In vivo evolution of antimicrobial resistance in a biofilm model of Pseudomonas aeruginosa lung infection

Abstract

The evolution of antimicrobial resistance (AMR) in biofilms has been repeatedly studied by experimental evolution in vitro, but rarely in vivo. The complex microenvironment at the infection site imposes selective pressures on the bacterial biofilms, potentially influencing the development of AMR. We report here the development of AMR in an in vivo mouse model of Pseudomonas aeruginosa biofilm lung infection. The P. aeruginosa embedded in seaweed alginate beads underwent four successive lung infection passages with or without ciprofloxacin (CIP) exposure. The development of CIP resistance was assessed at each passage by population analysis of the bacterial populations recovered from the lungs of CIP-treated and control mice, with subsequent whole-genome sequencing of selected isolates. As inflammation plays a crucial role in shaping the microenvironment at the infection site, its impact was explored through the measurement of cytokine levels in the lung homogenate. A rapid development of AMR was observed starting from the second passage in the CIP-treated mice. Genetic analysis revealed mutations in nfxB, efflux pumps (mexZ), and two-component systems (parS) contribution to CIP resistance. The control group isolates exhibited mutations in the dipA gene, likely associated with biofilm dispersion. In the initial two passages, the CIP-treated group exhibited an elevated inflammatory response compared to the control group. This increase may potentially contribute to the release of mutagenic reactive oxygen species and the development of AMR. In conclusion, this study illustrates the complex relationship between infection, antibiotic treatment, and immune response.

Keywords: Pseudomonas aeruginosa; antimicrobial resistance; biofilm; in vivo evolution; inflammatory response.

Figure 1

The experimental setup of the evolution of P. Aeruginosa in a biofilm lung infection model in BALB/c mice; the experimental setup started with an overnight culture from a single colony of PAO1-mCherry-P_CD-gfp+; on Day 0, bacteria were embedded in alginate beads; on Day 1, mice were infected by injecting the left lung with alginate beads; on Day 2, the mice were treated twice with either CIP 0.25 mg (11.35 mg/kg) ×2 or saline (Placebo); on Day 3, the mice were euthanized, and the lungs were collected, homogenized, and used for population analysis; on Day 4, colonies from CIP plates were selected for overnight cultures that were embedded in new alginate beads to infect a new group of mice (a new passage); the illustration was created with BioRender.com

Figure 2

Drug PK and modeling; PK of CIP in the mouse lung with the MIC indicated; (A) PK model fit of CIP dosed 1 mg (45.5 mg/kg) ×1 (SC using a one-compartment model, model prediction (red lines), and the observed data (points) following SC administration of 1 mg CIP); (B) simulations for concentration–time courses in the lung for treatment dosing regimens of 0.25 mg (11.35 mg/kg) SC CIP at 8-h intervals; the corresponding dosing schedule is presented in the header of the figure; the dotted line represents the MIC of 0.094 mg/l.

Figure 3

Bacterial load and population analysis in lung homogenates; (A) the bacterial load in the lungs (CFU/ml) of CIP-treated and control mice from the different passages (Passages 1–4); no statistically significant differences were observed between the two groups of animals (P > .05, t-test) in either of the passages; (B) results of the population analysis of the bacteria in the lungs of treated animals in the different passages [1-4] presented as a fraction of the bacterial population that survived on plates with different CIP concentrations (0.5, 1, and 2 mg/l); significant differences were observed between the first passage compared to other passages (t-test, P < .05).

Figure 4

nfxB mutants tracked by fluorescence imaging during the evolution experiment; (A) the fluorescence of PAO1-mCherry-P_CD-gfp + colonies of the bacterial populations from lungs of CIP-treated and control mice plated on LB agar plates without CIP; red fluorescence indicates the wild-type and green fluorescence indicates nfxB mutants; the ratio of the green colonies to red colonies in the bacterial population from different individuals was calculated and presented in (B), which shows the percentage of nfxB mutants in the entire bacterial population in the different passages; ^** first and second passages indicate P = .0017, whereas ^** first to fourth passages indicate P = .0019; fluorescence images were captured by a Zeiss AxioZoom microscope; at least three images were randomly captured per petri dish plate; the objectives and detectors were optimized for detecting GFP fluorescence (green) and mCherry fluorescence (red); the number of tiles was adjusted to cover the whole plate area in one image.

Figure 5

Genetic-phenotype overview; the selected isolates from the treated (T) and control (C) groups of mice in each passage (first row); from different concentrations of CIP plates ranging from 0.5,1.0, and 2.0 mg/l (second row), three to six bacterial isolates were purified per passage (as indicated on the x-axis label); for these isolates, the top panel shows both the MIC values (black bars) and the growth rate (gray bars); values for the parental PAO1-mCherry-P_CD-gfp + are shown in the horizontal dashed lines; the bottom panel shows mutations identified in the genome sequence of the different bacterial isolates in green color for C, blue for T, and red for shared mutations; detailed information on variant annotations is shown in Table S3.

Figure 6

Macroscopic examination of infected and uninfected lungs; the figure shows differences in macroscopic pathology between (A) uninfected lungs that were inoculated by bacteria-free alginate beads (background) and infected lungs that were inoculated by P. Aeruginosa embedded in seaweed alginate beads (10⁷CFU/ml); inflammation (marked with arrows) was observed in both the CIP-treated mice with CIP (C–F), and the control mice (G–J) in the different passages of the evolution experiment.

Figure 7

Measurement of cytokines in lungs; the concentration (pg/ml) of cytokines (A) IFN-γ, (B) TNF-α, (C) IL-1β, (D) CXCL2, (E) G-CSF, and (F) IL-5 in the supernatant of the lung homogenates of CIP treated (+CIP) and control mice (C) during the four passages; all the results are compared to the background mice which were inoculated in the lung with alginate beads without bacteria; ^*P < .05, ^**P < .01, ^***P < .001, ^****P < .0001.