IRS2 and FOXO1 may reduce osteoarthritis by controlling cartilage cell recycling and energy use

After obtaining approval from the Ethics Committee of Nanjing First Hospital, Nanjing Medical University (KY20230918-KS-01), we acquired twelve tibial plateaus from consenting osteoarthritis (OA) patients undergoing total knee replacement surgery. The medial portion of each knee tibial plateau was utilized for the damaged group, while the lateral tibial plateau with comparatively less cartilage damage served as the intact group. All experiments involving humans followed the Declaration of Helsinki [17]. Clinical trial number: not applicable.

We retrieved gene expression microarray datasets encompassing OA and normal human knee joint cartilage from the GEO public database and downloaded OA gene expression profiles GSE178557, GSE169077, GSE64394, and GSE57218. To integrate multiple datasets, we initially employed the R package inSilicoMerging for dataset merging [18]. Subsequently, we further applied the method proposed by Johnson WE et al. to remove batch effects, ultimately obtaining a batch-effect-corrected matrix [19].

The limma package was employed for the analysis of differentially expressed genes (DEGs) between patients and control groups, with filtering criteria set at p < 0.01 and |FC| >1.5 [20]. This facilitated the identification of DEGs in normal versus OA cartilage. The visualization of gene expression patterns between normal and OA groups was accomplished using the "ggplot2" and "pheatmap" R packages.

In this study, we employed the "glmnet" R package to perform regression analysis using the lasso-cox method. The regression analysis yielded five genes, with a Lambda value set at 0.183424729303045.

We utilized the CIBERSORT package in the R software to compute the proportions of 22 distinct immune cell types for patients exhibiting diverse immune profiles within the dataset [21]. Box plots were employed to visually represent the immune cell compositions among patients with varying immune patterns. The differences in immune cell proportions were assessed using the Wilcoxon rank-sum test, with a significance threshold set at p < 0.05. Statistical significance was considered when p < 0.05.

Primary mouse knee articular chondrocytes were meticulously isolated from mouse knee cartilage. The articular cartilage specimens were finely diced and subjected to a digestion process with 0.2% collagenase II (Biosharp, #BS164) at 37 °C for 10 h or more. Subsequently, the digested chondrocytes underwent filtration through 70-µm nylon filters (Biosharp, #BS-70-CS).

The cultured chondrocytes were nurtured in DMEM/F12 medium (Biosharp, #BL305A) supplemented with 10% fetal bovine serum (Excell, #FND500; v/v), 100 U/mL penicillin, and 100 µg/mL streptomycin (Beyotime, #C0222). The culture medium was diligently replaced every two days to maintain optimal conditions. Using collagen type II alpha 1 chain (COL2A1) immunofluorescence for phenotypic characterization of harvested primary mouse knee joint chondrocytes (FigureA). S2

Cultures were maintained in an incubator at 37 °C with a humidified atmosphere containing 5% CO₂. Upon reaching 70–80% confluency, tert-Butyl hydroperoxide (TBHP, Macklin, #B802372,) was introduced to induce reactive oxygen species (ROS) production.

The cells were subjected to infection with the designated multiplicity of infection (MOI) using either control adenovirus (Ad-C) or adenovirus expressing Irs2 (Ad-Irs2, #ADV-262384). The adenoviruses for mouse Foxo1 knockdown (Ad-m-Foxo1-shRNA, #shADV-259590) or the GFP shRNA silencing Adenovirus were procured from Vector Biolabs (Malvern, PA). A transfection efficiency of > 80% was considered successful and the cells were then used for subsequent experiments.

The ATDC5 cell line (Cat# 305427), derived from a mouse teratocarcinoma, was purchased from Cytion (Hamburg, Germany). This chondroprogenitor cell line was cultured and maintained in DMEM/F12 medium (Biosharp, #BL305A) supplemented with 10% fetal bovine serum (Excell, #FND500), 100 U/mL penicillin, and 100 µg/mL streptomycin (Beyotime, #C0222). Cells were maintained in a humidified incubator at 37 °C with 5% CO₂. The culture medium was replaced every two days.

