Rapid adaptation drives invasion of airway donor microbiota by Pseudomonas after lung transplantation

Abstract

In cystic fibrosis (CF) patients, chronic airway infection by Pseudomonas leads to progressive lung destruction ultimately requiring lung transplantation (LT). Following LT, CF-adapted Pseudomonas strains, potentially originating from the sinuses, may seed the allograft leading to infections and reduced allograft survival. We investigated whether CF-adapted Pseudomonas populations invade the donor microbiota and adapt to the non-CF allograft. We collected sequential Pseudomonas isolates and airway samples from a CF-lung transplant recipient during two years, and followed the dynamics of the microbiota and Pseudomonas populations. We show that Pseudomonas invaded the host microbiota within three days post-LT, in association with a reduction in richness and diversity. A dominant mucoid and hypermutator mutL lineage was replaced after 11 days by non-mucoid strains. Despite antibiotic therapy, Pseudomonas dominated the allograft microbiota until day 95. We observed positive selection of pre-LT variants and the appearance of novel mutations. Phenotypic adaptation resulted in increased biofilm formation and swimming motility capacities. Pseudomonas was replaced after 95 days by a microbiota dominated by Actinobacillus. In conclusion, mucoid Pseudomonas adapted to the CF-lung remained able to invade the allograft. Selection of both pre-existing non-mucoid subpopulations and of novel phenotypic traits suggests rapid adaptation of Pseudomonas to the non-CF allograft.

Figure 1. Evolution of the microbial airway composition after LT.

Shown are the sequential analysis of bronchial aspirates (BA) and bronchoalveolar lavages (BAL) collected pre- and post-LT. (a) Timing of antimicrobial and immunosuppressing therapies. (b) Bacterial cultures over time. Bacterial load (number of cells/mL) is represented as following: red square, >10⁵; orange square, 10³–10⁴; yellow square, <10³; white square, positive culture; crossed square, not determined. (c) Taxonomic analysis, cutoff set >1%. (d) Absolute bacterial abundance determined by performing qPCR on gDNA extracted from native airway samples. qPCR detection limit established at 10³ rspL copy number/mL of sample. (e) Microbial diversity and richness.

Figure 2. Genome comparison between two pre-LT and one late post-LT isolate.

(a) Venn diagram summarizing the number of mutations identified in each genome. (b) Description of the 15 mutations detected in the genome of the final post-LT isolate Z95B.

Figure 3. In patient evolution of Pseudomonas genome variants after LT.

(a) Relative abundance of mutated variants in the whole Pseudomonas population. Thirteen genes were amplified by PCR and amplicons were submitted to deep sequencing. Mutations previously identified in the Z95B genome were searched in the total pool of sequencing reads. The amount of wild-type and mutated versions of the gene are expressed as the percentage of mutated alleles in the total population for each gene. (b) Evolutionary history of Pseudomonas isolates and populations following LT. The vertical arrays show the status (wild-type or mutated) of each gene in strains isolated from bronchial aspirates. These data analyzed concomitantly to those obtained for the total population (panel A) allows to reconstruct the evolution of the isolates, and to determine their respective percentage of variants in the total Pseudomonas community. The status of the hypothetical intermediate genomes (deduced) and those of the most recent common ancestors (MRCA) were inferred from all data taken together. ex = extubation.

Figure 4. Phenotypic analysis of ten P. aeruginosa strains isolated from sequential airway samples.

(a) *indicate the antibiotic resistance (R) or susceptibility (S) phenotypes according to EUCAST breakpoints (minimal inhibitory concentrations are provided in Table S4). (b) Biofilms were stained with crystal violet after 24 hours of static growth. Optical densities were measured at 590 nm. Biofilm formation of the reference strain PAO1 is shown for comparison. Error bars showed standard deviations; statistical significance was calculated by an unpaired t-test. Swimming phenotype was evaluated after 22 h of growth at 37 °C in LB 0.2% agar.