Multidrug-resistant Pseudomonas aeruginosa is predisposed to lasR mutation through up-regulated activity of efflux pumps in non-cystic fibrosis bronchiectasis patients

Abstract

Background: Multidrug-resistant (MDR) Pseudomonas aeruginosa is a frequent opportunistic pathogen that causes significant mortality in patients with non-cystic fibrosis bronchiectasis (NCFB). Although the quorum sensing (QS) system is a potential target for treatment, lasR mutants that present with a QS-deficient phenotype have been frequently reported among clinical P. aeruginosa isolates. We aimed to investigate whether antibiotic resistance would select for lasR mutants during chronic P. aeruginosa lung infection and determine the mechanism underlying the phenomenon.

Methods: We prospectively evaluated episodes of chronic P. aeruginosa lung infections in NCFB patients over a 2-year period at two centers of our institution. QS phenotypic assessments and whole-genome sequencing (WGS) of P. aeruginosa isolates were performed. Evolution experiments were conducted to confirm the emergence of lasR mutants in clinical MDR P. aeruginosa cultures.

Results: We analyzed episodes of P. aeruginosa infection among 97 NCFB patients and found only prior carbapenem exposure independently predictive of the isolation of MDR P. aeruginosa strains. Compared with non-MDR isolates, MDR isolates presented significantly QS-deficient phenotypes, which could not be complemented by the exogenous addition of 3OC12-HSL. The paired isolates showed that their QS-phenotype deficiency occurred after MDR was developed. Whole-genome sequencing analysis revealed that lasR nonsynonymous mutations were significantly more frequent in MDR isolates, and positive correlations of mutation frequencies were observed between genes of lasR and negative-efflux-pump regulators (nalC and mexZ). The addition of the efflux pump inhibitor PAβN could not only promote QS phenotypes of these MDR isolates but also delay the early emergence of lasR mutants in evolution experiments.

Conclusions: Our data indicated that MDR P. aeruginosa was predisposed to lasR mutation through the upregulated activity of efflux pumps. These findings suggest that anti-QS therapy combined with efflux pump inhibitors might be a potential strategy for NCFB patients in the challenge of MDR P. aeruginosa infections.

Keywords: LasR activity; Pseudomonas aeruginosa; efflux pump; multidrug resistance; non-cystic fibrosis bronchiectasis.

Figure 1

Quorum sensing (QS) phenotypic assessments for P. aeruginosa isolates from non-cystic fibrosis bronchiectasis patients. P. aeruginosa isolates were classified into the non-MDR group (n = 63) and the MDR group (n = 34) according to their antimicrobial susceptibility test results. The production of QS signals (A, B) and QS-controlled virulence factors (C, D) was compared between MDR isolates and non-MDR isolates from different patients. The data were presented as median with an interquartile range. Another 10 strains were isolated within the sample from each of the 20 randomly chosen patients. Among these patients, 10 had MDR isolates, and the others had non-MDR isolates. The production of QS signals (E, F) and QS-controlled virulence factors (G, H) of MDR isolates or non-MDR isolates within a given patient is shown. PAO1 (blue) and LasR-null mutant (red) served as positive and negative controls, respectively. The whiskers of the boxplot mark the 5th and 95th percentiles, while the box contains the 25th percentile, median, and 75th percentiles. A volcano plot (I) was worked out to show the difference of these QS products between non-MDR isolates and MDR isolates. The dotted line presented the significant level of p = 0.01. The production of pyocyanin (J) and protease (K) was compared between MDR cultures with and without 3OC12-HSL from a given patient. The data are presented as mean ± SD. ^* p < 0.05. MDR, multidrug resistance; OD, optical density; RFU, relative fluorescence unit.

Figure 2

Quorum sensing phenotypes of 11 paired P. aeruginosa isolates. Each pair was obtained from a different patient. The time interval between the paired two isolates was 81 (35–98) days apart. In patients who had isolates shifting from non-MDR to MDR, quorum sensing phenotypes, including signal molecules 3OC12-HSL (A) and C4-HSL (B), and virulence factors pyocyanin (C) and protease (D), were significantly poorer in recurrent isolates than in initial ones; while in patients who had paired isolates maintaining non-MDR, no significant difference of QS phenotypes was found between initial and recurrent isolates. For initial isolates vs. recurrent isolates, ^* p < 0.05; ns, nonsignificant; MDR, multidrug resistance; OD, optical density; RFU, relative fluorescence unit.

Figure 3

Single-nucleotide variation analysis of P. aeruginosa clinical respiratory isolates selected for whole-genome sequencing. The P. aeruginosa PAO1 strain (GenBank sequence AE004091) was used as the reference. The paired isolates (15 and 30, 27 and 28, and 23 and 24) were obtained from the same patients.

Figure 4

Mutations in quorum sensing genes (lasR and rhlI) and negative regulator genes (mexR, mexZ, nalC, nalD, and nfxB) of efflux pumps from non-MDR and MDR P. aeruginosa isolates that underwent whole-genome sequencing. A comparison of synonymous mutations (A), nonsynonymous mutations (B), and Ka/Ks ratios (C) was analyzed between non-MDR isolates and MDR isolates. Data were presented as mean ± SD. Heatmaps showed the correlation of synonymous mutations (D), nonsynonymous mutations (E), and Ka/Ks ratios (F) between non-MDR isolates and MDR isolates. ^* p < 0.05; ^** p < 0.01. MDR, multidrug resistance. Ka/Ks ratio is an indicator of selective pressures on genes by calculating the ratio of nonsynonymous (Ka) to synonymous (Ks) nucleotide substitution rates.

Figure 5

Phenotypic assessment of efflux pump-induced quorum sensing (QS) dysfunctionality. Growth curves showed that bacterial densities were similar among these overnight cultures (18 h) with and without the efflux pump inhibitor PAβN (100 mg/L) (A). MICs of ciprofloxacin, ceftazidime, and meropenem (B) and the QS phenotypes, including QS signals (C) and virulence factors (D), were analyzed between non-MDR isolates and MDR isolates with or without PAβN. ^* p < 0.05; ^** p < 0.01; ns, nonsignificant; MDR, multidrug resistance; MIC, minimum inhibitory concentration; OD, optical density; RFU, relative fluorescence unit.

Figure 6

Evolutionary experiments for lasR mutants among MDR P. aeruginosa isolates in casein broth medium. Isolates 17, 20, and 24, the three QS-proficient efflux-pump-regulator mutants with a complete lasR sequence, were selected. PAO1 was used as a control. Three independent laboratory evolution experiments were conducted, and lasR mutants (protease-negative strains) were identified by patching 100 individual colonies on skim milk agar and observing if there was an absence of a clear halo around a colony after about 18 h. The percentage of mutant cells (A) and numbers of total cells (B) were calculated. These experiments were repeated by adding the efflux pump inhibitor PAβN (100 mg/L) to the casein broth medium (C, D). In all panels, wild-type PAO1 is indicated by black hexagons, the isolate 17 by red circles, the isolate 20 by blue rhombuses, and the isolate 24 by orange squares. The first experiments are shown by the solid lines, the second experiments by the dashed lines, and the third experiments by the dotted lines. MDR, multidrug resistance; QS, quorum sensing.