Td (C₂₇H₄₅NO₂, purity ≥ 98%) was obtained from MedchemExpress (Shanghai, China). Lyophilized native chicken type Ⅱ collagen (CⅡ) was purchased from Chengdu Herbpurify Co., Ltd. (Chengdu, Sichuan, China). Methotrexate (MTX) was from Shanghai Sine Pharmaceutical Co., Ltd. (Shanghai, China). Commercial kits for measurement of alanine aminotransferase (ALT) and aspartate aminotransferase (AST) were from Jiancheng Institute of Biotechnology (Nanjing, Jiangsu, China). IL-1β, IL-6 and TNFα ELISA kits were purchased from SenBeiJia Biological Technology Co., Ltd. (Nanjing, Jiangsu, China). Dulbecco’s modified Eagle’s medium (DMEM) was purchased from Gibco (Gibco BRL, Grand Island, NY, United States). Fetal bovine serum (FBS) was purchased from Invigentech (Sydney, Australia). Recombinant tumor necrosis factor alpha (TNFα) was purchased from Peprotech (PeproTech, Rocky Hill, NJ, United States). MTS reagents, Triton-X 100, DAPI solution and Freund’s complete adjuvant (CFA) were purchased from Sigma-Aldrich (St Louis, MO, United States). 5-Ethynyl-2ʹ-deoxyuridine (EdU) was obtained from Guangzhou RiBo Bio Co., Ltd. (Guangzhou, Guangdong, China). Annexin V-FITC apoptosis detection kit was from KeyGEN Biotech Co., Ltd. (Nanjing, Jiangsu, China). TRIzol reagent was from Invitrogen Life Technologies (Carlsbad, CA, United States). Primary antibodies targeting phosphorylated extracellular signal-regulated kinase 1/2 (p-ERK1/2), phosphorylated c-Jun N-terminal kinases (p-JNK), phosphorylated p38 (p-p38), phosphorylated inhibitor of NF-κB (p-IκBα), total ERK, total JNK, total p38, IκBα, p65, MMP-2, MMP-9 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were purchased from CST (Cell Signaling Technology, Inc., Beverly, MA, United States). Anti-RANKL antibody was obtained from Proteintech (Proteintech Group, Inc., Chicago, United States).

Arthritic FLS cells were isolated and cultured as previously reported (Liu Z. et al., 2018). Briefly, the synovial tissues were dissected from the knees of the vehicle-treated CIA rats, minced and incubated with 0.25% trypsase and 0.4% collagenase. Arthritic FLS cells were collected and cultured in DMEM supplemented with 10% FBS, 100 U·mL⁻¹ penicillin and 100 μg·mL⁻¹ streptomycin at 37°C in 5% CO₂. Cells from passages three to nine were used for all in vitro experiments.

Arthritic FLS cells were seeded into 96-well plates and cultured for 48 h with complete DMEM containing varying concentrations of Td (0, 2.5, 5, 10, 20, 40, 80 μM). MTS/PMS mixture was added and cells were incubated for another 4 h. The absorbance at 490 nm was measured using a microplate reader (Model 680, BioRAD, Hercules, CA, United States). The effect of Td on cell viability was expressed as percent cell viability with vehicle treated control cells set at 100%. GraphPad Prism (version 5.0c; San Diego, CA) was used to calculate the half maximal inhibitory concentration (IC₅₀).

Arthritic FLS cells were incubated with Td (5 μM) for 6 h and then stimulated with or without TNFα (50 ng·mL⁻¹) for 24 h. EdU (10 μM) was added to measure the newly synthesized DNA and Hoechst 33342 was used to counterstain cell nuclei. The cell proliferation rate was calculated as the proportion of nucleated cells incorporating EdU in five high-power fields per well.

Wound-healing assay was carried out as our previous descriptions (Liu et al., 2018b; Qiu et al., 2018). In brief, arthritic FLSs were seeded into 12-well plates and cultured to 80% confluency. A linear scratch was created using a sterile 200 ul pipette tip. After washing the suspended cell debris with PBS, arthritic FLSs were treated with different concentrations of Td (0, 2.5, 5 and 10 μM) for 1 h and then stimulated with TNFα (50 ng·mL⁻¹) for 24 h. The cell migration area was measured by comparing the remaining cell-free area in the identical fields using Image J software (National Institutes of Health, Bethesda, MD, United States).

Arthritic FLS cells were seeded into 6-well plates and incubated with varying concentrations of Td (0, 2.5, 5 and 10 μM) for 24 h. Cells were then harvested for staining with annexin V and propidium iodide (PI) solution. The cells were detected by a flow cytometer (Beckman Coulter, CA, United States) and the apoptotic cell percentage was analyzed using Cell Quest Software (FACScan; Becton Dickinson, Franklin Lakes, NJ, United States).

