Genomics accurately predicts antimicrobial resistance in Staphylococcus pseudintermedius collected as part of Vet-LIRN resistance monitoring

Abstract

Whole-genome sequencing (WGS) has changed our understanding of bacterial pathogens, aiding outbreak investigations and advancing our knowledge of their genetic features. However, there has been limited use of genomics to understand antimicrobial resistance of veterinary pathogens, which would help identify emerging resistance mechanisms and track their spread. The objectives of this study were to evaluate the correlation between resistance genotypes and phenotypes for Staphylococcus pseudintermedius, a major pathogen of companion animals, by comparing broth microdilution antimicrobial susceptibility testing and WGS. From 2017-2019, we conducted antimicrobial susceptibility testing and WGS on S. pseudintermedius isolates collected from dogs in the United States as a part of the Veterinary Laboratory Investigation and Response Network (Vet-LIRN) antimicrobial resistance monitoring program. Across thirteen antimicrobials in nine classes, resistance genotypes correlated with clinical resistance phenotypes 98.4 % of the time among a collection of 592 isolates. Our findings represent isolates from diverse lineages based on phylogenetic analyses, and these strong correlations are comparable to those from studies of several human pathogens such as Staphylococcus aureus and Salmonella enterica. We uncovered some important findings, including that 32.3 % of isolates had the mecA gene, which correlated with oxacillin resistance 97.0 % of the time. We also identified a novel rpoB mutation likely encoding rifampin resistance. These results show the value in using WGS to assess antimicrobial resistance in veterinary pathogens and to reveal putative new mechanisms of resistance.

Keywords: Antimicrobial resistance; Genomics; Staphylococcus pseudintermedius.

Fig. 1.

Resistance prevalence determined by phenotypic testing across fifteen drugs among S. pseudintermedius. Prevalence was based on phenotypic testing. Not every isolate was tested for susceptibility to each drug, which explains some slight differences in resistance prevalence among drugs in the same class. Table 2 shows the number of isolates with MIC data for each drug; see Table S1 for isolate-level details.

Fig. 2.

Prevalence of resistance genes by number of drug classes for individual isolates. The number of drug classes refers to all drug classes for which we had resistance mechanisms (Table 1). All isolates had resistance genes and mutations expected to confer resistance for zero to nine drug classes.