Antimicrobial resistance in companion animal pathogens in Australia and assessment of pradofloxacin on the gut microbiota

Abstract

Antimicrobials are not only considered as beneficial agents to human and animal health, but are also potentially harmful, both in terms of promoting antimicrobial resistance (AMR), and the potentially negative effects they may have on the gut microbiota. Focusing on these two major detrimental effects, this thesis aims to determine the frequency of AMR in companion animal pathogens isolated from Australia, as well as identify risk factors associated with infection my multidrug-resistant strains; and to describe the gut microbial community changes and other adverse effects in rabbits in response to administration of pradofloxacin, a new veterinary antimicrobial belonging to the critically important fluoroquinolone class. In the first nation-wide survey of AMR in companion animal pathogens, resistance to antimicrobials from more than ten different classes was described for a total 883 clinical Escherichia coli isolated from dogs (n=514), cats (n=341) and horses (n=28). Resistance to critically important antimicrobials not registered for use in animals (imipenem) was not detected in all clinical E. coli isolates. Resistance to amikacin, another critically important antimicrobial not registered for use in Australian animals, was low in dog and cat isolates (<2%) but moderately high in horse isolates (10.7%), reflecting its increased use in this species. Resistance to other third line antimicrobials such as third generation cephalosporins (3GC) and fluoroquinolones (FQN), was generally lower in cat isolates (3.8%-5%) compared with dog isolates (9.1%-10.9%) and horse isolates (25%-35.7%). A higher proportion of resistance was observed among the first and second line antimicrobials, ranging from 25%-100% for ampicillin and amoxicillin-clavulanic acid, mainly due to the new interpretation by veterinary-specific clinical breakpoints for dogs and cats which are much lower than human clinical breakpoints. The frequency of multidrug-resistant (MDR) E. coli was 18.1%, 11.7% and 42.9% in dog, cat and horse isolates respectively, which was positively associated with the chronicity of infection and prior antimicrobial treatment, in particular for urinary tract infections from dogs. Furthermore, among coagulase-positive staphylococci (CoPS) isolates from companion animals the frequency of methicillin-resistant Staphylococcus pseudintermedius (MRSP) and methicillin-resistant Staphylococcus aureus (MRSA) in the isolate collections was 11.8% (74/629) and 12.8% (15/117), respectively. Resistance to FQN in S. pseudintermedius ranged from 8.1%-8.8% and was highly associated with MRSP isolates from dogs (OR 287; 95%CI 91.2–1144.8). MRSA isolates were also co-resistant to FQN (OR 5.4, 95%CI 0.6–252.1), with a frequency ranging from 8.5%-11.8% of total S. aureus from dogs (n=47) and cats (n=14). By contrast MRSA isolates from horses, were most likely to be co- resistant to amikacin and rifampicin (OR 6.5, 95%CI 0.7–315.2) with a frequency of 9.4% of total S. aureus from horses (n=53). A risk factor analysis showed that MRSP isolates from dogs were significantly more likely in surgical site infections (SSI) and skin and soft tissue infections (SSTI), particularly if the animals had received prior antimicrobial treatment. Compared with other countries where similar studies have been undertaken, rates of AMR in E. coli, were generally lower. However, due to the abrupt emergence and spread of MRSP infections in Australia, rates of AMR in S. aureus and S. pseudintermedius were moderately higher. Faecal and gut microbiota profiling using V3-V4 16S rRNA gene Illumina MiSeq DNA sequence analysis confirmed that the veterinary fluoroquinolone pradofloxacin (which has a much broader spectrum of activity compared to earlier generations) had quite different effects on the microbial community in the rabbit stomach and caecum; compared to hard faeces, and soft faeces. Microbial richness and diversity decreased significantly in hard faeces at the end of the 3- day treatment but not in the lumen of the stomach and caecum. Pseudomonas spp. was depleted significantly in hard faeces while Anaeroplasma significantly diminished in the stomach and caecum. The abundance of several bacterial taxa from the Ruminococcaceae and Lachnospiraceae was significantly overrepresented in all samples. However, overproliferation of bacteria causing enteritis such as Clostridium spp. and coliforms was not detected in the two gut compartments nor hard and soft faeces. Further, we hypothesized that stomach microbiota would be less diverse compared with caecum microbiota but our data show that the stomach harboured higher levels of microbial diversity compared with the caecum, regardless of pradofloxacin treatment, most likely due to coprophagy. Although the short-term use of pradofloxacin appears safe in rabbits as we observed no clinically adverse effects, these results confirm that pradofloxacin did result in significant disturbance of the faecal microbial community and changes of metabolic functional diversity. In conclusion, the first nation-wide AMR survey provides an important snapshot of the current situation of AMR in companion animal pathogens in Australia. Further, this thesis underlined the public health (AMR) and individual health-related (dysbiosis) consequences which should increase the awareness of prudent use of antibiotics especially in animals and broadly in human medicine.