Alleviation of Synovial Inflammation of Juanbi-Tang on Collagen-Induced Arthritis and TNF-Tg Mice Model

Rheumatoid arthritis (RA) is a chronic autoimmune disease that is primarily characterized by synovial inflammation. In this study, we found that a traditional Chinese decoction, Juanbi-Tang (JBT), JBT attenuated the symptoms of collagen-induced arthritis (CIA) mice and in tumor necrosis factor transgenic (TNF-Tg) mice by attenuating the arthritis index and hind paw thickness. According to histopathological staining of ankle sections, JBT significantly decreased the area of inflammation and reduced bone destruction of ankle joints in both these two types of mice. Moreover, decreased tartaric acid phosphatase-positive osteoclasts were observed in the JBT group compared with those found in the control group. We also revealed that JBT suppressed monocytes and T cells as well as the production of CCL2, CCR6, and CXCR3 ligands. We next used high-performance liquid chromatography to investigate the components and pharmacological properties of this classical herbal medicine in traditional Chinese medicine. Based on network pharmacology, we performed computational prediction simulation of the potential targets of JBT, which indicated the NF-kappa B pathway as its target, which was confirmed in vitro. JBT suppressed the production of pro-inflammatory cytokines including interleukin-6 (IL-6) and IL-8, and inhibited the expression of matrix metalloproteinase 1 in fibroblast-like synoviocytes derived from RA patients (MH7A cells). Furthermore, JBT also suppressed the phosphorylation of p38, JNK, and p65 in TNF-α-treated MH7A cells. In summary, this study proved that JBT could inhibit synovial inflammation and bone destruction, possibly by blocking the phosphorylation of NF-kappa B pathway-mediated production of proinﬂammatory effectors.


BACKGROUND
Rheumatoid arthritis (RA), one of the major contributors to disability globally (Vos et al., 2017), is characterized by synovitis, pannus formation, adjacent bone erosion, and joint destruction (Rana et al., 2018), although its etiology remains largely unclear. Among the potential cellular participants in RA, fibroblast-like synoviocytes (FLSs) play a critical role by regulating the secretion of inflammatory mediators and expression of matrix metalloproteinases (MMPs), which cause changes in chondrocyte metabolism and matrix degradation (Korb-Pap et al., 2016). Treatment targeting RA-FLSs has been recognized as a novel approach with the potential to improve clinical outcomes while having limited impact on systemic immunity (Wangyang et al., 2018;Zhang and Zhou, 2018). Tumor necrosis factor-alpha (TNF-a) acts as a primary cytokine in the pathology of RA. TNF-a can rapidly induce MMP gene expression in cultured FLSs. Its inhibitors provide other options for RA patients in whom treatment with conventional diseasemodifying anti-rheumatic drugs (DMARDs) has failed (Luchetti et al., 2017;Ornbjerg, 2018). Given its immunogenicity and high cost, an increasing number of alternatives to DMARDs are being explored .
Juanbi-Tang (JBT), traditional Chinese herbal compounds, has been widely used for the treatment of RA in China. It has proven to relieve the symptoms of arthritis and activate joint function of patients (Li, 2011;Li and Niu, 2011;Ni and Fu, 2014). In addition, JBT has been reported to reduce serum TNF-alpha and IL-1 levels in rats with RA (Yu et al., 2015), downregulate IgG in serum  and prostaglandin E receptor 4 in synovial tissue , and relieve acute and chronic ear inflammation in rats (Niu et al., 2018a). It was also demonstrated that JBT could regulate immune balance by reducing nature killer cells and promoting T-reg cells of CIA rats (Niu et al., 2018b), and reduce the cAMP level in T lymphocytes in rats with adjuvant arthritis . However, the mechanism by which JBT reduces inflammatory cytokines and chemokines warrants further research.
With the increasing investment in research on traditional Chinese herbal medicine, it has become increasingly important to determine the efficacy, toxicology, and therapeutic targets of such medicine, for which network pharmacology is a useful approach (Jiang et al., 2015).
Our study involved a search for as many as 44 active c o m p o n e n t s o f J B T b y h i g h -p e r f o r m a n c e l i q u i d chromatography-quadrupole time-of-flight mass spectrometry (HPLC-Q-TOF). By applying a network pharmacology method, we also enumerated the possible targets of the herbal components to look for their pharmacological effects. We further performed experiments to verify these predictions.

