Collateral Sensitivity and Resistance Networks in Clinical Escherichia coli Isolates

Abstract

Because of an increasing development of antimicrobial resistance, treatment options for bacterial infectious diseases are narrowed. A strategy to reduce the evolution of resistance is to take advantage of antimicrobial collateral networks. In this thesis, we explored these collateral networks with a focus on antimicrobials used in Norway for the treatment of cystitis. Cystitis is the most common outcome of urinary tract infection. Five clinical strains from a collection of pan-susceptible Escherichia coli (E. coli) isolated from uncomplicated urinary tract infection (ECO•SENS) were used. Spontaneous, single step mutants selected on trimethoprim, ciprofloxacin and nitrofurantoin were generated and their susceptibility determined by E-test. They were confirmed to belong to E. coli species by performing random amplified polymorphism DNA polymerase chain reaction (RAPD PCR). The susceptibility of the mutants was thereafter tested towards seven different antimicrobials. These results were compared to the ancestral strains’ susceptibility profiles to the same antimicrobials. This method represented a scenario where the susceptibility of an isolate is decreased towards trimethoprim, ciprofloxacin or nitrofurantoin due to a spontaneous mutation, but at the same time, the mutants’ susceptibility can increase or decrease for a different antimicrobial. The collateral networks of mutants with decreased susceptibility to three antimicrobials were visualized in heat maps, where blue color displays collateral sensitivity and red color indicates drug combinations that show collateral resistance. The use of drug combinations that show collateral resistance should be avoided and instead, the use of drug combinations that display collateral sensitivity should be applied to hinder the increase of antimicrobial resistance. The presented data suggest that acquisition of trimethoprim resistance simultaneously increases susceptibility to aminoglycosides and mecillinam. Moreover, we provide evidence that the temporal order of antimicrobial consumption can be optimized for resistance to trimethoprim and that the use of nitrofurantoin may be carried out with caution.