Nationwide Monitoring of Antimicrobial-Resistant Escherichia coli and Enterococcus spp. Isolated From Diseased and Healthy Dogs and Cats in Japan

The Japanese Veterinary Antimicrobial Resistance Monitoring System (JVARM) was established for nationwide monitoring of antimicrobial-resistant bacteria isolated from animals. Here, antimicrobial resistance of Escherichia coli and Enterococcus spp. isolates from diseased and healthy dogs and cats was investigated. Isolates were collected from diseased dogs and cats and from healthy dogs and cats in 2018 to 2020. Minimum inhibitory concentrations were determined for 1873 E. coli and 1383 Enterococcus spp. isolates. E. coli isolates were most commonly resistant to nalidixic acid [diseased dog (DD), 62.1%; diseased cat (DC), 59.9%; healthy dog (HD), 23.5%; healthy cat (HC, 24.0%] and ampicillin (DD, 54.4%; DC, 64.1%; HD, 28.4%; HC, 25.2%), followed by ciprofloxacin (DD, 45.0%; DC, 44.0%; HD, 12.9%; HC, 10.4%). Enterococcus spp. isolates were most resistant to tetracycline (DD, 66.9%; DC, 67.8%; HD, 47.0%; HC, 52.0%), followed by erythromycin (DD, 43.2%; DC, 46.6%; HD, 27.8%; HC, 34.0%) and ciprofloxacin (DD, 27.9%; DC, 43.7%; HD, 9.7%; HC 12.9%). Only a few E. coli isolates were resistant to colistin and none were resistant to meropenem. Also, none of the Enterococcus spp. isolates we have tested were resistant to vancomycin. The significantly higher resistance rates of E. coli and Enterococcus spp. isolates from diseased, as opposed to healthy, dogs and cats against most of the tested antimicrobials indicates that the use of antimicrobials could select resistant E. coli and Enterococcus spp.


INTRODUCTION
The emergence and spread of antimicrobial-resistant (AMR) bacteria is widely recognized as a global health threat (1). The concept of "One Health" is crucial to address this issue, because humans, animals, foods, and environments are potential reservoirs of AMR bacteria and genes (2). In the veterinary field, globally, livestock is the main target for monitoring and risk assessment of AMR bacteria. The Japanese Veterinary Antimicrobial Resistance Monitoring System (JVARM) was established in 1999 to monitor AMR bacteria isolated from livestock but not companion animals (i.e., dogs and cats) as same as many countries (3). In some countries, including Sweden, France and Norway, monitor AMR bacteria in companion animals (4)(5)(6).
Although AMR is also major concern in companion animals, there are few monitoring results for risk assessment (7). Studies in Japan detected extended-spectrum beta-lactamase-producing Escherichia coli and bla CTX−M type β-lactamase genes in samples collected from dogs and cats (8,9). Additionally, greater rates of enrofloxacin-resistant Enterococcus spp. were confirmed in diseased, as opposed to healthy, dogs, and cats (10). These studies revealed potential treatment failure risks in dogs and cats and transmission risks of AMR bacteria from companion animals to humans and vice versa. Therefore, our group developed a national monitoring system for dogs and cats under the JVARM framework to assess the risks according to the strategy of the first national action plan on AMR in Japan (11).
Monitoring of targeted pathogens isolated from diseased dogs and cats was initiated from the veterinary medical perspective. In addition, as a fundamental data of AMR bacteria, E. coli and Enterococcus spp. from healthy animals is also important to monitor, since E. coli and Enterococcus spp. are globally known as a multi-sectoral indicator bacteria and often selected as target bacteria of AMR monitoring. Also, E. coli and Enterococcus spp. are commensal microorganisms causing opportunistic infections (12,13). Therefore, we collected E. coli and Enterococcus spp. from both diseased and healthy dogs and cats for this monitoring.
Monitoring of both diseased and healthy dogs and cats is almost unprecedented; however, the aim of this study was to summarize the results of nationwide monitoring of AMR E. coli and Enterococcus spp. from diseased and healthy dogs and cats in Japan.

