Phenotypic Assessment of Clinical Escherichia coli Isolates Predicts Uropathogenic Potential

Abstract

AbstractFor women in the United States, urinary tract infections (UTI) are the most frequent diagnosis in emergency departments, comprising 21.3% of total visits. Uropathogenic Escherichia coli (UPEC) causes ∼80% of uncomplicated UTI. To combat this public health issue, it is vital to characterize UPEC strains as well as differentiate them from commensal strains to reduce the overuse of antibiotics. Surprisingly, no genetic signature has been identified which clearly separates UPEC from other E. coli. Therefore, we examined whether phenotypic data could be predictive of uropathogenic potential. We screened 13 clinical strains of UPEC, isolated from cases of uncomplicated UTI in young otherwise healthy women, in a series of microbiological phenotypic assays using UPEC prototype strain CFT073 and non-pathogenic E. coli strain MG1655 K12 as controls. Phenotypes included adherence, iron acquisition, biofilm formation, human serum resistance, motility, and stress resistance. These data were able to predict the severity of bacterial burden in both the urine and bladders using a well-established experimental mouse model of UTI. Multiple linear regression using three different phenotypic assays, growth in minimal medium, siderophore production, and type 1 fimbrial expression, was predictive of bladder colonization (adjusted r2=0.6411). Growth in ex vivo human urine, hemagglutination of red blood cells, and motility modeled urine colonization (adjusted r2=0.4821). These results showcase the utility of phenotypic characterization to predict the severity of infection these strains may cause. We predict that these methods will also be applicable to other complex, genetically redundant, pathogens.ImportanceUrinary tract infections are the second leading infectious disease worldwide, occurring in over half of the female population during their lifetime. Most infections are caused by uropathogenic Escherichia coli (UPEC). These strains can commensally colonize the gut, but upon introduction to the urinary tract, can infect the host and cause disease. Clinically, it would be beneficial to screen patient E. coli strains to understand their pathogenic potential, which may lead to the administration of prophylactic antibiotic treatment for those with increased risk. Others have proposed the use of PCR-based genetic screening methods to detect UPEC and differentiate them from other E. coli pathotypes; however, this method has not yielded a consistent uropathogenic signature. Here, we have used phenotypic characteristics such as growth rate, siderophore production, and expression of fimbriae to successfully predict uropathogenic potential.