Immune cells with senescence-related transcriptional signatures orchestrate the inflammatory continuum in osteoarthritis synovium: a single-cell and machine learning study.

BACKGROUND

Synovitis is a key driver of osteoarthritis (OA), characterized by a chronic inflammatory microenvironment. However, the contribution of immune-cell senescence to synovial pathology and its therapeutic potential remains poorly understood. This study aims to map the senescent immune landscape and identify actionable targets for OA therapy.

METHODS

Single-cell RNA dataset (GSE152805) of OA synovium was used to identify immune subsets with senescence-related transcriptional signatures and subset-specific gene signatures. These gene signatures served as machine-learning features to train and validate diagnostic models in a combine dataset of 4 OA transcriptomic datasets (GSE55235, GSE55457, GSE82107, GSE169077) using Random Forest, Support Vector Machine, Gradient Boosting Machine, and eXtreme Gradient Boosting. Immunofluorescence staining and Western blotting of human synovial tissues validated macrophage-associated markers. Pseudotime trajectory and cell-cell communication analyses defined temporal dynamics and intercellular communication. Drug-gene interaction analysis and molecular docking were performed to assess therapeutic potential.

RESULTS

Macrophages, mast cells, and dendritic cells were identified as the immune populations with senescence-related transcriptional signatures. Pseudotime and communication analyses revealed a coordinated inflammatory continuum: macrophages initiated and sustained synovial inflammation, mast cells amplified intermediate inflammatory responses, and dendritic cells contributed to late-stage matrix remodeling. Machine-learning models identified hub biomarkers, with RHOB and PDK4 emerging as the macrophage-associated biomarkers with therapeutic potential. RHOB and PDK4 were markedly upregulated and colocalized with CD68+ macrophages in OA synovium. Drug prediction suggested CHEMBL1797159 as a potential ligand for RHOB and dichloroacetate as a repurposable modulator of PDK4.

CONCLUSION

This study characterizes the transcriptomic landscape of senescence-related features in the OA synovial microenvironment. We identify RHOB and PDK4 as key macrophage-associated therapeutic targets, providing novel insights for immunomodulatory strategies in OA treatment.