Introduction
Macrophage migration inhibitory factor (MIF) plays a crucial role in rheumatoid arthritis (RA) pathogenesis, linking the innate and adaptive immune responses [
1,
2]. As well as its role in inflammatory responses, MIF takes part in the destructive process in RA. In RA joint destruction, matrix metalloproteinases (MMP) are thought to play an important role in synovial invasion [
3,
4]. Various MMPs are upregulated in RA synovial fluid and synovium [
4‐
6], and MIF upregulates MMP-1, MMP-2, and MMP-3 expression in RA synovial fibroblasts [
4,
6]. MIF also induces MMP-9 and MMP-13 in rat osteoblasts [
7]. Besides the induction of MMPs, MIF participates indirectly in joint destruction by promoting angiogenesis in RA synovial fibroblasts [
8] and inducing many osteoclast (OC)-inducing molecules such as TNF-α, IL-1, IL-6, and prostaglandin E
2 (PGE
2) [
1,
2,
9,
10].
MIF-deficient mice are resistant to ovariectomy-induced bone loss and MIF transgenic mice have high-turnover osteoporosis, suggesting that MIF could mediate bone resorption during bone remodeling and balance [
11,
12]. MIF also upregulates the expression of receptor activator of nuclear factor-κB ligand (RANKL) mRNA in murine osteoblasts. MIF has no effect on bone formation
, indicating that it might play a role in the physiological or pathological metabolism of bone, especially in bone resorption [
12]. However, a recent study suggests that MIF inhibits osteoclastogenesis, based on the result that MIF inhibits OC formation in murine bone marrow (BM) cultures in the presence of RANKL. BM cells from MIF knockout mice had an increased capacity to form OC, and MIF knockout mice had decreased trabecular bone volume with low turnover [
13].
To date, the effects of MIF on osteoclastogenesis have not been studied in the context of human disease systems. Two clinical studies suggest that MIF might be involved in joint destruction in RA patients. Greater circulating MIF levels correlate with more severe radiographic joint damage [
14], and the MIF concentration of synovial fluid is significantly higher in RA patients with bony erosion than in those without [
8]. RA joint destruction is closely related to osteoclastogenesis and the major inducer of OC, RANKL. So, we hypothesized that MIF may play an important role in the process of bone destruction in RA patients through the induction of RANKL or direct involvement of osteoclastogenesis. Thus we needed a greater understanding of the relation between MIF and the pathogenesis of bony destruction in RA. In this study, we determined the effect of MIF on RANKL induction in human RA synovial fibroblasts, the relation of RANKL and MIF, and the role of MIF in OC differentiation in RA patients.
Materials and methods
Patients
Synovial fluids were obtained from 16 RA patients fulfilling the 1987 revised criteria of the American College of Rheumatology (formerly the American Rheumatism Association) [
15]. Informed consent was obtained from all patients, and the experimental protocol was approved by the Institutional Review Board for Human Research, Konkuk University Hospital (KUH1010186). Synovial tissues were isolated from eight RA patients (mean age 63.4 ± 4.6, range 38 to 76 years) undergoing total knee replacement surgery.
Isolation of synovial fibroblasts
Synovial fibroblasts were isolated by enzymatic digestion of synovial tissues obtained from RA patients undergoing total joint replacement surgery, as described previously [
16].
Reagents
Recombinant human (rh) MIF, rhRANKL and rh monocyte-colony stimulating factor (M-CSF) were purchased from R&D Systems (Minneapolis, MN, USA). Parthenolide, curcurmin and cyclosporin A were obtained from Sigma Chemical Co. (St. Louis, MO, USA). LY294002, SB203580, SP600125, PD98059, and AG490 were obtained from Calbiochem (Schwalbach, Germany). Anti-human IL-1β, TNF-α, IL-6, RANKL and MIF were purchased from R&D Systems (Minneapolis, MN, USA).
Determination of concentrations of soluble RANKL and MIF by sandwich ELISA
Concentrations of soluble (s) RANKL and MIF in sera and synovial fluids were measured by sandwich ELISA as described previously [
16].
