Sub-inhibitory concentrations of some antibiotics can drive diversification of Pseudomonas aeruginosa populations in artificial sputum medium

Abstract

Background: Pseudomonas aeruginosa populations within the cystic fibrosis lung exhibit extensive phenotypic and genetic diversification. The resultant population diversity is thought to be crucial to the persistence of infection and may underpin the progression of disease. However, because cystic fibrosis lungs represent ecologically complex and hostile environments, the selective forces driving this diversification in vivo remain unclear. We took an experimental evolution approach to test the hypothesis that sub-inhibitory antibiotics can drive diversification of P. aeruginosa populations. Replicate populations of P. aeruginosa LESB58 were cultured for seven days in artificial sputum medium with and without sub-inhibitory concentrations of various clinically relevant antibiotics. We then characterised diversification with respect to 13 phenotypic and genotypic characteristics.

Results: We observed that higher population diversity evolved in the presence of azithromycin, ceftazidime or colistin relative to antibiotic-free controls. Divergence occurred due to alterations in antimicrobial susceptibility profiles following exposure to azithromycin, ceftazidime and colistin. Alterations in colony morphology and pyocyanin production were observed following exposure to ceftazidime and colistin only. Diversification was not observed in the presence of meropenem.

Conclusions: Our study indicates that certain antibiotics can promote population diversification when present in sub-inhibitory concentrations. Hence, the choice of antibiotic may have previously unforeseen implications for the development of P. aeruginosa infections in the lungs of cystic fibrosis patients.

Figure 1

Diversification of LESB58 grown in the presence (closed circles) or absence (open circles) of antibiotics. Forty isolates of LESB58 from each culture were characterised using 13 traits (colony morphology, pyocyanin production, hypermutability, auxotrophy, susceptibility to 6 antibiotics and the presence/absence of 3 genomic regions). Therefore, 120 isolates were analysed for each experimental and control group across the 3 replicate populations. Isolates with different traits were identified as being a different haplotype. 3 replicate populations from each of the following treatments were analysed: LB (18 hours), ASM, and ASM with ceftazidime (+ CAZ), ASM with colistin (+CT), ASM with meropenem (+MEM), ASM with tobramycin (+TOBI), ASM with azithromycin (+AZT). (A) Number of novel haplotypes found within each replicate population. (B) Haplotype diversity found within each replicate population, defined as the probability of two randomly picked haplotypes being non-identical. (C) The colony forming units found within each replicate population following culture. P-values represent comparisons with ASM alone.

Figure 2

Population structure of LESB58 grown in ASM with and without sub-inhibitory concentrations of antibiotics. Each population structure of LESB58 was calculated using 13 traits (colony morphology, pyocyanin production, hypermutability, auxotrophy, susceptibility to 6 antibiotics and the presence/absence of 3 genomic regions) for the total 120 isolates by the eBurst algorithm. Each dot (and subsequent number) represents one novel haplotype, with dot size reflecting abundance. The larger the dot size, the more abundant that novel haplotype was in the 120 isolates that we characterised. Haplotypes designated with the number 1 represent isolates with the same characteristics as the P. aeruginosa LESB58 wild-type. The haplotypes representing isolates that had a straw-coloured colony morphology are circled in red; the haplotypes representing isolates that did not over-produce pyocyanin are circled in blue; and the one isolate that was hypermutable is circled in green. The remaining haplotypes differed in their antibiotic susceptibility profiles (All groups: n = 120).

Figure 3

Variations in zones of inhibition within LESB58 populations. The 120 LESB58 isolates obtained from the triplicate ASM cultures were assessed for susceptibility to six commonly used antibiotics (ceftazidime, ciprofloxacin, colistin, meropenem, tazobactam/piperacillin and tobramycin). Boxplots showing the range in the diameter of the zones of inhibition to these antibiotics are presented. 1. LB (18 hours) 2. ASM 3. ASM with ceftazidime 4. ASM with colistin 5. ASM with meropenem 6. ASM with tobramycin 7. ASM with azithromycin 8. Normal range of LESB58 (Groups 1–8: n = 120). The red line represents the cut-off for the sensitivity of P. aeruginosa to the antibiotics tested, in accordance with the guidelines of Andrews and Howe [37]. The values above the red line denote a higher sensitivity to antibiotics and the values below the line denote a higher resistance.

Figure 4

Summary of experimental design. The figure describes the steps involved in processing of the LESB58 populations cultured in ASM, with or without antibiotics, and the phenotypic and genotypic tests performed on individual isolates.