In this experiment, mice were housed under a constant temperature and humidity environment with regular 12-hour light/dark cycles as well as free access to a standard diet and water. All experimental procedures were approved by the Ethics Committee of Institute of Radiation Medicine Chinese Academy of Medical Sciences (approval number: IRM-DWLL-2018010).

OA was established in 10-week-old C57BL/6 mice by destabilization of the medial meniscus (DMM) of the right knee (Li et al., 2020a). Briefly, after anesthesia with 1.5% tribromoethanol 200 mg/kg body weight i.p. injection, surgery was completed by transection of the anteromedial meniscotibial ligament and the medial collateral ligament. They were randomly divided into the following groups: the control group, DMM group, DMM + low-dose Imrecoxib group (5 mg/ml/day), DMM + medium-dose Imrecoxib group (10 mg/ml/day), and DMM + high-dose Imrecoxib group (20 mg/m1/day), with 15 mice in each group. The day after surgery, the mice received Imrecoxib by oral gavage once a day for 12 weeks. We performed histological analysis and graded articular cartilage degeneration using the Osteoarthritis Research Society International (OARSI) guidelines.

Behavioral assessments including mechanical allodynia and thermal hyperalgesia were conducted using von Frey filaments and radiant heat tests, respectively, at pre-surgery and 1, 4, 8 and 12 weeks post-operation.

Mechanical Allodynia (von Frey test): Animals were acclimated in a mesh-bottomed cage for 3 h prior to testing. A graded series of von Frey filaments (forces: 2.44–4.74 g) were applied perpendicularly to the mid-plantar surface of the right hind paw. The minimal force eliciting a positive response (paw withdrawal or licking) was recorded using an up-down paradigm, with termination criteria defined as either five consecutive negative responses or four consecutive positive responses.

Thermal Hyperalgesia (Hargreaves' test): Mice were habituated in a transparent acrylic chamber mounted on a temperature-controlled glass plate (30°C) for 30 min. A focused radiant heat source (5 × 5 mm aperture) was directed to the plantar surface, and the paw withdrawal thermal latency (PWTL) was measured as the time from heat onset to withdrawal. A 20-s cutoff was implemented to prevent tissue injury. Three trials per animal were averaged with 6–8-minute inter-trial intervals.

Specimens were prepared as described previously. The fixed knee joints were decalcified for 30 days using a 10% EDTA solution. Subsequently, the specimens were subjected to dehydration, embedded in paraffin, and serially sectioned at a thickness of 5 μm to ensure the inclusion of the entire joint. For IHC analysis, sections were stained with primary antibodies: IL-6 (1:100, ab290735), TNF-α (1:100, ab183218), MMP3 (1:100, ab52915), IL-1β (1:100, ab315084). For immunofluorescence, sections were stained with primary antibodies: CD86 (1:100, Servicebio GB13585), CD206 (1:100, Servicebio GB11349) and FITC-labeled or CY3-labeled secondary antibodies (1:200, Servicebio GB22303; 1:200, Servicebio GB21303). The sections were mounted with medium containing DAPI and images were obtained using a fluorescence microscope (Nikon Eclipse Ti-SR). Three sagittal sections containing the lesion were selected for each group. In each section, four distinct areas within the lesion were chosen, photographed, and analyzed. The polarization of macrophages was then quantified using ImageJ software.

The entire fixed joint was subjected to Micro CT scanning (μCT 40; Scanco, Zurich, Switzerland) as to assess bone damage and repair. The scanning precision was 8.96 μm, with a filter selection of 0.5 mL Al and a 180° helical scan. The scan voltage and current were set at 60 kV and 368 μA, respectively. NRecon software was used for data reconstruction, and Data Viewer software was employed to observe the bone structure in coronal, sagittal, and transverse planes. CTan software was utilized for 3D quantitative analysis, and CTvox/CTvol software was employed to generate 3D effect images. The analysis primarily focused on bone trabecular data at the distal end of the femur and the tibial plateau. The ROI for trabecular bone at the distal end of the femur was selected within the range of 0.215 mm–1.94 mm from the growth plate. Key evaluation parameters included the ratio of trabecular bone volume to total bone volume (BV/TV), trabecular separation (Th.Sp), trabecular number (Tb.N), and trabecular thickness (Tb.Th).