Arthritic FLS cells were incubated with varying concentrations of Td (0, 2.5, 5 and 10 μM) for 1 h and then stimulated with TNFα (50 ng·mL⁻¹) for 24 h. The cells were harvested for real-time polymerase chain reaction (PCR) assay or Western blot analysis. The culture supernatants were collected for ELISA assay.

For real-time PCR assay, the total RNA was extracted using TRIzol reagent and subjected to cDNA synthesis according to the manufacturer’s instructions. Real-time PCR was performed using a SYBR Green PCR Kit (Vazyme Biotech) and run in Mastercycler ep realplex two systems (Eppendorf, Hamburg, Germany). The primer pairs for IL-1β, IL-6, TNFα, MMP-2, MMP-9, RANKL and β-actin were used as the previous descriptions (Ueland et al., 2005; Liu et al., 2018b). The expression of each target gene was normalized to the transcription level of β-actin gene.

For Western blot assay, the cells were lysed and the protein levels of MMP-2, MMP-9 and RANKL were examined by Western blot as described below.

For ELISA assay, the culture supernatants were centrifuged at 2000 rpm for 20 min to remove the particulate matter. The pro-inflammatory cytokines IL-1β, IL-6 and TNFα were determined using cytokine-specific ELISA kits according to the manufacturer’s instructions.

A total of 30 female Wistar rats (170–180 g, about 6–8 weeks) were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). All of the rats were housed under specific pathogen-free (SPF) conditions (22°C, 12 h/12 h light/dark, 50–55% humidity) and given free access to food and water. All the animal experiments were approved by the Experimental Committee of Nanjing Normal University (#2020065, approved date July 15, 2020).

The CIA model was induced as previously described (Liu et al., 2018a; Liu et al., 2018b; Qiu et al., 2018). Briefly, the rats were intradermally injected with 1 mg native chicken CII emulsified in Freund’s complete adjuvant. Seven days later, the rats were subcutaneously boosted with half the amount of CII emulsified in Freund’s incomplete adjuvant. The normal control rats were immunized with saline (n = 6). All the immunized rats developed CIA (clinical score ≥2) after a mean (±SD) interval of 14 ± 1 days. Rats with CIA were randomly divided into four groups as two Td-treated groups, MTX-treated group and vehicle-treated group (n = 6 per group). The MTX-treated group was administrated with MTX (3 mg·kg⁻¹ body weight) every 3 days (according to clinical usage). The Td-treated groups were intraperitoneally injected with different doses of Td (5 and 15 mg·kg⁻¹ body weight, respectively) and the vehicle-treated group was injected with 0.9% saline every day for a 14-day period. The doses of Td were determined according to the previous reports (Fujiwara et al., 2012; Dyle et al., 2014; Hu et al., 2019) with modification from our preliminary experiments. Clinical arthritic scores were evaluated using a scoring system of 0–4 for each limb: 0 = no swelling; 1 = swelling and/or redness of one to two interphalangeal joints; 2 = involvement of three to four interphalangeal joints or one larger joint; 3 = more than four joints red/swollen; 4 = severe arthritis of an entire paw. The reported arthritis score for each rat was the sum of the two hind paws. The volumes of the hind paws were measured using a volume displacement plethysmometer (YLS-7A, Facility Station of Shandong Academy of Medical Science, Shandong, China). Arthritis scoring and paw volume measurements were performed by two independent observers in a blinded manner. After a 14-days treatment, serum was collected for measurement of pro-inflammatory cytokines, ALT and AST. Joint tissues were harvested for radiographic and histopathological evaluation, cytokine and MMP analysis, and signaling pathway detection. In addition, we weighed the main organs including heart, liver, kidney, spleen and thymus, and calculated the organ coefficients, as our previous report (Liu et al., 2013). Also, the change in body weight of each individual CIA rat was calculated as the following formula: change in body weight = [(body weight of day 14 arthritis/body weight of day 1 arthritis)−1] *100%.

The left hind paws were separated, fixed and exposed under X-ray (MX-20; Faxitron X-ray Corporation, Wheeling, IL, United States). Bone destruction was evaluated using a scoring system of 0–3, as the previous report (Cai et al., 2007). The radiological evaluation was carried out by two independent observers in a blinded manner.

After X-ray examination, the left ankle joints were decalcified in 12% ethylenediaminetetraacetic acid (EDTA) and then embedded in paraffin. Sections with a thickness of 4 μm were stained with haematoxylin and eosin (H&E). Pathological changes in inflammation, pannus formation, synovial hyperplasia, cartilage destruction and bone erosion were scored using a semi-quantitative scale, as our previous descriptions (Liu et al., 2018a; Liu et al., 2018b; Qiu et al., 2018). Histopathological analysis was recorded by two independent observers in a blinded manner.

