Chronic wound microenvironment mediates selection of biofilm-forming multi drug resistant Staphylococcus epidermidis with capability to impair healing

Abstract

Venous leg ulcers (VLU) are the most common chronic wounds characterized by bacterial biofilms and perturbed microbiome. Staphylococcus epidermidis is primarily known as skin commensal beneficial for the host, however, some strains can form biofilms and cause infections. By employing shotgun metagenomic sequencing we show that genetic signatures of antimicrobial resistance, adhesion and biofilm formation in VLU isolates correlate with in vitro bacterial traits. We demonstrate that the capability of chronic wound isolates to form biofilms and elicit IL-8 and IL-1β expression in human ex vivo wounds, correlates with the non-healing outcomes in patients with VLU. In contrast, commensal strains were incapable of surviving in the human ex vivo wounds. We show that major fitness traits of S. epidermis from VLU involve genes for resistance to methicillin and mupirocin, while the biofilm formation relied on the minimal number of genetic elements responsible for bacterial binding to fibronectin and fibrinogen. This underscores the importance of the emergence of treatment resistant virulent lineages in patients with non-healing wounds.

Keywords: Staphylococcus epidermidis; antimicrobial resistance; biofilm; chronic wounds; wound healing.

Figure 1.. S. epidermidis isolates from VLUs show resistance to broad-spectrum antibiotics and higher capacity to form biofilm and bind to fibronectin, in vitro.

(a) Heat-map representing the antibiotic resistance gene (ABR) prevalence in healthy skin and chronic wound isolates followed by (b) MIC50 values of selected antibiotics showing susceptibility of S. epidermidis isolates from both origins. Data are presented from 3 independent experiments with 3 technical replicates. (c) Heat-map showing prevalence of known S. epidermidis adhesion and biofilm associated genes across all isolates. (d) The growth curves of S. epidermidis isolates showing better fitness of chronic wound isolates with the arrows pointing to exponential (8h) and stationary phases (24h) when the bacteria were collected for gene expression analysis. (e, f) Gene expression of the fibronectin-binding (embp) and the fibrinogen-binding (sdrG) genes through growth phases in all isolates assessed by qRT-qPCR; data is presented as mean ± SD from results obtained from three independent experiments (n=3–6). (g) Cristal violet assay showing biofilm forming potential of all isolates after 72h of incubation. (h) Percentage (%) of adhesion of all isolates to fibronectin; data are presented as the mean ± SD from results obtained from three independent experiments (n=6–12). One-way ANOVA followed by Dunnett’s post hoc test was used to compare the results of all isolates relative to CCN027 healthy skin strain (*p<0.05, **p<0.01, ***p<0.001).

Figure 2.. The ability of chronic wound S. epidermidis isolates to form pro-inflammatory biofilm in ex vivo human wounds associates with clinical outcomes in patients.

(a) Immunostaining of S. epidermidis biofilm on ex vivo human wounds day 4 post-wounding; control – no infection. Yellow arrows showing wound edge, white arrows indicate migrating epithelium visualized by DAPI staining of keratinocyte nuclei, with epidermis separated from dermis caused by VLU isolates (green arrows) (b) H&E staining of uninfected wound (control), and wounds infected with CW9, CW20, and CW48 showing wound edge (yellow arrows) and epithelial tongue (white arrows) on day 4 post-wounding. Bacterial aggregates stained by hematoxylin are indicated by dashed lines; scale bar = 200 μm. (c) Bacterial growth in ex vivo wounds on day 4. (d) Percentage of re-epithelization of ex vivo wounds non-infected and infected with VLU isolates; data is presented as mean ± SD from three independent experiments (n=3). One-way ANOVA followed by Dunnett’s post hoc test was used to compare the results of all isolates relative to uninfected control or CCN027 (*p<0.05, ***p<0.001). (e) Gene expression analysis of pro-inflammatory cytokines and antimicrobial peptides; data is presented as the mean ± SD from three independent experiments (n=3–8). 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). (e) Clinical healing trajectories of VLU study subjects correlate with the ex vivo biofilm formation of corresponding S. epidermidis isolates. Ulcer size of VLU patients was assessed weekly during 8 weeks of standard of care.