ATDC5 cells were seeded for differentiation at a density of 6.0 × 10⁴ cells/cm² in multi-well plates. After 24 h, the maintenance medium was replaced, and to induce a uniform inflammatory state, all cells were pre-treated for 24 h with 10 ng/mL of recombinant mouse IL-1β (PeproTech, Cat# 211-11B). Following this pre-treatment, cells were washed with Phosphate-Buffered Saline (PBS) and subsequently cultured for 14 days in one of the four following experimental groups: (i) a vehicle control group, treated with differentiation medium containing 0.1% Dimethyl Sulfoxide (DMSO); (ii) an ITS-G induction group, treated with differentiation medium supplemented with 1X ITS-G Supplement (100X) [22](Thermo Fisher Scientific, Cat# 41400045); (iii) an Ad-Irs2 treatment group, where cells were transduced with an adenovirus expressing mouse Insulin Receptor Substrate 2 (Ad-Irs2) at a pre-determined Multiplicity of Infection (MOI); and (iv) an adenovirus control group, where cells were transduced with a corresponding empty adenovirus vector at the same MOI. The respective treatment media for all four groups were replaced every two to three days throughout the 14-day culture period.

The accumulation of sulfated glycosaminoglycans (sGAGs) in the extracellular matrix was assessed by Alcian Blue staining. Cells were washed twice with PBS and fixed with 4% paraformaldehyde for 20 min at room temperature. After fixation, the cells were rinsed with distilled water and stained for 30 min at room temperature using an Alcian Blue Staining Solution (pH 2.5) (Solarbio, Cat# G1560). Following staining, the solution was aspirated, and the plates were washed three times with distilled water to remove excess stain. The stained plates were air-dried, and images of the cartilage nodules were captured using a light microscope.

RNA was extracted from the cells utilizing the TRIzol reagent (ThermoFisher, #15596018CN). Subsequently, a first-strand cDNA was synthesized from 0.5 µg of total RNA, employing the SweScript RT II First Strand cDNA Synthesis Kit (With gDNA Remover, Servicebio, #G3333-50). The real-time reverse transcriptase RT-qPCR reaction was executed using the ABI QuantStudio 5 system. To assess gene expression, the RT-qPCR data were quantified relative to the housekeeping gene Gapdh.

For the cell viability assessment, a cell suspension (100 µL/well) was seeded into a 96-well plate, and the culture plate was incubated in a CO₂ incubator for 24 h. Subsequently, the designated treatments were applied and incubated for either 24–48 h. Following the treatment, 10 µL of CCK-8 solution (Beyotime, #C0037) was introduced to each well. The culture plates were returned to the incubator and incubated for an additional 2 h. The optical density at 450 nm was quantified using a Sunrise microplate reader (TECAN, Männedorf, Switzerland).

Anti-IRS2 (ab134101, Abcam), Anti-MMP13 (18165-1-AP, Proteintech), Anti-ACAN (AHP0022, Invitrogen), Anti-GAPDH (60004-1-Ig, Proteintech), Anti-COL2A1 (sc-52658, Santa Cruz), Anti-SOX9 (ab185966, Abcam), Anti-Beclin-1 (A21191, Abclonal), Anti-ATG7 (10088-2-AP, Proteintech), Anti-LC3 (#2775S, Cell Signaling Technology), Anti-AKT (60203-2-Ig, 1: 5000, Proteintech), Anti-Phospho-AKT (#4060S, Cell Signaling Technology), Anti-FOXO1 (#2880, Cell Signaling Technology), Anti-Phospho-FOXO1 (RK05760, Abclonal), Anti-Histone-H3 (17168-1-AP, Proteintech), Anti-Mouse secondary antibody (BL001A, Biosharp), Anti-Rabbit secondary antibody (BL003A, Biosharp).

