Antibodies used for western blotting (WB) and immunostaining included mouse anti- COL2A1 (Novus; United States), rabbit anti-β-catenin (ABclonal; China), rabbit anti-BMAL1 (Absin; China), rabbit anti-MMP-3 (CST; United States), rabbit anti-MMP-13 (ABclonal; China), rabbit anti-ADAMTS-4 (ABclonal; United States), rabbit anti-Cyclooxygenase-2 (COX-2) (CST; United States), rabbit anti-β-catenin (ABclonal; China), rabbit anti-GSK-3β (ABclonal; China), mouse anti-p-β-catenin (SANTA; United States), mouse anti-p-GSK-3β (SANTA; United States), mouse anti-GAPDH (Zsbio; China), Alexa Fluor® 488-conjugated goat anti-mouse immunoglobulin G (IgG) (H + L) (Servicebio; China), Cy3 conjugated goat anti-rabbit IgG (H + L) (Servicebio; China), Horseradish Peroxidase (HRP) goat anti-mouse IgG (H + L) (Zsbio; China), and HRP goat anti-rabbit IgG (H + L) (Zsbio; China).

Primary chondrocytes were isolated from tibial platform of rats under sterile conditions, released by treating with 0.25% Trypsin-EDTA and 0.2% Type II collagenase (Gibco Company, United States) at 37°C (5% CO2) in Dulbecco’s Modified Eagle’s Medium (DMEM)/Hams F-12, supplemented with 10% fetal bovine serum (FBS) (BI, Israel) and 1% Penicillin/Streptomycin. After passage, a second generation of chondrocytes was used for the experiment. For cell transfection, chondrocytes in logarithmic growth phase were taken and transfected when the density reached 60–80%. GP-Transfect-Mate medium mixture (serum-free medium + GP-Transfect-Mate) and siRNA medium mixture (serum-free medium + siBMAL1 (20 μM)) (Genepharma, China) were prepared and transfected. Sequences for negative control and siBMAL1s are listed in Table 1. DMEM/F-12 containing 10% FBS was replaced after 4 h, and the result was tested by WB or qPCR at 72 h after transfection.

The secondary generation chondrocytes were diluted to 2×10⁴ cells/ml, and cultured in 6-well plates. Cells were fixed with 4% paraformaldehyde, permeabilized, and then blocked with 3% bovine serum albumin (BSA) at room temperature for 30 min. After removing the blocking solution, chondrocytes were incubated with primary antibody overnight at 4°C. COL2A1 (1:100) and DAPI for chondrocyte identification, β-catenin (1:100) for nuclear translocation. The cell climbing slides were washed and covered with fluorescently labeled secondary antibody for 50 min at room temperature. DAPI was incubated for 10 min at dark. Fluorescence was observed under a fluorescence microscope (Nikon Eclipse Ti-S; Japan) at 545 nm (red) and 340 nm (blue).

Ribonucleic acid (RNA) in the chondrocytes was extracted at 4°C using SV total RNA isolation system (Promega, United States), the concentration was measured at 260 and 280 nm, and then reverse transcribed blending with GoScript™ Reverse Transcription Mix, Oligo (dT) (Promega, United States) into cDNA (1 μg/ml). The mixture of 10 μl GoTaq®qPCR Master Mix, 0.6 μl forward and reverse primers, 2.5 μl cDNA and 6.3 μl RNase-free ddH₂O (Promega, United States) were subjected to 40 amplification cycles. Primer sequences are listed in Table 2. The intensity of fluorescence signal accumulation during the entire PCR process was monitored using LightCycler^®480 (Roche, Germany) to obtain the cycle threshold (Ct).

