Biomarkers for diabetic foot ulcers: single-cell transcriptomics bioinformatics analysis and experimental validation

Abstract

BACKGROUND: Factors such as infection, limb ischemia, and histiocyte activation are involved in diabetic foot ulcers, but the key cell subpopulations influencing diabetic foot ulcer healing remain unclear, and specific biomarkers for diabetic foot ulcers have yet to be identified. Gene Expression Omnibus (GEO) is a publicly accessible database managed by the National Center for Biotechnology Information that stores high-throughput gene expression data, allowing users to freely submit, share, query, and analyze data. Secondary analysis of published data can save research costs and uncover new research targets and ideas. OBJECTIVE: To screen biomarkers for diabetic foot ulcers using single-cell transcriptome and conventional transcriptome bioinformatics analysis, high-dimensional weighted gene co-expression network analysis (hdWGCNA), and weighted gene co-expression network analysis (WGCNA). METHODS: The single-cell transcriptome dataset GSE165816, containing non-healing ulcer tissue samples from diabetic foot ulcer patients and foot skin samples from diabetic patients, was downloaded from GEO. After data quality control, dimensionality reduction, differential analysis, cell type annotation, and pseudotime analysis, cell types spanning the entire course of diabetic foot ulcers were identified, and differentially expressed genes (DEGs) were obtained. hdWGCNA identified gene modules highly correlated with diabetic foot ulcers. Conventional transcriptome datasets GSE68183 and GSE80178, containing non-healing ulcer tissue samples from diabetic foot ulcer patients and foot skin samples from diabetic patients, were downloaded for differential analysis to screen DEGs, and WGCNA was used to identify diabetic foot ulcer-related gene modules. The DEGs from single-cell transcriptome, DEGs from conventional transcriptome samples, and module genes from WGCNA and hdWGCNA were integrated to screen biomarkers for diabetic foot ulcers. The GSE134431 dataset was downloaded as a validation conventional transcriptome dataset, and the expression levels of diabetic foot ulcer biomarkers were compared in single-cell transcriptome and validation conventional transcriptome datasets. Diabetic and diabetic foot ulcer rat models were replicated, wound tissue was collected, and immunohistochemistry and western blot were used to detect biomarker expression levels. RESULTS AND CONCLUSION: Single-cell transcriptome data analysis showed that epithelial cell differentiation spanned the entire pathological process of diabetic foot ulcers. A total of 146 DEGs were obtained from single-cell transcriptome between groups, including 59 upregulated and 87 downregulated DEGs. hdWGCNA identified 19 gene modules related to diabetic foot ulcers, containing 476 core genes. Conventional transcriptome data analysis yielded a total of 913 DEGs, including 343 upregulated and 570 downregulated DEGs. WGCNA obtained 19 diabetic foot ulcer-related gene modules, containing 887 genes. Two biomarkers for diabetic foot ulcers were screened: S100A14 and SFN. The expression levels of these two genes in diabetic foot ulcer samples were higher than those in diabetic foot skin samples in both single-cell transcriptome and validation conventional transcriptome datasets. Animal experiments showed that the expression levels of S100A14 and SFN in wound tissue of diabetic foot ulcer rats were higher than those in back skin tissue of diabetic rats. The results indicate that the pathological process of diabetic foot ulcers involves multiple cell types, among which epithelial cells are the key cell subpopulation. S100A14 and SFN are significantly upregulated in diabetic foot ulcer samples and are potential targets for the treatment of diabetic foot ulcers.

1. Introduction

Diabetic foot ulcer (DFU) is a common complication of diabetes, clinically manifested as foot infection, ulceration, or deep tissue destruction [1]. Up to 80% of amputations in diabetic patients are caused by DFU [2]. Data show that the 1-year mortality rate associated with DFU is approximately 5%, and the 5-year mortality rate is as high as 42% [3]. Wound healing is a complex and orderly biological process involving inflammation, re-epithelialization, angiogenesis, histiocyte activation, and extracellular matrix synthesis and degradation [4]. Due to hyperglycemia and inflammatory responses, DFU is difficult to heal and has a high recurrence rate [5]. Uncovering the potential mechanisms affecting DFU healing is of great clinical significance for improving patient prognosis and reducing amputation and mortality rates.