Pathogen-free male DBA/1 mice were purchased from Beijing Weitong Lihua Experimental Animal Co. Ltd. (Beijing, China) (SCXK2012-0001). The mice were housed in a laminar flow cabinet with a 12 h light/dark cycle and maintained on specific pathogen-free (SPF) laboratory chow and water ad libitum.

Preparation of type II collagen (CII) emulsion: chicken CII was dissolved to a concentration of 2 mg/mL in 0.1 μM acetic acid overnight at 4°C and mixed with complete Freund's adjuvant (C9301, Sigma, USA). Male DBA/1 mice (6–8 weeks old) were immunized with 100 μg of chicken CII (100 μL). The day of the first immunization was defined as day 0. The mice then received a booster with an equal amount of chicken CII emulsified in Freund's incomplete adjuvant on day 21 (31, 32). Mice with CIA were divided into six groups (n = 10 per group): a sham group, a model group, three DC32-treated groups (12.5, 25, 50 mg/kg/d), and a methotrexate (MTX) group (2 mg/kg/ every 3 days). From day 26 to 46, three doses of DC32 and MTX (dissolved in corn oil) or vehicles (corn oil) were administrated by oral gavage to the CIA mice. The sham group was also injected with chloral hydrate on day 0 and administrated with vehicle (corn oil) from day 26 to day 46.

From day 21 after the first immunization, clinical arthritis scores were determined every week using a scoring system of 0–4 for each limb: 0, normal; 1, definite redness and swelling of the ankle or one digit; 2, two joints involved; 3, more than two joints involved; and 4, severe arthritis of the entire paw and all digits. The footpad thicknesses were also measured every week. On day 46, the mice were killed, and joint or paw tissues were harvested for hematoxylin and eosin staining (33) and protein extraction. Then, the spleens were removed and weighed to evaluate splenomegaly. The spleen index (ratio of spleen weight to body weight × 10) was calculated.

Blood was collected from the retro-orbital venous plexus. After standing for 30 min at room temperature (RT), the samples were centrifuged at 5,000 rpm for 15 min. We evaluated the Nrf2 protein expression by western blotting of hind paw homogenates (34–36). Briefly, the mice were sacrificed, and the ankle joint was cut. Frozen hind paws were homogenized in liquid N₂. Then, the homogenate was suspended in 1 mL of ice-cold RIPA lysis buffer (Beyotime Biotechnology, China) with 10 μL of PMSF for 30 min for complete lysis. After centrifugation at 15,000 rpm for 10 min at 4°C, the supernatants were divided into aliquots and transferred to new tubes. All extraction procedures were performed on ice. Protein concentrations were then determined using a BCA protein assay kit (Thermo, USA).

The levels of HO-1 (ab204524, Abcam, UK), IgG (70-EK2712, Multisciences, China), and IgE (70-EK2752, Multisciences, China) in the serum were determined by ELISA kits.

Antibodies: antibodies against Nrf2, Keap1, lamin B, Akt, p-Akt, mTOR, p-mTOR, JNK, p-JNK, ERK, p-ERK, p38, and p-p38 were purchased from CST (USA); antibodies against LC3, HO-1, and p62 were purchased from Abcam (UK); and anti-GAPDH antibody was purchased from Sigma (USA). Proteins were resolved by 10% SDS–polyacrylamide gel electrophoresis and transferred to a PVDF membrane (0.22 μm, Merck Millipore). The membranes were treated with 5% BSA-Tris buffered saline (TBS) containing 0.1% Tween 20 for 2 h to block non-specific binding, rinsed, and incubated with primary antibodies diluted in 5% w/v non-fat dry milk (or BSA when indicated by the instructions), 1 × TBS, and 0.1% Tween 20 at 4°C with gentle shaking overnight. Signals were detected with HRP-conjugated anti-rabbit IgG using an enhanced chemiluminescence system (Bio-Rad, USA) (26).

The synovium from CIA mice was removed and incubated for 1 h with 80 rpm shaking and 1 mg/mL collagenase (Sigma–Aldrich) at 37°C. The cells were then washed and cultured in DMEM-10% FBS. The fibroblast-like synoviocytes (FLSs) were used after 3–5 passages. NIH-3T3 cells were cultured in DMEM-10% FBS with Penicillin-Streptomycin.

NIH-3T3 cells were seeded in 6-well plates at a density of 1 × 10⁵ cells/well overnight. LPS was added to the cells to generate a representative model of inflammation. After treatment with different concentrations of DC32 and LPS for 24 h, the cells were collected for protein extraction. The effect of DC32 on cell viability was determined by the MTT assay. For the detection of nuclear Nrf2 and Keap1 expression, nuclear and cytoplasmic fractions were separated by a nuclear and cytoplasmic protein extraction kit (Beyotime Biotechnology, China), according to the manufacturer's instructions. All steps were carried out on ice.

