Implicating Anoikis and CXCR4 in Osteoarthritis: Predictive Modeling and Therapeutic Insights

Abstract

Abstract Background Osteoarthritis (OA) is a globally prevalent joint disorder characterized by complex pathophysiology and diagnostic limitations. The present study aimed to elucidate the potential roles of anoikis-related genes in OA and develop a reliable diagnostic model. Additionally, we investigated the role of hub gene CXCR4 in immune response modulation within OA and explored potential therapeutic agents. Methods Gene expression datasets GSE207881 and GSE168505, were downloaded from Gene Expression Omnibus database and analyzed for potential disparities in immune cell infiltration between OA and control samples. Results We identified 94 DEGs involved in OA pathogenesis and pinpointed 6 hub genes (CXCR4, PECAM1, CCL5, VWF, MCAM, and CXCL12) with potential roles in OA. The machine learning model, which developed nine models and used SHAP method for interpreting the models, showed high predictive accuracy for OA risk, with CCL5, CXCL12, and CXCR4 emerging as key predictive factors. Notably, CXCR4, an immune system modulator, exhibited potential interactions with immune responses in OA. Furthermore, potential therapeutic compounds were examined using Connectivity Map (CMap) analysis and molecular docking. CMap analysis proposed potential therapeutic compounds, which require further validation. Conclusions Our study unveils the association between anoikis-related genes and OA, highlighting the utility of machine learning models for OA risk prediction. It also underscores the significance of CXCR4 in immune response modulation and potential therapeutic strategy in OA. These findings could spur the development of novel diagnostic tools and personalized therapies for OA, although further validations are necessary.