Dispersal of Epithelium-Associated Pseudomonas aeruginosa Biofilms

Abstract

Pseudomonas aeruginosa grows in highly antibiotic-tolerant biofilms during chronic airway infections. Dispersal of bacteria from biofilms may restore antibiotic susceptibility or improve host clearance. We describe models to study biofilm dispersal in the nutritionally complex environment of the human airway. P. aeruginosa was cocultured in the apical surface of airway epithelial cells (AECs) in a perfusion chamber. Dispersal, triggered by sodium nitrite, a nitric oxide (NO) donor, was tracked by live cell microscopy. Next, a static model was developed in which biofilms were grown on polarized AECs without flow. We observed that NO-triggered biofilm dispersal was an energy-dependent process. From the existing literature, NO-mediated biofilm dispersal is regulated by DipA, NbdA, RbdA, and MucR. Interestingly, altered signaling pathways appear to be used in this model, as deletion of these genes failed to block NO-induced biofilm dispersal. Similar results were observed using biofilms grown in an abiotic model on glass with iron-supplemented cell culture medium. In cystic fibrosis, airway mucus contributes to the growth environment, and a wide range of bacterial phenotypes are observed; therefore, we tested biofilm dispersal in a panel of late cystic fibrosis clinical isolates cocultured in the mucus overlying primary human AECs. Finally, we examined dispersal in combination with the clinically used antibiotics ciprofloxacin, aztreonam and tobramycin. In summary, we have validated models to study biofilm dispersal in environments that recapitulate key features of the airway and identified combinations of currently used antibiotics that may enhance the therapeutic effect of biofilm dispersal.IMPORTANCE During chronic lung infections, Pseudomonas aeruginosa grows in highly antibiotic-tolerant communities called biofilms that are difficult for the host to clear. We have developed models for studying P. aeruginosa biofilm dispersal in environments that replicate key features of the airway. We found that mechanisms of biofilm dispersal in these models may employ alternative or additional signaling mechanisms, highlighting the importance of the growth environment in dispersal events. We have adapted the models to accommodate apical fluid flow, bacterial clinical isolates, antibiotics, and primary human airway epithelial cells, all of which are relevant to understanding bacterial behaviors in the context of human disease. We also examined dispersal agents in combination with commonly used antipseudomonal antibiotics and saw improved clearance when nitrite was combined with the antibiotic aztreonam.

Keywords: Pseudomonas aeruginosa; biofilm; cyclic-di-GMP; cystic fibrosis; dispersal; dispersion.

FIG 1

Biotic biofilm dispersal in flow cells. (A) CFBE41o- airway epithelial cells are grown on glass coverslips and placed in perfusion chambers. The chambers are inoculated with PAO1-gfp (green). Bacteria attach for 60 min, and then the biofilm is grown for 4 more h. Then, from hours 5 to 6, the coculture is treated with NaCl (tonicity control) or 15 mM sodium nitrite. Images of identical points are taken at hours 5 and 6. (B) Volumetric projection of representative coculture after 6 h taken at ×40 magnification. Blue (DAPI) stains the epithelial nuclei beneath the PAO1 (green) biofilms. (C) Quantification of biomass before (hour 5) and after (hour 6) treatment. (D) Representative field at 5 h. (E) The same field imaged at hour 6 after 1 h of exposure to control sodium chloride. (F) Representative field at 5 h. (G) The same field at h 6 after 1 h of exposure to 15 mM sodium nitrite. P values from one-way analysis of variance (ANOVA) followed by Sidak’s test. Three replicates per condition with at least five paired z-stacks quantified per sample at each time point.

FIG 2

Characterization of static biotic biofilm dispersal. (A) Airway epithelial cells are cultured at the air-liquid interface, and then the apical surface is infected with P. aeruginosa. Once the coculture is mature, it is treated with a dispersal agent for 15 min, and the resulting populations can be studied further. Biofilms were grown on the apical surface of CFBE41o- airway epithelial cells for 6 h and dispersed for 15 min for all panels. Either samples were prepared for imaging or dispersed bacteria were counted by serial dilution. (B and C) Confocal scanning light microscopy images of the biotic biofilms after treatment with medium (B) or 75 mM nitrite (C). Green, PAO1-gfp. (D) Biomass quantified for at least 6 h; ×40 magnification fields from 3 samples per condition are shown; P value from unpaired, two-sided t test. (E to I) CFBE 41o- and PAO1 cocultures were treated with the indicated compounds, and the released bacteria were counted. (E) Dose response for bacterial release across 10-fold concentrations of sodium nitrite. P values from one-way ANOVA followed by post hoc Dunnett’s test. (F) Bacteria released from biofilm after treatment with 500 μM SNP versus MEM; P value from two-sided, unpaired t test. (G) Quantification of dispersed bacteria from biofilms treated with DPTA-NONOate, glutamate, succinate, or ammonium chloride. (H) Quantification of dispersed bacteria from biofilms treated with the proton ionophore CCCP and nitrite. (I) Quantification of dispersed bacteria from biofilms treated with nitrite and a protease inhibitor cocktail (cOmplete) or 50 mg/liter tetracycline. (G to I) Statistics are from one-way ANOVA followed by post hoc Dunnett’s test. Biologic replicates: 3 to 6 per condition tested.

