Single-cell transcriptomics and machine learning characterize LTA4H-linked lipid metabolism and inflammatory...

Knee osteoarthritis (KOA) is a whole-joint disease characterized by inflammation, immune dysregulation, and metabolic imbalance. To define lipid metabolism-related transcriptional heterogeneity within the infrapatellar fat pad (IPFP)/synovium compartment, we integrated publicly available single-cell transcriptomic data from IPFP/synovium-derived joint cells with bulk synovial transcriptomic datasets from patients with KOA and controls. Lipid metabolism-related scores were derived using multiple single-cell scoring algorithms, and candidate genes were identified through score-associated analyses, differential expression, protein-protein interaction analysis, and five machine-learning approaches, followed by external bulk validation and single-cell characterization. Myeloid cells showed the most consistent KOA-associated increases in lipid metabolism-related transcriptional scores. Fifty composite-score-associated genes were identified and were enriched in lipid metabolic, lysosomal, and PPAR-related pathways. Cross-screening prioritized PLIN2, LTA4H, and PPT1, with LTA4H showing the strongest association with lipid metabolism-related transcriptional scores. LTA4H was predominantly enriched in myeloid cells, particularly monocytes and inflammatory macrophages, and was increased in KOA myeloid cells and high-score cells. Cell-cell communication, myeloid subclustering, and pseudotime analyses linked LTA4H-positive myeloid cells to inflammatory signaling and the transition from monocytes toward inflammatory macrophages. In silico knockout analysis further implicated LTA4H in regulatory programs related to leukocyte migration, chemotaxis, phagocytosis, and cytokine receptor binding. In a rat KOA model, LTA4H protein expression was increased in knee synovial tissues and spatially colocalized with CD68-positive macrophages. These findings define a myeloid-cell-centered context for LTA4H in KOA and link its expression to lipid metabolism-related transcriptional features and inflammatory myeloid-cell states.