Total protein samples of chondrocytes were extracted following the instructions (Beyotime; China). The mixture of radio immunoprecipitation assay (RIPA) and Phenylmethylsulfonyl fluoride (PMSF) fully lysed the cells. Then equal 20 μg amounts of protein were separated onto 10% SDS-PAGE and transferred to nitrocellulose (NC) filter membrane. The membranes were blocked in 5% nonfat milk dissolved in Tris Buffered Saline with Tween®20 (TBST) buffer and probed with primary antibodies for BMAL1 (1:1,500), MMP-3 (1:1,000), MMP-13 (1:1,000), ADAMTS-4 (1:1,500), COX-2 (1:1,000), β-catenin (1:2,000), GSK-3β (1:1,500), p-β-catenin (1:1,000), p-GSK-3β (1:1,000), and GAPDH (1:3,000) at 4°C overnight followed by secondary antibody coupled with horseradish peroxidase at room temperature for 2 h. The blots were developed using the ECL system (Tanon-5200; China) and the gray value was quantified with Image J software (1.53c; United States).

Adolescent (8–10 week old) Sprague Dawley rats were purchased from Liaoning Changsheng biotechnology co.,Ltd. The applicable Johns Hopkins Bloomberg School of Public Health Guidelines for the care and use of animals were followed. The ARRIVE (Animal Research: Reporting In Vivo Experiments) guidelines to improve standards of reporting of animal experiments was used. The experiment was divided into four groups: Control group; OA group; Shift group; and Non-shift group.

For OA model, the anterior cruciate ligament excision transection (ACLT) to the experimental rats was used according to the modeling method summarized by Teeple et al. (2013). Rats were narcotized with isoflurane and immobilized in the supine position. An incision of 1–1.5 cm in length was cut in the longitudinal direction of the surgical department to expose articular cavity and the ACL was broken by micro-shearing without damaging the articular cartilage under a microsurgical microscope. After performing a drawer experiment to confirm the surgery, the joint capsule and skin were sutured. The control (sham) group only cut the skin and subcutaneous fascia in the surgical department, did not expose the joint cavity, and then sutured. Mice were sacrificed at the end of the sixth week. The entire knee joints were harvested and submitted for histological analyses. The serum was collected for subsequent ELISA tests.

The OA model was not performed in the chronic shift group and the non-shift group. The modeling method of chronic CRD summarized by Preuss et al. (Preuss et al., 2008) was adopted. The non-shift (control) group maintained a constant L/D (12/12) (L: 8 AM-8 PM; D: 8 PM-8 AM) cycle during the entire experiment. The chronic shift model group also maintained a constant L/D (12/12) cycle, but with a 12 h phase shift per week for 22 weeks ( Figure 1A). The rats were constantly entering new LD cycles, resulting in chronic CRD. The rat knee joints were harvested at the end of the 22nd week after sacrificed. Subsequent studies were consistent with those of OA rats.

Bones were dehydrated and waxed in a gradient alcohol. After the wax-soaked tissues were blocked, sliced the trimmed wax block in a paraffin section, 4 μm. The water is roasted with dry wax and stored at room temperature. Paraffin slides dewaxed as follows: Two changes of pure xylene-two changes of pure ethanol-75% ethanol. For special staining, the slides stained in fast green dye solution for 1–5 min, washed away the excess dye solution until the cartilage was colorless, and soaked in 1% hydrochloric acid and alcohol for 10 s. Then staining with saffron dye solution for 1–5 s, and rapid dehydration in absolute ethanol. The slides were immersed in xylene to transparent, sealing with neutral resin. Red or orange-red cartilage and green bone formation were observed under microscope (Nikon E100; Japan), and took images for analysis. The scoring criteria for OA cartilage pathology were based on the assessment system issued by Osteoarthritis Research Society International (OARSI) (Pritzker et al., 2006).

For immunohistochemical analysis, deparaffinzing and rehydrating paraffin sections were performed following by antigen retrieval, blocking endogenous peroxidase activity and serum sealing. Incubations were performed overnight at 4°C with primary antibodies for BMAL1 (1:100), MMP-3 (1:100), MMP-13 (1:100), ADAMTS-4 (1:200), β-catenin (1:100), Aggrecan (1:100), and COL2A1 (1:20) prepared with PBS. Then the tissues were covered with secondary antibody (HRP labeled) and incubated at room temperature for 1.5 h. After Diaminobenzidine (DAB) chromogenic reaction and nucleus counterstaining with hematoxylin stain solution, the slices were dehydrated and sealed following by examined under a microscope (Nikon E100; Japan) and analyzed.