Macrophages (10⁶ cells per well) seeded in a 6-well plate were stimulated with or without LPS (Sigma-Aldrich, 100 ng/mL) and various concentrations of Imrecoxib for 24 h. The experiment was divided into five groups: the macrophage group, M1-type macrophage group, M1 + low-dose Imrecoxib group (10 μmol/L), M1 + medium-dose Imrecoxib group (50 μmol/L), and M1 + high-dose Imrecoxib group (100 μmol/L). The qRT-PCR and immunofluorescene were conducted to detect the polarization transitions.

On the sixth day, when changing the culture medium, add 100 ng/mL of LPS to the existing culture medium. After 24 h of stimulation, RAW264.7 macrophages (Procell Life Science&Technology Co) polarized into M1 macrophages. Then, the new medium was changed to exclude the effect of residual LPS. The conditioned medium (CM) from macrophages was collected within 24 h of M1 macrophages. Centrifuge the medium at 1000 g for 5 min to remove cell debris and store the supernatant at −80°C for further experiments. The CM was diluted with serum-free culture medium at a 1:1 ratio and added to chondrocytes (Yi et al., 2024). The qRT-PCR and immunofluorescene were conducted to verify the protective role of Imrecoxib in chondrocyte.

Chondrocytes seeded in 6-well plates were stimulated with or without IL-1β (PeproTech, 10 ng/mL) and various concentrations of conditioned medium (CM) for 24 h. The rates of chondrocytes apoptosis were determined by flow cytometry analysis with an Annexin V-FITC/PI Apoptosis kit according to the manufacturer’s instructions (BD Biosciences).

Total RNA from primary chondrocytes and macrophages in different groups was isolated by the Total RNA kit (R6834–01, Omega). Total RNA integrity and quantity were determined using an Agilent 2,100 Bioanalyser (Agilent Technologies, San Jose, CA, USA). Only RNA with an A260/A280 ratio between 1.8 and 2.1 was used. and cDNA was transcribed reversely from total RNA. Reverse transcription was conducted using Transcription High Fidelity cDNA Synthesis Kit (5081963001, Roche) according to manufacturer’s instructions. Finally, we used the LightCycler^® 96 system to analyze the mRNA expression levels in cells. The primers used are listed in Table 1. The qRT-PCR conditions were: 95°C for 10 min, 40 cycles of 95°C for 10 s, 60°C for 20 s and 72°C for 15 s. The expression levels of each target gene were calculated with the 2^−ΔΔCT method and normalized to the internal control (GAPDH).

Pre-polarized RAW264.7 macrophages were treated with various concentrations of Imrecoxib for 24 h. The CM of macrophages was collected, and the concentrations of PGE2 were measured via ELISA kits according to the manufacturer’s guidelines (Enzyme-linked Biotechnology, ml028719).

Graphad Pro 8.0 was used for data analysis. The data were shown in mean ± standard error of mean (SEM). One-way ANOVA with Tukey’s post hoc test and two-way ANOVA with Bonferroni’s post hoc test were used for comparisons among multiple groups. P < 0.05 was considered statistically significant.

To investigate the sensitivity of mice to noxious stimulation after DMM surgery, the Von Frey and hot plate test was conducted. The results showed no significant differences in baseline paw withdrawal thresholds among the groups before surgery. From the 4th week postoperatively to the 12th week, the paw withdrawal thresholds in the DMM group began to slowly rise and were significantly lower than those in the Imrecoxib group at each time point (p < 0.01), indicating that Imrecoxib alleviates the pain caused by DMM surgery (Figure 2A).

Then, the mouse knee joints were subjected to micro-CT scanning to explore the impact of Imrecoxib on subchondral bone remodeling in OA. In the DMM group, the BV of subchondral bone beneath the tibial plateau was significantly decreased compared to the control group (p < 0.05). Nevertheless, mice treated with Imrecoxib at concentrations of 5, 10, and 20 mg/mL exhibited suppressed bone destruction, as evidenced by BV/TV, Tb.N, Tb.Th, and Tb.Sp (Figures 2B,C). The results indicated that administration of Imrecoxib in a dose-dependent manner significantly reduced bone loss, demonstrating increased integrity of knee joint bones and enhanced subchondral bone mass.