For tissue samples, after cutting into fine pieces with scissors, add lysate (Beyotime, #P0013B) containing 1mM PMSF (Beyotime, #ST506). Homogenize with a homogenizer until fully lysed. For cell samples, the method is as described previously [23]. Briefly, after removing the culture medium, add an appropriate amount of lysis solution directly, and lyse on ice for 5 min. Subsequent steps follow the guidelines provided by the reagent manufacturer. Proteins were separated through 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (Biosharp, #BL565A), and then transferred to polyvinylidene fluoride (PVDF) membranes (Millipore, #IPVH07850). After blocking with 5% skim milk for an hour, the membrane was incubated with primary antibodies overnight at 4 °C. After washing with Tris-buffered saline with 0.1% Tween 20 detergent (TBS-T) for 3 × 10 min, incubated at 37 °C for an hour with secondary antibodies. Immunoreactive proteins were visualized by enhanced chemiluminescence and light detection by Tanon 5200 (Tanon, China). Relative quantitative analysis was performed using ImageJ 1.54 F.

Mitochondrial transmembrane potential and mitochondrial morphology were assessed using MitoTracker Red CMXRos (Beyotime, #C1049B). the location of lysosomes using LysoTracker Green was visualized using LysoTracker Green (Beyotime, #C1047S). Mouse chondrocytes were stained with 50 nM MitoTracker Red and 50 nM LysoTracker Green for 30 min at 37 °C and nuclei were stained with Hoechst 33,258 (Beyotime, # C1011) for 10 min. Microscopic observation was performed with a fluorescence microscope (Zeiss, Germany), and fluorescence intensity was measured using ImageJ 1.54 F.

The destabilization of the medial meniscus (DMM) was performed on 12-week-old male C57BL/6 mice (purchased from Nanjing First Hospital) to induce osteoarthritis (OA). Knee joint tissues were harvested four weeks post-surgery [24]. Ethical approval for animal-based studies was obtained from the Ethics Committee of Nanjing First Hospital, Nanjing Medical University (DWSY-2101052). All animal experiments strictly adhered to the guidelines outlined in the National Institutes of Health's Guide for the Care and Use of Laboratory Animals. The mice were housed in specific pathogen-free conditions with a 12-hour light and 12-hour dark cycle at a temperature of approximately 22 ~ 23 °C.

For the OA model, animals were anesthetized with 75 mg/kg ketamine and 10 mg/kg xylazine before undergoing unilateral DMM procedures. The surgical protocol was meticulously performed in accordance with the methodology outlined by S.S. Glasson B.V.Sc [24].

To elucidate the involvement of IRS2 in OA, mice subjected to either sham or DMM surgery received intra-articular injections of 1 × 10⁹ plaque-forming units (pfu) of control adenovirus (Ad-C) or adenovirus expressing murine Irs2 (Ad-Irs2).

All laboratory mice were disposed in accordance with the American Veterinary Medical Association (AVMA) Guidelines for the Euthanasia of Laboratory Animals (2020).

Chondrocytes seeded in six-well plates were treated with IL-1β and Ad-Irs2 for 24 h, followed by incubation with 10 µM 2′,7′-dichlorofluorescein diacetate (DCFH-DA) at 37 °C for 20 min to assess intracellular ROS levels. We measured the fluorescence of dichlorofluorescein (DCF), the oxidized product of DCFH-DA, using an excitation wavelength of 488 nm and an emission wavelength of 525 nm. After staining, the cells were analyzed using flow cytometry (Thermo Fisher Scientific), and the resulting fluorescence intensity was quantified using FlowJo software version 10.7 (BD Life Sciences, Franklin Lakes, NJ, USA).

Micro-CT analysis was conducted to specifically evaluate structural changes in subchondral bone, such as osteophyte formation and subchondral bone sclerosis, which are critical pathological features in OA progression. Knee joints were extracted, fixed in 4% paraformaldehyde for 24 h, and scanned using SkyScan1176 (Bruker, Germany). The manufacturer's software was used for 3D reconstruction and analysis.