Immunohistochemistry of RA synovium and synovial fibroblasts
Immunohistochemical staining for RANKL and MIF was performed on sections of synovium. Briefly, synovium samples were obtained from patients, fixed in 4% paraformaldehyde solution overnight at 4°C, dehydrated with alcohol, washed, embedded in paraffin, and sectioned into slices 7 μm thick. The sections were depleted of endogenous peroxidase activity by adding methanolic H2O2 and were blocked with normal serum for 30 minutes. After overnight incubation at 4°C with polyclonal anti-human RANKL and anti-MIF antibodies (Santa Cruz Biotechnology, Santa Cruz, CA, USA), the samples were incubated with the appropriate secondary antibodies biotinylated anti-rabbit IgG or biotinylated anti-goat IgG for 20 minutes and then incubated with streptavidin-peroxidase (Vector, Peterborough, UK) for one hour followed by incubation with 3,3"-diaminobenzidine (Dako, Glostrup, Denmark) for five minutes. The sections were counterstained with hematoxylin. Samples were photographed with an Olympus photomicroscope (Tokyo, Japan). Synovial fibroblasts were grown in 150 mm dishes in DMEM complete medium, plated at a density of 1 × 104 cells/cm2 onto glass coverslips (12 mm diameter), and stimulated with rhMIF (0.1, 1, 5, and 10 ng/mL) (R&D Systems, Minneapolis, MN, USA). Cells were fixed in 4% paraformaldehyde for immunohistochemical analysis using anti-RANKL antibody 72 hours after the addition of rhMIF.
Expression of RANKL mRNA measured by real-time reverse transcription polymerase chain reaction amplification
RA synovial fibroblasts were stimulated with rhMIF (0.1, 1, 5, and 10 ng/mL). For signal pathway analysis of RANKL, synovial fibroblasts were incubated in the presence or absence of LY294002 (20 μM), SB203580 (10 μM), SP600125 (1 μM), PD98059 (10 μM), AG490 (50 μM), cyclosporin A (100 nM), parthenolide (10 μM), or curcumin (10 μM) for one hour before the addition of rhMIF. After incubation for 72 hours, mRNA was extracted using RNAzol B (Biotex Laboratories, Houston, TX, USA) according to the manufacturer's instructions. RT-PCR of 2 μg of total mRNA was carried out at 42°C using the SuperScript™ reverse transcription system (Takara, Shiga, Japan). PCR was performed in a 20 μl final volume in capillary tubes in a LightCycler instrument (Roche Diagnostics, Mannheim, Germany). The reaction mixture contained 2 μl of LightCycler FastStart DNA MasterMix for SYBR® Green I (Roche Diagnostics, Mannheim, Germany), 0.5 μM of each primer, 4 mM MgCl2, and 2 μl of template DNA. All capillaries were sealed, centrifuged at 500 g for five seconds and then amplified in a LightCycler instrument, with activation of polymerase (95°C for 10 minutes), followed by 45 cycles of 10 seconds at 95°C, 10 seconds at 60°C (for β-actin control) and at 59°C (RANKL), and 10 seconds at 72°C. The temperature transition rate was 20°C/second for all steps. The double-stranded PCR product was measured during the 72°C extension step by detection of fluorescence associated with the binding of SYBR Green I to the product. Fluorescence curves were analyzed with LightCycler software (v. 3.0; Roche Diagnostics, Mannheim, Germany). The relative expression level of each sample was calculated as the level of RANKL, tartrate-resistant acid phosphatase (TRAP), cathepsin K, calcitonin receptor, or MMP-9 normalized to the endogenously expressed housekeeping gene for β-actin. Melting curve analysis was performed immediately after the amplification protocol under the following conditions: 0 seconds (hold time) at 95°C, 15 seconds at 71°C, and 0 seconds (hold time) at 95°C. The rate of temperature change was 20°C/second, except for 0.1°C/second in the final step. The melting peak generated represented the amount of specific amplified product. The crossing point (Cp) was defined as the maximum of the second derivative from the fluorescence curve. Negative controls were included and contained all elements of the reaction mixture except template DNA. All samples were processed in duplicate.
