Circulating Metabolites Treat Human TMJ‐OA by Eliminating Senescent Chondrocytes via the C1QBP/C1q/p14ARF Axis

ABSTRACT

Temporomandibular joint osteoarthritis (TMJ‐OA) is a progressive degenerative disorder, for which therapeutic interventions remain limited. The disruption of metabolic homeostasis plays a critical role in the pathogenesis and advancement of TMJ‐OA. However, it remains unclear whether extracellular vesicles (EVs) as cellular metabolites are correlated with the pathogenesis, treatment and diagnosis of TMJ‐OA. In this study, we demonstrated that autologous circulating extracellular vesicles (C‐EVs) possessed significant therapeutic potential for TMJ‐OA through the targeted removal of senescent chondrocytes. In a randomized clinical trial (ChiCTR2200063153), C‐EV administration was found to significantly enhance condylar bone regeneration and alleviate symptoms relative to hyaluronic acid controls, without eliciting any adverse effects. Comparative analysis revealed that joint cavity‐derived EVs from TMJ‐OA patients (OA‐EVs) exhibited structural abnormalities, diminished expression of canonical EV markers, and pro‐inflammatory characteristics. In contrast, C‐EVs were significantly enriched with functional proteins C1q binding protein (C1QBP). And the level of C1QBP‐positive EVs was positively correlated with therapeutic outcomes, thereby establishing C1QBP as a potential predictive biomarker for TMJ‐OA. Furthermore, C‐EVs reestablished joint homeostasis by regulating the immune microenvironment and tissue regeneration capacity. Mechanistically, C1QBP^high C‐EVs upregulated the expression of membrane C1q on senescent chondrocytes, thereby initiating C1q–C1QBP binding, p14ARF translocation to mitochondria, and subsequent cytochrome C/caspase‐3‐dependent apoptosis. Our findings demonstrate that C‐EVs serve a dual therapeutic role by facilitating the clearance of senescent cells via the C1QBP/C1q/p14ARF axis, while promoting tissue regeneration and regulating metabolites homeostasis, offering a novel biological strategy for TMJ‐OA treatment.

1. Introduction

Temporomandibular joint osteoarthritis (TMJ‐OA) is a degenerative condition characterized by inflammation of the synovial membrane, degradation of cartilage, and alterations in bone structure, which collectively result in persistent pain and impaired joint function (Jones et al. 2019, Katz et al. 2021, Kloppenburg et al. 2025). The intricate biomechanics of the joint render it especially vulnerable to injury, exhibiting a higher prevalence compared to numerous other joint pathologies (Glyn‐Jones et al. 2015, Kloppenburg et al. 2025). Existing therapeutic approaches, including nonsteroidal anti‐inflammatory drugs (NSAIDs), hyaluronic acid (HA) injections, and joint replacement surgery, primarily provide symptomatic relief without effectively arresting the progression of the disease (Yang et al. 2023). A primary contributing factor to TMJ‐OA is the accumulation of senescent chondrocytes. These dysfunctional cells evade apoptosis and secrete pro‐inflammatory mediators, thereby perpetuating a deleterious cycle of cartilage degradation (Andia and Maffulli 2013, Coryell et al. 2021). The selective elimination of senescent cells constitutes a potentially effective therapeutic strategy for addressing this complex condition.

Disrupted metabolic homeostasis significantly contributes to the progression of osteoarthritis (Mobasheri et al. 2017). Significantly, the properties of extracellular vesicles (EVs) isolated from osteoarthritic synovial fluid (OA‐EVs), especially in terms of their function as cellular metabolites and their involvement in the pathogenesis of osteoarthritis, remain incompletely understood. Current therapeutic strategies have demonstrated limited efficacy in restoring metabolic homeostasis in degenerated chondrocytes. In contrast, EVs obtained from healthy sources represent a promising therapeutic approach for the management of OA. These nanoscale phospholipid bilayer particles inherently transport bioactive molecules, including proteins, nucleic acids, and lipids, and play a crucial role in intercellular communication by enabling receptor‐mediated uptake (Maacha et al. 2019, van Niel et al. 2018). Our prior research demonstrated that autologous circulating extracellular vesicles (C‐EVs), derived from plasma and peripheral blood mononuclear cells (PBMCs) of healthy donors, have the capacity to mitigate inflammatory bone loss associated with diseases such as osteoporosis and rheumatoid arthritis (Liu et al. 2018, Qu et al. 2024, Wang et al. 2021, Wang et al. 2023, Xue et al. 2024). Parabiosis studies have further elucidated that C‐EVs act as intrinsic anti‐aging agents, mitigating the progression of age‐associated tissue degeneration (Loffredo et al. 2013, Prattichizzo et al. 2019, Ruckh et al. 2012, Sahu et al. 2021, Villeda et al. 2014, Wagner et al. 2024, Yoshida et al. 2019, Zhang et al. 2023). Recent developments in EV research have underscored their potential utility as diagnostic biomarkers for osteoarthritis (Kumar et al. 2024); however, the therapeutic significance of alterations in EV composition associated with the disease has not been extensively investigated. This deficiency is especially pronounced within the domain of cellular senescence, wherein existing senolytic agents that target anti‐apoptotic pathways (SCAPs) in senescent cells exhibit considerable limitations, such as inadequate tissue specificity, safety issues, and an absence of regenerative potential (Childs et al. 2017, Kirkland and Tchkonia 2020, Lelarge et al. 2024, Li et al. 2019, Novais et al. 2021, Xu et al. 2018, Zhang et al. 2019, Zhu et al. 2015).

This study undertook a prospective clinical trial (ChiCTR2200063153) to evaluate the therapeutic efficacy of autologous C‐EVs in TMJ‐OA and to elucidate the associated underlying mechanisms. Our findings indicated that OA‐EVs displayed pronounced structural abnormalities relative to those derived from healthy controls, suggesting that they may play an active role in the pathogenesis of osteoarthritis. The localized administration of C‐EVs demonstrated a markedly superior efficacy compared to conventional HA therapy in enhancing condylar bone regeneration and alleviating associated symptoms. A substantial 17.9‐fold increase in the expression of C1q binding protein (C1QBP) was observed in therapeutic C‐EVs relative to pathogenic OA‐EVs. C1QBP^high C‐EVs selectively eliminate senescent chondrocytes, thereby facilitating tissue repair in TMJ‐OA. We elucidate a mechanism through which C‐EVs reestablish joint homeostasis by transporting C1QBP, thereby inducing C1q‐dependent mitochondrial translocation of p14ARF and subsequently activating apoptosis in senescent cells. Additionally, we identify C1QBP as a predictive EV biomarker for the early detection of OA and for monitoring therapeutic responses.

2. Results

2.1. EVs Within the Joint Cavity Exhibit Pathogenic Properties in TMJ‐OA Patients

To examine the involvement of EVs in the progression of TMJ‐OA, EVs were isolated from three distinct sources: (1) synovial fluid obtained from the joint cavity of TMJ‐OA patients (OA‐EVs), (2) autologous plasma (Plasma‐EVs), and (3) peripheral blood mononuclear cells (PBMC‐EVs) (Figure 1A). The comparative analysis demonstrated significant variations in the morphology and composition of EVs. Transmission electron microscopy (TEM) images revealed that OA‐EVs exhibited disrupted or incomplete membrane structures, whereas plasma‐EVs and PBMC‐EVs consistently presented intact bilayered membranes (Figure 1B). PKH26‐labeled immunofluorescence staining exhibited a similar phenotype and OA‐EVs were more likely to aggregate (Figure 1C). Concurrently, nanoparticle tracking analysis (NTA) indicated that these OA‐EVs were smaller in size (Figure 1D) and displayed a decreased zeta potential (Figure 1E) relative to circulating EVs. These results indicate that membrane abnormalities in OA‐EVs may result in the leakage of their contents and subsequent functional deficits, given that membrane integrity is critical for the proper activity of extracellular vesicles (Hallal et al. 2022).

FIGURE 1.

The characteristics of Plasma‐derived EVs (Plasma‐EV), PBMC‐derived EVs (PBMC‐EV), and joint liquid‐derived EVs from TMJ‐OA patients (OA‐EV). (A) Schematic presentation of separation and purification for Plasma‐EV and PBMC‐EV. (B) Representative transmission electron micrographs (TEM) images of Plasma‐EV, PBMC‐EV, and OA‐EV from TMJ‐OA patients. Scale bar: 50 nm. n = 10. (C) Representative super‐resolution structured illumination microscopy (SIM) images of PKH26‐labelled Plasma‐EV, PBMC‐EV, and OA‐EV. Scale bar: 150 nm. n = 10. (D, E) Representative particle size (C) and membrane potential (D) of Plasma‐EV, PBMC‐EV, and OA‐EV as assessed by nanoparticle tracking analysis (NTA). n = 10. (E) Western blotting analysis depicted the expression of EV markers CD9, CD63, Alix, and TSG101 in Plasma‐EV, PBMC‐EV, and OA‐EV. n = 5. (G–J) Representative super‐resolution SIM images and semi‐quantification of the expression of EV markers CD9, CD63, Alix, and TSG101 in Plasma‐EV, PBMC‐EV, and OA‐EV. Scale bar: 100 nm. n = 10. (K) Nano flow cytometric analysis (NanoFCM) revealed the expression of EV markers CD9, CD63, Alix, and TSG101 in Plasma‐EV, PBMC‐EV, and OA‐EV. n = 10. Error bars are mean ± SD. Data were analysed by one‐way ANOVA with a Bonferroni test for comparison of multiple groups. ****p < 0.0001.