To confirm the role of autophagy in the regulation of Nrf2 and Keap1, we used the autophagy inhibitors 3-methyladenine (3-MA) and chloroquine (CQ) and the proteasome inhibitor MG-132 to block the degradation of Keap1 in NIH-3T3 cells. Cells were seeded in 6-well plates and pretreated with or without 5 mM 3-MA, 10 μM CQ, or 1 μM MG-132 before stimulation by LPS for 2 h. Then, DC32 (3 μM) was added to the corresponding wells.

The mRNA extraction kit (TransZol Up, ET111, Transgen, China) and reverse transcription kit (HiScript II Reverse Transcriptase, R223-01, Vazyme, China) were used according to the manufacturer's instructions.

qPCR was performed according to the manufacturer's instruction (AQ131, Transgen, China). All values were expressed relative to the expression of GAPDH.

The following primers were used:

F(Nrf2):TAGATGACCATGAGTCGCTTG

R(Nrf2):GCCAAACTTGCTCCATGTCC;

F(HO-1): CCGCCTTCCTGCTCAACAT

R(HO-1):GCCACATTGGACAGAGTTCAC;

F(Keap1): CCGCAGAATGTTACTATCCAGAG

R(Keap1): CGCTCCACACTGTTCAACTG;

F(p62): ATGTGGAACATGGAGGGAAGA

R(p62): GGAGTTCACCTGTAGATGGGT;

F(GA): TGATGGGTGTGAACCACGAG

R(GA): GCCCTTCCACAATGCCAAAG;

The translocation of Nrf2 was determined using an immunofluorescence assay (IF). The FLSs (5,000 cells/well in a 24-well plate) were seeded on glass coverslips overnight. Then, 1 mg/mL LPS and DC32 (1, 3, 10 μM) were added to the corresponding groups and cultured for 24 h. The cells were then fixed with 4% paraformaldehyde, washed and incubated with blocking buffer for 2 h at RT. The cells were incubated with the Nrf2 antibody (1:100) for 16–19 h at RT and then washed and incubated with biotin-conjugated anti-rabbit IgG (1:300) for 2 h. The cells were washed again and incubated with DAPI for 5 min. The coverslips were fixed on slides with 50% glycerin.

Predesigned small interfering RNA (siRNA) for p62 (sense: GGACCCAUCUACAGGUGAAT, antisense: UUCACCUGUAGAUGGGUCCTT') and Nrf2 (sense: GCAAGUUUGGCAGGAGCU-ATT, antisense: UAGCUCCUGCCAAACUUGCTT) and negative-control siRNA were purchased (TranSheepBio, China). Transfection mixes were prepared using Namipo. Cells at 30% confluence were transfected and then cultured for 24 h before treatment with DC32 or LPS.

SPSS (version 15.0) and GraphPad Prism software (Version 6.0) were used to conduct the statistical analyses. After a Kolmogorov-Smirnov non-parametric test for normality, data were statistically evaluated by one-way analysis of variance (ANOVA) and two-way ANOVA followed by Bonferroni's post-hoc tests. P < 0.05 was considered significant. Data are presented as the mean ± SEM.

DC32 and MTX were given to mice with CIA from day 26 to 46, and the clinical arthritis scores and footpad thickness were measured every week to evaluate the therapeutic effect of DC32. DC32 significantly reduced the clinical arthritis scores and footpad thickness (Figures 1B,C) without influencing the body weight (Figure S1F). The spleen index was measured after the mice were sacrificed. The results suggested that DC32 attenuated the splenomegaly caused by immunologic abnormalities (Figure 1D), the footpad swelling (Figure 1E) and the inflammation in CIA mice. Radiographs of the hind paws in the model group showed articular destruction and degradation of the bone matrices. However, bone destruction was obviously attenuated in the DC32-treated mice (Figure 1F). The HE sections revealed that the ankle joints in CIA mice were seriously damaged by synovial proliferation and inflammatory cell infiltration. DC32 treatment led to a remarkable improvement in inflammation and joint destruction (Figure 1G).

TNF-α was significantly reduced in all treatment groups, demonstrating the anti-inflammatory activity (Figure 2A). The rheumatoid factors (RFs) IgG and IgE were significantly upregulated in the model group, indicating that humoral immunity was activated. IgG and IgE were remarkably decreased in the DC32 groups in a dose-independent manner (Figures 2B,C).

