We purchased IL-1β, JNK, p-JNK, ERK, p-ERK, c-Jun, and p-c-Jun primary antibodies from Santa Cruz Biotechnology, myostatin-specific anti-rabbit polyclonal antibody from Abcam and recombinant human myostatin from PeproTech. R&D Systems supplied the myostatin and IL-1β enzyme-linked immunosorbent assay (ELISA) kits. Calbiochem supplied the inhibitors for ALK4/5/7 (SB431542), ERK (U0126), JNK (SP600125), and AP-1 (Curcumin and Tanshinone IIA). All other chemicals were purchased from Sigma-Aldrich.

We cultured the human rheumatoid cell line (MH7A; Riken cell bank, Ibaraki, Japan) in RPMI-1640 medium that was supplemented with 10% fetal bovine serum (FBS) and penicillin/streptomycin 100 U/mL. Cell incubation was performed in a humidified atmosphere of 37°C, 5% CO₂. We changed the culture medium on alternate days and passaged the cells when they had grown to 80% confluence.

The inhibitors for ALK4/5/7 (SB431542), ERK (U0126), JNK (SP600125), and AP-1 (curcumin and tanshinone IIA) were used to study the signaling transduction pathway. We cultured cells in 6-well plates with RPMI-1640 media containing 10% FBS. Cells were grown to 70% confluence, then incubated overnight in serum-free medium and treated for 30 min with or without the inhibitors for ALK4/5/7, ERK, JNK, and AP-1 (all at the concentration of 10 µM), followed by myostatin (10 ng/mL). Cells were harvested and examined by western blot, ELISA, and immunofluorescence assays.

MH7A cells were incubated with myostatin, harvested with centrifugation then resuspended for 10 min in hypotonic buffer A (21) on ice. Nuclei were isolated by centrifugation at 12,000g for 5 min. The nuclear pellet was suspended in buffer C (21) on ice for 30 min then vortexed for 15 s. The mixture was centrifuged at 12,000g for 10 min to collect supernatants containing nuclei proteins, then stored at −70°C.

Study approval was granted by China Medical University’s ethics committee and all subjects submitted written informed consent prior to entering the study. The study protocol was approved by China Medical University Hospital’s Institutional Review Board. Samples of synovial fluids were taken during total knee arthroplasty in 15 patients aged 40–60 years [10 patients with RA and 5 with osteoarthritis (OA)], then stored at −80°C until assay by myostatin and IL-1β ELISA kits.

We used a TRIzol kit (MDBio) to extract total RNA from MH7A cells and the Mir-XTM miRNA First Strand Synthesis Kit to quantify miRNAs; we followed the manufacturers’ protocols for both procedures. Each sample was evaluated for RNA quality and yield, using the A260/A280 ratio of 2:1, with the NanoVue spectrophotometer (GE Healthcare). Using the M-MLV Reverse Transcriptase kit (Invitrogen), we synthesized complementary DNA from 1 µg of total RNA and the KAPA SYBR^® FAST qPCR Kit (Applied Biosystems) for real-time quantitative polymerase chain reaction (qPCR) analysis under the following cycling conditions: 10 min at 95°C, then 40 cycles at 95°C for 15 s, and 60°C for 60 s. The relative expression of IL-1β and miRNA was normalized to endogenous GAPDH and snRNA U6, respectively, as internal controls. We used the threshold cycle (CT) to calculate the relative expression levels (22).

Downstream signaling pathways involved in myostatin stimulation were investigated. Human MH7A cells were cultured and pretreated with the pathway inhibitors (as indicated earlier), followed by myostatin treatment. We collected the supernatant medium as conditioned medium (CM) and stored it at −80°C until assay. The IL-1β ELISA kit assayed IL-1β in the CM, as per protocol.

Proteins were electrophoresed in sodium dodecyl sulfate-polyacrylamide resolving gel and transferred to Immobilon polyvinyldifluoride membranes, which were incubated in 4% bovine serum albumin (BSA) as blocking buffer for 1 h at room temperature, then probed overnight with primary antibodies (1:3,000) at 4°C. After three washes, the membranes were incubated (for 1 h at room temperature) with anti-rabbit peroxidase-conjugated secondary antibody (1:3,000). The membranes were developed via enhanced chemiluminescence then detected using a LAS-4000 image reader (FujiFILM).

The DNA fragment was constructed by RT-PCR with the following thermal cycling conditions: 3 min at 94°C once, then 35 cycles at 94°C for 1 min, 54°C for 1 min, and 72°C for 2 min. The primers 5′-AACCGCTTCCCTATTTATTT-3′ and 5′-GCTCATTTATAAATATTCCC-3′ were used to construct a DNA fragment of human IL-1β 3′-UTR containing the miR-21-5p binding site. This was inserted into a pGLO-2 basic vector and the construct was confirmed by DNA sequencing (Applied Biosystems 3730xl DNA analyzer). The plasmid containing mutant-IL-1β 3′-UTR was purchased fromGENEWIZ.

