CILP is downregulated in the damaged areas of articular cartilage in clinical patients. (A) Digital radiograph and radiomics of patients with knee osteoarthritis (OA) (n = 16). Red ROI area to blue area is trend of cartilage thickness from thick to thin and from intact to damaged, anterior to posterior (A-P). (B) Western blot of CILP, typeⅡcollagen, typeⅠcollagen, Sox9, and MMP13 in undamaged (U) and damaged (D) areas. (C) Relative protein levels of CILP, type II collagen, type I collagen, Sox9, and MMP13 in undamaged and damaged areas, with β-actin as control. (D) Gross image and H&E, toluidine blue, and IHC staining of knee articular cartilage. Blue box indicates undamaged area, red box indicates undamaged area. (E) OARSI score of histology evaluation. Statistical analysis of CILP, collagen II/I, MMP13, and Sox9. *P< 0.001). (F) Quantification with heat maps for relative protein level of CILP in undamaged (U) and damaged (D) areas, with β-actin as endogenous control (n = 16). (G) Pearson r: correlation of CILP data.

Moderate-intensity treadmill exercise can increase the expression of CILP in the cartilage of SD rats and can restore the imaging and histological characteristics and subjective and objective function of the knee joint in SD rats. (A) Digital radiograph, gross image, H&E staining, toluidine blue, and IHC evaluation of rats. (B) Track and heatmap of open field test of rats. (C) Gait analysis of rats, including step sequence, footprint time, and footprint pressure. LF: left forelimb; LH: left hindlimb; RF: right forelimb; RH: right hindlimb. (D) Western blot of rat cartilage protein expression, including CILP, typeⅡcollagen, typeⅠcollagen, Sox9,α-SMA expression, and anti-ferroptosis proteins SLC7A11, HO-1, GPX4, and SOD-1. (E) Macroscopic and OARSI scores of rats. (F) Statistical analysis of IHC evaluation of rats. (G) Open field test of rats. Left axis shows distance and average speed. Right Y axis shows central zone exploration times. (H) Statistical analysis of gait. Significant differences were found between CG and OA groups (*P< 0.001; ANOVA) and OA and OAM groups (#P< 0.001, #a P = 0.0213; ANOVA). (I) Statistical analysis of rat cartilage protein. CG and OA groups (*a P = 0.0066; *b P = 0.0178; *c P = 0.0468; *d P = 0.0022; ANOVA); OA and OAM groups (#a P = 0.0073; #b P = 0.0065; #c P = 0.0387; #d P = 0.0425; #e P = 0.0354; #f P = 0.0043; #g P = 0.0026; ANOVA). Data are expressed as the mean ± 95% confidence interval; n = 3 per group. (J) IF and IHC evaluations of CILP expression in superficial, intermediate, and deep zones in CG, OA, and OAM groups. Significant differences were found between the superficial and intermediate zone groups (*P< 0.001; ANOVA) and intermediate and deep zone groups (#P< 0.001; ANOVA).

The cartilage intermediate zone acts as a mechanical-stimulation-sensitive zone in exercise therapy. (A) Using multiplex fluorescent immunohistochemical staining, we divided the cartilage into three zones (superficial, intermediate, and deep). (B) Multiple fluorescent immunohistochemical staining of CG, OA, and OAM groups; green: collagen II; yellow: collagen I; aggrecan: orange; red: sox9; blue: DAPI. (C) Statistical analysis of the relative fluorescence intensity differences between zones (superficial, intermediate, and deep). Significant differences were found between superficial and intermediate zones (*P< 0.001; *a P = 0.033; ANOVA) and intermediate and deep zones (#P< 0.001, #a P = 0.045; ANOVA). (D) Statistical analysis of relative fluorescence intensity of superficial zone between groups. Significant difference comparison gave the same result as in Figure 2. #a P = 0.0041. (E) Statistical analysis of relative fluorescence intensities of intermediate zones between groups. (F) Statistical analysis of relative fluorescence intensities of deep zones between groups. Significant differences were found between the CG and OA groups (*P< 0.001; ANOVA) and OA and OAM groups (#P< 0.001, #a P = 0.0014; ANOVA). Data are expressed as mean ± 95% confidence interval; n = 3 per group.