A thorough biomarker analysis demonstrated a marked reduction in the expression of canonical extracellular vesicle (EV) markers, including CD9, CD63, Alix, and TSG101, in OA‐EVs as evidenced by multiple detection techniques such as western blotting, immunofluorescence, and flow cytometry (Figure 1F–K). Among the three EV populations examined, those PBMC‐EVs consistently exhibited superior quality parameters (Figure 1B–K), suggesting their potential therapeutic advantage. The abnormal features of OA‐EVs, such as structural abnormalities, increased tendency to aggregate, and modified molecular composition, provide compelling evidence for their involvement in the pathogenesis of TMJ‐OA. These results prompt important inquiries regarding the mechanistic involvement of OA‐EVs in the progression of the disease, as well as the potential of C‐EVs, including plasma‐derived EVs and PBMC‐derived EVs, to function as innovative biological therapies aimed at mitigating these pathological processes.

2.2. Clinical Effectiveness and Safety Profile of C‐EVs in the Management of TMJ‐OA

To assess the therapeutic efficacy of C‐EVs in the treatment of TMJ‐OA, we conducted a prospective clinical trial registered with the WHO International Clinical Trials Registry (ChiCTR2200063153). From October 2022 to December 2024, a total of 72 eligible patients diagnosed with TMJ‐OA were enrolled. Of these, 59 patients completed the entire treatment regimen and subsequent follow‐up assessments. To determine the effective and safe dosage for EV administration, dosing regimens of 20 and 50 mL of circulating EVs were used. These regimens were formulated in accordance with the principles of equivalent dose conversion between rats and humans, as well as the guidelines of dose escalation design. The cohort included 20 patients assigned to the hyaluronic acid (HA) control group, 20 patients receiving 20 mL of C‐EV treatment, and 19 patients administered 50 mL of C‐EV treatment. The remaining 13 participants were lost to follow‐up for reasons outlined in Figure S1. Extensive baseline evaluations demonstrated the absence of statistically significant differences in demographic variables, disease severity, and clinical parameters across the treatment groups, thereby confirming the comparability of the study cohorts (Table S1, Figure 2A–F).

FIGURE 2.

C1QBP^high C‐EVs alleviate condyle injury and clinical symptoms in TMJ‐OA patients. (A‐C) Representative condyle CBCT images and semi‐quantification analysis of condyle from TMJ‐OA patients after HA and different doses of C‐EV treatment at 0 month, 3 months, and 6 months. CBCT scores were used to statistically evaluate the intergroup (B) and intragroup comparison (C). (0 M, before treatment; 3 M, 3 months after treatment; 6 M, 6 months after treatment) (D) Fricton Index analysis showed the mandibular function score of TMJ‐OA patients in HA group, 20 mL‐C‐EV group and 50 mL‐C‐EV group at 0 M, 1 M, 3 M and 6 M. (E) NRS Index analysis showed the pain evaluation of TMJ‐OA patients in HA group, 20 mL‐C‐EV group and 50 mL‐C‐EV group at 0 M, 1 M, 3 M and 6 M. (F) OHIP‐TMDs analysis showed the life quality assessment of TMJ‐OA patients in HA group, 20 mL‐C‐EV group and 50ml‐C‐EV group at 0 M, 1 M, 3 M and 6 M. Error bars are mean ± SD. Data were analysed by one‐way ANOVA with a Bonferroni test for comparison of multiple groups. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Serial cone‐beam computed tomography (CBCT) imaging conducted at baseline (0 months), 3 months, and 6 months post‐treatment revealed progressive regeneration of the condyle across all experimental groups (Figure 2A–C). Quantitative analysis of CBCT scores indicated significantly enhanced bone repair in both C‐EV treatment groups relative to the HA control group by the 3‐month time point (Table S2). Notably, the 50‐mL dose of C‐EV demonstrated superior efficacy in addressing severe osteolytic lesions compared to the 20‐mL dose (Figure 2B,C).

Standardized assessments utilizing the Fricton index, the Numerical Rating Scale (NRS), and the Oral Health Impact Profile for Temporomandibular Disorders (OHIP‐TMDs) revealed a uniform reduction in symptom severity across all experimental cohorts (Figure 2D–F). Both C‐EV groups demonstrated significantly greater improvement compared to the HA control group at 6 months; however, no significant differences were observed between the different C‐EV dosage groups (Figure 2E,F). Extensive monitoring encompassing complete blood counts (Fig. S2A–D), serum biochemical analyses (Fig. S2E–H), and coagulation assessments (Fig. S2I–L) indicated no statistically significant differences between the treatment cohorts, thereby affirming the safety of C‐EV administration. These findings support the potential of C‐EVs as a therapeutic intervention for TMJ‐OA, demonstrating substantial structural restoration alongside symptomatic improvement and a favorable safety profile.

2.3. C‐EVs Reestablish Joint Cavity Homeostasis in TMJ‐OA Patients via the Enrichment of C1QBP

Proteomic analysis of EVs derived from patients with TMJ‐OA demonstrated notable distinctions between C‐EVs and OA‐EVs. A total of 7,360 proteins were identified in C‐EVs, whereas 6,370 proteins were detected in OA‐EVs (Figure 3A). Among these, 2,664 proteins exhibited differential expression that satisfied the established significance criteria (fold change >1.2, Q‐value <0.05). Specifically, 1,667 proteins were upregulated and 997 proteins were downregulated in C‐EVs relative to OA‐EVs (Figure 3B,C; Table S3). Functional annotation revealed that the proteins upregulated in C‐EVs were predominantly associated with key biological pathways, including the regulation of cell growth and death, signal transduction, immune modulation, and tissue regeneration (Figure 3D). These findings indicate that C‐EVs contain a diverse array of functionally active molecules that potentially contribute to mitigating the pathological processes of osteoarthritis.

FIGURE 3.

The different protein expression patterns of C‐EV and OA‐EV from TMJ‐OA patients. (A)Venn diagram analysis showed overlapping proteins between C‐EV and OA‐EV. (B‐C) Volcano plot and cluster heatmap analysis showing the differentially expressed proteins (DEPs) distribution between C‐EV and OA‐EV. (D) KEGG classified annotation of 1667 highly expressed proteins in the C‐EV group than in the OA‐EV group. (E) GO cellular component, molecular function and biological process (GO‐CFP) classified annotation of 1667 highly expressed proteins in the C‐EV group than in the OA‐EV group. (F) Western blotting showed the expression of C1QBP in OA‐EVs under 50 mL‐C‐EV treatment at 0 W, 2 W, and 4 W. (0 W, before treatment; 2 W, 2 weeks after treatment; 4 W, 4 weeks after treatment.) n = 5. (G) Representative super‐resolution SIM microscopy images and semi‐quantification analysis showed the expression of C1QBP in OA‐EVs under 50 mL‐C‐EV treatment at 0 W, 2 W, and 4 W. C‐EVs from TMJ‐OA patients were used as positive controls.Scale bar: 100 nm. n = 10. (H, I) Representative particle size (H) and membrane potential (I) of OA‐EVs under 50ml‐C‐EV treatment at 0 W, 2 W, and 4 W. n = 10. (J) NanoFCM revealed the expression of C1QBP in OA‐EVs under 50ml‐C‐EV treatment at 0 W, 2 W, and 4 W. n = 10. (K) The level of C1QBP in the joint fluid of TMJ‐OA patients under HA, 20ml‐C‐EV and 50ml‐C‐EV treatment at 0 W, 2 W, and 4 W. n = 10. Error bars are mean ± SD. Data were analysed by one‐way ANOVA with a Bonferroni test for comparison of multiple groups. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

A notably significant observation was the pronounced 17.9‐fold increase in the expression of C1QBP in C‐EVs relative to OA‐EVs (Table S3). This differential expression was validated using several independent methodologies, including western blotting, immunofluorescence, and nanoflow cytometry, all of which consistently revealed markedly lower levels of C1QBP in baseline OA‐EVs (0 W) compared to C‐EVs (Figure 3F,G,J). A gradual recovery of C1QBP expression within EVs derived from the joint cavity was observed throughout the four‐week C‐EV treatment duration (Figure 3F,G,J). This restoration was concomitant with the normalization of critical EV properties, including membrane integrity, particle size distribution, and membrane potential (Figure 3G–I).

The therapeutic significance of these findings was further corroborated by ELISA quantification, which revealed concurrent elevations in joint fluid C1QBP levels following administration of C‐EVs (Figure 3K). Collectively, these results indicate that C‐EVs exert their clinical effects through multiple complementary mechanisms: the restoration of deficient C1QBP expression in diseased OA‐EVs, normalization of EV biophysical characteristics, and reestablishment of balance within the synovial EV proteome. The strong association between C1QBP restoration and clinical improvement underscores its role as a critical therapeutic mediator and suggests its potential utility as a predictive biomarker for treatment response in TMJ‐OA.

2.4. C1QBP as a Predictive EV Biomarker for Early Detection and Therapeutic Assessment of TMJ‐OA

To evaluate the effects of C‐EVs on the senescence‐associated secretory phenotype (SASP) in TMJ‐OA, we conducted analyses of joint lavage fluid employing cytokine array and ELISA. The cytokine heatmap indicated that administration of a high dose of C‐EVs significantly diminished levels of pro‐inflammatory cytokines while concurrently enhancing the expression of anti‐inflammatory mediators (Figure 4A). These findings were corroborated by ELISA results, which demonstrated notable reductions in IL‐1β, IL‐6, TNF‐α, and RANTES, accompanied by increases in IL‐10 and TGF‐β1 at both 2 and 4 weeks post‐treatment with C‐EVs (Figure 4B).

FIGURE 4.

