Thirty four serum samples of patients with arthritis were recruited at the Second Affiliated Hospital of Guangzhou University of Traditional Chinese Medicine (Guangzhou, China) from January 2018 to June 2018. Thirty serum samples of healthy donors were also collected as healthy controls (HCs). The patients and controls were well-matched by age and gender. The study was reviewed and approved by the Medical Ethics Committee of Southern Medical University. Before the collection of the blood sample, informed consent for taking part in the study was obtained from each participant.

WT C57BL/6J mice were obtained from the Laboratory Animal Center of Southern Medical University (Guangzhou, China). MBL^−/− mice were purchased from the Jackson Laboratory (Bar Harbor, ME, USA). The mice were housed under a specific pathogen-free condition, on a 12-h light-dark cycle, and with food and water ad libitum. All animal experiments in this study were approved by the Welfare and Ethical Committee for Experimental Animal Care of Southern Medical University (Approval number: L2016014).

MBL protein was prepared as previously described (16). Recombinant human RANKL (310-01C) and M-CSF (300-25) were purchased from Peprotech (London, UK). Recombinant murine RANKL (315-11) and M-CSF (315-02) were purchased from Peprotech (London, UK). Anti-c-fos antibody (26192-1-AP), anti-GAPDH antibody (10494-1-AP), and anti-CTSK antibody (11239-1-AP) were purchased from proteintech (Chicago, IL, USA). The anti-NFATc1 antibody was purchased from Abclonal Technology (Wuhan, China). Polyclonal antibodies against p38 (8690), phospho-p38 (4511), and Histone H3 (4499) were obtained from Cell Signaling Technology (Cambridge, MA, USA). The TRAP staining kit (387A) was from Sigma-Aldrich (St. Louis, MO, USA). Safranin O-solid green cartilage staining solution (G1317), Hematoxylin-Eosin solution (G4520), Masson's trichrome staining kit (G1340), and Toluidine blue staining solution (G2543) were purchased from Solarbio (Beijing, China).

Adjuvant arthritis was induced on day 0 of the experiment by subcutaneous injection of 0.1 ml of complete Freund's adjuvant (CFA) (4 mg/ml heat-killed Mycobacterium tuberculosis, Chondrex, Redmond, WA, USA). Mice were injected with 20 μl of incomplete Freund's adjuvant (IFA) into the knee joint space under general anesthesia on day 14.

Human monocytes were purified from EDTA-blood of healthy donors using CD14 Microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany) following the manufacturers' instructions and cultured in α-minimum Eagle's medium (α-MEM) supplemented with 10% heat-inactivated fetal bovine serum (FBS), 1% penicillin/ streptomycin, 50 ng/ml of recombinant human M-CSF and 100 ng/ml of RANKL for indicated days.

To identify osteoclasts, the differentiated cells were fixed in 4% paraformaldehyde for 20 min and then stained with the TRAP staining kit (Sigma-Aldrich) according to the manufacturer's instructions. Dark-red cells containing three or more nuclei were considered TRAP⁺ multinucleated cells. The total number of TRAP-positive osteoclasts was calculated using Osteomeasure software (OsteoMetrics, Inc., Decatur, GA, USA).

The resorptive function of mature osteoclasts was analyzed on bovine bone slices (Immunodiagnostic Systems, London, England). Briefly, cells were cultured in the differentiated medium in the presence or absence of MBL protein for 8 days on bone slices. Then the slices were washed twice with PBS, and the resorption pits were stained with toluidine blue (Sigma-Aldrich) for 5 min. The resorption area was analyzed using the Olympus image system.

Cells were grown in confocal dishes, fixed in 4% formaldehyde for 15 min at room temperature and permeabilized with 0.25% Triton X-100 for 10 min at room temperature. After blocking with 5% FBS for 1 h, cells were incubated with primary antibodies overnight at 4°C, rinsed, and incubated with fluorescent-conjugated secondary antibodies for 1 h in the dark. Finally, cells were counterstained with 4′,6-diamidino-2-phenylindole (DAPI, Beyotime).

