Colistin (polymyxin E) is a last-resort antibiotic against multidrug-resistant isolates of Pseudomonas aeruginosa. However, the nephro-toxicity of colistin limits its use, spurring the interest in novel antimicrobial peptides (AMP). Here, we show that the synthetic AMP-dendrimer G3KL (MW 4,531.38 Da, 15 positive charges, MIC = 8 mg/liter) showed faster killing than polymyxin B (Pmx-B) with no detectable resistance selection in P. aeruginosa strain PA14. Spontaneous mutants selected on Pmx-B, harboring loss of function mutations in the PhoQ sensor kinase gene, showed increased Pmx-B MICs and arnB operon expression (4-amino-l-arabinose addition to lipid A), but remained susceptible to dendrimers. Two mutants carrying a missense mutation in the periplasmic loop of the PmrB sensor kinase showed increased MICs for Pmx-B (8-fold) and G3KL (4-fold) but not for the dendrimer T7 (MW 4,885.64 Da, 16 positive charges, MIC = 8 mg/liter). The pmrB mutants showed increased expression of the arnB operon as well as of the speD2-speE2-PA4775 operon, located upstream of pmrAB, and involved in polyamine biosynthesis. Exogenous supplementation with the polyamines spermine and norspermine increased G3KL and T7 MICs in a phoQ mutant background but not in the PA14 wild type. This suggests that both addition of 4-amino-l-arabinose and secretion of polyamines are required to reduce susceptibility to dendrimers, probably neutralizing the negative charges present on the lipid A and the 2-keto-3-deoxyoctulosonic acid (KDO) sugars of the lipopolysaccharide (LPS), respectively. We further show by transcriptome analysis that the dendrimers G3KL and T7 induce adaptive responses through the CprRS two-component system in PA14.

With the increase of infections due to multidrug resistant bacterial pathogens and the shortage of antimicrobial molecules with novel targets, interest in bacteriophages as a therapeutic option has regained much attraction. Before the launch of future clinical trials, in vitro studies are required to better evaluate the efficacies and potential pitfalls of such therapies. Here we studied in an ex vivo human airway epithelial cell line model the efficacy of phage and ciprofloxacin alone and in combination to treat infection by Pseudomonas aeruginosa. The Calu-3 cell line and the isogenic CFTR knock down cell line (cftr-) infected apically with P. aeruginosa strain PAO1 showed a progressive reduction in transepithelial resistance during 24 h. Administration at 6 h p.i. of single phage, phage cocktails or ciprofloxacin alone prevented epithelial layer destruction at 24 h p.i. Bacterial regrowth, due to phage resistant mutants harboring mutations in LPS synthesis genes, occurred thereafter both in vitro and ex vivo. However, coadministration of two phages combined with ciprofloxacin efficiently prevented PAO1 regrowth and maintained epithelial cell integrity at 72 p.i. The phage/ciprofloxacin treatment did not induce an inflammatory response in the tested cell lines as determined by nanoString R gene expression analysis. We conclude that combination of phage and ciprofloxacin efficiently protects wild type and cftr-epithelial cells from infection by P. aeruginosa and emergence of phage resistant mutants without inducing an inflammatory response. Hence, phage-antibiotic combination should be a safe and promising anti-Pseudomonas therapy for future clinical trials potentially including cystic fibrosis patients.