C1QBP is a newly predicted EV marker of early detection and monitoring the progression of TMJ‐OA. (A) The heatmap of the cytokine chip showed differentially expressed cytokines in the joint fluid of TMJ‐OA patients under HA and 50ml‐C‐EV treatment at 6 months. (B) The inflammation biomarkers in the joint fluid of TMJ‐OA patients under HA, 20ml‐C‐EV, and 50ml‐C‐EV treatment at 0, 2, and 4 weeks, including IL‐1 β, IL‐6, TNF‐ α, IL‐10, TGF‐ β1, and RANTES. n = 6. (C‐H) Linear regression plots showing the correlation between the percentages of C1QBP⁺ EVs and cytokine concentration in joint fluid under HA, 20 mL‐C‐EV, and 50 mL‐C‐EV treatment at 4 weeks. Error bars are mean ± SD. Data were analysed by one‐way ANOVA with Bonferroni test for comparison of multiple groups. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Significantly, linear regression analysis revealed a robust negative correlation between C1QBP‐positive EVs and pro‐inflammatory cytokines, including IL‐1β, IL‐6, TNF‐α, and RANTES, alongside a positive correlation with anti‐inflammatory cytokines such as IL‐10 and TGF‐β1 (Figure 4C–H). These results identify C1QBP as a novel predictive marker within EVs for the early detection of TMJ‐OA and for monitoring the effectiveness of therapeutic interventions.

2.5. Therapeutic Efficacy of C1QBP^high C‐EVs in Rat TMJ‐OA

To elucidate the mechanisms by which C‐EVs facilitate condylar repair, our study concentrated on mandibular condylar chondrocytes (MCCs), the principal functional cells of the temporomandibular joint. MCCs isolated from rat models of TMJ‐OA were subjected to treatment with either C‐EVs or OA‐EVs, while untreated MCCs served as the control group. We found that diseased MCCs from rat TMJ‐OA models expressed less CLO2A1 and more ADAMTS5 compared to normal MCCs (Fig. S3A, S3B). RNA sequencing analysis demonstrated distinct transcriptional signatures across the three experimental groups, as evidenced by principal component analysis (PCA) which revealed clear segregation (Figure 5A). Comparative evaluation identified 3,220 differentially expressed genes between MCCs treated with C‐EV and those treated with OA‐EV (fold change >1.0, Q‐value <0.05), comprising 1,377 genes upregulated and 1,843 genes downregulated in the C‐EV‐treated cells (Figure 5B, C; Table S4).

FIGURE 5.

The pleiotropic effects of EV therapy on regulating mandibular condyle chondrocytes (MCCs) from rat TMJ‐OA models. (A) Principal component analysis (PCA) of transcriptome data of MCCs from rat TMJ‐OA under C‐EV and OA‐EV treatment. MCCs with no treatment were used as negative controls. (B) A Venn diagram showed common and differential gene expression of MCCs under EV treatment. (C) The heatmap showed the differential gene expression pattern of MCCs under EV treatment. (D) GO biological process enrichment analysis of differential gene expression of MCCs under EV treatment. (E) KEGG pathway enrichment analysis of differential gene expression of MCCs under EV treatment.

Functional annotation of genes upregulated in C‐EV‐treated MCCs demonstrated significant enrichment in essential biological processes, including chondrocyte proliferation, tissue regeneration, ossification, and anti‐apoptotic mechanisms (Figure 5D). Additionally, KEGG pathway analysis identified the participation of these genes in critical signaling pathways, notably the p53 signaling pathway, cellular senescence, and cell cycle regulation (Figure 5E). In contrast, MCCs treated with OA‐EVs demonstrated an increased expression of inhibitors related to inflammation and cell proliferation, notably involving tumor necrosis factor and interleukin‐1 signaling pathways (Figure 5D). These observations were corroborated by gene set enrichment analysis (GSEA) of proteomic data, which revealed that C‐EV treatment was associated with the upregulation of pathways that promote cellular growth, such as PPAR, MAPK, and mTOR signaling cascades (Fig. S4A–Y).

The local administration of human C‐EVs into the temporomandibular joints of rats exhibited a favorable safety profile, as no significant differences were detected in body weight, hematological indices, or organ histopathology across the treatment groups (Fig. S5A–L). Micro‐computed tomography (micro‐CT) analysis demonstrated that treatment with C‐EVs effectively mitigated condylar bone loss, as indicated by enhanced bone microarchitecture parameters, including increased bone volume fraction (BV/TV) and trabecular thickness (Tb.Th), alongside reduced trabecular separation (Tb.Sp) (Figure 6A; Table S5). These results were corroborated by histological assessments, which revealed that C‐EVs treatment maintained cartilage integrity and exhibited diminished synovial inflammation relative to both HA controls and OA‐EV groups (Figure 6B–F; Tables S6, S7).

FIGURE 6.

Circulating EV Therapy promotes cartilage regeneration and bone repair in rat TMJ‐OA. (A) Representative gross morphology and micro‐CT images of the condyle from rat TMJ‐OA treated with HA, C‐EV, and OA‐EV. Geobel scoring, BV/TV (%), BS/BV (%), Tb.Th (mm) and Tb.Sp (mm) were used to assess bone repair. Scale bar: 1 mm. n = 5. Between groups, each group is compared with the OA group. (B) Representative H&E staining images of the condyle from rat TMJ‐OA treated by HA, C‐EV, and OA‐EV. Scale bar: 250 µm. n = 5. (C) Representative Modified Saffron O‐Fast Green staining images were used to evaluate cartilage injury and regeneration status of the condyle from rat TMJ‐OA treated by HA, C‐EV, and OA‐EV. Scale bar: 250 µm. n = 5. (D) Representative Modified Masson's Trichrome images were used to evaluate the condyle regeneration and bone fiber formation of the condyle from rat TMJ‐OA treated by HA, C‐EV, and OA‐EV. Red staining areas indicate unmineralized bone (osteoid), and blue staining areas indicate mature bone in subchondral bone. Scale bar: 250 µm. n = 5. (E) Representative TRAP staining images were used to evaluate osteoclast activity in the condyle from rat TMJ‐OA treated by HA, C‐EV, and OA‐EV. Scale bar: 250 µm. n = 5. (F) The condyle was statistically analysed by the thickness of fibrocartilage (B), Mankin scoring (C), osteoid areas (D), and TRAP⁺ cells (E). n = 5. (G) Representative Crystal violet staining images showed the cell proliferation of MCCs from rat TMJ‐OA under EV treatment. n = 6. (H) Representative Toluidine blue staining images and quantification analysis showed the chondrogenesis of MCCs under EV treatment. Scale bar: 200 µm. n = 6. (I) Western blotting analysis and semi‐quantification analysis showed the expression of proliferation biomarkers PCNA and SOX9 in MCCs from rat TMJ‐OA under EV treatment. n = 3. (J) Western blotting analysis and semi‐quantification analysis showed the expression of chondrogenic markers COL2A1 and ACAN in MCCs from rat TMJ‐OA under EV treatment. n = 3. (K, L) Representative immunofluorescence images and semi‐quantification analysis of the expression of PCNA (K) and COL2A1 (L) in MCCs from rat TMJ‐OA under EV treatment. Scale bar: 5 µm. n = 6. Error bars are mean ± SD. Data were analysed by one‐way ANOVA with a Bonferroni test for comparison of multiple groups. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

At the cellular level, treatment with C‐EVs significantly enhanced the proliferation of MCCs and promoted their chondrogenic differentiation, as evidenced by the upregulated expression of PCNA, SOX9, COL2A1, and ACAN (Figure 6G–L). These pro‐regenerative effects were in stark contrast to the degenerative phenotype elicited by OA‐EVs, underscoring the pivotal role of EV quality in influencing therapeutic efficacy. Our data indicate that C1QBP^high C‐EVs possess robust chondroprotective properties. The pronounced differential effects observed between C‐EVs and OA‐EVs highlight the therapeutic promise of circulating EVs, while simultaneously emphasizing the necessity for stringent quality control measures in EV‐based interventions. Collectively, these findings provide compelling preclinical evidence supporting the potential application of C‐EVs as a novel biological therapy for TMJ‐OA in clinical settings.

2.6. C1QBP^high C‐EVs Eliminate Senescent Chondrocytes Through p14ARF Mitochondrial Translocation‐Mediated Apoptosis

Based on our discovery of cellular senescence pathways involved in TMJ‐OA development, we examined the senolytic impact of C‐EVs on MCCs. SA‐β‐Gal staining and flow cytometry analysis showed that C‐EVs treatment significantly decreased the number of SA‐β‐Gal‐positive senescent MCCs, whereas OA‐EVs, in contrast, increased the accumulation of senescent cells (Figure 7A,B). Rat C‐EVs and OA‐EVs show comparable clearance efficiencies of senescent MCCs to human C‐EVs and OA‐EVs (Fig. S6A). In addition, neither the C‐EV group nor the OA‐EV group could induce apoptosis in normal MCCs, but OA‐EV increased SA‐β‐Gal⁺ senescent MCC. And C‐EV could increase the cell activity of normal MCCs, while OA‐EV inhibited their activity (Fig. S6B–D). Western blot and immunofluorescence analyses reconfirmed that C‐EVs decreased the levels of the senescence markers P16 and P21, whereas OA‐EVs increased their expression (Figure 7C, S6E, S6F). Additionally, ELISA analysis showed that C‐EVs decreased the levels of SASP factors such as IL‐1β, IL‐6, and TNF‐α in the culture supernatants, while OA‐EVs increased the level of these inflammatory mediators (Figure 7D).

FIGURE 7.

