Integrating bulk and single-cell RNA sequencing identifies and validates lipid metabolism-associated genes in diabetic foot ulcer

Abstract Diabetic foot ulcer (DFU) is characterized by persistent inflammation and impaired wound healing. Although macrophages are central regulators of this process, the relationship between inflammatory activation and lipid-associated transcriptional programs in DFU macrophages remains incompletely understood. We integrated bulk and single-cell transcriptomic analyses to identify macrophage-associated biomarkers and characterize DFU-associated inflammatory and lipid-associated gene-signature activity. Public bulk and single-cell RNA sequencing datasets from GEO were analyzed. Single-cell data were processed using Seurat, and lipid-associated gene (LAG) signature activity was quantified at single-cell resolution using AUCell. Candidate biomarkers were identified using LASSO, SVM-RFE, and Boruta, followed by model construction and external validation. Functional analyses included GO/KEGG enrichment, GSVA, GSEA, immune infiltration analysis, in silico perturbation analysis, CellChat-based communication inference, DSigDB-based compound prioritization, and molecular docking. Expression patterns of key biomarkers were further assessed by qRT-PCR and western blotting in DFU and diabetic foot skin (DFS) tissues. Single-cell analysis identified macrophages as a major cell population with elevated LAG signature activity in DFU. Pro-inflammatory macrophages were selectively expanded and showed the highest LAG activity among macrophage subsets. Integrated single-cell and bulk transcriptomic analyses identified three candidate biomarkers, JUNB , LDLR , and TREM1 , which showed good diagnostic performance and were associated with inflammatory and lipid-metabolism-related transcriptional programs. Pseudotime, in silico perturbation, and pathway analyses further linked these biomarkers to Pro-inflammatory macrophage activation, lipid-associated gene-signature activity, and inferred endothelial communication. CellChat analysis suggested enhanced inferred communication between Pro-inflammatory macrophage_high cells and endothelial cells, mainly involving the CXCL signaling pathway. DSigDB screening prioritized ciclopirox as a candidate compound, and molecular docking predicted potential binding to all three corresponding protein targets. qRT-PCR and western blotting supported increased expression of these biomarkers in DFU tissues. This study identifies a DFU-associated macrophage program centered on Pro-inflammatory macrophages and characterized by inflammatory activation, lipid-associated gene-signature activity, and inferred endothelial communication. These findings provide a hypothesis-generating framework for understanding macrophage heterogeneity in DFU and highlight candidate biomarkers and computationally prioritized compounds for future functional and translational studies.