Total RNA was extracted using Trizol (TransGene Biotech, Beijing, China) and then transcribed into cDNA using TranScript All-in-One First-Strand cDNA Synthesis SuperMix (TransGene Biotech), as instructed by the manufacturer. Real-time PCR was performed with an Eppendorf Realplex PCR system using TransStart Tip Green qPCR SuperMix (TransGene Biotech). The expression was normalized to the expression of the housekeeping gene GAPDH. The primer sequences used in the experiment are shown in Table 1.

Mice blood samples were obtained from the carotid artery and centrifuged at 3,500 × g for 15 min, and then the supernatant was collected and set aside at −80°C for serum cytokine analysis. Cytokine levels in the sera were assessed using commercial ELISA kits purchased from eBioscience (San Diego, CA, USA).

Protein lysate was prepared in RIPA buffer (Beyotime, Hangzhou, China) supplemented with protease and phosphatase inhibitor cocktails (Beyotime). Nuclear and cytoplasmic proteins were extracted according to the manufacturer's protocol (Beyotime). Protein samples were fractionated by SDS-PAGE and electrophoretically transferred onto polyvinylidene fluoride (PVDF) membranes (Millipore, Billerica, MA, USA). After blocking with bovine serum albumin (BSA, 5%) for 1 h at room temperature, the membranes were incubated overnight at 4°C with primary antibodies. Subsequently, the membranes were incubated with the horseradish peroxidase-conjugated corresponding secondary antibody for 1 h at room temperature. Finally, detection of the target protein was conducted with enhanced chemiluminescence (Thermo Fisher, Carlsbad, CA, USA) according to the manufacturer's protocol.

Hind limb samples were collected from arthritis mice and then fixed in 4% paraformaldehyde, embedded in paraffin, and stained with hematoxylin and eosin (H&E), TRAP, Masson's trichrome, safranin O-fast green, and toluidine blue according to the manufacturer's instruction.

For immunohistochemical (IHC) staining, antigen retrieval was performed in a citrate buffer (pH 6.0) at 120°C for 10 min and endogenous peroxidase activity was blocked by exposure to 3% H₂O₂ for 15 min. Sections were then incubated with primary antibodies at 4°C overnight. Immuno-reactivity was detected using the corresponding HRP-conjugated secondary antibody and visualized using diaminobenzidine kit (Beyotime Biotechnology, Shanghai, China).

Knee joints were scanned using a skyscan micro-CT scanner (Bruker, Karlsruhe, Germany) with X-ray beam settings of 50 kV and 50 μA. Trabecular bone regions of interest were selected by highlighting trabecular bone regions for cross-sectional slices of the hind limb and bone architecture determined by quantifying trabecular bone parameters using CTAN software.

All values were expressed as mean ± SD. One-way ANOVA was used for comparisons among multiple groups. Differences between two groups in the experiments were analyzed by Student's t-test. Value of p < 0.05 was considered statistically significant.

To explore the pathogenic role of MBL in inflammatory arthritis disease, we developed AIA in MBL^−/− and paired WT mice. Histological analysis of knee joints after adjuvant injection demonstrated that MBL^−/− mice showed severe joint damage compared with WT counterparts (Figure 1A). Besides, MBL^−/− mice exhibited more osteoporosis and collagen deposition than WT mice during the development of arthritis (Figures 1B,C). Moreover, there was significantly more destruction of the bone and cartilage in the adjuvant-treated MBL^−/− mice than those in WT controls (Figures 1D,E). We also examined and compared the expression of inflammatory cytokines between adjuvant-treated WT and MBL^−/− mice. The result showed that aggravated joint damage in MBL^−/− mice was accompanied by a significant elevation of pro-inflammatory cytokines (i.e., IL-1β, TNF-α, and IL-6) and reduction of anti-inflammatory cytokine IL-10 in hind paw (Figure 1F).