Generation 3 chondrocytes (G3) could mimic early fibrosis of articular cartilage with early OA, and moderate CTS alleviates early fibrosis of chondrocytes by upregulating CILP. (A) Gross image of adherent morphology of chondrocytes during passage. (B) Western blot of hyaline cartilage versus fibrocartilage phenotype in different passage chondrocytes. G0: primary chondrocyte; G3: Generation 3 chondrocyte; and so on. (C) Immunofluorescence of collagen II in different passage chondrocytes. (D) Statistical analysis of the hyaline cartilage versus fibrocartilage phenotype in different passage chondrocytes. Significant differences were found between G0 and G3 (*P< 0.001; ANOVA), G3 and G5 (#P< 0.001, #a P = 0.019; #b P = 0.021; #c P = 0.004; ANOVA), G3 and G7 (+P< 0.001; +a P = 0.002; +b P = 0.003; ANOVA), and G3 and G9 groups (^P< 0.001; ANOVA). (E) Statistical analysis of relative fluorescence intensity in different passage chondrocytes (#a P = 0.015; ANOVA). (F) Western blot of expression of hyaline cartilage versus fibrocartilage phenotype in G3 chondrocytes and TGF-β1-induced chondrocyte fibrosis models. (G) Western blot analysis of phenotypic proteins such as typeⅡcollagen, typeⅠcollagen, CEMIP, CILP, Sox9, and GPX4 of G3 chondrocytes after CTS treatment. (H) Statistical analysis of G3 chondrocytes and TGF-β1-induced chondrocyte fibrosis models. G0 and G3 groups (*P< 0.001; *a P = 0.0052; ANOVA); G3 and TGF-β1 10 ng groups (#P< 0.001; #a P = 0.0122; ANOVA); G3 and TGF-β1 20 ng group (+P< 0.001; +a P = 0.0293; ANOVA). (I) Statistical analysis of G3 chondrocytes after CTS treatment. G0 and G3 groups (*P< 0.001; *a P = 0.006; ANOVA); G3 and G3+CTS (low) groups (#P< 0.001; #a P = 0.007; #a P = 0.008; ANOVA); G3 and CTS (medium) groups (+P< 0.001; +a P = 0.0015; ANOVA); G3 and CTS (high) groups (^P< 0.001; ANOVA). Data are expressed as mean ± 95% confidence interval; n = 3 per group.

CILP competitively binds to sites of Keap1 protein and reduces stability of Keap1-Nrf2 dimer, decreasing Nrf2 ubiquitination and promoting Nrf2 nuclear translocation. (A) Yeast one-hybrid assays to detect potential binding sites for Keap1 and CILP. (B) OE/KD-CILP group CILP immunoprecipitated from chondrocytes with anti-Keap1 and anti-CILP antibodies, respectively, was analyzed using immunoblotting. (C) OE/KD-CILP group Keap1 immunoprecipitated from chondrocytes with anti-Keap1 and anti-CILP antibodies, respectively, were analyzed using immunoblotting. (D) OE/KD-CILP group Keap1 immunoprecipitated from chondrocytes with anti-Keap1 and anti-Nrf2 antibodies, respectively, were analyzed using immunoblotting. (E) OE/KD-CILP group Nrf2 immunoprecipitated from chondrocytes with anti-Keap1 and anti-CILP antibodies, respectively, were analyzed using immunoblotting. (F) OE/KD-CILP group Nrf2 immunoprecipitated from chondrocytes with anti-ubiquitination antibodies. (G) Western blot of Nrf2 in nucleus and cytoplasm of chondrocytes in OE/KD-group. Significant differences were found between negative control and OE-CILP, KD-02 (*P< 0.001; *a P = 0.0099; ANOVA), and OE-CILP and KD-02 groups (#P< 0.001; ANOVA). β-actin and Histone H3 were used for negative control.

