Abstract
BACKGROUND: Diabetic foot ulcers (DFUs) involve chronic inflammation and impaired healing. Emerging evidence implicates copper-dependent regulated cell death (cuproptosis) in immune microenvironment regulation, but its role in DFU fibroblast-immune crosstalk remains unknown. METHODS: We analyzed scRNA-seq datasets (GSE231643, GSE134431) from DFU/healthy skin. After quality control (25,198 cells retained) and clustering, fibroblasts with elevated cuproptosis scores were identified via ssGSEA. Fibroblast-specific DEGs were screened (Wilcoxon rank-sum; |log(2)FC|>1, p<0.05) and functionally annotated (KEGG/GO). Machine learning (LASSO, SVM-RFE, Random Forest) pinpointed diagnostic biomarkers, validated by ROC analysis. Immune infiltration (CIBERSORT) and cellular communication (CellChat) were assessed in high/low biomarker groups. In vivo validation used nine SD rats (normal/wound/diabetic wound groups; n=3); GK mRNA/protein levels were measured via PCR/Western blot. RESULTS: Through scRNA-seq analysis of 33,095 cells (25,198 post-quality control), fibroblasts exhibited the highest cuproptosis-related pathway scores. A total of 543 fibroblast-specific DEGs were identified, significantly enriched in pathways such as PI3K-Akt, chemokine signaling, NF-κB, IL-17 signaling, and AGE-RAGE signaling. GO analysis revealed enrichment in processes associated with extracellular matrix organization, inflammation regulation, and immune signaling. Integrated analyses involving bulk transcriptomics and machine learning identified GK as a robust cuproptosis-associated diagnostic biomarker (AUC=0.929, 95% CI: 0.714-1.000). GK expression positively correlated with neutrophil infiltration (R=0.95, p=0.00081) and negatively correlated with CD8(+) T cells (R=-0.77, p=0.044) and follicular helper T cells (R=-0.89, p=0.0073). CellChat analysis demonstrated that high GK expression significantly enhanced fibroblast-mediated communication with immune cells via critical signaling pathways (MK, HGF, IGFBP, KIT), reshaping the immune-regulatory network within DFU tissues. Animal experiments further validated these findings: Quantitative PCR demonstrated that GK mRNA levels were significantly elevated in both wound model and diabetic wound model groups compared with the normal group, with the diabetic wound group showing significantly higher expression than the wound model. Consistent with mRNA findings, Western blot analysis revealed a progressive increase in GK protein expression across the normal, wound model, and diabetic wound model groups. CONCLUSIONS: Our data indicate that the cuproptosis-related gene GK is associated with alterations in DFU immune communication through fibroblast-centric networks, as reflected by correlations with immune infiltration and predicted ligand-receptor crosstalk. GK therefore represents a promising candidate diagnostic biomarker and a potential therapeutic target for DFUs that warrants further functional validation.