Multi-drug resistant Staphylococcus epidermidis from chronic wounds impair healing in human wound model

Abstract

Venous leg ulcers (VLUs) represent one of the most prevalent types of chronic wounds characterised by perturbed microbiome and biofilm-forming bacteria. As one of the most abundant skin-commensal, Staphylococcus epidermidis is known as beneficial for the host, however, some strains can form biofilms and hinder wound healing. In this study, S. epidermidis distribution in VLUs and associated resistome were analysed in ulcer tissue from patients. Virulence of S. epidermidis isolates from VLUs were evaluated by whole genome sequencing, antimicrobial susceptibility testing, in vitro biofilm and binding assays, and assessment of biofilm-forming capability and pro-inflammatory potential using human ex vivo wound model. We demonstrated that S. epidermidis isolates from VLUs inhibit re-epithelialization through biofilm-dependent induction of IL-1β, IL-8, and IL-6 which was in accordance with impaired healing outcomes observed in patients. High extracellular matrix binding ability of VLU isolates was associated with antimicrobial resistance and expression levels of the embp and sdrG, responsible for bacterial binding to fibrinogen and fibrin, respectively. Finally, we showed that S. epidermidis from VLUs demonstrate pathogenic features with ability to impair healing which underscores the emergence of treatment-resistant virulent lineages in patients with chronic ulcers.

Keywords: infection; keratinocytes; translational research; venous leg ulcer; wound healing.

FIGURE 1

Staphylococcus epidermidis isolates from VLUs show high biofilm formation ability in vitro and in ex vivo human wounds. (A) Distribution of S. epidermidis in VLU patients based on detection of S. epidermidis specific RpoB gene. (B) Crystal violet assay showing biofilm forming potential of all isolates after 72 h of incubation. (C) Colony forming units (CFU/mL) of S. epidermidis isolates recovered from ex vivo wounds on day 4. (D) Immunostaining of S. epidermidis (green) attached to ex vivo wounds on day 4 post‐wounding. Yellow arrowheads showing the wound edge with initial wound size of 3 mm, while white arrowheads indicate migrating epithelial tongue visualised by phalloidin staining (red) for Actin cytoskeleton. The dashed white line indicates epidermal/dermal boundary and white arrows indicate detachment of epidermis caused by VLU isolates (scale bar = 200 μm). Data are presented as the mean ± SD from results obtained from three independent experiments. One‐way ANOVA followed by Dunnett's post hoc test was used to compare the results of all isolates relative to beneficial S. epidermidis CCN027 used as control (*p <0.05, ***p <0.001).

FIGURE 2

Staphylococcus epidermidis isolates from chronic wounds inhibit re‐epithelialization which associated with healing outcomes in VLU patients. (A) Wound samples of control uninfected and wounds infected with CW9, CW20, and CW48 from with H&E, showing wound edge with the initial wound size of 3 mm (yellow arrows) and epithelial tongue (white arrows). Bacterial aggregates are stained by haematoxylin (Scale bar = 200 μm). (B) Percentage of re‐epithelization of ex vivo wounds infected with VLU isolates. (C) Wound healing rates of patients undergoing VLU debridement correlate with the ex vivo biofilm formation of corresponding S. epidermidis isolates. Data are presented as the means ± SD from results obtained from three independent experiments. One‐way ANOVA followed by Dunnett's post hoc test was used to compare the results of all isolates relative to uninfected control (***p <0.001).

FIGURE 3

Staphylococcus epidermidis aggregates on human ex vivo wounds upregulate pro‐inflammatory signals. Gene expression of pro‐inflammatory cytokines (A)–(D) and antimicrobial peptides (E), (F) assessed by qRT‐qPCR. Data are presented as the means ± SD from results obtained from three independent experiments. One‐way ANOVA followed by Dunnett's post hoc test was used to compare the gene expression levels relative to untreated control (*p <0.05, **p <0.01, ***p <0.001).

FIGURE 4

Expression of Staphylococcus epidermidis adhesion genes configure in vitro bacterial binding to extracellular matrix. (A) Heat‐map showing presence of known S. epidermidis adhesion and biofilm‐associated genes across all strains. (B) The growth curves of S. epidermidis strains with the arrows showing exponential (8 h) and stationary phases (24 h) when the bacteria were collected for gene expression analysis. (C), (D) Changes of mRNA levels of the fibronectin‐binding (embp) and the fibrinogen‐binding (sdrG) genes through growth phases in all strains assessed by qRT‐qPCR. (E), (F) Percentage (%) of adhesion of all strains to immobilised fibronectin and collagen, in vitro. Data are presented as the means ± SD from results obtained from three independent experiments. One‐way ANOVA followed by Dunnett's post hoc test was used to compare the gene expression levels and binding percentage of all strains relative to CCN027 control (*p <0.05, **p <0.01, ***p <0.001).

FIGURE 5

Staphylococcu epidermidis isolates from VLU show high level of resistance to antibiotics and antimicrobials. (A) Heat‐map representing the antibiotic resistance gene (ARG) prevalence in tested S. epidermidis strains. (B) MIC50 values of selected antibiotics showing susceptibility of S. epidermidis strains. Data are presented from three independent experiments with three technical replicates. Diagrams showing (C) prevalence of mupA and mecA in VLU patient samples (n = 24) and (D) distribution of S. epidermidis rpoB gene in the wound linked with the presence of mupA and mecA genes.