Four weeks post-DMM, mouse long bones were harvested with the knee joint preserved and fixed in 10% neutral buffered formalin for 24 h at room temperature. Following fixation, the specimens underwent decalcification in 10% EDTA (pH 7.4) for 4 weeks at room temperature with gentle agitation, with the EDTA solution changed every 3–4 days to ensure effective decalcification[25]. Subsequent processing, including paraffin embedding, and sectioning, was conducted by ServiceBio in Wuhan, China. The specimens underwent Haematoxylin and Eosin staining, immunohistochemical staining, and Safranin O/Fast green staining (Sigma Aldrich, Germany), following the instructions provided by the respective reagent manufacturers.

For immunohistochemical staining, the Super Plus™ High Sensitive and Rapid Immunohistochemical Kit (Elabscience, #E-IR-R220) was utilized for subsequent procedures. Briefly, antigen retrieval coincided with heat restoration, followed by peroxidase inactivation. The primary antibody was incubated overnight at 4 ℃, and the corresponding secondary antibody was applied at room temperature for 30 min. After DAB color development, haematoxylin staining was performed, followed by dehydration, transparency, and mounting.

In tissue immunofluorescence, knee joint sections were deparaffinized, washed, and blocked with 10% goat serum for 1 h at room temperature. Primary antibodies were applied in 0.1% Tween 20 and incubated overnight at 4 °C, followed by secondary antibody using Alexa Fluor488, 568, 647-conjugated secondary antibody (Abcam, #ab150077, #ab175473, #ab15011) for 1 h at room temperature, and stained with DAPI for 3–5 min.

In cellular immunofluorescence, mouse chondrocytes were rinsed with Phosphate-Buffered Saline (PBS), fixed in 4% paraformaldehyde, and permeabilized in 0.1% Triton X-100 for 15 min. Subsequently, cells were blocked with 5% bovine serum albumin for 30 min at 37 °C, washed with PBS, and incubated overnight at 4 °C. Following washing, cells were incubated with Alexa Fluor488-conjugated secondary antibody (Abcam, #ab150077, #ab175473, #ab150115) for 1 h at room temperature and stained with DAPI for 3–5 min. Fluorescence signals were detected and captured using Axio Imager.A2 (Zeiss, Germany).

Quantitative analysis of Western blots, immunofluorescence, and immunohistochemistry was performed using ImageJ software (National Institutes of Health, USA). All data are presented as mean ± standard deviation (mean ± S.D). Statistical comparisons between the two groups were conducted using an unpaired, two-tailed Student's t-test. For comparisons among three groups, a one-way analysis of variance (ANOVA) followed by Dunnett's multiple comparison tests was applied. For all statistical analyses, differences with p values < 0.05 were considered statistically significant, and experiments were repeated for indicted times in the figure legend. All statistical analyses were performed using GraphPad Prism software version 9.5 (GraphPad Software, Inc.).

GSE178557, GSE169077, GSE64394, and GSE57218 were obtained from Gene Expression Omnibus (GEO). These samples were from healthy populations and osteoarthritis patients. We utilized the R package inSilicoMerging to merge the datasets. As is shown in the density graphic and box plots, the sample distributions of the datasets were quite different (FigureA, B, and C). The empirical Bayes method was used to adjust the batch effects of the merged datasets and the results indicated that the batch effect was removed effectively. S1

After the merging and standardization of the datasets, 86 DEGs (up = 46, down = 40) were identified using the "limma" package (Table S2, Fig. 1A, B). Among these DEGs, IRS2 exhibits the most significant differential expression, ranking first (p = 5.1 × 10⁻⁷) (Fig. 1C). The LASSO regression analysis yielded 5 genes (IRS2, ARHGDIB, BEST1, SERPING1 and VPS37B) from the optimal model (Fig. 1D, S1D, E). Therefore, we hypothesized that IRS2 may be a key target in the OA process. Furthermore, KEGG pathway enrichment analysis of the differentially expressed genes was performed (Fig. 1E), revealing a significant enrichment of the PI3K-AKT signaling pathway.

