Gram-Negative Bacteria and Lipopolysaccharides as Risk Factors for the Occurrence of Diabetic Foot

Abstract

Context: Imbalance of the skin microbial community could impair skin immune homeostasis and thus trigger skin lesions. Dysbiosis of skin microbiome may be involved in the early pathogenesis of diabetic foot (DF). However, the potential mechanism remains unclear.

Objective: To investigate the dynamic composition and function of the foot skin microbiome with risk stratification for DF and assess whether dysbiosis of the skin microbiome induces diabetic skin lesions.

Methods: We enrolled 90 consecutive subjects who were divided into 5 groups based on DF risk stratification: very low, low, moderate, and high risk for ulcers and a healthy control group. Integrated analysis of 16S ribosomal RNA and metagenomic sequencing of cotton swab samples was applied to identify the foot skin microbiome composition and functions in subjects. Then a mouse model of microbiota transplantation was used to evaluate the effects of the skin microbiome on diabetic skin lesions.

Results: The results demonstrated that, with the progression of diabetic complications, the proportion of gram-negative bacteria in plantar skin increased. At the species level, metagenome sequencing analyses showed Moraxella osloensis to be a representative core strain in the high-risk group. The major microbial metabolites affecting diabetic skin lesions were increased amino acid metabolites, and antibiotic resistance genes in microorganisms were abundant. Skin microbiota from high-risk patients induced more inflammatory cell infiltration, similar to the lipopolysaccharide (LPS)-stimulated response, which was inhibited by Toll-like receptor 4 (TLR4) antagonists.

Conclusions: The skin microbiome in patients with diabetes undergoes dynamic changes at taxonomic and functional levels with the progression of diabetic complications. The increase in gram-negative bacteria on the skin surface through LPS-TLR4 signal transduction could induce inflammatory response in early diabetic skin lesions.

Trial registration: ClinicalTrials.gov NCT04916457.

Keywords: 16S rRNA sequencing; diabetic foot; gram-negative bacteria; lipopolysaccharide; metagenome sequencing; skin microbiome.

Figure 1.

Dysbiosis of skin microbiota induce skin inflammatory response. (A) Bar plots represent the relative abundance of the top 10 phyla and changes in G-positive and G-negative bacteria among the groups. (B) Linear discriminant analysis effect size column and cladogram demonstrated the remarkable biomarkers with statistical differences among the 5 groups. (C) Mice were topically inoculated with the skin microbiota from the patients in 5 different groups every other day 7 times before sacrifice on the 15th day.

Figure 2.

Metagenome sequencing of the skin microbiome. (A) Linear discriminant analysis effect size distributed in column and cladogram of the metagenomic analysis demonstrates remarkable biomarkers with significant differences among the 3 groups. (B) Heatmap of Spearman correlation analysis showing the correlation between dominant species in the skin microbiome and clinical indexes. The value corresponding to the picture is the Spearman related coefficient (r). (C) Statistical chart of the gene numbers of KEGG pathway annotation. Abbreviations: DM_y, duration of diabetes (years); KEGG, Kyoto Encyclopedia of Genes and Genomes; Ulcer.r, risk for foot ulcer.

Figure 3.

Topical application of Resatorvid attenuated the inflammatory response induced by LPS or the skin bacteria of high-risk patients. Matched biopsies of BALB/c mice from group ND, HD, HD+ Resatorvid, LPS, or LPS+ Resatorvid, were stained for H&E, TLR4, and CD14 expression. Scale bar 100 µm. Abbreviations: H&E, hematoxylin and eosin; LPS, lipopolysaccharide; TLR4, Toll-like receptor 4.