Preparation of JBT
Herbs in JBT ( Table 1) were authenticated by a pharmacognosist of Longhua Hospital affiliated to Shanghai University of Traditional Chinese Medicine, in accordance with standard protocols. In line with standard methods of Chinese Pharmacopoeia, all crude drugs except Cinnamomi cortex were soaked in 12 volumes of water for 40 min and boiled for 40 min (C. cassia was added at 35 min). The extracts were filtered, and the filter residue was boiled in eight volumes of water for another 40 min and the solution was filtered again. Both batches of the filtrate of the drugs were mixed and concentrated to 0.557 kg/L, which were prepared for intragastric administration. The filtrate was vacuum-freeze-dehydrated, prepared to treat cells after highspeed centrifugation three times, and stored at −80°C overnight.

Model Establishment and Evaluation of Collagen-Induced Arthritis
Specific pathogen-free, DBA/1J male mice (7-8-week-old) were purchased from Vital River (Beijing, China). The CIA model was established in accordance with a previously reported protocol (Brand et al., 2007). Briefly, the mice were immunized intradermally with 100 mg of bovine type II collagen emulsified in complete Freund's adjuvant. To ensure a high incidence of RA induction in the CIA model, a booster immunization of bovine type II collagen emulsified in incomplete Freund's adjuvant was used at 21 d after the primary immunization. Typically, the first signs of arthritis appeared in this model at 21-28 d after immunization. CIA was considered to have successfully developed when swelling was observed in at least one paw.

TNF-Tg Mice
The 3647 line of TNF-Tg mice in a C57BL/6 background was originally obtained from Dr. G. Kollias's lab (Keffer et al., 1991) (Institute of Immunology, Alexander Fleming Biomedical Sciences Research Center, Vari, Greece) and was maintained as heterozygotes by backcrossing with C57BL/6 mice. Nontransgenic littermates were used as aged-matched wild-type (WT) controls. This line of TNF-Tg mice carries one copy of the human TNF transgene, overexpresses human TNF-alpha, and develops erosive polyarthritis with many characteristics also observed in RA patients (Jiang et al., 2015). TNF-Tg mice develop mild ankle joint inflammation and bone erosion at 3 months old, which become more severe with aging (Li et al., 2004;Guo et al., 2009;Chen et al., 2016). Therefore, in this study, we chose 3-month-old TNF-Tg mice and WT littermates.

Experimental Groups
Forty DBA/1J mice were randomly divided into four groups (10 mice/group): group 1, control (normal, nonimmunized and untreated); group 2, CIA-Veh [positive control, PBS, 0.2 ml/ day, intraperitoneally]; group 3, CIA-MTX (MTX, 0.1 mg/kg/3 d, intraperitoneally); and group 4, CIA-JBT (JBT, 12 g/kg/day, orally). All of the mice from these groups received additional treatments between day 22 and day 50. Seven 3-month-old TNF-Tg mice were orally administered JBT (12 g/kg) or the same volume of physiological saline once every day for 12 weeks. Seven 3-month-old WT littermates were treated with the same volume of physiological saline as a negative control. Twelve weeks later, all mice were sacrificed and their ankle joints were harvested for histological staining and data analysis.
All animal procedures were approved by Longhua Hospital Animal Ethics Committee and were performed in accordance with the Guiding Principles for the Care and Use of Laboratory Animals approved by the Animal Regulations of National Science and Technology Committee of China.

Arthritis Assessment
After the CIA model had been successfully established, we measured the arthritis index and hind paw thickness every 5 d in a blinded manner. Each paw was scored on a scale of 0-4 by visual evaluation as follows: 0, no evidence of erythema and swelling; 1, erythema and mild swelling confined to the tarsals or ankle joint; 2, erythema and mild swelling extending from the ankle to the tarsals; 3, erythema and moderate swelling extending from ankle to metatarsal joints; and 4, erythema and severe swelling encompassing the ankle, foot, and digits, or ankylosis of the limb. The final score for each mouse was the sum of the scale score from four paws. TNF-Tg mice were measured the deformity score every week in a blinded manner for 12 weeks. Each paw was scored on a scale of 0-4 by visual evaluation as follows: 0, no deformity; 1, mild deformity; 2, obvious deformity; 3, moderate deformity, 4, severe deformity. The final score for each mouse was the sum of the scale score from four paws. Thickness of the ankle was measured separately by two individuals with digital calipers placed across the ankle joint at its widest point.