In datasets using human-derived samples, we verified that IRS2 expression was significantly reduced in OA articular cartilage (Fig. 2A). We subsequently performed histological staining on the tibial plateau cartilage obtained from individuals undergoing human knee arthroplasty, revealing a significant reduction in the protein level of IRS2 in OA (Fig. 2B-E). We employed a unilateral destabilization of the medial meniscus (DMM) to establish a murine knee osteoarthritis (OA) model (Fig. 2F, G) and, through qPCR analysis, observed a significant decrease in IRS2 mRNA levels in the articular cartilage of DMM-induced OA knee joints (Fig. 2H). After histological staining of mouse knee joint sections, we further confirmed a significant reduction in IRS2 protein levels in DMM-induced osteoarthritis (OA), accompanied by an increase in matrix metalloproteinases 13 (MMP13) and a decrease in Aggrecan (ACAN) (Fig. 2I, J). To further elucidate the role of IRS2 in chondrocytes, we initially validated the decreased expression of IRS2 in primary mouse chondrocytes subjected to IL-1β intervention using the GSE104793 dataset. Subsequently, we employed in vitro stimulation of mouse chondrocytes with 10 ng/ml of IL-1β, confirming our findings at both the transcriptional and protein levels. (Fig. 2K-N). The results of immunofluorescence demonstrated a significant increase in MMP13 expression and a significant decrease in collagen type II alpha 1 chain (COL2A1) expression in mouse chondrocytes under IL-1β stimulation. This was concomitant with a marked reduction in IRS2 expression (Fig. 2O, P). The results of these in vivo and in vitro experiments further confirm the potential key role played by IRS2 in OA.

To further validate the role of IRS2 in chondrocytes, we utilized adenovirus transfection to overexpress IRS2 in primary mouse chondrocytes. qPCR results revealed that the elevated levels of Mmp13 and the decreased levels of the extracellular Acan and Col2a1 induced by IL-1β stimulation were alleviated in mouse chondrocytes overexpressing IRS2 (Fig. 3A). The Western blot results were consistent with the qPCR findings, demonstrating that the expression levels of cartilage functional differentiation and extracellular matrix synthesis proteins, including sry-box transcription factor 9 (SOX9), could be restored with the overexpression of IRS2. Conversely, the extracellular matrix degradation protein MMP13 was reduced due to IRS2 overexpression. These findings confirmed that IRS2 overexpression could restore the functional loss of chondrocytes induced by IL-1β (Fig. 3B, C). Markedly increased EdU-positive chondrocytes evidenced that IRS2 overexpression could restore the proliferative capacity of chondrocytes (Fig. 3D, E). Additionally, the proliferation capacity of chondrocytes was assessed using CCK-8, and the results were consistent with the EdU findings. The overexpression of IRS2 reversed the reduced proliferative ability of chondrocytes induced by IL-1β (Fig. 3F). According to the Annexin V/PI method, the cell apoptosis rate of the control group, IL-1β group, IL-1β + Ad-C group, and IL-1β + Ad-Irs2 group was observed. After IL-1β treatment for 24 h, the chondrocyte apoptosis rate in the IL-1β group and IL-1β + Ad-C group was significantly increased compared with the control group. The chondrocyte apoptosis rate in the IL-1β + Ad-Irs2 group was significantly lower than that in the IL-1β + Ad-C group (Fig. 3G).

