Bacterial fitness in chronic wounds appears to be mediated by the capacity for high-density growth, not virulence or biofilm functions

Abstract

While much is known about acute infection pathogenesis, the understanding of chronic infections has lagged. Here we sought to identify the genes and functions that mediate fitness of the pathogen Pseudomonas aeruginosa in chronic wound infections, and to better understand the selective environment in wounds. We found that clinical isolates from chronic human wounds were frequently defective in virulence functions and biofilm formation, and that many virulence and biofilm formation genes were not required for bacterial fitness in experimental mouse wounds. In contrast, genes involved in anaerobic growth, some metabolic and energy pathways, and membrane integrity were critical. Consistent with these findings, the fitness characteristics of some wound impaired-mutants could be represented by anaerobic, oxidative, and membrane-stress conditions ex vivo, and more comprehensively by high-density bacterial growth conditions, in the absence of a host. These data shed light on the bacterial functions needed in chronic wound infections, the nature of stresses applied to bacteria at chronic infection sites, and suggest therapeutic targets that might compromise wound infection pathogenesis.

Fig 1. Wound infections initiated with P. aeruginosa liquid cultures delay wound healing.

A. H&E staining of excised wounds show original wound edges (black arrows) and the extent of epithelial migration after 5 or 11 days of healing (green arrows). B. The extent of wound closure (healing) was delayed in wounds challenged with P. aeruginosa. C. Total CFU recovered from wounds shows P. aeruginosa-challenged wounds contained high bacterial densities, and unchallenged wounds harbored low densities of bacteria (likely due to spontaneous infection Staphlococcus sp. as previously observed [52]). For A and B, each point represents data from an individual mouse. Significance determined by Mann-Whitney analysis.

Fig 2. Loss of acute virulence and biofilm formation functions does not decrease fitness in chronic wound infections.

A. Mutants lacking indicated acute virulence genes were competed 1:1 against wild-type P. aeruginosa in murine wounds, and competitive index was measured in scabs (circles) and wound beds (triangles). Line indicates mean values. B. Mutants defective and enhanced in biofilm formation (top) were competed 1:1 against wild-type P. aeruginosa in murine wounds (bottom). Scab and wound bed were harvested together (filled circle). Wild-type P. aeruginosa was assigned a competitive index of 1 (star). * indicates competitive index significantly different than 1 (p<0.01) by one sample t test. ^# indicates significant difference from wild type (p<0.05) by one-way ANOVA followed by Dunnett’s test.

Fig 3. Identification of gene inactivations that compromise wound fitness.

Tn-seq measured the relative change in abundance of mutants in the inoculum (x-axis) and after infection (y-axis). Each point represents a different gene inactivation. Gray points indicate gene inactivations with low input abundance; orange points indicate virulence genes [29]; blue, green and red points indicate gene inactivations exhibiting significant (P<0.01) decreases in abundance after infection. Green and red points indicate gene inactivations tested in 1:1 experiments (see Fig 4); gene inactivations indicated by green points were verified to be wound-defective, while gene inactivations indicated by red points were not defective in verification experiments. Gene inactivations indicated by blue points were not tested in 1:1 competitions.

Fig 4. Confirmation of gene inactivations that compromise wound fitness.

A. Mutants in 30 genes showing Tn-seq defects were tested in 1:1 competition with wild type, and mutant fitness in wound beds (triangles) or scabs (circle) was measured after five-days of infection. Dashed line indicates > two-fold defects in competition; * indicates competitive index significantly different than 1 (p<0.01) by one sample t-test. B. Fitness of individual transposon mutants in the 13 genes whose inactivation caused >10-fold impairments in wound fitness. Every gene verified to produce wound fitness defects was inactivated by at least five unique transposon mutants (points represent individual mutants). Almost every mutant in each gene exhibited similar fitness defects indicating that spontaneous secondary mutations were not likely responsible.

Fig 5. In vitro stress sensitivity of mutants with high and low wound fitness.

Mutants with high (yellow bars) and low (blue bars, darker blue indicates more severe defects) wound fitness were tested for anaerobic growth capacity (A); and sensitivity to paraquat (B), polymixin B (C), and EDTA (D). Data is shown as box plots indicating the median, and 25^th and 75^th percentiles of at least three experiments (performed in duplicate with error bars showing the range. *indicates p<0.05 compared to wild type by non-parametric ANOVA (Kruskal-Wallis and Dunnet’s test).

Fig 6. High-density growth and wound fitness correlate.

(A and B) Relationship between the competitive index of mutants in colonies (A) or stationary phase cultures (B) and in wound infections. Spearman correlation and significance values were calculated using mean competitive index of each mutant. C. The wound competitive index of the 100 mutants exhibiting the lowest and highest stationary phase culture fitness (as measured by Tn-seq). Box plots indicate median, and 25^th and 75^th percentiles; error bars show the range of wound competitive index. Significance determined by Mann-Whitney analysis.