Histopathological Examination
After final treatment, the mice were sacrificed. Ankle joints were carefully dissected and cleared of adjacent muscle, followed by being fixed in 4% paraformaldehyde at room temperature for 48 h, decalcified in 10% EDTA for 20 d, dehydrated through a series of ethanol, and embedded in paraffin. A total of 30 4-mmthick consecutive sections from one ankle joint in the sagittal position were collected and mounted on common slides, and were divided into three levels. Each level was 40 mm from the previous level. One section from each of the three levels was subjected to hematoxylin and eosin (H&E) staining, Alcian Blue/ Orange G (ABOG) staining, tartrate-resistant acid phosphatase ( T R A P ) s t a i n i n g , a n d i m m u n o fl u o r e s c e n c e a n d immunohistochemical staining. After the full-length scan by Olympus VS120-SL, the target area of the ankle talus was analyzed with Olympus VS120 image analysis software. We traced the area of interest, which was then measured by the software (in mm 2 ). The inflammatory area (multinuclear dark zone) and bone area (astragalus bone zone) were analyzed on HE-stained sections. Cartilage area (blue zone) at astragalus bone was measured on ABOG-stained sections, and TRAP-positive area was evaluated on TRAP-stained sections. The data are presented as the mean from three levels from each ankle joint sample.

Micro-Computed Tomography (Micro-CT) Analysis
After excision from DBA/1J mice, right legs were fixed in 4% paraformaldehyde for 24 h, washed in PBS for 2 h, and then immersed in 75% ethanol, until a micro-CT (Scanco VIVA CT80, SCANCO Medical AG, Bassersdorf, Zurich, Switzerland Switzerland) scan. The X-ray tube voltage was 55 kV and tube current was 72 mA, with a pixel size of 9 mm and 200 ms integration time. The cross-sectional images were then reconstructed and realigned in 3D. A density threshold was set from 370 to 1,000 as bone using the µCT Evaluation program V6.6 (Scanco Medical AG). A stack of 340-441 cross sections was reconstructed, with an interslice distance of 1 pixel (15.6 µm), corresponding to a reconstructed height of 5.3-6.9 mm, recreating the ankle joint.

Network Pharmacological Analysis of JBT
The 44 active chemical components deriving from 11 components of JBT were collected by HPLC-MS. The Chinese Medicine System Pharmacology Database and Analysis Platform and Swiss Target Prediction databases were used to screen the targets of these 44 compounds. The DisGeNET database was used to find the targets related to RA. Cytoscape 3.6.0 was used to analyze the targets of JBT in the treatment of RA, and the DAVID database was used to analyze the specific mechanism of action of JBT in the treatment of RA. Twenty-four of the pharmacological components (see Figure 3 for the chemical formulas) that satisfied the above conditions were selected. Then, we used Drug Bank, TCMID, Gene Cards, and STITCH databases to validate 194 RA targets to construct a compound target network using Cytoscape 3.3.0 software. To explain the participation of targets in the progression of RA, we established a network between target and function using Cytoscape 3.3.0. Multiple targets presaged an integral function of JBT in RA and shared the synergistic targets of the different compounds.

Cell Culture
MH7A, a human RA-FLS cell line (Miyazawa et al., 1998;Jia et al., 2019), was a gift from the Chinese Academy of Sciences, Shenzhen, which had been purchased from the Riken Cell Bank (Tsukuba, Japan). Cells were maintained in Roswell Park Memorial Institute 1640 medium (Hyclone, USA), supplemented with 10% fetal bovine serum (Gibco, USA) and penicillin/streptomycin (1:100; Sigma, St. Louis, MO), in a humidified atmosphere of 95% air and 5% CO 2 in an incubator at 37°C.

Cell Viability Assay
Cell viability was determined using the CCK-8 assay. Briefly, MH7A cells were seeded in 96-well culture plates at 5000 cells/ well for 72 h and then treated with various concentrations of JBT (0, 0.25, 0.5, 1, and 2 mg/ml) for 24 h. Then, CCK-8 (Dojindo, Japan) was added and incubated at 37°C for 1.5 h. The optical density was detected at 450 nm with a microplate reader (BioTek Synergy 3).

