Unraveling the oxidative stress landscape in diabetic foot ulcers: insights from bulk RNA and single-cell RNA sequencing data.

Abstract

Background: Oxidative stress plays a crucial role in the development of diabetic foot ulcers (DFU). However, its underlying mechanisms are not fully understood. The purpose of this study was to use bioinformatics and preliminary validation methods to preliminarily reveal the oxidative stress landscape in DFU.

Methods: Based on the single-cell and bulk RNA sequencing data of DFU, we conducted differential genes screening, machine learning, PPI network construction, immune infiltration analysis, drug prediction, TF-mRNA-miRNA network, cell-cell interaction, pseudotime trajectory analysis, external cohort validation, and in vitro experiments to develop the oxidative stress landscape in DFU.

Results: Bulk RNA-seq analysis identified 63 oxidative stress-related genes of DFU (DORGs), and the top 59 genes were screened out for key nodes with close functional associations. Functional enrichment analysis showed significant involvement in oxidative stress response. Drug prediction highlighted Thymoquinone and Erlotinib as potential therapeutic candidates. Machine learning algorithms (SVM-RFE, LASSO and RF) identified BCL2 and FOXP2 as candidate hub DORGs for DFU diagnosis. Immune cell infiltration analysis indicated a significant presence of naive B cells and CD8 T cells in DFU. The analysis of single-cell RNA sequencing identified a total of 31,787 cells across 10 distinct clusters, with a notably lower proportion of fibroblasts in DFU group than that in the control group. The expression patterns of BCL2 and FOXP2 across the different groups were consistent with findings from bulk RNA sequencing analysis. Notably, fibroblasts derived from DFU patients exhibited the highest oxidative stress scores. Intercellular signaling analysis indicated that fibroblasts serve as crucial communication cells, primarily engaged in COLLAGEN signaling network. Additionally, fibroblasts are categorized into five distinct clusters. Among these, COL6A5+ fibroblasts constitute the predominant cluster in DFU and exhibit low differentiation potential. Furthermore, in vitro experiments successfully established a DFU oxidative stress model of fibroblasts, revealing reduced migration ability in the absence of cell death. Both in vitro findings and external data corroborated the decreased expression levels of BCL2andFOXP2in DFU.

Conclusion: The oxidative stress-related genes BCL2 and FOXP2 could serve as diagnostic markers for DFU. Furthermore, we identified the novel pathogenic mechanism associated with oxidative stress in DFU fibroblasts. This study may offer new insights for the diagnosis and treatment of DFU.