C1QBP^high C‐EVs eliminate senescent MCCs via p14ARF/cleaved caspase 3‐mediated mitochondria‐dependent apoptosis. (A) Representative SA‐β‐Gal staining images and quantification analysis of SA‐β‐Gal⁺ MCCs from rat TMJ‐OA under C‐EV and OA‐EV treatment. MCCs with no treatment were used as negative controls. Scale bar: 50 µm. n = 6. (B) Flow cytometry and quantification analysis of the total cell apoptosis ratio of MCCs from rat TMJ‐OA under EV treatment. n = 4. (C) Western blotting analysis showed the expression of age‐related markers P16 (C) and P21 (D) in MCCs from rat TMJ‐OA under EV treatment. Scale bar: 5 µm. n = 3. (D) ELISA analysis showing the levels of IL‐1β, IL‐6 and TNF‐α in the culture supernatant of MCCs from rat TMJ‐OA under EV treatment. n=6. (E) GO biological process enrichment analysis of 1667 highly expressed proteins in the C‐EV group than in the OA‐EV group. (F) KEGG pathway enrichment analysis of 1667 highly expressed proteins in the C‐EV group than in the OA‐EV group. (G) Representative super‐resolution SIM microscopy images and semi‐quantification analysis showed that MCCs from rat TMJ‐OA engulfed PKH26‐labeled C1QBP^high C‐EV at different time points. Scale bar: 20 µm. n = 6. (H) Representative super‐resolution SIM microscopy images and semi‐quantification analysis showed that time‐course changes of JC‐1‐marked mitochondrial membrane potential (MMP) in MCCs from rat TMJ‐OA under C1QBP^high C‐EV treatment. Red for J‐aggregates indicated viable cells with high potential. Green for J‐monomers indicated early apoptosis and low potential. Scale bar: 5 µm. n = 6. (I, J) Representative super‐resolution SIM microscopy images and semi‐quantification analysis showed the distribution and co‐location of C1QBP, p14ARF, and mitochondria in MCCs from rat TMJ‐OA under C1QBP^high C‐EV treatment. siRNA: p14ARF siRNA treatment. Scale bar: 20 µm. n = 6. (K) Western blotting analysis showed the expression of C1QBP, p14ARF, cyto‐C, and cl‐caspase3 in MCCs from rat TMJ‐OA under C1QBP^high C‐EV treatment. n = 3. (L) Flow cytometry analysis of the total cell apoptosis ratio of MCCs from rat TMJ‐OA under C1QBP^high C‐EV treatment. n = 4. Error bars are mean ± SD. Data were analysed by one‐way ANOVA with a Bonferroni test for comparison of multiple groups. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Proteomic pathway analysis of the 1,667 proteins upregulated in C‐EVs demonstrated a significant enrichment of proteins involved in mitochondrial regulation (Figure 7E,F). Time‐course experiments utilizing PKH26‐labeled C‐EVs revealed a progressive internalization by MCCs over a 24‐hour period (Figure 7G), which corresponded with a peak in mitochondrial membrane potential dissipation, as assessed by JC‐1 staining (Figure 7H). Furthermore, qPCR analyses demonstrated a reduction in C1QBP mRNA expression in TMJ‐OA rat MCCs. Correspondingly, ELISA analyses demonstrated that synovial fluid C1QBP level showed a decreased trend (Figure S6G,S6H). Immunofluorescence analysis revealed that C‐EVs enhanced the colocalisation of C1QBP with p14ARF and facilitated the translocation of p14ARF from the nucleus to the mitochondria (Figure 7I,J). Furthermore, Co‐IP assay confirmed that C1QBP could bind with p14ARF (Fig. S6I). Additionally, qPCR analysis showed that C‐EV treatment significantly upregulated the expression of endogenous C1QBP mRNA in senescent MCCs (Fig. S6J). These alterations correlated with an upregulation of apoptotic markers, specifically cytochrome C and cleaved caspase 3, as demonstrated by western blot analysis and flow cytometry (Figure 7K,L). p14ARF‐siRNA knockdown experiments effectively inhibited both the mitochondrial translocation of p14ARF and the ensuing apoptotic responses (Figure 7I–L).

The findings indicate that C1QBP^high C‐EVs selectively induce apoptosis in senescent chondrocytes via a mitochondria‐dependent pathway, which is triggered by C1QBP‐facilitated translocation of p14ARF. The synchronized timing of C‐EV internalization, p14ARF redistribution, and subsequent caspase activation elucidates the mechanistic underpinnings of the senolytic activity observed, thereby accounting for the specificity of C‐EVs toward senescent cells. This targeted removal of dysfunctional chondrocytes complements the previously demonstrated pro‐regenerative properties of C‐EVs, collectively establishing a dual therapeutic mechanism for the treatment of TMJ‐OA.

2.7. C1q Expression Is Essential for C‐EVs‐Mediated Senescent Chondrocyte Clearance

The interaction between C1q and its receptor C1QBP constitutes a critical mechanism underlying the clearance of apoptotic cells in normal tissues (Kerdidani et al. 2022). To elucidate the specific function of this pathway in the therapeutic effects of C1QBP^high C‐EVs, we conducted a targeted knockdown of C1q in senescent MCCs utilizing siRNA technology. Immunofluorescence analysis demonstrated that untreated senescent MCCs displayed minimal membrane‐associated C1q expression, whereas C‐EVs treatment markedly increased C1q localization on the cell surface (Figure 8A). C1q upregulation was concomitant with the characteristic nuclear‐to‐cytoplasmic translocation of p14ARF, a pivotal regulator of mitochondria‐mediated apoptosis (Figure 8A). Significantly, C1q‐siRNA knockdown treatment resulted in the complete inhibition of both C1QBP expression and p14ARF translocation, thereby highlighting the critical function of membrane‐bound C1q in facilitating the uptake of C‐EVs and subsequent intracellular signaling pathways (Figure 8A).

FIGURE 8.

C1q/C1QBP binding is critical for mitochondria‐dependent apoptosis. (A) Representative super‐resolution SIM microscopy images and semi‐quantification analysis showed the distribution and co‐location of C1q, C1QBP and p14ARF in MCCs from rat TMJ‐OA under C1QBP^high C‐EV treatment. siRNA: C1q siRNA treatment. Scale bar: 20 µm. n = 6. (B) Western blotting analysis showed the expression of C1q, C1QBP, p14ARF, cyto‐C, and cleaved caspase 3 in MCCs from rat TMJ‐OA under C1QBP^high C‐EV treatment. n = 3. (C, D) Representative super‐resolution SIM microscopy images and semi‐quantification analysis showed the distribution and co‐location of p14ARF, cytochrome C and cleaved caspase 3 in MCCs from rat TMJ‐OA under C1QBP^high C‐EV treatment. siRNA: p14ARF siRNA treatment/ C1q siRNA treatment. Scale bar: 20 µm. n = 6. (E) Representative super‐resolution SIM microscopy images showed that the changes of JC‐1‐marked mitochondrial membrane potential (MMP) in MCCs from rat TMJ‐OA under C1QBP^high C‐EV treatment. Red for J‐aggregates indicated viable cells with high potential. Green for J‐monomers indicated early apoptosis and low potential. Scale bar: 5 µm. n = 6. (F) Representative TEM images showed the structure of mitochondria in MCCs from rat TMJ‐OA under C1QBP^high C‐EV treatment. Scale bar: 100 nm. n = 4. (G) The semi‐quantification analysis showed that the changes of JC‐1‐marked mitochondrial membrane potential (MMP) in MCCs from rat TMJ‐OA under C1QBP^high C‐EV treatment. n = 6. Error bars are mean ± SD. Data were analysed by one‐way ANOVA with a Bonferroni test for comparison of multiple groups. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Western blot analysis provided molecular validation of these findings, demonstrating that C‐EVs treatment led to an upregulation in the expression levels of C1q, C1QBP, p14ARF, cytochrome C, and cleaved caspase 3 in senescent MCCs (Figure 8B). In contrast, C1q‐siRNA knockdown treatment counteracted these effects, resulting in a reduction in the expression of all examined apoptotic markers (Figure 8B). Further examination of the apoptotic signaling pathway revealed that genetic suppression of either p14ARF or C1q completely abolished the C‐EV‐induced expression of cytochrome C and cleaved caspase 3 (Figure 8C,D), suggesting that both proteins are indispensable for the initiation of apoptosis. The functional evaluation of mitochondrial integrity using JC‐1 staining revealed that C‐EVs treatment resulted in a marked loss of mitochondrial membrane potential. This effect was entirely abrogated by the knockdown of either p14ARF or C1q (Figure 8E,G). Ultrastructural analysis via transmission electron microscopy corroborated these results, demonstrating that mitochondrial morphology remained intact in the siRNA‐treated groups, in contrast to the typical mitochondrial disruption observed in senescent MCCs exposed to C‐EVs (Figure 8F). In addition, C‐EV had no effect on the mitochondrial structure of normal MCC (Figure S7).

The findings presented herein elucidate a mechanism involving C1QBP^high C‐EVs, which promote the upregulation of membrane‐bound C1q on senescent chondrocytes. This upregulation facilitates the binding and internalization of the C1q–C1QBP complex, subsequently activating p14ARF‐dependent mitochondrial apoptosis characterized by cytochrome C release and caspase 3 activation. The abolition of apoptosis following knockdown of either C1q or p14ARF substantiates the essential role of this pathway. The increased susceptibility of senescent cells to C‐EV‐induced C1q expression likely accounts for the observed therapeutic specificity. Collectively, these results provide insight into a targeted senolytic mechanism relevant to TMJ‐OA, thereby complementing the established pro‐regenerative properties of C‐EVs.

2.8. C1QBP^high C‐EVs Rescue Condyle Bone Loss via Eliminating Senescent Cells

To assess the therapeutic potential of C‐EVs in TMJ‐OA, PKH26‐labeled extracellular vesicles were injected into the joint cavities of rat TMJ‐OA models. Immunofluorescence analysis demonstrated extensive uptake of C‐EVs within both the synovial tissue and condylar cartilage, with a markedly greater accumulation relative to OA‐EVs (Figure 9A,B). In addition, C‐EV accumulated in the subchondral bone compartment (Fig. S8). The efficient delivery was associated with significant enhancements in joint homeostasis. Histological analysis revealed that C‐EVs treatment substantially diminished cellular senescence, as indicated by reduced P16 expression, and mitigated inflammatory responses, evidenced by decreased IL‐1β levels. Simultaneously, C‐EVs promoted tissue regeneration, demonstrated by increased expression of PCNA, a marker of cellular proliferation, and elevated levels of functional matrix components such as COL2 and OCN. Additionally, cartilage degradation was inhibited, as reflected by the downregulation of MMP13 (Figure 9C–H).