The MH7A cells were transfected with plasmid (1 µg), using Lipofectamine 2000 (Invitrogen), as per protocol. After applying myostatin for 24 h, we added reporter lysis buffer (Promega) to each well and scraped the cells, which were centrifuged at 13,000 rpm for 2 min to collect the supernatant. Aliquots of cell lysates were plated into a 96-well opaque, black plate; luciferase substrate was added to all aliquots and luminescence was recorded using a microplate luminometer.

We cultured MH7A cells on 12-mm coverslips in 24-well plates in RPMI-1640 supplemented with 10% FBS. At 70% confluence, cells were incubated as previously indicated then fixed with 3.7% paraformaldehyde at room temperature for 20 min and washed twice with PBS. Triton X-100 (0.1%) in PBS was added to the mixture for 5 min at room temperature, then the cells were washed three times with PBS and incubated with anti-c-Jun (1:100) at 4°C and left overnight. After three PBS washes, the cells were incubated for 1 h with fluorescein isothiocyanate conjugated secondary antibody. Nuclei were stained with DAPI and cells were mounted and detected by fluorescence microscopy (Zeiss).

Cells were cultured in RPMI-1640 with 10% FBS. At 70% confluence, cells were incubated overnight in serum-free medium then treated with myostatin. DNA was immunoprecipitated with c-Jun antibody and purified, extracted with phenol–chloroform, and subjected to RT-PCR. Agarose gel electrophoresis (1.5%) visualized PCR products under ultraviolet light. The IL-1β promoter region was amplified by the RT-PCR primers 5′-TCTGGTTCATCCATGAGATT-3′ and 5′-GGATAAAATGGGTACAATGAA-3′.

Animal procedures were conducted according to approved protocols issued by the Institutional Animal Care and Use Committee at China Medical University (Taichung, Taiwan) (IACUC Approval No. 104-154-N). The study protocol for clinical sample collection was approved by the Institutional Review Board of China Medical University Hospital (CMUH104-REC2-055). All patients completed written informed consent prior to study entry.

C57BL/6J mice, age: 8–10 weeks, obtained from the National Laboratory Animal Centre in Taipei, were maintained according to animal welfare conditions approved by China Medical University. Six mice formed the control group; the remaining mice were allocated to the CIA group. CIA was induced according to published protocols (23). Within 6 weeks after the primary immunization, 95% of the CIA group had developed severe arthritis. At 6 weeks, clinical severity was assessed in each knee by plethysmometer measurements.

8-µm sections were prepared from paraffin-embedded tissues, deparaffinized in xylene, rehydrated in a graded alcohol series, and washed in de-ionized water. The antigens were retrieved in boiling 10 mM sodium citrate, pH 6.0 for 12 min; intrinsic peroxidase activity was blocked by incubation with 3% hydrogen peroxide. 3% BSA in PBS blocked nonspecific antibody-binding sites. Sections were incubated overnight at 4°C in diluted primary antibodies specific for myostatin or IL-1β. After PBST washing, the secondary antibody (HRP-labeled anti-rabbit IgG) was applied for 30 min at room temperature. Stained sections were examined with 3,3′-diaminobenzidine tetrahydrochloride, counterstained with hematoxylin and eosin and observed under a light microscope. Safranin O-fast Green staining of the same specimens determined bone erosion.

SigmaPlot software was used to quantify results, which are expressed as the mean ± SEM of at least three experiments. We used the Student’s t-test to compare the means between groups. For comparisons of more than two groups, we used two-factor analysis of variance with the Bonferroni post hoc test. Results were considered statistically significant when p < 0.05.

Activated RASFs produce pro-inflammatory cytokines that infiltrate into synovial fluid and thereby initiate and perpetuate RA (24). Myostatin positively regulates IL-6, TNF-α, and IL-17 (18). We, therefore, analyzed myostatin and IL-1β expression profiles in human synovial fluids, using samples obtained during total knee arthroplasty from RA and OA patients. Levels of myostatin and IL-1β were significantly higher in RA synovial samples compared with OA samples (Figures 1A,B). Moreover, the expression patterns of myostatin and IL-1β were highly and positively correlated (Figure 1C). These observations suggest that myostatin and IL-1β might play a role in RA pathogenesis.

To further characterize the role of myostatin inducing RASF activation via IL-1β expression, we examined IL-1β production after myostatin treatment in human MH7A cells. qPCR and ELISA assays revealed dose-dependent increases with myostatin in IL-1β mRNA expression and protein secretion (Figures 2A,B). However, treatment with myostatin for 6 or 12 h only slightly increased IL-1β mRNA expression (Figure S1 in Supplementary Material). The evidence demonstrates that myostatin binds to the activin type IIB receptor and then interacts with either ALK4 or ALK5 to stimulate downstream signaling (25, 26). When MH7A cells were pretreated with SB431542 (a potent ALK receptor inhibitor), myostatin-induced IL-1β expression was significantly reduced (Figures 2C,D). Thus, IL-1β expression is increased by myostatin via the ALK receptor.