Serum was collected by centrifugation (5000 rpm for 15 min) and stored at −80°C until analysis. The levels of IL-1β, IL-6 and TNFα were measured using ELISA kits according to the standard method. In addition, serum ALT and AST were detected using commercial kits following the protocol of manufacturer’s instructions.

The right hind paws of rats were removed and homogenized, as our previous report (Liu et al., 2018a; Liu et al., 2018b; Qiu et al., 2018). The homogenates were used to detect the protein levels of IL-1β, IL-6, TNFα and MMP-9, RANKL by ELISA and Western blot, respectively. In addition, the phosphorylation changes of MAPKs (ERK, JNK and p38) and the protein level of IκBα were examined by Western blot.

Arthritic cells or joint tissues were homogenized in radioimmunoprecipitation assay buffer and the supernatants containing proteins were collected by centrifugation. Protein was separated by sodium dodecyl sulphate–polyacrylamide gelelectrophoresis (SDS-PAGE) and then transferred to nitrocellulose membranes. The membranes were incubated with various kinds of primary antibodies including p-JNK, JNK, p-ERK, ERK, p-p38, p38, p-IκBα, IκBα, MMP-2, MMP-9, RANKL and GAPDH. The immunoreactivity was visualized using enhanced chemiluminescence reagents (Labgic Technology) according to the manufacturer’s instructions. Three independent experiments were performed and the intensity of each band was analyzed using ImageJ software.

Arthritic FLS cells were seeded on sterile cover slips in a 24-well plate and cultured overnight at 37°C. Following serum starvation, the cells were treated with Td (10 μM) for 4 h, and then stimulated with TNFα (50 ng·mL⁻¹) for 30 min. After washing with PBS, the cells were fixed by methanol and permeabilized by 0.5% Triton-X 100. The arthritic FLS cells were blocked with 10% goat serum and then incubated with NF-κB p65 antibody. DAPI solution were used to stain cell nuclear. The nuclear translocation of p65 was imaged using a Nikon A1R resonance scanning confocal microscope with spectral detector (Nikon, Tokyo, Japan).

All data were expressed as the mean ± SD of results obtained from three or more experiments. Multiple comparisons were performed using one-way analysis of variance (ANOVA), followed by Tukey’s post hoc analysis. Comparisons between two groups were made using Student’s t-test. P < 0.05 was considered statistically significant.

Arthritic FLSs are the main effector cells of RA. Inhibiting the over-activated properties of FLSs has become a potential approach to RA therapies. In this study, we firstly used an MTS assay to quantify the potential cytotoxicity of Td on FLSs. As shown in Figure 1B, up to 20 μM for 2 days treatment, arthritic FLSs did not show a significant decline in cell viability. And the calculated IC₅₀ was 34.31 μM (Figure 1C). In the subsequent in vitro experiments, concentrations of Td used did not exceed 10 μM. We here investigated the effects of Td on TNFα-induced proliferation, migration and apoptosis in arthritic FLSs. The EdU assay showed that TNFα stimulation could dramatically enhance FLS’s proliferative potential, however, this enhancement was significantly suppressed by 5 μM Td (Figures 1D,E). The wound healing assay demonstrated that Td at doses of 2.5–10 μM could significantly and dose-dependently inhibit the migration rate of arthritic FLSs (Figures 1F,G). Since Td was able to induce apoptosis in several tumor cells (Song et al., 2018), we analyzed the apoptosis-inducing effect of Td on arthritic FLSs. As shown in Figure 1H, the given doses of Td had little effect on the apoptotic rate of arthritic FLSs, suggesting that the suppression of Td on proliferation and migration was not due to its apoptosis-inducing action.

It is well known that RA is an immune-driven inflammation disease and manifests as cartilage and bone erosion occurs the later stage of the disease. Efforts should be made to inhibit the expressions of pro-inflammatory cytokines as well as genes involved cartilage and bone destruction in RA treatment. Our studies revealed that TNFα stimulation could dramatically upregulate the mRNA levels of pro-inflammatory cytokines including IL-1β, IL-6 and TNFα. However, these upregulations were significantly inhibited by Td treatment (Figure 2A). Consistent with the real-time PCR data, ELISA assay showed that the protein levels of IL-1β, IL-6 and TNFα were also decreased by Td (Figure 2B), which further confirmed the anti-inflammatory action of Td. Rheumatoid FLSs contribute to the degradation of connective tissue by secreting substantial amounts of MMPs, which results in the breakdown of cartilage destruction. In this study, the data of real-time PCR and Western blot both demonstrated that Td had no obvious effect on the expression of MMP-2 but distinctively inhibited MMP-9 production (Figures 2A,C). RANKL has been regarded as the strongest inducer in osteoclast differentiation and activity (Yasuda et al., 1998; Kobayashi et al., 2000; Liu et al., 2018a). In our study, RANKL was significantly suppressed by Td at both the mRNA and protein levels (Figures 2A,C), indicating that Td might play an important protective role in bone destruction.

