ameliorate TNF-α-induced paw edema. These results suggest that 5-LOX inhibitor can reduce TNF-α-induced chemokine production and paw edema in vivo.