The ERK, p38, and JNK signaling pathways have been implicated as participating in myostatin signal transduction (26–28). We sought to determine the relative contributions of p38, JNK, and ERK activation upon myostatin-mediated regulation of IL-1β expression. Within the first 2 h of myostatin treatment, we found no evidence of p38, JNK, and ERK activation (data not shown). In contrast, an analysis of the expression of JNK and ERK proteins during prolonged myostatin treatment revealed that myostatin-promoted JNK and ERK phosphorylation in a dose- and time-dependent manner (Figures 3A,B), whereas no such effect was seen with p38 signaling (Figure S2 in Supplementary Material). We next observed the role of JNK and ERK in the mediation of myostatin-induced IL-1β expression, by using SP600125 (a broad-spectrum JNK inhibitor) and U0126 (a highly selective ERK inhibitor). Both agents abolished myostatin-induced increases in IL-1β expression (Figures 3C,D). Moreover, we confirmed a relationship between the ALK receptor and JNK/ERK signaling pathways and the myostatin pathway. Incubating the cells with the ALK receptor inhibitor reduced myostatin-induced increases in JNK and ERK phosphorylation (Figure 3E). These data demonstrate that myostatin induces IL-1β production in RASFs via the ALK receptor and JNK/ERK signaling pathways.

The transcription factor AP-1 is a well-known downstream target of JNK and ERK (29–31). We, therefore, sought to determine whether myostatin disrupts AP-1 signaling. We found that incubating the cells with myostatin dose-dependently promoted c-Jun phosphorylation (Figure 4A). Moreover, when we separated nuclear extract after myostatin treatment, we observed that myostatin induced c-Jun translocation into the nucleus (Figure 4B). We then found that the AP-1 inhibitors curcumin and tanshinone IIA reduced myostatin-promoted IL-1β expression (Figures 4C,D). We also confirmed that the ALK receptor inhibitor, JNK inhibitor, and ERK inhibitor all reduced myostatin-promoted phosphorylation of c-Jun and c-Jun accumulation in the nuclei (Figures 4E–H). Results of a ChIP assay confirmed that c-Jun binds to the AP-1 binding site in the IL-1β promoter after myostatin stimulation, and this was attenuated by inhibition of the ALK receptor and JNK/ERK signaling (Figure 4I). These data suggest that myostatin not only acts through the ALK receptor and JNK, ERK, AP-1 signaling pathways but also promotes IL-1β expression in RASFs.

miRNAs are involved in classical epigenetic mechanisms, such as downregulation of pro-inflammatory expression at the posttranscriptional level (10, 11, 32, 33). Most importantly, miRNAs have been identified as having therapeutic potential in arthritic diseases (34). To understand the mechanism underlying miRNA-mediated myostatin promotion of IL-1β expression, we used three target prediction programs (TargetScan, miRDB, and miRanda) and identified potential miRNA binding sites involving the IL-1β 3′ UTR sequence (Figure 5A). We found that myostatin dose-dependently downregulated miR-21-5p (Figures 5B,C). Notably, miR-21-5p levels were much lower in RA than in OA synovial fluid (Figure 5D). Next, we used the miR-21-5p mimic to investigate miR-21-5p involvement in myostatin-enhanced IL-1β expression. Transfection of cells with the miR-21-5p mimic weakened myostatin-induced IL-1β mRNA expression and protein secretion (Figures 5E,F). Pretreatment of cells with SB431542 significantly reversed myostatin-induced inhibition of miR-21-5p expression (Figure 5G). To examine whether miR-21-5p mediates the IL-1β 3′ UTR sequence, we constructed a luciferase reporter vector harboring this sequence that contained the miR-21-5p binding site (Figure 5H). Luciferase activity was increased by myostatin in the wild-type (wt)-IL-1β 3′ UTR plasmid; no such change was seen with the mutant (mut)-IL-1β 3′ UTR plasmid, while the ALK receptor inhibitor reduced myostatin-mediated wt-IL-1β 3′ UTR activity (Figures 5I,J). These data suggest that miR-21-5p directly targets IL-1β and thus downregulates IL-1β expression.

Paw swelling was significantly pronounced in CIA mice compared with controls (Figure 6A). IHC staining of murine tissue revealed substantially higher levels of myostatin and IL-1β expression from CIA mice compared with controls (Figures 6B–D). IHC data also revealed a high, positive correlation between myostatin and IL-1β (Figure 6E). These findings indicate that myostatin and IL-1β expression influence the progression of RA disease.