In the control group, the articular cartilage surface was smooth and intact, with a well-maintained cartilage structure and tidemark. In contrast, the articular cartilage staining in the DMM group revealed a significant reduction in thickness of articular cartilage and disorganized arrangement. OARSI scores indicated a significant increase in the DMM group compared to the blank control group (p < 0.01). Compared to the DMM group, the Imrecoxib treatment group showed a significant increase in cartilage matrix and joint thickness, with a dose-dependent decrease in OARSI scores (low dose: 5.33 ± 1.15; medium dose: 4.67 ± 1.15; high dose: 2.67 ± 1.15) (p < 0.05).

Immunohistochemical staining results demonstrated that, compared to the control group, the inflammatory factors IL-6, TNF-α, MMP3, and IL-1β significantly increased in the synovial tissue of the DMM group (p < 0.01). In comparison, the Imrecoxib treatment group exhibited a significant reduction in inflammatory reactions (p < 0.05). In summary, Imrecoxib could protect cartilage and delay the progression of OA (Figure 3).

Immunofluorescence staining of synovial macrophages revealed that the proportions of CD86⁺ and CD206+ cells in the DMM group were 26.27% ± 0.64% and 8.53% ± 1.29%, respectively. Compared to the DMM group, the Imrecoxib treatment group showed a significant decrease in the proportion of CD86⁺ cells (low dose: 19.50% ± 1.50%; medium dose: 16.10% ± 0.85%; high dose: 9.07% ± 1.90%), while the proportion of CD206+ cells was significantly upregulated (low dose: 9.50% ± 1.32%; medium dose: 19.70% ± 1.54%; high dose: 28.07% ± 2.53%) (Figure 4). These results indicate that Imrecoxib in OA mice can reduce the proportion of M1-type synovial macrophages, increase the proportion of M2-type synovial macrophages.

Moreover, Immunofluorescence revealed that Imrecoxib inhibited CD86⁺ cells (M1 phenotype) and promoted the expression of CD206+ cells (M2 phenotype) (Figures 5A–D). Imrecoxib exhibited a dose-dependent effect on RAW264.7 macrophages polarization, which reduced the expression of M1-related genes, including IL-6, IL-1β, TNF-α, IL-12, and iNOS, while increasing the expression of M2-related genes, such as Arg-1 and CD206. These differences were statistically significant (p < 0.01) (Figure 5E). These results indicate that Imrecoxib promotes polarization of M1 macrophages toward M2 phenotype and suppresses the inflammatory response of synovial macrophages.

To explore whether Imrecoxib could protect chondrocytes by modulating macrophage polarization, our study evaluated the impact of conditioned medium (CM) from differently polarized macrophages (including M0, M1 and M1+differerent dose of Imrecoxib, collected within 24 h) on chondrocyte apoptosis and inflammatory factor release. Compared to the M1-CM group, Imrecoxib-CM reduced the apoptosis rate of chondrocytes (Low-dose CM: 19.50% ± 1.50%; Medium-dose CM: 16.10% ± 0.85%; High-dose CM: 9.07% ± 1.90%, p < 0.01) (Figure 6A). RT-qPCR results indicated that after intervention with CM from M1 macrophages, the mRNA expression of inflammatory factors (IL-6, IL-1β, TNF-α, and Cox-2) and degenerative indicators (MMP-3 and ADAMTS5) significantly increased, while in the Imrecoxib CM group, the expression levels of these indicators decreased in a dose-dependent manner (Figure 6C). These results suggest that Imrecoxib plays important roles in inhibiting chondrocytes apoptosis, reducing inflammation and delaying degeneration through modulating macrophage polarization.

Immunofluorescence staining indicates a significant increase in COX-2 protein content in M1-type synovial macrophages. Following Imrecoxib intervention, the expression of COX-2 in M1-type synovial macrophages significantly decreased. (Figures 7A,B). After LPS stimulation, ELISA results demonstrated that Imrecoxib significantly reduced the LPS-induced secretion of PGE2 by RAW264.7 macrophages (p < 0.01) (Figure 7C). Taken together, the results suggests that Imrecoxib may modulate macrophage polarization through the COX-2/PGE2 pathway.