The serum and extracellular secretions (ECS) was extracted after centrifugation (2,500 rpm/min, 20 min). ELISA was performed following the instructions strictly by using rat tumor necrosis factor α (TNF-α) (MEIMIAN, China), Interleukin-6 (IL-6) (MEIMIAN, China), C-terminal crosslinking telopeptide of type Ⅱ collagen (CTX-Ⅱ) (MEIMIAN, China), cartilage oligomeric matrix protein (COMP) (MEIMIAN, China), inducible Nitric Oxide synthase (iNOS), prostaglandin E2 (PGE2), and cyclooxygenase-2 (COX-2) (NJJCBIO, China) ELISA kit. The absorbance of each tube was measured with a microplate reader at 450 nm reflecting the content and activity. ELISAcalc software was used to calculate the regression equation of the standard curve linear according to the concentration and OD value and four-parameter logistic curve was selected for fitting model.

Data are presented as mean ± SD in vivo and mean ± S.E.M in vitro. Unpaired t test was performed between the two groups. When three or more groups are compared, IBM SPSS Statistics 25.0 with one-way ANOVA and Tukey test for post hoc test were used. Graphs were drawn using the GraphPad Prism (version 8). Values with p < 0.05 were considered statistically significant, and p < 0.01 extremely significant.

Epidemiological investigation showed that long-term shift workers were more prone to obesity (Kubo et al., 2011), and obesity is also one of the major causes of OA. In this study, we monitored the body weight of rats, and found abnormal increase in shift group. No significant difference in body weight gain was found between control and OA group in 6 weeks. However, from week 10 to week 22, the weight gain of shift group was significantly higher than that of non-shift group (p < 0.01) (Figure 1B). By evaluating the changes of pro-inflammatory cytokines in circulating blood, the levels of IL-1β, IL-6, TNF-α, iNOS, COX-2 and PGE2 of OA rats were significantly increased compared with the control group (p < 0.05, p < 0.01, and p < 0.0001). Similarly, compared with non-shift rats, the shift group showed a similar upward trend of these inflammatory markers (p < 0.05, p < 0.01, and p < 0.001) (Figure 1C). It suggests that CRD can induce inflammation in rats.

To assess the effect of CRD on the articular cartilage, histological staining and immunological assays of the tibial plateau were performed. As shown in Figure 2A, the surface of cartilage in the control and non-shift group was smooth, and the matrix was evenly red stained. In the OA and Shift groups, the surface layer was rough, and red staining was decreased (asterisk), indicating the reduction of cartilage matrix synthesis or increase of degradation. However, hypertrophic and vacuolar chondrocytes could be seen (arrows) in OA cartilage but not in shift group. Additionally, OA and shift rats showed significant hyperplasia of synovium lining cell layer compared with control and non-shift rats, and inflammatory cells accumulated in small quantities (Figure 2B). Meanwhile, the OA biomarkers COMP and CTX-Ⅱ were significantly increased in OA rats and shift rats (Figure 2C). These indicate that the OA-like pathological changes of cartilage can be induced by altering the LD cycle for a long time.

Additionally, the significant increase of some hydrolytic proteases (MMP-3, MMP-13 and ADAMTS-4) (p < 0.001 and p < 0.0001) and the significant decrease of matrix fibers (Aggrecan and COL2A1) (p < 0.001 and p < 0.0001) in OA cartilage indicated that the inflammatory response of OA cartilage was severe. Interestingly, similar changes in protein expression were observed in the shift group (Figures 2D,E) (p < 0.05, p < 0.01, p < 0.001, and p < 0.0001), which suggested that LD cycle disturbance bias the rate of cartilage degradation over matrix synthesis. Both ACLT and CRD significantly reduced the expression of BMAL1 in cartilage (Figures 2D,E) (p < 0.05 and p < 0.001), indicating the dysregulation of cartilage clock.