Western blot analysis
Synovial fibroblasts were incubated with rhMIF for 30 minutes, a whole cell lysate was prepared from about 2 × 105 cells by homogenization in the lysis buffer, and the lysate was centrifuged at 14,000 rpm for 15 minutes. The protein concentration in the supernatant was determined using the Bradford method (BioRad, Hercules, CA, USA). Protein samples were separated on 10% SDS-PAGE and transferred to a nitrocellulose membrane (Amersham Pharmacia Biotech, Uppsala, Sweden). For western hybridization, the membrane was preincubated with 0.5% skim milk in Tris-buffered saline (TBS) with 0.1% Tween 20 (TTBS) at room temperature for two hours. The primary antibody to phospho-Akt, phospho-STAT3, phospho-IκBα, phospho-c-Jun (Cell Signaling Technology Inc, Danvers, MA, USA) or phospho-p38 mitogen-activated protein kinase (MAPK; Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA) diluted 1:1000 in 5% bovine serum albumin, TTBS, was added and incubated overnight at 4°C. The membrane was washed four times with TTBS, horseradish peroxidase-conjugated secondary antibody was added, and the membrane was incubated for one hour at room temperature. After TTBS washing, the hybridized bands were detected using an ECL detection kit and Hyperfilm-ECL reagents (Amersham Pharmacia Biotech, Uppsala, Sweden).
Monocyte isolation
Peripheral blood mononuclear cells (PBMC) were separated by Ficoll-Hypaque (Sigma Chemicals, Poole, Dorset, UK) density gradient centrifugation from buffy coats obtained from healthy volunteers. The cells were washed three times with sterile phosphate-buffered saline (PBS) and resuspended in RPMI 1640 (Life Technologies, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (FBS), 2 mM l-glutamine, and 1% penicillin/streptomycin, henceforth called complete medium. Freshly isolated PBMCs were incubated at 37°C in complete medium and allowed to adhere for 45 minutes. The nonadherent cells were removed and the adherent cells were washed with sterile PBS, harvested with a rubber policeman, and stained with monocyte-specific anti-CD14 monoclonal antibody to assess the purity of the preparation. Of the isolated cells, 90% expressed CD14.
RA synovial fibroblasts were seeded into 12-well multiwell dishes (5 × 10
3 cells/well) and stimulated with rhMIF for three days. As described above, isolated human monocytes (5 × 10
4 cells/well) were added to the stimulated fibroblasts with fresh media. The cells were cocultured for three weeks in α-minimal essential medium (MEM) and 10% heat-inactivated FBS in the presence of 25 ng/mL of rhM-CSF. The medium was changed on day three and then every other day. The addition of rhRANKL protein, prepared as described previously [
17], was used as a positive control. On day 21, TRAP-positive cells were identified using a leukocyte acid phosphatase kit according to the manufacturer's recommended protocol (Sigma-Aldrich, Poole, Dorset, UK) [
18].
Statistical analysis
Data are expressed as the mean ± standard deviation (SD). Statistical analysis was performed using the Mann-Whitney U test for independent samples and the Wilcoxon signed rank test for related samples. P values less than 0.05 were considered significant.
Discussion
Synovitis and bony destruction are pathophysiological characteristics of RA, and marginal bony erosion, periarticular osteopenia, and joint space narrowing are the radiographic hallmarks of RA [
22,
23]. Synovial inflammation and bony destruction are closely related processes [
24], but contrary to synovitis, the bony changes are usually irreversible and accumulate with time, and can bring about joint dysfunction and an unfavorable disease outcome [
25,
26]. As a result, RA causes significant socioeconomic impact because of physically disabled and unemployed people [
27,
28].
Both cellular mechanisms and various inflammatory mediators are involved in the pathogenesis of bone erosion in RA, forming complex networks [
24,
29]. Among these process, OCs are the essential cells involved in the cellular mechanisms of the process of bony erosion [
30‐
32]. In RA synovium, OCs are found at the pannus-bone and pannus-subchondral bone junctions of arthritic joints, forming erosive pits in the bone [
30,
31]. Two additional cells play important roles in osteoclastogenesis: synovial fibroblasts and activated T cells. They express RANKL in the inflamed synovium, which promotes osteoclastogenesis, and also express cathepsin K at sites of synovial bone destruction [
33]. RANKL is the key molecule in OC differentiation and the augmentation of activity and survival of these cells, and is often called OC differentiation factor (ODF). In the serum transfer model of arthritis in the RANKL knockout mouse, the synovial inflammation and cartilage erosions are similar to those in wild-type mice, but the degree of bony erosion is significantly reduced [
34]. This result confirms the essential role of RANKL in the pathogenesis of bone erosion, regardless of inflammation or cartilage damage. The expression of RANKL is regulated by proinflammatory mediators such as TNF-α, IL-1, IL-6, IL-17, and PGE
2
[
35]. These inflammatory molecules are abundant in RA synovium, so the inflamed synovium supplies an optimal environment for RANKL activation.