Quantitative Real-Time PCR Analysis
MH7A cells were pretreated with various concentrations of JBT lyophilized powder (0, 0.5, 1, and 1.5 mg/ml) for 2 h, and then incubated for another 24 h with or without stimulation by 10 ng/ m TNF-a. Total RNA was extracted using Trizol reagent, in accordance with the manufacturer's instructions, and each sample was reverse-transcribed using the cDNA synthesis kit, in accordance with the manufacturer's protocol. Real-time PCR analysis was performed using SYBR Green PCR Premix Ex Taq II reagents on a CFX96 real-time system (Thermo, USA). Relative gene expression was calculated by the −DD Ct method. The sequence of primers (forward and reverse) was as follows: actin, 5′-CGTTGACATCCGTAAAGACC-3′ and 5′- a n d 5 ′ -TTCCTTGGGGTCCAGACAGA-3′; and MMP-1, 5′-C T C A A T T T C A C T T C T G T T T T C T G -3 ′ a n d 5 ′ -CATCTCTGTCGGCAAATTCGT-3′. The efficiencies of the primers were 90%-95%. The PCR reaction system was 20 µl: SYBR 10 µl, water 7 µl, cDNA 1 µl, front and rear primers 1 µl. The thermocycling condition: 95°C for 5 min, 95°C for 30 sec!60°C for 30 sec!72°C for 40 sec, 30 cycles, 72°C for 7 min, and stored at 4°C.

Enzyme-Linked Immunosorbent Assay
MH7A cells were seeded in six-well plates (1*10 6 cells/well) for 24 h, and then pretreated or left untreated with the indicated concentrations of JBT for 2 h, followed by incubation for another 24 h with or without TNF-a (10 ng/ml). Cell culture supernatants were collected and stored at −80°C until analysis. The concentrations of the cytokines in culture supernatants were determined by ELISA using a commercial kit, in accordance with the manufacturer's instructions.

Western Blot Analysis
MH7A cells were treated with or without TNF-a (10 ng/ml) in the presence or absence of JBT (0.5, 1, and 1.5 mg/ml) for 24 or 0.5 h. The protein was collected, and 30 mg of it from each sample was separated by 10% SDS-PAGE and transferred to a polyvinylidene fluoride membrane. After blocking with 5% bovine serum albumin in TBST at room temperature for 1 h, the membranes were incubated with the corresponding primary antibodies overnight at 4°C. After washing with TBST three times, the membranes were incubated with the secondary antibodies. Proteins were scanned and analyzed by the chemiluminescence system and autoradiography.

Statistical Analysis
The data are presented as mean ± standard deviation (SD) from at least three separate experiments. Statistical analysis was performed using GraphPad Prism 7.0 software (GraphPad, La Jolla, CA, USA). One-way ANOVA followed by Dunnett's t-test was used to determine the significance of differences between groups. In cases of three groups treated at different times points, we applied two-way ANOVA for comparison. We used Fisher's test in the analysis of categorical data. P < 0.05 was considered statistically significant.

JBT Clearly Attenuated Symptoms, Synovitis, Cartilage, and Bone Damage of CIA Mice
To investigate the therapeutic effects of JBT on RA in vivo, we established a CIA mouse model. Compared with the WT group, CIA mice showed higher scores for the arthritis index and greater hind paw thickness. JBT treatment had improved the function of ankle by attenuating the arthritis index ( Figure 1A) and hind paw thickness ( Figure 1B). HE, ABOG, and TRAP staining at ankle joints was scanned using an Olympus VS120-SL ( Figure 1C). Treatment of mice with JBT was initiated at 22 d after the primary immunization. Both JBT and MTX treatments had positive effects on attenuating the inflammatory area ( Figure 1D) and restoring bone ( Figure 1E) and cartilage area ( Figure 1F) of CIA mice. JBT treatment significantly reduced the number of osteoclasts with TRAP-positive staining ( Figure 1G). Micro-CT showed that the JBT group, as well as the MTX group, suffered less bone damage in their ankle joints than the control group ( Figure 1H). These results indicated that JBT effectively inhibited the development of arthritis in CIA mice.
JBT Attenuated Symptoms, Inflammatory Synovial Volume and Bone Erosion at Ankle Joint of TNF-Tg Mice To investigate the effect of JBT on arthritis, TNF-Tg mice were orally administered JBT once a day for 12 weeks. Compared with the WT group, TNF-Tg mice showed higher deformity score and greater hind paw thickness. JBT treatment had improved the function of ankle by attenuating the deformity score ( Figure 2A) and hind paw thickness ( Figure 2B). HE, ABOG, and TRAP staining at the ankle joints indicated normal integrity of the ankle joint without synovial inflammation and bone erosion in WT mice ( Figure 2C). In TNF-Tg mice, there was a large amount of synovial inflammation in the ankle joint, with the presence of a large number of TRAP-positive osteoclasts, and severe bone erosion was observed at the ankle joint, while cartilage with Alcian Blue-positive staining almost disappeared. JBT treatment significantly reduced the inflammatory synovial volume, bone erosion ( Figures 2D, E), and osteoclasts positively stained for TRAP ( Figure 2F), while increasing cartilage area at the ankle joints ( Figure 2G). Micro-CT showed that the JBT group suffered less bone damage in the ankle joints ( Figure 2H). These results indicated that JBT effectively inhibited the development of arthritis in TNF-Tg mice.