Using the established murine model of AIA, the impact of MBL on bone architecture was evaluated by micro-CT analysis. Qualitative analysis of the three-dimensional reconstruction of the knee joints (Figure 2A) confirmed that MBL had been able to prevent the external focal erosion on the periarticular surfaces. Similarly, longitudinal mid-sections (Figure 2B) and transaxial (Figure 2C) images of knee joint showed a reduction in trabecular bone mass in MBL^−/− mice upon adjuvant injection. Compared with WT AIA mice, the MBL^−/− AIA mice exhibited decreased bone volume/tissue volume (BV/TV) (Figure 2D), reduced trabecular number (Tb.N) (Figure 2E) and declined trabecular thickness (Tb.Th) (Figure 2F). Concomitantly, the trabecular spacing (Tb.Sp) of MBL^−/− AIA mice was significantly higher than that of WT counterparts (Figure 2G). Together, these results indicate a potential protective role of MBL in the pathogenesis of experimental AIA.

Accumulating evidence pointed out that the increased osteoclastic activity is responsible for bone loss or joint destruction during the development of inflammatory arthritis (21). To determine the role of MBL in osteoclastogenesis, TRAP staining was performed on the bone sections isolated from adjuvant-treated WT and MBL^−/− mice. The result demonstrated that a massive increase in numbers of osteoclasts in the knee joints from MBL^−/− mice compared with those from WT mice (Figure 3A). CTSK is a novel cysteine protease previously reported to be predominantly expressed by osteoclasts (22). Immunolocalization of CTSK revealed a higher amount of CSTK-positive osteoclasts in the joint from adjuvant-treated MBL^−/− mice than that from WT controls (Figure 3B). Osteoprotegerin (OPG), produced by osteoblasts, is an essential regulator in osteoclast formation via inhibiting both differentiation and function of osteoclasts (23). We observed that OPG expression was significantly lower in MBL^−/− AIA mice than WT controls by the immunohistochemical staining (Figure 3C). Besides, bone morphogenetic proteins (BMPs, e.g., BMP2 and BMP4), potent mediators for osteoblast differentiation (24), were strongly over-expressed in the tissues from MBL^−/− arthritis mice compared to those from WT counterparts (Figures 3D,E). Moreover, the levels of serum PINP, a biochemical marker of bone formation, was significantly higher in MBL^−/− mice with arthritis than WT counterparts (Figure 3F). The adjuvant-treated MBL^−/− mice also displayed a markedly increase in the serum level of β-CTX, a marker of bone resorption, compared to WT controls (Figure 3F). Collectively, these data suggest that MBL may modulate the osteoclast formation and activity in the pathogenesis of inflammatory arthritis.

We compared the osteoclast formation in bone marrow cells from WT and MBL^−/− mice. Upon cultivation with M-CSF and RANKL, bone marrow cells from MBL^−/− mice more efficiently differentiated into mature TRAP-positive multinucleated osteoclasts than those from WT littermates (Supplementary Figure 1A). Consistently, osteoclast cultures of MBL^−/− mice displayed an elevated resorption activity (Supplementary Figure 1B). The data indicate that MBL generated by osteoclast precursors and/or mature osteoclasts initiates an autocrine negative feedback loop to regulate osteoclastogenesis.