CILP ameliorates chondrocyte ferroptosis through Keap1-Nrf2 pathway in vitro. (A) Western blot of OE –CILP G3 chondrocyte, including CILPP<P<, typeⅡcollagen, typeⅠcollagen, Sox9, α-SMA expression, and anti-ferroptosis protein SLC7A11, HO-1, GPX4, and SOD-1 expression. (B) Statistical analysis of proteins of G3 chondrocytes after OE-CILP treatment in G0 and G3 groups (*P< 0.001; ANOVA); G3 and OE-CILP groups (#P< 0.001; #a P = 0.041; ANOVA). (C) Western blot of KD–CILP (KD-01, KD-02) G3 chondrocyte protein expression, including CILP, typeⅡcollagen, typeⅠcollagen, Sox9, α-SMA expression, and anti-ferroptosis proteins SLC7A11, HO-1, GPX4 and SOD-1. (D) Statistical analysis of G3 chondrocytes after KD-01, KD-02 treatment. G0 and G3 (*P< 0.001; *a P = 0.0271; *b P = 0.0013; ANOVA); G3 and KD-01 (+P< 0.001; +a P = 0.002; +b P = 0.0146; ANOVA); and G3 and KD-02 groups (^P< 0.001; ^a P = 0.0313; ^b P = 0.0498; ^c P = 0.0068; ^d P = 0.0015; ANOVA). (F) Confocal microscope showed intracellular ROS levels, and flow cytometry analysis showed cellular ROS result. (F) Statistical analysis of G3 chondrocytes after OE-CILP, KD-01, and KD-02 treatment. G0 and G3 (*P< 0.001; ANOVA); G3 and OE-CILP (#P< 0.001; #a P = 0.011; ANOVA); G3 and KD-01, and KD-02 groups (+P< 0.001; ANOVA). (G) Results of incubation with JC-1 fluorescent and FerroOrange probes (I) and corresponding statistical analysis. (H) Statistical analysis of relative fluorescence intensity of JC-1 (+a P = 0.041; ANOVA). (I) Statistical analysis of relative fluorescence intensity of Fe2+ (*a P = 0.008; #a P = 0.0015; +a P = 0.0012; ANOVA). Data are expressed as mean ± 95% confidence interval; n = 3 per group. (J) Transmission electron microscopy of mitochondrial morphology changes in chondrocytes. Blue arrows indicate normal mitochondria, red arrows indicate ferroptosis-changed mitochondria (a reduction in mitochondrial volume, disappearance of cristae, and membrane disruption).

Exercise-CILP-Keap1-Nrf2 axis ameliorates early OA cartilage fibrosis by inhibiting chondrocyte ferroptosis. (A) Content of Malondialdehyde (MDA) in chondrocytes. Significant differences were found between G0 and G3 (*P< 0.001; *a P = 0.0018; ANOVA); OE-CILP and G3 (#P< 0.001; #a P = 0.0016; ANOVA); and G3 and KD-02 groups (+P< 0.001; ANOVA). (B) Content of GSH in chondrocytes. Significant differences were found between G0 and G3 (*P< 0.001; *a P = 0.006; ANOVA); OE-CILP and G3 (#P< 0.001; #a P = 0.0023; ANOVA); and G3 and KD-02 groups (+P< 0.001; +a P = 0.0056; ANOVA). (C) Western blot of ML385 as ferroptosis activator to antagonize anti-fibrotic and anti-ferroptosis effects of CILP. (D) Statistical analysis of relative protein expression after ML385 treatment. Significant differences were found between NC and OE-CILP (*P< 0.001; *a P = 0.0018; ANOVA) and OE-CILP and OE-CILP +ML385 groups (#P< 0.001; #a P = 0.0264; #a P = 0.042; ANOVA). (E) Exercise-induced cartilage intermediate zone and CILP-KEAP1-Nrf2 axis inhibit hyaline cartilage fibrosis and chondrocyte ferroptosis to alleviate early OA.

CILP and hyaline cartilage fibrosis are crucial in chondrocytes during early OA, with radiomics, single-cell transcriptomic, pseudo-time trajectory, and chondrocyte proteomic analysis. (A) Radiomics, single-cell transcriptomics, and chondrocytes proteomics diagram. (B) t-distributed stochastic neighbor embedding (t-SNE) of cell type of whole joint of SD rats between OAM and OA groups. (C) Differentially expressed genes in chondrocytes between OAM and OA groups were displayed in volcano plot and t-SNE. (D) Pseudo-temporal analysis of pathological process of OA chondrocytes. Starting from node 1, chondrocytes showed some phenotypic changes. (E) Identified chondrocyte subtypes throughout OA pathology; (F) Fc subtype was more expressed in OA group, and cluster was in early stage of OA chondrocytes. (G) Gene expression of each subtype of chondrocytes and type Ⅰ collagen content of Fc subgroup were increased, indicating cartilage fibrosis.

Proteomic analysis of CILP in treatment of early OA through Keap1-Nrf2 and chondrocyte ferroptosis. (A) Differentially expressed protein in chondrocytes between OE-CILP and G3 groups displayed in volcano plot. (B) Bioinformatic analysis of differential proteins associated with ferroptosis (displayed in heatmap). (C) Subcellular location showed Keap1 enriched in chondrocyte cytoplasm. (D) Differentially expressed protein cluster analysis. (E) Bioinformatic analysis of enrichment function of different proteins between OE-CILP and G3 groups. (F) Differentially enriched protein Keap1 interacts with string network of ferroptosis.