Background: Our recent study identified chondroitin sulfate proteoglycan (CSPG4) as a key surface marker of extracellular vesicles (EVs) in synovial fluid (SF) of patients with knee osteoarthritis (OA) [1]. CSPG4 was strongly positively correlated with approximately 60% of the total SF EV peptides, as well as over 75% of those classified as pathogenic due to their positive Spearman correlation with knee radiographic (r)OA severity [1]. Peptides of fibronectin (FN1), a known driver of OA, comprised nearly 20% of the pathogenic SF EV peptides [1]. Additionally, interleukin-11 (IL-11), a known risk factor in OA pathogenesis and progression [2], induces CSPG4 expression in cancer cells [3].

Objectives: This study aimed to detect the pathogenic cargo of CSPG4 ⁺ EVs from OA joint fluid and determine if CSPG4 ⁺ plasma EVs are new biomarkers associated with knee rOA severity and predictors of knee rOA progression using independent study cohorts.

Methods: Knee rOA severity was assessed using Kellgren-Lawrence (K/L), joint space narrowing (JSN), and osteophyte (OST) scores. Symptomatic OA severity was evaluated using a Knee Injury and Osteoarthritis Outcome Score (KOOS). Knee rOA progression was defined by any increase in K/L, JSN, or OST scores from baseline to follow-up. For EV analyses, in the discovery cohorts (Duke), plasma (n=81) and SF (n=25) samples were collected from 106 knee rOA patients (defined as K/L ≥ 1); 30 patients had follow-up data (1-9 years). For the validation cohorts (Harvard), plasma samples were collected from 92 symptomatic/radiographic knee OA patients (defined as modified KOOS ≥ 2/12 and JSN or OST ≥ 1); 62 patients had follow-up data (2 years). EVs were isolated and characterized by size, concentration, surface markers using flow cytometry, and morphology using transmission electron microscopy. Flow cytometry quantified SF EVs co-expressing CSPG4 with FN1 or IL-11, and plasma EVs expressing CSPG4 and tetraspanins (CD9, CD81, CD63). Multivariable linear regression models, adjusted for age, BMI, and sex, assessed associations of CSPG4 ⁺ plasma EVs with knee rOA severity and progression in discovery and validation cohorts. The ability of CSPG4 ⁺ plasma EVs to predict knee rOA progression was evaluated using AUC and 95% bootstrap bias-corrected confidence intervals (CIs). Gene co-expression of CSPG4 with FN1 and IL-11 in OA joint tissue cells was analyzed using published single-cell sequencing data (scRNA-seq, GSE152805).

Results: Analysis of scRNA-seq data from knee OA joint tissue cells revealed that CSPG4 ⁺ cells, particularly CSPG4 ⁺ chondrocytes from damaged cartilage, co-expressed the IL11 gene. CSPG4 ⁺ chondrocytes and synoviocytes also co-expressed genes encoding IL-11 receptor subunits (IL11RA and IL6ST), as well as FN1 and its receptor subunits—ITGB1 and ITGA5. Consistent with these findings, the proportions of IL-11 ⁺ and FN1 ⁺ EV subpopulations were dramatically higher (both p<0.0001) in CSPG4 ⁺ compared to CSPG4 ^- EVs in OA SF (n=25). In OA plasma, a “minimally invasive” biospecimen, the discovery cohort (n=51) frequency of CSPG4 ⁺ large EVs (LEVs) was positively associated with knee rOA severity; these findings were confirmed in the independent validation cohort (n=92) in which the frequency of CSPG4 ⁺ total EVs, LEVs, and medium-sized EVs (MEVs) was significantly positively associated with knee rOA severity. The baseline frequency of individual CSPG4 ⁺ plasma EV subpopulation (total EVs, LEVs, MEVs, and small EVs [SEVs]) predicted knee rOA progression in the discovery cohort (n=30) with AUCs (95% CI) of 0.6065 (0.4150, 0.8065), 0.5556 (0.4083, 0.6944), 0.6181 (0.4107, 0.8306), 0.5926 (0.4220, 0.7804), respectively; and in the validation cohort (n=62) 0.6266 (0.5005, 0.7643), 0.5224 (0.4667, 0.5831), 0.6870 (0.5318, 0.8204), 0.6800 (0.5440, 0.8087), respectively. Despite modest AUCs for some predictors, the predictive pattern (highest AUC from CSPG4 ⁺ MEVs) was consistent in both cohorts. Notably, the combination of multiple individual predictors, including the frequencies of CSPG4 ⁺ subpopulations in LEVs and MEVs, as well as tetraspanin ⁺ subpopulations in LEVs, MEVs, and SEVs, achieved an AUC of 0.8843 (95% CI: 0.6797, 1.0) in the discovery cohort and 0.7618 (95% CI: 0.5884, 0.8378) in the validation cohort.

Conclusion: CSPG4 ⁺ cells isolated from OA joint tissues co-express the FN1 and IL11 genes. Similarly, CSPG4 ⁺ SF EVs from knee OA patients carry FN1 and IL-11 proteins as part of their cargo. In OA plasma, the frequency of CSPG4 ⁺ EV subpopulations represents a new biomarker of knee rOA severity and progression, although associations between some predictors and rOA progression were not strong. These findings provide a basis for future studies examining the mechanistic roles of CSPG4 ⁺ EVs in the pathogenesis of OA. Developing strategies to suppress or clear CSPG4 ⁺ EVs and their pathogenic cargo may have therapeutic potential for OA.

REFERENCES: [1] Zhang, X., S. Ma, S. I. Naz, V. Jain, E. J. Soderblom, C. Aliferis and V. B. Kraus (2023). “Comprehensive characterization of pathogenic synovial fluid extracellular vesicles from knee osteoarthritis.” Clin Immunol 257 : 109812.

[2] Chou, C. H., M. T. Lee, I. W. Song, L. S. Lu, H. C. Shen, C. H. Lee, J. Y. Wu, Y. T. Chen, V. B. Kraus and C. C. Wu (2015). “Insights into osteoarthritis progression revealed by analyses of both knee tibiofemoral compartments.” Osteoarthritis Cartilage 23 (4): 571-580.

[3] Winship, A., M. Van Sinderen, A. Heffernan-Marks and E. Dimitriadis (2017). “Chondroitin sulfate proteoglycan protein is stimulated by interleukin 11 and promotes endometrial epithelial cancer cell proliferation and migration.” Int J Oncol 50 (3): 798-804.

Acknowledgements: The authors wish to acknowledge funding support from National Institute on Aging grants R01AG070146 (VBK and XZ); all participants who donated specimens for this study; the Duke Cancer Institute Flow Cytometry Shared Resource for providing access to the flow cytometers; and the Duke Human Vaccine Institute Research Flow Cytometry Shared Resource Facility for providing the FCS Express 5 software.

Disclosure of Interests: Xin Zhang: None declared, Lindsey Adair MacFarlane: None declared, Janet L Huebner: None declared, Jason Jinkun Tao: None declared, Elena Losina: None declared, Jeffrey Katz Dr. Jeffrey Katz is PI of the the Osteoarthritis Registry of Biomarker and Imaging Trajectories (ORBIT) study which was funded by Biosplice. Funding ended in 2023, Virginia B Kraus: None declared.

© The Authors 2025. This abstract is an open access article published in Annals of Rheumatic Diseases under the CC BY-NC-ND license ( http://creativecommons.org/licenses/by-nc-nd/4.0/ ). Neither EULAR nor the publisher make any representation as to the accuracy of the content. The authors are solely responsible for the content in their abstract including accuracy of the facts, statements, results, conclusion, citing resources etc.