An investigation of antimicrobial susceptibility and genotypes of Mycoplasma bovis isolates derived from western Canadian feedlot cattle

Mycoplasma bovis poses a significant threat to the Canadian beef industry, particularly the feedlot sector where it is associated with bovine respiratory disease (BRD) and chronic pneumonia and polyarthritis syndrome (CPPS). Furthermore, its characteristic chronic infections which are refractory to antimicrobials results in animal welfare and economic concerns. Currently, antimicrobials are the primary therapeutic option for treatment and control of M. bovis infections due the absence of a vaccine. This is salient given the global concern regarding antimicrobial use (AMU), antimicrobial resistance (AMR), and increased levels of AMR reported for M. bovis worldwide. Due to the current reliance on antimicrobials for prevention, control, and treatment, continual surveillance of antimicrobial susceptibility is crucial for antimicrobial stewardship and therapeutic treatment of M. bovis related disease. Not only is there a reduced arsenal of antimicrobials for the prevention, control and treatment of M. bovis, but its fastidious nature and difficulties associated with culturing emphasize a need for antimicrobial susceptibility testing (AST) using rapid, accurate, molecular methods. Whole genome sequencing (WGS) provides a plethora of information that can be interrogated to investigate different aspects of M. bovis pathogenesis, including antimicrobial susceptibility and virulence, such as multilocus sequence typing (MLST) for classifying bacterial strains. In conclusion, an epidemiological, genotypic and phenotypic investigation into the antimicrobial susceptibility of M. bovis will contribute to the body of knowledge needed to assist in evidence-based decisions for the treatment and control of M. bovis infections in cattle and provide the rationale for the studies outlined in this thesis. M. bovis isolates derived from western Canadian feedlot cattle sampled over a 12-year period (2006 - 2018) were used to address three objectives: 1), describe the AMU and AMR profiles of isolates derived from M. bovis mortalities in feedlot cattle (Chapter 2); 2), investigate the genotypic basis for macrolide resistance by assessing single nucleotide polymorphisms (SNPs) in 23S rRNA gene alleles and ribosomal proteins L4 and L22 (Chapter 3); and 3), assess the application of four genotyping methods for M. bovis (Chapter 4). M. bovis was cultured from deep nasopharyngeal swabs as well as lung and joint tissue from western Canadian feedlot cattle (cattle, n = 134; isolates, n = 183). Antimicrobial susceptibility testing (AST) was performed using a microbroth dilution assay and a customized panel of nine antimicrobials, representing four drug classes, most commonly administered to feedlot cattle in western Canada. Furthermore, M. bovis isolates (n = 129) underwent WGS utilizing Illumina technology. Although M. bovis isolates were derived from western Canadian feedlot cattle, Chapter 2 was not intended as a representative study of western Canadian feedlots but rather to provide context on AMR and AMU in feedlots and the background of the cattle and isolates represented in this thesis. In chapter 2, over 90% of cattle had received antimicrobial metaphylaxis, with tulathromycin accounting for 94.2 % of treatments. On average, cattle received three antimicrobial classes prior to dying of a mycoplasma-related pneumonia. The most commonly administered classes were macrolides (93.2%), phenicols (78.4%), and fluoroquinolones (67.6%). Isolates had the least resistance to florfenicol, with 89.9% classified as susceptible. Nearly all isolates were resistant to all five macrolides (gamithromycin, tildipirosin, tilmicosin, tulathromycin, tylosin) assessed. The study described in Chapter 3 found that mutations in both domains II and V of the 23S rRNA gene alleles were found to be associated with resistance to all five macrolides. Isolates with a mutation in domain II and the L4 and L22 ribosomal proteins were also resistant to all macrolides, except tulathromycin. Lastly, in Chapter 4, four in silico genotyping methods were applied to M. bovis isolates and the Simpson’s Diversity Index (D) was used to assess resolution of each method. MLST had the lowest resolution (D=0.932) but was the easiest to implement and apply; contrastingly, whole genome single nucleotide variant (wgSNV) yielded the highest resolution (D=1.000), but also involved the most complicated analysis. Application of core genome MLST (cgMLST) and core genome SNV (cgSNV) had a similar resolution of 0.987 and 0.984, respectively. No association between genotype and phenotype was resolved. Overall, M. bovis isolates in western Canadian feedlot cattle were predominately susceptible to a single antimicrobial, florfenicol, and commonly administered macrolides for BRD metaphylaxis. AMR was observed to all macrolides tested, and the accumulation of SNPs in genes associated with macrolide resistance correlated to a macrolide resistant phenotype using AST. Despite being unable to associate genotype and phenotype, the typing methods yielded comparable phylogenetic relationships. Furthermore, the diversity of strain types highlighted the structure of the Canadian cattle industry and how cattle are procured for western Canadian feedlots with some strain types being dominant over geographical areas and time