FIGURE 9.

C1QBP^high C‐EVs restore rat TMJ‐OA by eliminating senescent cells. (A, B) Representative super‐resolution SIM microscopy images and semi‐quantification analysis showed that PKH26‐labeled EVs were endocytosed with synovial tissue (A) and condyle tissues (B). Scale bar: 5 µm. n = 5. (C–J) Representative immunohistochemical images and semi‐quantification analysis of biomarkers in condyle from rat TMJ‐OA under HA, C‐EV, C1QBP inhibitor M36 (50 µM), and OA‐EV treatment, including age‐related markers P16 and inflammatory marker IL‐1 β (C, D), proliferative marker PCNA (E), chondrogenic marker COL2A1 (F), matrix degradation marker MMP13 (G), osteogenic marker OCN (H). Scale bar: 100 µm. n = 5. Between groups, each group is compared with the OA group. Error bars are mean ± SD. Data were analysed by one‐way ANOVA with a Bonferroni test for comparison of multiple groups. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

To elucidate the mechanistic function of C1QBP, we conducted a co‐administration of the C1QBP inhibitor M36 alongside C‐EVs (Yenugonda et al. 2017). Notably, M36 treatment negated the therapeutic advantages observed, reversing the decreases in cellular senescence and inflammation, and attenuating the regenerative outcomes (Figure 9C–J). These findings substantiate that C1QBP^high C‐EVs mediate their anti‐osteoarthritic effects via C1QBP‐dependent clearance of senescent cells. Significantly, OA‐EVs intensified disease progression by upregulating P16 and IL‐1β expression while downregulating PCNA and anabolic markers. Although HA partially alleviated these effects, its efficacy was markedly inferior to that of C‐EVs (Figure 9C–J). These results demonstrate that C1QBP^high C‐EVs selectively target and eliminate senescent chondrocytes through the C1QBP pathway, thereby mitigating condylar bone loss and facilitating cartilage regeneration. Conversely, OA‐EVs exacerbate TMJ‐OA pathology, highlighting the critical influence of extracellular vesicle origin on therapeutic efficacy.

In summary, this study provides evidence that C‐EVs alleviate TMJ‐OA through the targeted elimination of senescent cells mediated by the C1QBP/C1q/p14ARF signaling pathway, which induces mitochondrial apoptosis (Fig. S9). Furthermore, EVs enriched with C1QBP represent a potential biomarker for assessing disease progression and therapeutic efficacy.

3. Discussion

Temporomandibular joint disorders (TMDs) impact approximately 75% of the adult population, with a higher prevalence observed among adolescent and young female individuals (Alzarea 2015, Schmitter et al. 2010, Wang et al. 2015, Zhao et al. 2011). TMJ‐OA constitutes a notably debilitating condition marked by the gradual deterioration of the joint and persistent orofacial pain (Alzarea 2015, Schmitter et al. 2010, Wang et al. 2015, Zhao et al. 2011). TMJ‐OA is highly prevalent, yet existing therapeutic approaches are primarily palliative, underscoring the critical need for targeted interventions. In this study, we identify two contrasting EV‐mediated mechanisms implicated in TMJ‐OA pathogenesis: (1) OA‐EVs characterized by low expression of C1QBP and associated structural abnormalities actively contribute to disease progression; and (2) circulating EVs with high C1QBP expression facilitate tissue repair by promoting the clearance of senescent chondrocytes and modulating immune responses. Clinical evaluation of 72 TMJ‐OA patients indicates that administration of C‐EVs is safe and effectively restores condylar bone architecture and joint function, concomitant with the replenishment of C1QBP‐deficient EVs in synovial fluid. Collectively, these results position C1QBP^high C‐EVs as a promising novel biological therapy for TMJ‐OA and identify C1QBP as both a therapeutic effector and a predictive biomarker for treatment efficacy.

Although EVs are traditionally regarded as nanocarriers enclosed by phospholipid membranes (Dixson et al. 2023, van Niel et al. 2022), our investigation demonstrates that their structural integrity plays a crucial role in modulating their functional efficacy in the pathogenesis of TMJ‐OA. In contrast to prior research that concentrated exclusively on cargo heterogeneity (Mustonen and Nieminen 2021, Yin et al. 2022, Zhang et al. 2021), our study identified a substantial presence of membrane‐compromised EVs within the synovial fluid of TMJ‐OA patients (Figure 1B–E). This membrane damage is likely to promote the leakage of pathogenic cargo. Histopathological examinations verified that these aberrant EVs facilitate inflammation, tissue degradation, and enzymatic breakdown. These effects may be exacerbated by the atypical membrane thickening observed through TEM, which is likely attributable to the adherence of leaked intracellular contents. Importantly, this pathological EV phenotype corresponds with recent findings concerning synoviocyte‐derived EVs in arthritis (Filali et al. 2022). Extracellular vesicle (EV) exchange constitutes a fundamental mechanism for maintaining tissue homeostasis. Previous studies have demonstrated that the composition and membrane structural integrity of EVs influence cellular uptake efficiency (Mulcahy et al. 2014). Due to their limited bioavailability, OA‐EVs may not undergo metabolic processing within the body and could persist within the joint cavity for extended periods, potentially leading to sustained damage to the joint environment. Importantly, C‐EVs not only reinstated the inherent characteristics of native EVs, such as membrane integrity and biophysical properties, but also ameliorated clinical symptoms. This underscores the preservation of structural integrity as a fundamental requirement for the effective therapeutic application of EVs in the treatment of TMJ‐OA.

C1QBP plays diverse and essential roles in maintaining cellular homeostasis, including the regulation of mitochondrial function, apoptosis, and stress responses (Ghebrehiwet and Peerschke 2004, Ghebrehiwet et al. 2019, McGee et al. 2011). These functions collectively establish C1QBP as a pivotal factor in the pathogenesis of TMJ‐OA. Our results elucidate a previously uncharacterized mechanism in which C1QBP‐enriched C‐EVs are taken up by condylar chondrocytes, facilitating the mitochondrial colocalisation of C1QBP and p14ARF (Figure 7I, J). This interaction subsequently triggers the clearance of senescent cells via p14ARF‐dependent apoptotic pathways. This signaling pathway holds particular significance in TMJ‐OA, given that p14ARF, encoded by the CDKN2A gene and recognized as a canonical marker of cellular senescence, generally acts as an anti‐aging regulator through its interaction with the MDM2/p53 axis (Baker et al. 2016, Carrasco‐Garcia et al. 2017, Kim and Sharpless 2006). The observed dissipation of mitochondrial membrane potential (Figure 7H) substantiates the ability of C1QBP to trigger p14ARF‐dependent apoptotic pathways in senescent chondrocytes, as previously reported (Itahana and Zhang 2008). This finding aligns with the well‐documented functions of C1QBP in regulating energy metabolism and cellular differentiation (Ghosh et al. 2002, Sünderhauf et al. 2021, Thakur and Datta 2008). Importantly, deficiency of C1QBP was found to accelerate the progression of TMJ‐OA, whereas C‐EVs treatment effectively restored joint homeostasis through the supplementation of functional C1QBP within synovial EVs. These findings not only clarify a specific senolytic mechanism but also identify C1QBP as a potential therapeutic agent and predictive biomarker for assessing treatment response in TMJ‐OA.

The interaction between C1q and C1QBP, recognized for its essential function in the clearance of apoptotic cells, facilitates a novel senolytic mechanism in TMJ‐OA. Our study demonstrates that C‐EVs utilize this pathway to selectively target and eliminate senescent chondrocytes through C1q‐dependent activation. Although C1QBP (also known as p32/gC1qR) was originally identified for its immunoregulatory role mediated by complement binding (Ghebrehiwet et al. 2019, Wang et al. 2022), our results highlight its critical involvement in the removal of senescent cells. The CDKN2A gene encodes senescence‐associated proteins p16INK4a and p14ARF, which serve as highly specific markers of cellular senescence (Baker et al. 2016). Notably, co‐localization of C1QBP and p14ARF within mitochondria may induce a loss of mitochondrial membrane potential, subsequently leading to p14ARF‐mediated apoptosis (Itahana and Zhang 2008). In the present study, we showed direct involvement of p14ARF in the elimination of senescent cells and identified a novel C1QBP‐p14ARF signaling axis in C‐EV treatment. These findings may provide a critical theoretical foundation for the advancement of innovative senolytic therapeutic approaches. Specifically, the binding of C1QBP to C1q initiates the translocation of p14ARF from the nucleus to the cytoplasm (Figure 8A), which subsequently leads to mitochondrial membrane destabilization and a reduction in membrane potential (Figure 8E–G). This sequence of events is critical for apoptosis mediated by C‐EVs. Additionally, this interaction promotes the phagocytosis of C‐EVs by senescent cells, thereby establishing a dual mechanism for the clearance of these cells. Collectively, these findings identify the C1q–C1QBP axis as both a potential therapeutic target and a specificity determinant in EV‐based treatments for TMJ‐OA.