In this study, we determined the relation between bony erosion and MIF in human RA. In the previous studies, MIF induces TNF-α, IL-1, IL-6, and PGE
2, which in turn promote RANKL expression [
1,
2,
9,
10,
36], and the synovial MIF concentration is higher in RA patients with bony erosion than in those without [
8]. Based on these results, we hypothesized that MIF might have a role in the pathogenesis of bone erosion, that is, it could have a direct effect on OC differentiation and an indirect effect on the induction of other inflammatory mediators that induce RANKL expression. First, we measured the synovial concentrations of MIF and RANKL in RA patients. Synovial fluid MIF concentration was higher in RA patients than in controls, as in our previous study [
8], but the synovial RANKL concentration did not differ between RA patients and controls. In previous studies, serum and synovial RANKL levels were higher in RA patients than in controls [
37], but the RANKL level was not related to any measures for disease activity [
38]. In contrast, we found that the serum and synovial MIF concentration was well correlated with RA disease activity [
8,
14]. Compared with previous studies, the patients enrolled in this study had longer disease duration and less active disease [
37], so MIF may reflect disease activity more closely than does RANKL. In this study, synovial RANKL concentration was significantly correlated with synovial MIF concentration, and this observation led us to investigate their close relation in the RA synovial tissues.
We investigated the effect of MIF on RANKL expression in RA synovial fibroblasts. Synovial fibroblasts, such as activated T cells, are major sources of the RANKL that promotes OC differentiation and bone erosion [
33]. Like other proinflammatory cytokines, MIF stimulates the expression of RANKL mRNA and protein in RA synovial fibroblasts, but there was no additive effect with other proinflammatory cytokines such as TNF-α and IL-1β. After blocking IL-1β, MIF-induced RANKL expression was partially decreased. This result suggests that RANKL expression was directly induced by MIF and also that it was indirectly stimulated by MIF-induced IL-1β. IL-1β has the potential to induce OC differentiation and RANKL expression, and overexpressed MIF could induce some inflammatory mediators, such as IL-1β in RA synovium, resulting in upregulation of RANKL and promotion of OC differentiation. Therefore, the MIF-IL-1β-RANKL interaction could be a major axis involved in RA bone erosion.
We investigated the effect of MIF on OC differentiation. We substituted MIF for RANKL in the traditional culture system for OC differentiation. After isolated PBMC were cultured with rhMIF and M-CSF, the numbers of TRAP-positive multinucleated cells were counted. OC developed in this new system without RANKL, but the degree of OC differentiation by MIF was less than that of RANKL. This result showed that MIF is one of the inflammatory cytokines involved in osteoclastogenesis, even if RANKL is the major molecule that induces OC differentiation. We also demonstrated that MIF-prestimulated RA synovial fibroblasts have a potential effect on osteoclastogenesis when the cells are co-cultured with PBMC. This culture system is more practical in an in vitro system similar to human RA synovium. RA synovial fibroblasts are exposed to a variety of cytokines that promote inflammation, and when these ailing cells encounter OC precursors, they could induce osteoclastogenesis by cytokine production or direct interaction between cells. This study was focused on the indirect osteoclastogenic effect mediated by RA synovial fibroblasts and RANKL, but MIF could directly enhance osteoclastogenesis from monocytes in the absence of additional RANKL. These two pathways imply more distinct and reinforced mechanisms for MIF-induced osteoclastogenesis, and a tipping point such as MIF production could be a potential therapeutic target.