In order to further explore the underlying mechanism of cartilage clock disorder, we knocked down the BMAL1 protein in second-generation chondrocytes (Supplementary Figure S1). All three siBMAL1s significantly reduced the expression of BMAL1 protein in chondrocytes, with interference efficiency between 30 and 40% (Supplementary Figure S2A). And compared with the control group, no significant change of BMAL1 expression in negative control (mock) group was found. In order to improve the efficiency of siRNA interference, a hybrid siBMAL1 interference (siRNA614:siRNA1083:siRNA1125 = 1:1:1) was performed. Protein silencing efficiency was about 50% (p < 0.0001) (Supplementary Figure S2B), and gene silencing efficiency was about 80% (p < 0.0001) (Supplementary Figure S2C). In vitro OA chondrocyte model was established by treating chondrocytes with 10 ng/ml IL-1β (Wu et al., 2018). The secretion of IL-6, TNF-α, and PGE2 was significantly increased after treatment with IL-1β and siBMAL1 compared with the control group (Figure 3A). It is suggested that inhibition of BMAL1 expression can produce an inflammatory response similar to IL-1β stimulation.

Two studies have shown that IL-1β activates the Wnt/β-catenin signaling pathway in chondrocytes (Wu et al., 2018; Xi et al., 2020). In this study, the expression of β-catenin protein and mRNA in chondrocytes treated with IL-1β were significantly increased (p < 0.0001), and the expression of p-β-catenin was significantly decreased (p < 0.0001), compared with the control group. Similarly, a similar trend occurred after silencing BMAL1 (p < 0.0001) (Figure 3B). The protein and mRNA contents of GSK-3β were significantly dampened after IL-1β stimulation (p < 0.0001), BMAL1-KD also accelerated the phosphorylation of GSK-3β (p < 0.01) (Figure 3C). Moreover, IL-1β and siBMAL1 both induced different degrees of nuclear translocation of β-catenin in chondrocytes (Figure 3D). These indicate that the down-regulation of BMAL1 expression activates the Wnt/β-catenin signaling pathway.

In vivo, the percentage of β-catenin positive cells in the OA and Shift groups was extremely significantly ascended compared to the control group (Figure 3E). It is suggested that the disturbance of the cartilage clock may induce the expression of a series of downstream inflammatory factors and matrix degradation enzymes by activating the Wnt/β-catenin signaling pathway, thus accelerating the degradation of cartilage.

As can be seen from Figures 4A,B, compared with the control group and mock group, the expression levels of MMP-3, MMP-13, ADAMTS-4, and COX-2 protein and corresponding mRNA levels in chondrocytes treated with IL-1β and siBMAL1 were significantly increased (p < 0.01, p < 0.001, and p < 0.0001). This corresponds to the results in vivo. The Wnt/β-catenin pathway specific inhibitor XAV-939 has been shown to be significantly effective in improving an OA-mouse model (Lietman et al., 2018). After co-treatment with XAV-939, the expression levels of these proteins and mRNAs were significantly decreased compared with siBMAL1 group (p < 0.05 and p < 0.0001) (Figures 4A,B). XAV-939 also significantly dampened the increase of COMP caused by siBMAL1 (p < 0.05), but had no obvious effect on CTX-Ⅱ (Figure 4C). These findings suggest that the canonical Wnt/β-catenin pathway plays an important role in the mechanism of accelerated matrix degradation caused by disruption of cartilage clock.

Additionally, the contents of COL2A1 and Aggrecan mRNA in cells were extremely decreased after stimulation of IL-1β and transfection of siBMAL1 (p < 0.001 and p < 0.0001). XAV-939 significantly recovered the reduced COL2A1 and Aggrecan mRNA (p < 0.05) (Figure 4D), promoting chondrocyte matrix synthesis. In addition, levels of pro-inflammatory factors in ECS following inhibition of the Wnt/β-catenin pathway were also detected. The results showed that XAV-939 significantly reduced TNF-α secretion, but was not sensitive to IL-6 and PGE2 secretion (Figure 4E).