JBT Suppressed Inflammatory Cell Infiltration and the Production of CCL2, CXCR3, and CCR6
To further investigate the effect of JBT on inflammatory cell infiltration, immunofluorescence staining was performed to mark macrophages with F4/80/iNOS and T cells with CD3. Immunohistochemical staining was performed to label CCL2, CXCR3 and CCR6. The results showed that JBT inhibited F4/80/ iNOS+ macrophages ( Figures 3A, B and Supplementary 5A, B) and CD3+ T cells ( Figures 3C, D) in the ankle joints of CIA and TNF-Tg mice, and reduced the expression of CCL2, CXCR3, and CCR6 ( Figures 3E, F and Supplementary 5C, D).

HPLC-Q-TOF Detected 44 Compounds in JBT Decoction
With the aid of accurate masses measured by HPLC-Q-TOF, a total of 44 detected peaks were identified in the JBT decoction, as shown in Table 2. Among them, there were 24 compounds (Figure 4 and Supplementary 4) definitively identified by comparison with the standard compounds, while the remaining 20 compounds were tentatively characterized a c c o r d i n g t o t h e i r f o r m u l a , r e t e n t i o n t i m e , a n d fragmentation patterns.

Network of 44 Potential Compounds Predicted to Have 194 Potential Targets
We obtained the SMILE number of 44 compounds through the PubChem database (https://pubchem.ncbi.nlm.nih.gov/); then, we used the Swisst Target Prediction database (http:// swisstargetprediction.ch/) to query the target sites of the 44 compounds. After removing the duplicates, a total of 194 targets were obtained. After querying the JBT target in the Swiss Target Prediction database, the protein interaction network related to JBT was constructed with the STRING database. Using DisGeNET (http://www.disgenet.org/) database, we found 1,848 genes related to RA. Network of compounds in JBT and their potential targets. JBT (yellow rectangle), JBT 11 herbs (H, red diamond), and network of 44 potential compounds (C, blue ellipse) predicted to have 194 potential targets (green triangle). It comprised 252 nodes and 702 edges, as shown in Figure 5A. We use the DAVID database to analyze the GO enrichment and KEGG pathways of JBT genes in RA. To make the research more targeted, we used the JBT targets applied as queries in DisGeNET to map the RA genes, and constructed the common genes through the STRING database to obtain the in vivo network of responses of RA to JBT. The JBT protein interaction network was visualized in Cytoscape 3.6.0 software ( Figure 5B). Each protein was represented by a node and the size of the nodes is proportional to the degree of the node. Relative pathways were analyzed according to the targets. As shown in Figure 5C, NF-kappa B pathways was one of the most relative pathways. More specific targets were marked with pentagram (Supplementary 1).

JBT Suppressed the mRNA and Protein
Expression of MMP-1, IL-6, and IL-8 in TNF-a-Induced MH7A Cells To assess the potential cytotoxicity of JBT, cell viability was evaluated by CCK-8 assay. As shown in Figure 6A, JBT at concentrations in the range of 0.25-2 mg/ml did not affect cell viability after incubation for 72 h. To determine the protective effect of JBT against the expression of inflammatory cytokines, MH7A cells were incubated for 24 h with TNF-a. TNF-a significantly increased IL-6, IL-8, and MMP-1 expression at the mRNA level ( Figures 6B-D) and the levels excreted in conditioned medium ( Figures 6E-G). JBT significantly decreased the mRNA and protein expression of MMP-1 in TNF-a-induced MH7A cells in a dosedependent manner, compared with that in the control group treated with TNF-a alone ( Figure 6H). However, JBT did not show a significant effect on the mRNA expression of MMP-3 (Supplementary 2). Furthermore, JBT inhibited osteoclastogenesis (Supplementary 3).

