Efficacy of species-specific protein antibiotics in a murine model of acute Pseudomonas aeruginosa lung infection

Abstract

Protein antibiotics, known as bacteriocins, are widely produced by bacteria for intraspecies competition. The potency and targeted action of bacteriocins suggests that they could be developed into clinically useful antibiotics against highly drug resistant Gram-negative pathogens for which there are few therapeutic options. Here we show that Pseudomonas aeruginosa specific bacteriocins, known as pyocins, show strong efficacy in a murine model of P. aeruginosa lung infection, with the concentration of pyocin S5 required to afford protection from a lethal infection at least 100-fold lower than the most commonly used inhaled antibiotic tobramycin. Additionally, pyocins are stable in the lung, poorly immunogenic at high concentrations and efficacy is maintained in the presence of pyocin specific antibodies after repeated pyocin administration. Bacteriocin encoding genes are frequently found in microbial genomes and could therefore offer a ready supply of highly targeted and potent antibiotics active against problematic Gram-negative pathogens.

Figure 1. Pyocins are stable and do not cause inflammation or tissue damage in the murine lung.

(a) Spot tests to determine stability of pyocins (75 μg) in lung tissue (24 h after administration) against P. aeruginosa P8 (P. aeruginosa P17 for pyocin S2). Five microlitres of homogenised post-caval lobe lung section in 100 μl of PBS was spotted onto a growing lawn of P. aeruginosa. The presence of clear zones indicates pyocin activity in the lung tissue and that pyocins are stable and still active in vivo for 24 h. (b) Hematoxylin and eosin (H&E) staining of paraffin-embedded sections of pyocin (75 μg) treated lung. Lack of inflammation or neutrophil influx are noted. All magnifications ×10.

Figure 2. P. aeruginosa P8 bacterial recovery from pyocin treated mice.

All mice were given 75 μg of pyocin. Bacterial counts were determined by CFU counts of homogenised lungs. (a) Mice treated with pyocin 1 h post-infection, all mice culled 4.5 h post-infection. (b) Mice treated with pyocin 1 h post-infection, pyocin treated mice survived to 24 h. No colonies were recovered from pyocin S5 treated mice in (b). Bars represent Mean ± SEM. *Denotes statistical significance for comparison of treatment versus control by a one-sided Mann-Whitney U test with Bonferroni correction applied.

Figure 3. Acquired tolerance to pyocins can be overcome by treating with a range of pyocins.

(a) Activity of pyocins S5, AP41 and L1 against WT P8, P8AP41T (a pyocin AP41 tolerant P8 strain) and P8AP41T* (P8AP41T recovered from untreated control mice shown in b). Purified protein at 200 μg ml⁻¹ was spotted onto a growing lawn of bacteria. Clear zones indicate pyocin cytotoxicity. (b) Bacterial counts for mice infected with P8AP41T and treated 1 h post-infection with pyocins (75 μg). Pyocin treated mice survived to 24 h. No colonies were recovered from pyocin S5 treated mice. Bars represent Mean ± SEM. *Denotes statistical significance for comparison of treatment versus control by a one-sided Mann-Whitney U test with Bonferroni correction applied. (c) P8 infected mice treated 1 h post-infection with pyocin combinations (7.5 μg of each pyocin); pyocin treated mice survived to 24 h. Bacterial counts were determined by CFU counts from homogenised lungs. Bars represent Mean ± SEM. *Denotes statistical significance for comparison of treatment versus control by a one-sided Mann-Whitney U test with Bonferroni correction applied.

Figure 4. Comparison of pyocin S5 and tobramycin efficacy.

Bacterial counts were determined by CFU counts of homogenised lungs. (a) Mice treated 1 h post-infection, S5–750 pg and tobramycin–7.5 μg mice survived to 24 h. All other mice culled 5.5 h post-infection. Bars represent Mean ± SEM. *Denotes statistical significance for comparison of treatment versus control by a one-sided Mann-Whitney U test with Bonferroni correction applied.

Figure 5. Pyocin S5 can afford protection against lethal P. aeruginosa infections in the presence of pyocin S5 antibodies.

(a) Bacterial counts were determined by CFU counts from homogenised lungs. Multiple doses of pyocin S5 (75 μg/dose) were administered intranasally three times, two weeks apart over four weeks. At thirteen weeks, mice were infected with P. aeruginosa P8 and treated with pyocin S5 (75 μg) or PBS intranasally 1 h post-infection. Bars represent Mean ± SEM of counts from 5 animals. *Denotes statistical significance for comparison of treatment versus control by a one-sided Mann-Whitney U test with Bonferroni correction applied. (b) Pyocin S5-specific IgG and IgA serum levels for mice repeatedly exposed to pyocin S5 or PBS (as described in a). The control group were immunized subcutaneously (S.C.) with pyocin S5 (75 ug/dose) in Freunds complete/incomplete adjuvant on three occasions, two weeks apart. No pyocin S5-specific IgA was detected in any of the animals tested. Bars represent Mean ± SEM calculated from the serum of 5 animals per group. (c,d) as for (a,b) except mice were repeatedly exposed to pyocin S5 via the intraperitoneal (I.P.) route prior to intranasal pyocin S5 or PBS treatment. The pyocin S5-specific IgG levels in (d) were very low in the pyocin S5 only group (1000-fold less than the Freunds complete/incomplete control group) and no pyocin S5-specific IgA was detected.