To assess the anti-arthritic effect of Td, a CIA model in Wistar rats was used. When the CIA model was firmly established (clinical score ≥2), Td with different doses or vehicle was intraperitoneally administrated once a day and continued for 14 days. As shown in Figure 3A, arthritis symptoms such as swelling and erythema were obviously observed in the vehicle-treated CIA rats, while Td or MTX treatment significantly attenuated the severity of the disease. Compared to MTX or 5 mg·kg⁻¹ of Td which significantly suppressed the development of CIA, 15 mg·kg⁻¹ of Td showed a stronger protective action as assessed by clinical score and paw swelling (Figures 3B,C).

To further confirm the therapeutic effect of Td, radiographic and histopathological assessments were performed on the rat hind paws. The X-ray images showed that the typical RA changes, such as joint destruction and joint space narrowing, appeared in the vehicle-treated CIA rats without exception (Figure 3D). However, these destructions were significantly improved in MTX- or Td-treated groups. Furthermore, this improvement was more obvious in the high-dose Td group (Figure 3E). In line with the data of radiographic scoring and the clinical scoring, H&E staining and histopathological assessment showed that Td significantly attenuated the pathological characteristics of CIA rats including inflammatory cell infiltration, synovial hyperplasia, pannus formation, cartilage erosion, and bone erosion, which further demonstrated the inhibitory effect of Td on RA (Figures 4A,B).

To evaluate the potential adverse effect of Td on CIA rats, a preliminary study was carried out to analyze the weight loss, organ indexes and serum markers of liver injury (ALT and AST). As shown in Table 1, compared to the normal control, the vehicle-treated CIA rats showed a significant weight loss. However, this weight loss was significantly suppressed by Td treatment. Except that the kidney index was decreased by Td, the other organ indexes as well as serum ALT and AST have no significant differences between the Td-treated groups and the vehicle-treated CIA group, further supporting that the given doses of Td used in this study had no significant side effects.

We have demonstrated that, in primary arthritic FLSs, Td could inhibit the expressions of multiple pro-inflammatory cytokines. To further confirm the in vitro results, serum and joint tissues of each group were collected and the protein expressions of IL-1β, IL-6 and TNFα were examined using ELISA. As shown in Figures 5A,B, the local and systemic levels of IL-1β, IL-6 and TNFα in the Td-treated groups were much lower than those in the vehicle-treated group. In addition, compared with the vehicle-treated rats, those treated with Td had significantly decreased production of MMP-9 and RANKL (Figure 5C), which further confirmed the data from arthritic FLSs.

To reveal the mechanisms through which Td plays an inhibitory action on CIA model and arthritic FLS cells, we used immunoblotting to examine the activations of MAPKs and NF-κB, which play vital roles in RA pathogenesis. In cultured primary arthritic FLSs, TNFα stimulation could rapidly increase the phosphorylation levels of the MAPK family members (ERK, JNK and p38) (Figures 6A,B). Td treatment significantly suppressed TNFα-induced phosphorylations of ERK and JNK but without effect on p38 (Figures 6A,B). In addition, TNFα stimulation initiated phosphorylation and degradation of IκBα. When the cells were treated with Td (10 μM), IκBα phosphorylation and degradation were both significantly suppressed (Figures 6A,B). It is well known that IκB degradation can liberate NF-κB p65 protein into the nucleus and trigger its downstream target gene expression (Makarov, 2001). Thus, we used immunofluorescence microscopy to detect the effect of Td on p65 nuclear translocation. As shown in Figure 6C, TNFα incubation for 30 min clearly promoted NF-κB p65 translocation from the cytoplasm to the nuclei. However, this translocation process was substantially blocked when the cell was treated with 10 μM Td, as evidenced by the cytoplasmic retention of p65 proteins (Figure 6C).

To confirm the in vitro results, we isolated the protein from the joint tissues of each group and examined the activations of MAPKs and NF-κB using Western blot. As expected, Td had no obvious influence on p38 phosphorylation. However Td significantly decreased the phosphorylation levels of ERK and JNK compared with the vehicle-treated group (Figure 6D). Additionally, the degradation of IκBα was also suppressed by high-dose Td treatment (Figure 6D).