JBT Inhibited TNF-a-Induced Activation of NF-Kappa B Pathways
To further investigate the mechanism by which JBT inhibits the production of inflammatory cytokines, we examined its effect on the  Table 2 and Figure 4. Among them, 24 compounds were definitively identified by comparison with the standard compounds, while the remaining 20 compounds were tentatively characterized according to their formula, retention time, and fragmentation patterns. NF-kappa B-mediated pathways in TNF-a-induced MH7A cells. MH7A cells were pretreated with the indicated concentrations of JBT for 2 h, before stimulation with TNF-a (10 ng/ml) for another 30 min. Then, the protein levels and phosphorylation levels of p38, ERK, JNK, and p65 were determined. We found that the phosphorylation of p38, JNK, and p65 was markedly increased after TNF-a stimulation. JBT treatment significantly inhibited the phosphorylation of p38, JNK, p65 dose-dependently (0.5, 1, and 1.5 mg/ml, Figures 7A-C). We also used immunofluorescence analysis to determine the intracellular localization of p65. We pretreated MH7A cells with JBT (1.0 mg/ml) for 2 h or left them untreated, and then exposed them to TNF-a (10 ng/ml) for 0.5 h. The immunofluorescence assay revealed that JBT significantly reduced the level of p65 in the nucleus of MH7A cells ( Figures 7D, E).