Staurosporine‐induced extracellular vesicle production may encounter substantial challenges, such as contamination from residual drug compounds and ethical issues (Hao et al. 2025, Huang et al. 2024, Liu et al. 2018, Ou et al. 2022, Ou et al. 2024, Sui et al. 2024, Wang et al. 2021, Wang et al. 2023, Zhang et al. 2022, Zheng et al. 2021). Hypobaric‐pressure‐induced extracellular vesicle production technique could avoid these issues. Hypobaric‐pressure‐induced peripheral blood mononuclear cells (PBMCs) extracellular vesicles exhibited superior quality and yield. Accordingly, this investigation encompassed the concurrent isolation of circulating extracellular vesicles from both plasma and PBMCs, offering a promising strategy for advancing novel clinical‐grade EV production. The results from the differential protein GSEA enrichment analysis and differential gene GO biological process enrichment analysis suggest that circulating EVs play a significant role in modulating immune microenvironment and promoting proliferation of bone‐related cells as well as tissue regeneration. These findings imply that circulating EVs contain multiple functional molecules that contribute to the therapeutic effect. Consequently, it is necessary to further validate diverse biological functions of circulating EVs in futures study. Compared to traditional senolytic agents, such as the dasatinib and quercetin combination (Xu et al. 2018), extracellular vesicles offer several distinct advantages. This approach achieves greater specificity in the clearance of senescent cells through intrinsic targeting mechanisms, and its biocompatibility may contribute to an enhanced safety profile. Importantly, it facilitates simultaneous execution of dual functions: the removal of senescent cells and delivery of regenerative signals to tissues. Therefore, this strategy signifies a paradigm shift from exclusively anti‐aging interventions toward approaches that prioritize tissue regeneration.

3.1. Limitations of Study

This study presents compelling evidence supporting the therapeutic potential of C1QBP^high C‐EVs in the treatment of TMJ‐OA; however, several limitations warrant consideration. Firstly, although the pivotal role of the C1QBP‐C1q‐p14ARF axis in the clearance of senescent chondrocytes has been established, the comprehensive mechanistic framework underlying this process remains incompletely understood. Secondly, despite the encouraging outcomes observed in our clinical trial, the single‐center design and relatively modest sample size (n = 72) may constrain the generalizability of the findings. Consequently, larger‐scale, multicenter trials with extended follow‐up durations exceeding six months are necessary to substantiate the long‐term efficacy and safety of C‐EV therapy. Furthermore, while C1QBP has been identified as a promising predictive biomarker, its clinical applicability requires validation in independent patient cohorts prior to integration into diagnostic protocols. Lastly, although the study underscores the benefits of EV‐based senolytic approaches, direct comparative analyses with established pharmacological senolytics, such as dasatinib combined with quercetin, were not performed. Conducting such comparative investigations would be instrumental in benchmarking the relative efficacy and safety profiles of C‐EV therapy against current treatment options. Addressing these limitations in future research endeavors will be essential for optimizing C‐EV‐based therapeutic strategies and facilitating their clinical translation in the management of TMJ‐OA.

4. Conclusion

In summary, EVs play a pivotal role as mediators of intercellular communication, acting as biological carriers that preserve tissue homeostasis. Our study reveals that in TMJ‐OA, EVs derived from diseased tissue exhibit structural and functional deficiencies that contribute to disease progression. Conversely, normal C‐EVs demonstrate therapeutic potential by restoring joint homeostasis. Notably, C1QBP^high C‐EVs facilitate the clearance of senescent chondrocytes through C1QBP/C1q/p14ARF pathway mediated apoptosis, while concurrently promoting tissue regeneration. Clinical evidence further supports that C‐EVs treatment safely enhances condylar bone architecture and mitigates symptoms in patients with TMJ‐OA. These findings identify C1QBP^high C‐EVs as a promising bifunctional therapeutic strategy that addresses both cellular senescence and regenerative requirements in TMJ‐OA. Additionally, C1QBP is proposed as a potential predictive biomarker for therapeutic responsiveness. This research lays the groundwork for the development of EV‐based interventions for TMJ‐OA and other degenerative joint diseases, underscoring a novel avenue for precision medicine approaches in osteoarthritis treatment.

5. Materials and Methods

5.1. Antibodies and Reagents

All antibodies and reagents used in this study are listed in Table S8.

Isolation and characterization of plasma‐derived EVs (plasma‐EVs), PBMC‐derived EVs (PBMC‐EV) and joint synovial fluid derived EVs from TMJ‐OA patients (OA‐EVs)

EVs were prepared by ultracentrifugation according to the protocol we reported previously (Huang et al. 2024, Kou et al. 2018, Liu et al. 2018, Ou et al. 2024, Sui et al. 2024, Wang et al. 2023, Zhang et al. 2022). In the clinical department, 20 or 50 mL of autologous peripheral blood was drawn from the patients’ elbow vein, assigned to sterile tubes coated with an anti‐coagulant (acid‐citrate‐dextrose, 3.2% sodium citrate) and then sent to the department of clinical laboratory for further operation. These tubes were centrifuged in 800 g for 10 min, 2000 g for 10 min at room temperature to obtain 2 layers: Plasma in the upper part and blood cells in the bottom part. Plasma was diluted 10‐fold by sterile saline and then ultracentrifuged at 120,000 g for 2 h at 4°C to obtain Plasma‐EVs (Plasma‐EVs). 0.5‐mL sterile saline resuspension of Plasma‐EVs for later use. Peripheral blood mononuclear cells (PBMCs) were isolated from blood cells by Ficoll‐hypaque (polysucrose‐pantethine‐glucosamine) density gradient centrifugation. Briefly, blood cells were diluted twice with sterile saline, slowly added to a centrifuge tube containing sample density separation solution and then placed in a centrifuge at 800 g for 25 min on acceleration of 4 up and 0 down at room temperature. After centrifugation the tubes were seen to be divided into 4 layers: the uppermost layer was yellowish diluted plasma; the second layer was a cloudy white membrane layer of PBMCs; the third layer was a layer of separation solution; and the bottom layer consisted of granulocytes and erythrocytes. The white membrane layer was aspirated to a centrifuge tube filled with saline at 800 g for 10 min, washing away the residual plasma and the separation solution. Cells were resuspended in RPMI Medium 1640 basic. The medium was placed in a self‐developed hypobaric pressure chamber and the parameters were adjusted: 37°C, 20 kPa, 4 h (Meng et al. 2025). After 4‐h induction, PBMC‐EVs (Hypobaric extracellular vesicle‐derived from PBMCs) were isolated from the medium of PBMCs using sequential centrifugation. Briefly, after sequential centrifugation at 800 g for 10 min and at 2000 g for 10 min at 4°C, cell debris was removed. The supernatant was further collected and centrifuged at 16,000 g for 30 min at 4°C to obtain PBMC‐EVs. 0.5 mL sterile saline resuspension of PBMC‐EVs for later use. The mixture of the above two types of vesicles is compound extracellular vesicles (C‐EVs). Synovial fluid samples were collected during arthrocentesis of the TMJs. All samples were centrifuged at 1500 g for 10 min at 4°C to remove cells and debris. The supernatant was further collected and centrifuged at 120,000 g for 2 h at 4°C to obtain OA‐EVs (OA‐EVs).

5.2. Nanoparticle Tracking Analysis (NTA) and Transmission Electron Microscopy (TEM)

The particle size distribution and potential of Plasma‐EVs, C‐EVs, and OA‐EVs were measured by using ZetaView PMX120 (Particle Metrix, Germany) and the data were analysed using the ZetaView analysis software (Version 8.02.31). The particle morphology and membrane integrity of the extracted EVs mentioned above were measured by a Hitachi TEM system (Hitachi, Ltd., Japan), according to the user manual.

5.3. Nano Flow Cytometry (nFCM)

According to the manufacturer's protocol, nanoflow cytometry (NanoFCM, Xiamen, China) was used to analyse surface molecules of EVs. For detecting the surface markers, EVs were stained with primary antibodies, then stained with secondary antibodies. The total number of particles and number of positive particles were calculated by using NanoFCM software (NanoFCM Professional Suite V1.15).

5.4. Western Blotting

Samples were lysed with the RIPA Lysis Buffer System to extract proteins. After quantification of protein concentration with the BCA Protein Assay Kit, an aliquot of protein sample was loaded onto NuPAGE 10% Minutesi Protein Gel (Invitrogen, USA) and transferred to PVDF membranes. The membranes were blocked with 5% BSA for 1 h and then incubated with primary antibodies at 4°C overnight. After washing 3 times with TBST, the membrane was incubated with HRP‐conjugated secondary antibodies for 1 h. Finally, bands were visualized with the SuperSignal West Pico PLUS Chemiluminutesescent Substrate Kit or the SuperSignal West Femto Maximum Sensitivity Substrate kit and then evaluated with the ChemiDoc MP Imaging System (BIO‐RAD, USA).

5.5. Immunofluorescent Staining

EVs were fixed in 4% PFA for 15 min and blocked with 5% BSA with 0.1% Triton X‐100 for 1 h. Then, EVs were incubated with primary antibodies overnight at 4°C. The fluorescent secondary antibodies were used to visualize the corresponding subsets. Further, EVs were sealed with CellMask. EVs were observed under a super‐resolution structured illumination microscopy (SIM).

5.6. Animal Experiments

All animal experimental procedures were performed according to the Institutional Animal Care and Use Committee at Sun Yat‐Sen University under protocol number: SYSU‐IACUC‐2023‐B1044. Eight‐week‐old male and female Sprague Dawley (SD) rats (provided by the Sun Yat‐Sen University Animal Center) underwent sham surgery or TMJ‐OA model were randomly and averagely divided into five groups: (1) Sham group, (2) OA group, (3) HA group, (4) C‐EVs group, and (5) OA‐EVs group. The TMJ‐OA model was constructed by intra‐articular injection of Monosodium iodoacetate (MIA), as previously described (Wang et al. 2012). Briefly, all rats were anesthetized with isoflurane and each of their TMJ received a single intra‐articular injection of 0.5‐mg MIA dissolved in 50‐µL saline using an ultra‐fine insulin syringe. The TMJ was palpated to feel its condylar position and movement, while the mandible was manipulated. The needle insertion point was 5 mm in front of the lower edge of the external auditory canal and 5–10 mm behind the outer canthus of the eye. When the mandibular condyle was identified, the needle was inserted from a posterosuperior direction with an angle of 30°–40° to the sagittal plane under the zygomatic arch. The needle was inserted 3–5 mm, until it came into contact with the TMJ fossa bone wall (Gül Koca et al. 2023). The content of the syringe was injected after negative aspiration. All injections were performed by a single researcher using the same technique for both inducing OA and injecting the experimental agents. After two weeks, 50 µL of 0.9% sterile saline, Hyaluronic acid sodium (HA) (Köhnke et al. 2021), C‐EVs (1×10¹⁰ particles/ml), or OA‐EVs (1×10¹⁰ particles/ml), was weekly injected into the TMJ cavity of rats by using an ultra‐fine insulin syringe, and it lasted for 4 weeks. Inject once every week for a total of four injections. The rats in the Sham group were given only needle pricks. All animals were housed under controlled temperature with a 12‐hour light/12‐hour dark cycle and allowed free movement and access to food and water.