FIG 3

Model adaptation for primary human AECs and CF bacterial strains. (A) Biofilms were grown on the apical surface of mucus-producing primary human airway epithelial cells for 6 h. A panel of 10 CF clinical isolates was used. Biofilms were dispersed with 50 mM NO₂⁻, and the number of dispersed bacteria were quantified by serial dilution. Several isolates did not disperse (blue), while some showed great dispersal (yellow). (B) Three to six biologic replicates of individual clinical isolates were grown as biofilms on the surface of CFBE4lo- cells and dispersed, and the dispersed bacteria were quantified. *, P < 0.05 by one-way ANOVA followed by Sidak’s test.

FIG 4

Phosphodiesterase signaling in biotic biofilm dispersal. (A) In the PAO1 background, biofilms were grown for the indicated deletion strains on CFBE41o- AECs. All deletion strains were dispersed with DPTA-NONOate. (B) In the PA14 background, biotic biofilms were grown for the indicated deletions strains. All strains were dispersed with sodium nitrite. *, P < 0.05 by one-way ANOVA followed by post hoc Sidak’s test; at least three replicates were done for each strain.

FIG 5

Dose response of NO-induced dispersal for biofilms grown in Fe-MEM. PAO1-gfp was grown on a glass surface in MEM supplemented with transferrin and hemoglobin and imaged after 6 h (Fe-MEM). Either sodium chloride (control) or sodium nitrite was added directly to the culture for 15 min prior to fixation. Z-stack images were collected at ×20 magnification to capture the geographic diversity of the mounts. Mean biomass ± SD is shown on each image. Three complete technical replicates were done with 6 to 10 fields imaged per replicate. (A) 15 mM sodium chloride; (B) 15 mM sodium nitrite; (C) 1.5 mM sodium nitrite; (D) 150 μM sodium nitrite.

FIG 6

NbdA-DipA compound deletion strain dispersals. (A) Biotic biofilms of ΔnbdA ΔdipA expressing gfp were grown on CFBE41o- AECs. (B) When the biotic biofilm was treated with DPTA for 15 min, biomass decreased. Representative ×20 magnification z-stacks are shown. At least 8 fields were taken per condition. (C) Dispersed bacteria from cocultures were quantified by serial dilution; P values from one-way ANOVA followed by Sidak’s test. (D) ΔnbdA ΔdipA cultured on glass in Fe-MEM and rinsed once with 15 mM NaCl. (E) Abiotic culture treated with 15 mM nitrite for 15 min. Representative ×40 magnification fields. Mean biomass ± standard deviation (SD) are shown on graphs. Comparisons between conditions were significant, with P < 0.05 by unpaired, two-way t test. At least 3 biologic replicates were done for each condition.

FIG 7

Hyper-biofilm strain dispersed with NO. (A) PAO1-pJ220-WspR grown on glass for 6 h prior to imaging; the strain expresses gfp. (B) The addition of rhamnose after the first 60 min leads to formation of distinct mounds and increased biomass. (C) 15 mM nitrite treatment of PAO1-pJM220-WspR strain with rhamnose. Changes in biomass between panels A, B, C: P < 0.05 by one-way ANOVA. (D) CFU counts from the indicated strains grown on AECs and dispersed with DPTA. (E) Representative image of PAO1-pJ220-WspR grown on AECs with rhamnose added after the 60-min attachment period. (F) After 15 min of exposure to DPTA, biomass drops (P < 0.05 by unpaired two-sided t test). At least three biologic replicates were done per condition.

FIG 8

Interaction between NO-triggered dispersal and antibiotics. Biotic biofilms were grown for 6 h on CFBE41o- cells. Cocultures were dispersed for 15 min with DPTA-NONOate, and then antibiotics were added for an additional 20 h. Adherent bacteria were counted by serial dilution and analyzed with one-way ANOVA followed by Dunnett’s test, with at least three replicates per condition.