Next, we investigated whether exogenous MBL protein affected human osteoclast differentiation in vitro. Human purified monocytes were treated with 100 ng/ml of RANKL and 50 ng/ml of M-CSF, which is known to induce osteoclast formation (4, 25), in the presence or absence of varying concentrations of MBL protein. As shown in Figure 4A, MBL treatment significantly reduced the number of TRAP-positive multinucleated osteoclasts. Besides, MBL protein inhibited the bone resorption ability of osteoclasts in a dose-dependent manner (Figure 4B). Consistent with the limitation of osteoclastogenesis, the expression levels of several established osteoclast marker genes, including dendritic cell-specific transmembrane protein (DC-STAMP), CTSK and calcitonin receptor (CTR), were sharply reduced upon MBL administration (Figure 4C). Besides, MBL also showed an inhibitory effect on the induction of c-fos and NFATc1, two of the essential osteoclasts-specific transcription factors, during RANKL-stimulated osteoclast formation (Figure 4D). Western blot analysis further confirmed the reduced protein level of osteoclast marker and transcription factors in MBL-treatment group compared to the control group (Figure 4E). Notably, the nuclear localizations of NFATc1 and c-fos are essential for the osteoclast differentiation (3, 25). We performed immunofluorescence staining assay to confirm whether MBL inhibited the expressions and nuclear localizations of NFATc1 and c-fos in the osteoclasts. The result showed that MBL treatment significantly decreased the expressions and nuclear accumulations of NFATc1 and c-fos in M-CSF and RANKL-stimulated human monocytes (Figure 4F). MMPs are a large group of enzymes responsible for matrix degradation (26). Among the MMPs, MMP-9, uniquely expressed by osteoclasts (27), play a crucial role in joint destruction of inflammatory arthritis (26, 28). Additional study demonstrated that MBL inhibited the expression of MMP-9 in differentiated human osteoclasts (Figure 4G). Taken together, these data indicate that MBL impairs osteoclast differentiation in vitro.

All the three MAPK family members (i.e., extracellular signal-regulated kinase (ERK), c-Jun N-terminal kinase (JNK), and p38) are rapidly phosphorylated and activated following RANKL stimulation of osteoclast precursor cells (29). Among them, the role of p38 in osteoclast differentiation and function was extensively investigated (25, 30). Inactivation of the p38 signaling pathway completely blocked the induction of c-fos and NFATc1 with concomitant inhibition of RANKL-induced osteoclastogenesis (25). We, therefore, investigate the activation of the p38 signaling pathway in RANKL-induced osteoclast formation. Enhanced p38 phosphorylation was observed in MBL ^−/− mice-derived osteoclasts compared with WT controls (Supplementary Figure 2). Also, MBL protein efficiently inhibited RANKL-induced phosphorylation of p38 in human osteoclast precursors (Figure 5A). Intriguingly, the subsequent nuclear translocations of NFATc1 and c-fos were substantially limited in the presence of MBL (Figure 5B). To further evaluate the effect of p38 activation on MBL-mediated regulation of osteoclast differentiation, osteoclast precursors were pre-treated with p38 inhibitor, SB203580, before RANKL stimulation. The levels of osteoclast markers and osteoclasts-specific transcription factors were strongly reduced and comparable between RANKL-induced osteoclast differentiations with or without MBL incubation upon p38 blockade (Figures 5C–E). These results suggest that the regulation by MBL of osteoclast differentiation is dependent on the p38 signaling pathway.

Rheumatoid arthritis (RA) is characterized by joint inflammation and progressive joint damage, and bone destruction in RA is mainly attributable to the abnormal activation of osteoclasts (1). We, therefore, explored the role of MBL in the progression of RA as well as the process of osteoclastogenesis. In line with previous reports (12, 31), we found that the sera from RA patients showed the MBL serum levels in these patients were significantly lower than those in HCs (Figure 6A). We also analyzed the association of MBL serum level with the disease activity. The results showed that the serum level of MBL was negatively correlated with rheumatoid factor (RF) (Figure 6B) and erythrocyte sedimentation rate (ESR) (Figure 6C). Furthermore, MBL levels were found to be negatively correlated with serum levels of PINP and β-CTX in patients with RA (Figures 6D,E). These data indicate that circulating levels of MBL was significantly associated with the disease activity and osteoclastogenesis in patients with inflammatory arthritis.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.