5.7. In Vivo Safety Evaluation

Body weight and feed consumption were recorded weekly. Before euthanasia, blood samples from all rats were collected into an anticoagulant tube via the orbital venous plexus, followed by routine blood test, coagulation, and liver and renal function. Spleens, livers, kidneys, hearts, brains and lungs were analysed by H&E staining.

5.8. . Micro‐Computed Tomography Evaluation

All rats were euthanized and TMJs were harvested for photography. The dissected TMJ condyles in all groups were fixed with 4% neutral PFA for 24 h and analysed by a high resolution micro‐computed tomography (Micro‐CT) system (VENUS, China). Samples were scanned coronally at 90 kV and 65 µA with a 20 µm‐effective pixel size. The sagittal and top images of the condyle were reconstructed. A stack of 1000 coronal slides across the entire condyle were selected for measurements of the percentage of bone volume over total volume (BV/TV, %), bone surface over bone volume (BS/BV, %), trabecular thickness (Tb.Th, mm) and trabecular separation (Tb.Sp, mm), using CTAn version 1.13 (Bruker microCT, Kontich, Belgium).

5.9. Histological Evaluation and Immunohistochemistry

Following micro‐CT detection procedure, the fixed TMJ specimens were decalcified in 10% EDTA (pH 7.4) at room temperature for two months. The decalcified samples were embedded in paraffin and sectioned at 5‐µm thickness. Serial sections were stained with the Hematoxylin‐Eosin Stain Kit (Solarbio, China) for general morphology, the Modified Safranin O‐Fast Green Stain Kit (Solarbio, China) for the proteoglycan deposition in articular cartilage, the Masson's Trichrome Stain Kit (Solarbio, China) for the newly formed osteoid, and the Tartrate‐Resistant Acid Phosphatase Stain Kit (Solarbio, China) for the number of osteoclasts, according to the manufacturer's protocol. For immunohistochemistry analysis, the specimen sections in each group were deparaffinized in xylene, hydrated through an ethanol series, and then repaired using sodium citrate antigen retrieval solution. Subsequently, all the sections were blocked with 5% BSA for 1 h and incubated with primary antibodies overnight at 4°C. Then, each section was incubated with the HRP‐labelled secondary antibodies for 1 h, and DAB was used as HRP‐specific substrate. The images were observed and captured by a light inverted microscopy (Leica Microsystems, Wetzlar, Germany). The positive area (%) of each marker in each group was calculated, and then the change of each marker in each group was normalized against the Sham or OA group.

5.10. Study Design and Subject Selection

In the clinical setting of the Department of Temporomandibular Joint, Affiliated Stomatology Hospital of Guangzhou Medical University, Guangzhou, China, a prospective randomized clinical trial was conducted that involved patients with TMJ‐OA who underwent one of the three different treatment protocols between October 2022 and December 2024. This study has been conducted by the principles of the Helsinki Declaration and approved by the Institutional Review Board (IRB ID: LCYJ2022026) of the local ethics committee and registered on clinicaltrials.gov (registration ID: ChiCTR2200063153). The study inclusion criteria were as follows: (1) between the ages of 18 and 65 years, not limited to gender, Body Mass Index (BMI) between 18.5 and 35 kg/m²; (2) patients diagnosed with TMJ‐OA according to the Diagnostic Criteria for Temporomandibular Disorders (DC/TMD) Axis l algorithms; (3) localized pain of the affected joint in palpation, chewing, and jaw movements; (4) CBCT showed that the affected joint was undergoing degenerative changes; (5) adequate existing clinical and CBCT data at baseline and the post‐operation time interval (follow‐up). Exclusion criteria included any of the following: (1) systemic disease, severe cardiac, pulmonary, hepatic and renal dysfunction; (2) infectious diseases, positive HBs antigen, positive HCV antibody, positive HIV antibody or positive syphilis tests; (3) blood system diseases or abnormalities of the coagulation functions; (4) hypersensitivity diseases; (5) rheumatic or rheumatoid osteoarthritis, ankylosing spondylitis; (6) malignant disease of the head and neck; (7) temporomandibular joint dislocation, trauma, ankylosis, cysts, tumors; (8) apparent psychosomatic disorders; (9) lactating or pregnant women. A total of 72 TMJ‐OA patients enrolled in clinical trial were randomly assigned to 3 treatment groups using a random number table: Group I (HA), Group II (low‐dose extracellular vesicle group, 20‐mL EVs), and Group III (high‐dose extracellular vesicle group, 50‐mL EVs). Group I: These patients received 3 intra‐articular injections of 1 mL of high molecular weight hyaluronic acid (Medical Sodium Hyaluronate Gel, 10 mg/mL, Zhejiang Jingjia Medical Technology Co., Ltd, China) once fortnightly for three consecutive injections following intra‐articular flushing. Group II or III: These patients received three intra‐articular injections of 20 or 50 mL of autologous peripheral blood derived circulating vesicles (C‐EVs) once fortnightly for three consecutive injections following intra‐articular flushing. All participants provided written consent after being informed of the study's aims and design. Clinical data collection and analysis were performed using a double‐blind methodology prior to comprehensive evaluation of efficacy and safety. Patients were assessed through medical history, physical detection, cone‐beam computed tomography (CBCT), and hematology related tests with follow‐up visits at the first, third, and sixth months after treatment.

5.11. Application of Arthrocentesis Injection Technique

The arthrocentesis procedure was carried out using the standard technique described in a previous study by Nitzan et al. (Cascone et al. 1998). The patients or volunteers were placed in a semi‐supine position with the head facing the healthy side and instructed to open their mouths wide. After the skin surface was disinfected with iodophor (Lircon, China), a 5‐mL syringe with 3.5‐mL 0.9% sterile saline was inserted into the skin depression 1‐cm anterior to the tragus in an upward and inward direction, thus entering the supra‐articular cavity. Subsequently, the joint cavity was repeatedly washed out with saline, and then the joint flushing liquid was collected (All samples were centrifuged at 1500 g for 10 min at 4°C, stored in a refrigerator at ‐80°C, and used for subsequent analysis). Retaining the needle in the joint cavity, the syringe barrel was removed and replaced with a 2‐mL syringe containing 1‐mL HA or C‐EVs, which was promptly injected into the joint cavity. Once the injection was completed, the patient's TMJ was repositioned manually.

5.12. Study Outcomes

The main outcome variables included the Fricton Index, numerical rating scale (NRS), Cone‐beam computed tomography (CBCT) findings, and Oral Health Impact Profile‐Temporomandibular Disorders (OHIP‐TMDs). The outcome variables were recorded preoperatively and at the first, third, and sixth months postoperatively. This study was blinded to both the clinical and imaging evaluators, and the evaluators were blinded to the patient grouping and treatment measures.

5.13. Fricton Index

The Fricton Index can be used clinically to evaluate the therapeutic effect of TMD treatment described in a previous study by Fricton et al. (Fricton and Schiffman 1986). Furthermore, it was developed to provide a standardized measure of the severity of problems in mandibular movement, TMJ noise, and muscle and joint tenderness for use in clinical outcome studies. Consequently, the clinical signs and symptoms were scored for every participant based on the Fricton Index, including the Dysfunction Index (DI) and Palpation Index (PI) (Table S9).

5.14. Numerical Rating Scale (NRS)

NRS can be used to subjectively describe a patient's own pain intensity during resting, chewing, and jaw movement (Karcioglu et al. 2018). The patients were instructed to rate their pain levels on NRS, which ranged from a score of 0 –10, with 0 indicating no pain and 10 reflecting the worst possible pain. The pain scores were a gradient scale and interpreted as: 0 = no pain; 1–3 = mild pain; 4–6 = moderate pain; 7–10 = severe pain as described in Table S10.

5.15. Oral Health Impact Profile‐Temporomandibular Disorders (OHIP‐TMDs)

OHIP‐TMDs was a scale developed by British scholar Durham et al. (Durham et al. 2011) specifically for evaluating the quality of life associated with temporomandibular joint disorders. This scale included seven dimensions and a total of 22 items as described in Table S11. The evaluation of each item includes 5 options: 0 = none, 1 = very mild, 2 = occasional, 3 = frequent, 4 = very frequent. The higher the score, the greater the impact of temporomandibular joint disorders on quality of life.

5.16. Cone‐Beam Computed Tomography (CBCT)

CBCT has been recognized as a reliable method for the assessment of the TMJ bony structures. CBCT assessments of the radiographic osseous changes in the TMJs were carried out with a scoring system described by Cheng Li et al. (Li et al. 2015) in 2015. The damaged condyle at baseline included flattening, erosion, osteophytes, sclerosis, and cysts. Remodeling changes to the damaged condyle at follow‐ups included the profile of the condyle, new bone formation, dissolution of osteophytes, and a decrease in sclerosis and cysts. A point rating scale was used to evaluate the osseous changes of the condyle (Table S12). The sum of the scores of the different imaging manifestations was the total score of the radiographic osseous changes in the TMJ on CBCT. All patients’ radiological assessments were scored based on CBCT imaging of the changes found in the TMJ and were recorded by the same researcher (SK) to ensure the reliability of the analysis.

