Potential mechanisms of acupuncture treatment for rheumatoid arthritis: a study based on network topology and machine learning.

Abstract

Background: Rheumatoid arthritis (RA) is a systemic autoimmune disease that requires multitarget therapeutic strategies. Acupuncture, an integrative therapy of traditional Chinese medicine (TCM), has shown efficacy in the clinical treatment of RA, but its molecular mechanisms remain unclear.

Purpose: This study systematically elucidated the holistic regulatory effects of acupuncture on RA by integrating network topology with machine learning approaches.

Methods: Data on the interactions between acupuncture-affected endogenous compounds and RA-related targets were extracted from databases, and a multidimensional interaction network was constructed to map the interactions between acupuncture and RA. screened RA-related differentially expressed genes (DEGs) from the GEOdatabase that intersected with acupuncture-responsive genes. The clusterProfiler was used for KEGG/GO enrichment analysis of these DEGs, and the immune microenvironment was analyzed via the CIBERSORTx and xCell algorithms. ConsensusClusterPlus (R package) was used for unsupervised clustering to obtain DEGs. Subsequently, key genes were identified via an ensemble machine learning model (GLM/SVM/XGB/RF), and nomograms were created. Two-sample MR and colocalization analyses were applied to validate the causal relationship between core acupuncture-affected DEGs and RA risk.

Results: This study identified 10 acupuncture-regulated endogenous compounds and 49 RA-related DEGs. KEGG analysis revealed that the DEGs enriched in immune pathways included the JAK/STAT pathway, which mediates inflammatory responses, the T-cell receptor signaling pathway, which is involved in T-cell differentiation, and the TNF signaling pathway. Immunome profiling via the CIBERSORT algorithm revealed that the DEGs were enriched primarily in key immune cell subpopulations, such as M1 macrophages, activated CD4⁺ T cells, Tregs, and B lymphocytes. Machine learning identified five key genes associated with immune infiltration (STAT1, GAPDH, JAK2, PTGS2, and MDM2). MR/colocalization confirmed that acupuncture-regulated STAT1 expression was positively correlated with RA genetic susceptibility, highlighting that the STAT1-mediated JAK/STAT pathway is involved in immune remodeling.

Conclusion: STAT1, GAPDH, JAK2, PTGS2, and MDM2 may be potential targets for the acupuncture treatment of RA. Acupuncture may achieve systemic immune regulation by synergistically targeting multiple pathways (JAK/STAT, TNF) and immune cells (M1 macrophages, CD4⁺ T cells). This initiative integrates the holistic philosophy of TCM with the precision of AI-driven medical science.