Abstract
Background:
Osteoarthritis (OA) pathogenesis involves age-related immune dysregulation, yet non-invasive diagnostic tools and mechanistic insights remain limited.
Methods:
We integrated transcriptomic profiling of four OA-affected joint tissues with machine learning to identify potential peripheral blood biomarkers. Weighted gene co-expression network analysis was employed to explore gene modules and pathways associated with OA. Single-cell RNA sequencing was performed on 217 983 joint cells to delineate B cell differentiation trajectories. Flow cytometry was used to validate age-associated B cell imbalances in peripheral blood.
Results:
We identified five peripheral blood biomarkers: MAPK1, MAP3K8, ING1, LDLR, and NUP153 in distinguishing OA patients from controls (the area under the curve [AUC] = 0.966). Importantly, these markers exhibited age-specific expression profiles; ING1, NUP153, and MAP3K8 were elevated, while MAPK1 was reduced in elderly compared to younger OA patients. A refined predictive model based on these age-specific markers demonstrated superior performance specifically for elderly Knee OA (KOA, AUC: 0.8 vs. 0.7 for younger KOA). These biomarkers correlated with immune cell infiltration and inflammatory cytokines. In osteoarthritic joint tissues, B cells predominantly originated from subchondral bone and synovium. Single-cell analysis identified age-specific B cell differentiation patterns, with elderly KOA patients enriched in an activated B cell cluster (C1). Furthermore, B cells from elderly KOA patients showed altered energy metabolism and increased proportions in peripheral blood, and functionally promoted chondrocyte damage.
Conclusion:
Our findings establish a novel blood-based diagnostic framework for OA and uncover aging-driven B cell remodeling as a key contributor to elderly OA pathogenesis. These findings offer non-invasive diagnostics and immunomodulatory targets for age-specific OA therapy.