The Nrf2/HO-1 antioxidant pathway could be activated by inflammation and ROS (37). Consistent with the results in RA patients, the expression levels of HO-1 and Nrf2 in CIA mice were slightly increased (7, 19). We found that DC32 could further activate the Nrf2/HO-1 antioxidant pathway and relieve inflammatory symptoms in mice with CIA. The HO-1 concentration in serum was measured by ELISA (Figure 3A). The HO-1 level in the DC32 group obviously increased. The Nrf2 level was upregulated significantly in the DC32 groups in a slightly dose-dependent manner (Figure 3B). The above results have proven that Nrf2/HO-1 may play an important role the DC32-amelioration of CIA, which was in harmony with the previous view that Nrf2/HO-1 is a therapeutic target in RA. We evaluated the effects of DC32 on Nrf2/HO-1 in NIH-3T3 cells and the cytotoxicity was analyzed (Figure S1B). When DC32 (1, 3, 10 μM) was given to NIH-3T3 cells for 24 h, both the protein and mRNA levels of HO-1 were significantly upregulated (Figure 3C). Moreover, Nrf2 expression was also increased, whereas the level of Keap1 was oppositely regulated. To clarify the mechanism underlying this effect, we measured the mRNA levels of Nrf2 and Keap1 at 24 h (Figure 3D). The results showed that DC32 activated Nrf2/HO-1 in CIA mice. The transcription of Nrf2 was increased, and the transcription of HO-1 was 10 times higher than that in the control NIH-3T3 cells. However, the mRNA level of Keap1 was not different between the groups. Therefore, the decrease in Keap1 was not caused by inhibiting its transcription.

Nrf2 translocates into the nucleus and activates expression of its related genes. To determine whether DC32 increased HO-1 expression by facilitating Nrf2 translocation, nuclear, and cytoplasmic protein extraction and IF were carried out. The western blot results showed that DC32 significantly increased Nrf2 nuclear translocation and reduced Keap1 levels in the cytoplasm (Figure 4A). Nrf2 nuclear translocation was further confirmed by an immunofluorescence (IF) assay (Figure 4B). In the control group, Nrf2 protein was mainly localized in the cytoplasm, but a low density was observed in the nucleus. After treatment with DC32, Nrf2 was exclusively found in the nucleus, which was consistent with the nuclear and cytoplasmic protein separation results. Additionally, our results further suggested that 10 μM DC32 resulted in a significant increase in the nuclear expression of Nrf2 and a concomitant decrease in its cytoplasmic levels.

To further evaluate the relationship between p62 and Keap1 degradation, the effect of DC32 on p62 expression and transcription was determined by western blotting and qPCR. DC32 significantly increased the p62 mRNA levels in CIA mice (Figure 5A), as well as the p62 expression and transcription in NIH-3T3 cells (Figure 5B). Moreover, opposite trends for p62 and Keap1 were observed in each group, which hinted that DC32 regulated Keap1 in a p62-dependent manner. To confirm the role of p62 in Keap1 degradation, predesigned siRNA for p62 was transfected into NIH-3T3 cells. The siRNA-mediated knockdown of p62 (Figure S1C) increased the Keap1 protein level, which indicated that the DC32-induced degradation of Keap1 was mediated by p62 (Figure 5C). Our results indicated that DC32 could enhance the transcription and expression of p62 and facilitate the degradation of Keap1. The degradation of Keap1 could be inhibited by the siRNA interference of p62, and the effect of DC32 on Keap1 could be reversed by the siRNA for p62. This result suggested that the degradation of Keap1 induced by DC32 was mediated by p62, which blocked the proteasomal degradation of Nrf2.

Selective autophagy is closely connected to p62 and Keap1. Autophagy is reported to be regulated by the MAPK pathway. Thus, the phosphorylation of JNK, ERK, and p38 was analyzed by western blotting. The results showed that DC32 significantly decreased ERK phosphorylation (Figure 6A) but had no influence on JNK and p38 phosphorylation (Figure S1D). Moreover, DC32 dramatically enhanced autophagy by increasing LC3II/LC3I (Figure 6A). These results revealed that DC32 remarkably induced autophagy by inhibiting ERK phosphorylation. DC32 significantly inhibited the phosphorylation of Akt. However, DC32 reduced the phosphorylation of mTOR by inhibiting its expression (Figure 6B). The autophagy inhibitors 3-MA and CQ were administered to NIH-3T3 cells, and the change in protein levels was determined by western blotting. Keap1 expression was decreased significantly with 3 μM DC32, and this phenomenon was partly abolished by 3-MA. MG-132, a proteasome inhibitor, was given to NIH-3T3 to determine whether the proteasome was involved in this Keap1 degradation. Both CQ and MG-132 caused aggravated degradation of Keap1 and accumulation of p62 Figure 6C, Figure S1E).

Nrf2 regulates p62 transcriptionally and may further enhance the accumulation of Nrf2. To verify the role of Nrf2 in the expression of p62, siRNA for Nrf2 was transfected into NIH-3T3 cells. The change in the protein level of p62 was determined by western blotting. The results showed that the increase in p62 induced by DC32 was attenuated by Nrf2 siRNA (Figure 7A, Figure S1C). In addition, the effects of DC32 on the expression and transcription of Nrf2 and p62 at 2, 4, 8, 16, and 24 h were measured by western blotting and qPCR (Figure 7B). The expression and transcription of Nrf2 were consistently increased with time, whereas the expression and transcription of p62 first decreased and then increased at 16 h. These results confirmed that Nrf2 is upstream of this DC32-induced Nrf2-Keap1-p62 feedback loop.

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.