Adequate autophagic levels play a crucial role in maintaining the physiological homeostasis of chondrocytes [15]. We utilized RT-qPCR technology to analyze the impact of overexpressed IRS2 on the activity of the autophagy pathway in chondrocytes under interleukin treatment. The results revealed that following IL-1β treatment, the transcription levels of autophagy initiation protein Beclin-1, as well as autophagy-related 7 (ATG7), an E1-like enzyme and microtubule-associated protein 1 light chain 3 II (LC3-II), were significantly reduced. However, upon overexpression of IRS2, the transcription levels of these autophagy-related genes showed a partial recovery (Fig. 4A). Western blot analysis corroborated these findings at the protein level. Treatment with IL-1β decreased the expression of Beclin-1, ATG7, and the conversion to LC3-II, while causing an accumulation of the autophagy substrate P62, indicating a blockage of autophagic flux. Notably, the addition of chloroquine, an autophagy inhibitor that blocks lysosomal degradation, led to a greater accumulation of LC3-II in IRS2-overexpressing cells compared to the IL-1β-treated group, confirming that IRS2 restores the autophagic flux rather than just inducing autophagosome formation. Overexpression of IRS2 successfully reversed these effects, increasing Beclin-1, ATG7, and LC3-II levels while reducing P62 accumulation (Fig. 4B, C). In a destabilization of the medial meniscus (DMM)-induced mouse osteoarthritis model, tissue immunofluorescence technique similarly indicated a decrease in chondrocyte autophagic activity in OA, including ATG7, Beclin-1, and LC3 (Fig. 4D, E). To further confirm these findings at a cellular level, we performed immunofluorescence on cultured chondrocytes. The results showed that IL-1β treatment diminished the formation of distinct fluorescent puncta representing ATG7, Beclin-1, and LC3. In contrast, overexpression of IRS2 significantly restored the formation of these autophagosome-associated puncta, indicating a recovery of autophagosome generation (Fig. 4G, H). Mitochondria are strongly associated with the normal development and maintenance of cartilage. Chondrocytes can regulate the morphology and function of mitochondria through autophagy. Using MitoTracker staining, we observed that IL-1β promoted the fragmentation and fission of chondrocyte mitochondria, along with a corresponding decrease in mitochondrial membrane potential. However, overexpression of IRS2 partially restored a more filamentous mitochondrial morphology and rescued the membrane potential (Fig. 4F).

FOXO1 is strongly associated with the autophagic flux in chondrocytes. Therefore, we hypothesized that IRS2 regulates autophagy by modulating the expression of FOXO1. In both human and destabilization of the medial meniscus (DMM)-induced mouse osteoarthritis models, we observed that the protein levels of FOXO1 did not show significant changes compared to normal cartilage, while the phosphorylation levels of FOXO1 were significantly reduced in OA cartilage (Fig. 5A-D). Oxidative stress-related apoptosis and autophagy play crucial roles in the development of OA, a progressive cartilage degenerative disease with multifactorial etiologies. Therefore, we further validated the relationship between FOXO1 and IRS2 in mouse primary chondrocytes cultured in vitro. Western blot results demonstrated that the protein levels of p-FOXO1 significantly recovered in chondrocytes after overexpression of IRS2 compared to TBHP treatment. Additionally, to rule out other possible post-translational modifications of FOXO1 (such as acetylation), we used an AKT inhibitor to suppress FOXO1 phosphorylation without affecting other post-translational modifications. The results showed that when chondrocytes were treated with the AKT inhibitor, the protein levels of FOXO1 did not change significantly, and overexpression of IRS2 did not increase the phosphorylation levels of FOXO1 (Fig. 5E, F). To investigate whether FOXO1 was a key protein regulated by IRS2 in chondrocyte autophagy, we knocked down the expression of FOXO1 using sh-Foxo1. Western blot results showed that after FOXO1 knockdown, overexpression of IRS2 could not restore the autophagic activity of chondrocytes (Fig. 5G, H).

