The impact of antibiotic pressure on the phenotypic evolution of clinical antibiotic resistant Pseudomonas aeruginosa in a Morbidostat device

Abstract

Pseudomonas aeruginosa is one of the most dangerous opportunistic bacteria in hospital acquired infections and ICU patients in recent times. With its variety of virulence factors and its ability to maintain in moist environments and to form biofilms it can cause wound infections, pneumonia and septicemia. Especially the rising number of multi-drug-resistant (MDR) strains poses a threat to healthcare facilities, which is why P. aeruginosa is listed as one of the critical priority pathogens with by the WHO. For this reason, colistin, an antibiotic that had almost disappeared as a form of treatment, requires increasing use as a last- resort antibiotic. The impact of subinhibitory concentrations of colistin on bacterial traits like virulence and biofilm formation in clinical P.aeruginosa isolates is largely unexplored. Moreover, the effect of metronidazole, a drug with no bactericidal effect on Pseudomonas but has been shown to induce mutagenesis, could be important for the phenotypic evolution of the bacteria. We applied a morbidostat as a culture device, that can automatically adapt the required concentrations of antibiotics to allow consistent growth while still challenging the bacteria, with the three conditions colistin, metronidazole and a combination of the two antibiotics. Over a time period of 21 days of incubation performed in two replicates we were able to acquire a total of 315 strains of which 105 have been used for the phenotypic experiments. Most strains exposed to colistin became highly antibiotic resistant after 21 days of incubation, even though the majority in the combination condition reached this threshold already by day 14. A series of assays were performed to elucidate the changes in bacterial phenotype. Exposure to colistin resulted in irregular alterations in the overall biomass of the biofilm with rising and falling amounts along the incubation, but also a significant increase in viable cells in the biofilm. Correlating the cell density of the biofilm to the level of colistin resistance mainly showed positive correlation, although one exceptional development indicates that exposure to colistin may be key to phenotypic changes not resistance alone. This suggests that the exposure to colistin might lead to changes in the composition in the biofilm with more cells 66 in relatively less biomatrix. While in most cases virulence decreased with colistin exposure, not all strains displayed the same pattern and extent of attenuation, which makes it very difficult to predict the strains ́ development in individual cases. The loss in virulence could not be attributed to an impaired growth potential of the strains, as there was no noticeable difference in growth kinetics before and after the incubation. The addition of metronidazole caused faster antibiotic resistance, without this being statistically verifiable with our analysis and without distinct variations in other phenotypes. Concluding, the morbidostat culture device enabled us to simulate a clinical scenario in which P. aeruginosa strains spread from a compartment within the human body where the dosage of the antibiotic colistin does not exceed sublethal concentrations. The constant antimicrobial pressure led not only to the emergence of colistin resistance, but also to specific alterations in other phenotypic features. In view of our results in this study, we suppose that antibiotic treatment with colistin may result in strain-specific, variable bacterial features in the course of an infection, which might make it extremely difficult to treat the individual infection. Further investigation of these strains’ specific evolutions on a genetic level and the search for potential biomarkers that could help predict the bacterial behavior, could be key to an optimized treatment of a severe infection with P. aeruginosa.