Biomarkers for Highly Virulent Pseudomonas aeruginosa Strains

A substantial amount of effort has been devoted to the identification of host factors that might allow prediction of clinical outcomes in patients with bacterial infections, and sophisticated models and scoring systems have been developed for this purpose. These models, however, do not account for possible differences in virulence between the bacterial strains causing the infection. It stands to reason that if some strains within a particular bacterial species are more virulent than others, patients infected with these strain would be at increased risk for poor outcomes. We have used the gram-negative bacterium Pseudomonas aeruginosa (PA) as an example to investigate these questions. We used a mouse model of bacteremia to measure the 50% lethal dose (LD50) values of 100 PA isolates from patients with bloodstream infections. These LD50 values differed by three orders of magnitude. We then sequenced the genomes of all 100 PA isolates and used comparative genomic approaches to identify 15 accessory genomic elements (AGEs – DNA sequences found in some strains within a species but not others) over-represented in the highly virulent isolates. We individually deleted each of these15 AGEs and compared the virulence of the parental and deletion strains in the mouse model of bacteremia. We found that 13 of the 15 AGEs contributed to virulence and thus encoded factors important for pathogenesis. Our hypothesis is that AGEs such as these serve as biomarkers for PA strains capable of causing especially severe infections in people. To test this hypothesis, we will determine whether PA AGEs that cause increased virulence in mice are also associated with worse outcomes in people. We will also perform a second approach that does not depend on mouse model results but relies only on outcomes in patients. In this unbiased comparative genomics approach, we will identify all AGEs in a large collection of PA isolates from patients with bacteremia and determine which of these AGEs are associated with increased mortality. Findings from both approaches will be confirmed in the second cohort of patients with PA bacteremia. Completion of this proposal will yield a set of AGEs that are associated with PA strains capable of causing especially severe infections associated with high mortality rates. Such information will be useful in identifying patients at high risk for poor outcomes and will allow clinicians to target these patients with particularly aggressive therapies.