Immunofluorescence results showed that TBHP stimulation caused FOXO1 to translocate into the cell nucleus, and overexpression of IRS2 could reduce TBHP-induced nuclear translocation of FOXO1 (Fig. 5I, J). These results further confirmed that FOXO1 is an intermediate protein through which IRS2 regulated chondrocytes autophagy. We used DCFH-DA for chondrocyte reactive oxygen species (ROS) staining, and the results showed that overexpression of IRS2 could reduce the IL-1β-induced increase in ROS (Fig. 5K-M). These findings suggested that IRS2 played a regulatory role in chondrocyte autophagy and oxidative stress through the phosphorylation of FOXO1.

We initially stimulated chondrocytes with a gradient of TBHP for 24 h and observed that with increasing TBHP concentrations, FOXO1 gradually translocated into the nucleus, consistent with previous research findings (Fig. 6A, B) [26]. Literature has reported phosphorylated cytosolic FoxO1 associates with Atg7 to induce autophagy in iNKs[27, 28]. Thus, we hypothesized that cytosolic p-FOXO1 could facilitate chondrocyte autophagy by binding to ATG7. Immunoprecipitation results showed that TBHP treatment reduced the interaction between FOXO1 and ATG7, in line with previous results, partially explaining the downregulation of the chondrocyte autophagy pathway induced by TBHP. However, after treatment with the AKT activator SC79, the interaction level between FOXO1 and ATG7 significantly increased. Since activated AKT leads to the phosphorylation and nuclear export of FOXO1, this suggested that p-FOXO1, not Ac-FOXO1, participated in the autophagy regulation induced by ATG7 in chondrocytes (Fig. 6C, D).

Levels of the antioxidant substances GSH and SOD were measured in chondrocytes cultured in vitro, revealing that SC79 significantly increased intracellular GSH and SOD levels (Fig. 6E, F). Transmission electron microscopy was used to observe the mitochondrial morphology of chondrocytes, showing that TBHP stimulation caused mitochondrial swelling and fragmentation, while SC79 improved mitochondrial morphology (Fig. 6G). Immunofluorescence staining further supported our hypothesis: after treatment with SC79, FOXO1 translocated out of the nucleus, and the interaction between p-FOXO1 and ATG7 was significantly restored (Fig. 6H). Staining of chondrocyte mitochondria and lysosomes using MitoTracker Red and LysoTracker Green, respectively, similarly indicated that SC79 treatment restored mitochondrial autophagy in chondrocytes (Fig. 6I).

To delve into the functional role of IRS2 in mouse knee joint cartilage, we employed a strategy of intra-articular injection of adenovirus expressing Irs2. PCR validation revealed that intra-articular injection of adenovirus overexpressing Irs2 in DMM mice significantly increased Irs2 expression in chondrocytes (Fig. 7A). Hematoxylin and eosin (HE) staining results demonstrated an improvement in the cartilage damage on the surface of the mouse knee joint injected with Ad-Irs2. Safranin O/Fast Green staining similarly indicated that Ad-Irs2 injection could restore the loss of sulfated matrix and surface irregularities induced by destabilization of the medial meniscus (DMM) (Fig. 7B, C). Immunohistochemistry results revealed that Ad-Irs2 injection increased the content of p-FOXO1 protein in the knee joint cartilage layer. Additionally, the protein expression levels of IRS2, COL2A1 and ACAN were partially restored, while MMP13 was reduced compared to the DMM group (Fig. 7D, E).

Results from micro-CT indicated that Ad-Irs2 improved the formation of osteophytes and subchondral bone sclerosis caused by DMM, as well as the decreased smoothness of the joint surface (Fig. 7F). Calculated parameters such as Bone Area to Tissue Area ratio (B.Ar/T.Ar) and Trabecular Pattern Factor (Tb.pf) based on micro-CT results suggested that Ad-Irs2 ameliorated the changes in knee joint osteoproliferation, trabecular thickening, and trabecular structure induced by DMM (Fig. 7G). Additionally, the hot plate and knee extension tests showed that intra-articular of Ad-Irs2 significantly alleviates OA-induced knee pain in mice (Figure S4). These findings suggest that intra-articular injection of Ad-Irs2 significantly ameliorates DMM-induced knee arthritis in mice.