5.17. Hematological Test

All subjects were drawn peripheral blood followed by blood routine test, coagulation function and biochemical test. This serves as an indicator of security. This measurement was carried out preoperatively, during the treatment process, and then postoperatively.

5.18. Cytokine Profiles of the Patient's Temporomandibular Joint Synovial Fluid

Cytokine expression levels in TMJ synovial fluid were detected using Human Cytokine Array C1000 kits (AAH‐CYT‐1000, Raybiotech, USA) following the manufacturer's instructions.

5.19. Proteomic Analysis

Protein lysates of C‐EVs and OA‐EVs were prepared and analysed by LC‐MS/MS. Raw data were analysed using Proteome Discoverer software. Proteins were identified by comparison with the UniProt database by setting the false discovery rate (FDR) of peptides and proteins to 1%. Proteins with significant differential expression (fold change greater than 1.2, p value less than 0.05) were screened for further functional analysis according to the Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases.

5.20. Cell Culture

Induction of the TMJ‐OA model by injecting MIA into the temporomandibular joint of rats. Condylar cartilage of TMJ‐OA rats was dissected and cut into 1 mm³, and then the tissue was digested at 0.2% (w/v) collagenase II for 4 h at 37°C. Cells were passed through 70‐µm cell strainer to obtain single cells in DMEM/F‐12 basic supplemented with 10% FBS and 1% PS. When chondrocytes were 80%‐90% confluent, the cells were digested with Trypsin (Invitrogen, USA) and passaged to passage 2 for subsequent experiments. All cells were cultured at 37°C in a humidified 5 % CO₂ atmosphere.

5.21. Colony Formation Assay

Cells were seeded into a 6‐well plate, and then EVs were added to the culture medium after adhesion. After 5 days, cells were washed with PBS, fixed with 4% PFA and stained with crystal violet staining solution according to the manufacturer's instructions. The number of colonies was calculated using an optical microscopy.

5.22. Toluidine Blue Staining

Cells were seeded into 48‐well plate and were cultured with EVs, fixed with 4% PFA and stained with toluidine blue staining solution according to the manufacturer's instructions. The shades of stained GAG were visualized using an optical microscopy.

5.23. SA‐β‐Galactosidase Staining

The cellular senescence of chondrocytes was assessed according to the manufacturer's protocol for the SA‐β‐galactosidase detection Kit. The percentage of senescent cells was evaluated by the ratio of positive cells to the total number of cells.

5.24. Enzyme‐Linked Immunosorbent Assay (ELISA)

The TMJ synovial fluid and culture supernatants were collected and centrifuged at 1500 g for 10 min at 4°C to remove dead cells and debris. The concentrations of human IL‐1β, TNF‐α, IL‐6, IL‐10, TGF‐β1, RANTES, and C1QBP in TMJ synovial fluid were measured using human ELISA kits, and the concentrations of rat IL‐1β, TNF‐α, and IL‐6 in culture supernatants were measured using rat ELISA kits, according to the manufacturer's protocol.

5.25. Flow Cytometry

Chondrocytes were washed twice with 1× Annexin V binding buffer, and incubated with Annexin V and 7AAD for 15 min at room temperature. The flow cytometry analysis was performed using flow cytometry (ACEA NovoCyteTM).

5.26. Cellular and Organizational Uptake

EVs were resuspended with diluent C solution mixed with an equal amount of diluent C containing PKH26 Dye staining solution. After incubation at room temperature for 5 min, it was added to serum to terminate the staining and then centrifuged at 16,000 g for 30 min. The supernatant is removed, leaving a precipitate of PKH‐26‐labeled EVs, and the EVs were resuspended in 1 mL of PBS for subsequent experiments. Cells were seeded onto confocal dishes and cultured with PKH‐26‐labeled EVs. Then, the cells were fixed with 4 % PFA, the cytoskeleton was stained with Actin and the nuclei were stained with DAPI. Rats were injected with PBS containing PKH‐26‐labeled EVs into the joint cavity, and one week later, TMJ tissues were fixed, and cryosections were made. Frozen sections were hydrated through an ethanol series and then repaired by using sodium citrate antigen retrieval solution. Subsequently, the sections were blocked with 5% BSA for 1 h and incubated with primary antibodies overnight at 4°C. The fluorescent secondary antibodies were used to visualize the corresponding subsets. Further, samples were sealed with a DAPI sealing agent. Fluorescence imaging was acquired via confocal fluorescence microscopy (Axio Observer Z1M; Zeiss, Germany).

5.27. Immunofluorescent Staining

Cells were fixed in 4% PFA for 20 min and blocked with 5% BSA with 0.1% Triton X‐100 for 1 h. Then, cells were incubated with primary antibodies overnight at 4°C. The fluorescent secondary antibodies were used to visualize the corresponding subsets. Further, cells were sealed with a DAPI sealing agent; cells were observed under a confocal fluorescence microscope (Axio Observer Z1M; Zeiss, Germany).

5.28. Mitochondrial Membrane Potentials Assay

Detection of early apoptosis in chondrocytes using mitochondrial membrane potential assay kit with JC‐1 according to the manufacturer's instructions.

5.29. Transcriptome Sequencing of Chondrocytes

Chondrocytes were stimulated with EVs for 24 h to collect RNA for gene transcriptome microarray analysis. The protocol was modified from the Agilent monochrome microarray gene expression analysis protocol (Agilent Technologies). To reduce dimensionality, uniform cluster approximation and cast (UMAP) analysis was performed using Sangerbox (https://vip.sangerbox.com/). Differentially expressed genes were screened for further analysis, and Venn diagrams were constructed based on the biological functions of these factors. KEGG and GO analyses were performed using the online software program DAVID. OmicStudio tools (https://www.omicstudio.cn/tool) were used for data visualization.

5.30. siRNA Knockdown

Cells were transfected with siRNA‐C1q (si‐C1q), siRNA‐p14ARF (si‐p14ARF) (RiboBio, China) using a transfection kit (RiboBio, China) according to the manufacturer's instructions (Table S13). Non‐targeting control siRNAs (si‐control) (RiboBio, China) were used as a negative control. Transfection efficiency was detected by quantitative real time polymerase chain reaction (qRT‐PCR) and western blotting at 48‐h post‐transfection.

5.31. Statistical Analysis

All experiments were performed in biological triplicate, and data were expressed as mean ± standard deviation (SD). Statistical and graph analysis were performed by GraphPad Prism 7 (GraphPad Software, USA). Multiple group comparisons were assessed by one‐way ANOVA analysis with a Bonferroni test. For two‐group comparisons, significance was analysed using independent unpaired two‐tailed Student's t tests. Values of p < 0.05 were considered statistically significant.

Author Contributions

Bowen Meng: conceptualisation, methodology, data curation, investigation, validation, supervision, project administration, writing – original draft, writing – review and editing. Xin Li: methodology, data curation, investigation, validation, formal analysis, project administration, writing – original draft. Benyi Yang: methodology, software, data curation, investigation, validation, project administration, writing – original draft, writing – review and editing. Yan Qu: conceptualisation, methodology, data curation, validation, funding acquisition, project administration, writing – original draft, writing – review and editing. Yifan He: project administration, data curation. Chaoran Fu: data curation, project administration. Zhe An: data curation, project administration. Antong Wu: data curation, project administration. Yuzhuo Hei: data curation, project administration. Rong Zhang: data curation, project administration. Wenyi Cai: data curation, project administration. Lingyunbo Kong: data curation, project administration. Rui Li: data curation, project administration. Meng Hao: data curation, project administration. Zeyuan Cao: data curation, project administration. Xueli Mao: data curation, project administration. Janak Lal Pathak: writing – original draft, methodology. Yang Cao: conceptualisation, investigation, supervision, funding acquisition, resources, writing – original draft, writing – review and editing. Songtao Shi: conceptualisation, methodology, investigation, supervision, funding acquisition, resources, writing – original draft, writing – review and editing. Qingbin Zhang: conceptualisation, methodology, investigation, supervision, funding acquisition, project administration, resources, writing – original draft, writing – review and editing.

Conflicts of Interest

The authors declare that they have no competing interests.

Clinical Trial Registration

ClinicalTrials.gov Identifier (ChiCTR2200063153).

Supporting information

Supplementary Table: jev270224‐sup‐0001‐TableS1.xlsx

Supplementary Table: jev270224‐sup‐0002‐TableS2.xlsx

Supplementary Table: jev270224‐sup‐0003‐TableS3.xlsx

Supplementary Table: jev270224‐sup‐0004‐TableS4.xlsx

Supplementary Table: jev270224‐sup‐0005‐TableS5.docx

Supplementary Table: jev270224‐sup‐0006‐TableS6.docx

Supplementary Table: jev270224‐sup‐0007‐TableS7.docx

Supplementary Table: jev270224‐sup‐0008‐TableS8.docx

Supplementary Table: jev270224‐sup‐0009‐TableS9.docx

Supplementary Table: jev270224‐sup‐00010‐TableS10.docx

Supplementary Table: jev270224‐sup‐00011‐TableS11.docx

Supplementary Table: jev270224‐sup‐00012‐TableS12.docx

Supplementary Table: jev270224‐sup‐00013‐TableS13.docx

Supplementary Figures: jev270224‐sup‐00014‐Figures.docx

Supplementary Materials: jev270224‐sup‐00015‐SuppMat.pdf

Meng, B. , Li X., Yang B., et al. 2026. “Circulating Metabolites Treat Human TMJ‐OA by Eliminating Senescent Chondrocytes via the C1QBP/C1q/p14ARF Axis.” Journal of Extracellular Vesicles 15, no. 2: e70224. 10.1002/jev2.70224

Data Availability Statement

All data are available in the main text or the supplementary materials.