In contrast to our results, a recent study suggests that MIF inhibits osteoclastogenesis [
13]. Although MIF enhances the expression of RANKL mRNA in murine osteoblasts and the expression of RANKL mRNA is enhanced in MIF transgenic mice, MIF inhibits OC formation in bone marrow cultures by decreasing fusion and decreasing the number of nuclei. The number of TRAP-positive OC is greater in MIF-deficient mice than in wild type mice, and the addition of MIF to the cells decreased TRAP-positive OC formation. Therefore, it appears that MIF plays an inhibitory role in bone resorption. The discrepancy between two studies could be explained by several differences in study systems. First, our study used human PBMC, whereas the former study used osteoclast precursor cells from MIF knockout mice. MIF inhibits osteoclast formation
in vitro in wild type mice bone marrow cell cultures and in the RAW264.7 macrophage cell line. Based on these data, MIF appears to directly inhibit osteoclastogenesis
in vitro but its effects on osteoclasts
in vivo are complex and may result from decreased RANKL expression in the osteoclast precursor cells from MIF knockout mice that were exposed to low levels of RANKL
in vivo and as a result these cells have increased sensitivity to RANKL
in vitro when cultured at high density.The MIF knockout mice that they used, had a marked resistance to lipopolysaccharide-induced endotoxic shock, and decreased TNFα production in response to lipopolysaccharide treatment. TNF-a also acts directly on the osteoclast precursor to potentiate RANKL-induced osteoclastogenesis, even in the absence of elevated levels of RANKL. MIF knockout mice were used in the previous paper, and had inhibited TNF production. Thus, osteoclast formation may have been inhibited. Second, we put the focus on an actual inflammatory disease of humans. In human RA synovial fibroblasts, the over-expressed MIF induces other inflammatory mediators, and then the inflammatory mediators, such as RANKL and IL-1β, enhance and potentiate osteoclastogenesis. Third, the former study treated RANKL with MIF in the OC differentiation system, but we did not treat RANKL in the culture system. More intensive study will be needed for explaining these conflicting results. We hypothesize that MIF might play an essential role in normal bone remodeling; however, over-expressed MIF might have an osteoclastogenic effect on bone metabolism in inflammatory diseases.
We found that MIF-induced RANKL expression in RA synovial fibroblasts was decreased by inhibition of NF-κB, PI3K, STAT3, AP-1, and p38 MAPK, but not ERK and calcineurin. Of the three MAP kinase pathways, only p38 MAPK was involved in MIF-induced RANKL production. In addition, MIF-induced osteoclastogenesis was suppressed by inhibition of NF-κB, PI3K, AP-1, and p38 MAPK, but not by inhibition of JAK/STAT3. These results suggest that there are different signal pathways involved in MIF-induced osteoclastogenesis. Considering that AP-1 is a downstream molecule, MIF seems to induce RANKL production by synovial fibroblasts mainly via NF-κB, PI3K, STAT3, and p38 MAPK, while it promotes OC differentiation from monocyte precursors via NF-κB, PI3K, and p38 MAPK. In recent years, numerous studies have attempted to define the signal transduction pathways of inflammatory cells activated by MIF in RA synovial fluid. MIF promotes cyclooxygenase-2, PGE
2, and IL-6 expression via p38 MAPK [
39]. MIF also upregulates IL-8 and IL-1β via tyrosine kinase-, protein kinase C (PKC)-, AP-1-, and NF-κB-dependent pathways [
40]. MIF controls the proliferation of RA synovial fibroblasts, mediated by ERK [
36]. The upregulation of MMP-2 by MIF is dependent on PKC, JNK, and Src signal pathways [
4]. MIF also upregulates other MMPs including MMP-1 and MMP-3 via tyrosine kinase-, PKC-, and AP-1-dependent pathways [
6]. Through the various intracellular signal transduction pathways, MIF activates RA synovial fibroblasts to promote inflammation, cartilage degradation, and bony destruction. In our previous study, we found the induction of MIF is mediated by p38 MAPK pathway when RA synovial fibroblasts are stimulated by conA, IFN-γ, CD40 ligand, IL-15, TGF-β, as well as IL-1β and TNF-α [
41]. Among these data, intracellular signal pathways are deeply involved in the pathogenesis of RA. Clinical studies for RA treatment using the inhibitors of different signal pathways such as Syk, p38 MAP, and JAK have been performed until now, and successful results are expected [
42‐
44]. Beyond the inhibition of cytokines or immune cells, oral inhibitors of intracellular molecules will be another choice for refractory RA.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
HR Kim and KW Kim designed and performed all experiments and drafted the manuscript. HG Jung, HJ Oh and KS Yoon assisted in designing the study. ML Cho and SH Lee conceived the study and drafted and edited the manuscript. All authors read and approved the final manuscript.