DISCUSSION
In this study, we demonstrated that JBT could effectively inhibit joint symptoms, synovial inflammation, and joint destruction in CIA and TNF-Tg mice. We also revealed that JBT suppressed monocytes and T cells as well as the production of CCL2, CCR6, and CXCR3 ligands. In vitro data indicated that JBT could also profoundly inhibit the TNF-a-induced production of IL-6, IL-8, Primary studies have shown that JBT could be used for the treatment of chronic lower back pain, external humeral epicondylitis, synovitis of the hip in children, ankylosing spondylitis, gout, and other bone diseases (Li, 2011;Li and Niu, 2011;Ni and Fu, 2014). It could also reduce inflammatory cytokines such as TNF-a and IL-1 in serum of model rats induced by Freund's adjuvant (Yu et al., 2015). Against this background, we adopted two animal models so as to verify the effect of JBT during the early stage of RA. One is the most widely accepted, the CIA mouse model, and the other is the TNF-Tg mice model with TNF-a overexpression, both of which showed clear joint inflammation, similar to RA, before undergoing treatment. We found that JBT significantly reduced joint synovitis and the articular cartilage and bone defects in both animal models, thereby clarifying its efficacy in vivo. In other words, it was found that JBT can effectively reduce RA joint inflammation.
To clarify the pharmacological mechanism of JBT, we employed HPLC to detect its active components. We found 44 components in it, for the first time. To further analyze the mechanisms of action of these components, we then applied a network pharmacological method. Most previous studies on the mechanism of action of JBT tended to obtain the effective components based on retrieval of the associated literature from databases. For the first time, we here instead performed this based on the components detected by HPLC, which boosted the objectivity and credibility of our experimental results. Based on the network pharmacological analysis, we found that the targets of JBT components are mainly concentrated in the NF-kappa B pathway, as well as in cancer pathways, cancer proteoglycans, cancer microRNAs, viral carcinogenesis, and arachidonic acid metabolism.
Next, we conducted in vitro experiments for exploring the underlying mechanism. To simulate the in vivo environment of TNF-Tg mice and the high expression of TNF-a in the lesion area of RA, in the current study, we treated synovial cells with TNF-a (10 ng/ml) to observe the changes of inflammatory factors produced by them upon inflammatory stimulation. We found that JBT could also inhibit the TNF-a-induced production of IL-6, IL-8, and MMP-1 in synovial cells, restoring the structure of joints.
IL-6 is highly expressed in fibroblasts during the early stage of RA (Filer et al., 2017) and its inhibitor has started being used for clinical treatment (McInnes and Schett, 2017), as well as TNF-a. IL-8 also acts as a marker of synovitis (Alam et al., 2017). Among the compounds in JBT, ferulic acid was reported to have ameliorative potential for mitigating the effects of IL-6 associated with vincristine-induced painful neuropathy (Vashistha et al., 2017). Chlorogenic acid combined with chondrocytes on an alginate scaffold decreased MMP expression and improved the recovery of damaged articular cartilage . Caffeic acid repressed IL-6 and TNF-a in RA-FLS by blocking the phosphorylation of IkB and IkB kinase . Moreover, sweroside inhibited IL-1b-stimulated MMP-1, MMP-3, MMP-13, and ADAMTS-5 mRNA expression in rat articular chondrocytes by suppressing NF-kappa B and mTORC1 signaling (Zhang et al., 2019). Furthermore, ursolic acid reduced the incidence and severity of CIA-induced arthritis, accompanied by the decreased expression of TNF-a, IL-1b, IL-6, IL-21, and IL-17 in arthritic joints (Baek et al., 2014). Osthol was also shown to inhibit MMP-1, MMP-3, MMP-13, IL-6, and TNF-a in IL-1b-stimulated SW982 cells, as well as NF-kappa B and MAPK pathways . Although much research on the ingredients of JBT has been performed, as far as we know, we are the first to find that JBT acts as an NF-kappa B inhibitor. However, further confirmation of this mechanism via more in-depth studies is still required. The application of inhibitors and positive controls, such as iguratimod, should also be considered in future research.
We here applied network pharmacology to identify signaling pathways associated with RA. The NF-kappa B pathway was shown to be activated by many stimuli, including TNF-a (Wen et al., 2018). P65 participates in the classical NF-kappa B pathway. The inhibition of p65 phosphorylation was proven to suppress the pathogenesis of RA (Roman-Blas and Jimenez, 2006;Kang et al., 2018), as JBT did in this study. In the meantime, the effect inhibiting p38-MAPK was found. As a detectable component of JBT, gentiopicroside also exhibited a potent protective effect on the IL-1b-induced inflammatory response and the release of MMPs in rat articular chondrocytes by the p38, ERK, and JNK pathways (Zhao et al., 2015). Chlorogenic acid was reported to reverse IL-1binduced increases in IL-6, MMP-13, and COX-2/PGE2 production in human SW1353 chondrocytes, partially via the p65 NF-kappa B signaling pathway , and to induce apoptosis to inhibit the inflammatory proliferation of IL-6-induced FLSs through modulating the activation of the JAK/STAT and NFkappa B signaling pathways (Lou et al., 2016). Moreover, isoliquiritigenin treatment inhibited IL-1b-induced MMP production and NF-kappa B activation both in vitro and in vivo . Isoliquiritigenin was reported to suppress IL-1b-induced apoptosis and inflammation in chondrocyte-like ATDC5 cells by inhibiting NF-kappa B and to exert chondroprotective effects on a mouse model of anterior cruciate ligament transection (Ji et al., 2017). Xanthotoxin inhibited the expression of Runx2 and MMP-13 by reducing the activation of p38-MAPK . In contrast to previous studies, this study involved the first time that both HPLC-Q-TOF and network pharmacology have been applied to detect the effects of this classical herbal compound and determine the mechanisms involved. However, further studies are needed to clarify the pharmacokinetics of JBT, so as to provide references for clinical indications and the optimal dose. In the process of empirical revision, there are too many vague factors in this TCM compounds, and even the proportion of the compounds themselves needs to be optimized. Therefore, we must firstly prove the effectiveness of JBT in vitro and in vivo studies. This is the foundation for all proceeding research. Therefore, we tested the active ingredients using HPLC, with the aim of finding possible targets of JBT. In this way the follow-up work could be guaranteed. In future work, we also plan to continue to conduct experiments to explore other signal pathways by network pharmacology.
In conclusion, this study indicates that JBT could markedly ameliorate synovial inflammation and joint destruction in CIA and TNF-Tg mice. The NF-kappa B pathways were found to be closely involved in the inhibitory effects.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation, to any qualified researcher.

ETHICS STATEMENT
The animal study was reviewed and approved by Longhua Hospital -Animal Ethics Committee.