Abstract
The rise in antibiotic resistance among bacterial pathogens has prompted the exploitation of alternative antibacterial strategies, such as antivirulence therapy. By inhibiting virulence traits, antivirulence drugs are expected to lessen pathogenicity without affecting bacterial growth, therefore avoiding the spread of resistance. However, some studies argued against this assumption, and the lack of antivirulence drugs in clinical use hampers the empirical assessment of this concept. Here we compared the mode of action and range of activity of two drugs which have been proposed for repurposing as quorum sensing and pyoverdine inhibitors in the human pathogen Pseudomonas aeruginosa: the anticancer drug 5-fluorouracil (5-FU) and the antimycotic drug 5-fluorocytosine (5-FC), respectively. The effect on bacterial growth, emergence and spread of resistance, and activity against clinical isolates were assessed. Our results confirm that 5-FU has growth inhibitory activity on reference strains and can rapidly select for spontaneous resistant mutants with loss-of-function mutations in the upp gene, responsible for uracil conversion into UMP. These mutants were also insensitive to the anti-pyoverdine effect of 5-FC. Conversely, 5-FC did not cause relevant growth inhibition, likely because of poor enzymatic conversion into 5-FU by cytosine deaminase. However, coculturing experiments showed that 5-FU resistant mutants can outcompete sensitive cells in mixed populations, in the presence of not only 5-FU but also 5-FC. Moreover, we observed that serial passages of wild-type cells in 5-FC-containing medium leads to the appearance and spread of 5-FC insensitive sub-populations of 5-FU resistant cells. The different effect on growth of 5-FU and 5-FC was overall conserved in a large collection of cystic fibrosis (CF) isolates, corresponding to different infection stages and antibiotic resistance profiles, although high variability was observed among strains. Notably, this analysis also revealed a significant number of pyoverdine-deficient isolates, whose proportion apparently increases over the course of the CF infection. This study demonstrates that the efficacy of an antivirulence drug with no apparent effect on growth can be significantly influenced by the emergence of insensitive mutants, and highlights the importance of the assessment of resistance-associated fitness cost and activity on clinical isolates for the development of "resistance-proof" antivirulence drugs.