Sampling
In order to determine an appropriate number of samples to avoid bias, all prefectures of Japan were divided into blocks and the numbers of samples were calculated based on the number of small animal clinics in each block. Hence, collection of only one E. coli and/or Enterococcus spp. isolate from each clinic, which was to be in total 200 isolates, suggest that each bacterial species should be collected from dogs and cats per year.
E. coli strains were isolated from clinical urine or genital tract samples and Enterococcus spp. strains were isolated from clinical urine or ear samples of diseased dogs and cats. All of E. coli and Enterococcus spp. isolated from diseased dogs and cats are isolated and identified in clinical laboratories from clinical samples.
In addition, E. coli and Enterococcus spp. were isolated from rectal swabs of healthy dogs and cats brought to small animal clinics for either medical checkups or vaccinations, but not treatment.

Informed Consent for Sampling
Clinical samples from diseased dogs and cats, submitted to clinical laboratories, were sourced from veterinarians and owners and used under the agreement of the use for research.
The owners were explained the purpose of the surveillance and requirement of isolates from healthy dogs and cats, and written informed consent was obtained prior to sample collection.
All of the isolates were anonymized. For healthy dogs and cats, suspected E. coli isolates were tested using DHL agar, then cultured on triple sugar iron agar, and subjected to the IMViC test. Suspected Enterococcus spp. isolates were smeared on enterococcosel agar and three colonies were randomly sampled and subjected to Gram staining, the catalase test and the pyrrolidonyl arylamidase test, and then cultured in heart infusion broth with NaCl. The species of the suspected Enterococcus spp. isolates were identified using the Rapid ID32 Strep kit.

Antimicrobial Susceptibility Testing
The tested antimicrobials were selected from two perspectives, in order to enable comparison with existing data of livestock in JVARM and to grasp resistant rates of antimicrobials frequently used in small animal clinical practices.
The Enterococcus spp. isolates were tested for MICs of ampicillin, gentamicin, erythromycin, azithromycin, ciprofloxacin, chloramphenicol, and tetracycline. MICs of vancomycin was assessed for isolates from diseased dogs and cats collected in 2019 and 2020, and all isolates from healthy dogs and cats.
MICs were calculated using a standardized microdilution method in accordance with the Clinical and Laboratory Standards Institute (CLSI) standard (14) using "Dry Plate 'Eiken' 192" bacterial drug sensitivity testing reagent (EIKEN Chemical Co., Ltd., Japan). The breakpoints as listed in CLSI document M100 (15) and VET01S (16) were applied. Considering the purpose of nationwide continual monitoring, break points set in M100 were adopted instead of break points set by each samples' origins in VET01S. Escherichia coli ATCC 25922, Pseudomonas aeruginosa ATCC 27853, and Enterococcus faecalis ATCC29212 were used as control strains.

Statistics Analysis
The Fisher's exact test was used to identify differences in the resistance rates among diseased and healthy dogs and cats. A probability p < 0.05 was considered statistically significant.

RESULTS
In total, 1,873 E. coli and 1,383 Enterococcus spp. isolates were collected. Eight hundred ninety one E. coli isolates were collected from diseased dogs and cats and 982 from healthy dogs and cats, whereas 695 Enterococcus spp. isolates were collected from diseased dogs and cats and 688 from healthy dogs and cats. All of the isolates were collected in 2018, 2019, and 2020.

DISCUSSION
Overall, resistance to most of the tested antimicrobials was significantly higher in diseased, as opposed to healthy, dogs and cats. In this study, background information of the samples to trace antimicrobial use was limited; however, diseased dogs and cats are more likely to have been treated with antimicrobials. According to Nippon AMR One Health Report (17), the highest volume of veterinary antimicrobials estimated sales for dogs and cats were cephalosporins, especially the first generation cephalosporins followed by penicillins. Conversely, almost none carbapenems were sold. It is may be correlated with resistant rates of E. coli isolated from diseased dogs and cats against cefazolin, cefalexin, cefotaxime, and ampicillin, which were more than 30%, and none were resistant to meropenem. This results indicate that among diseased dogs and cats that were administered antimicrobials, the use of antimicrobials is may be responsible for the selective pressure of bacterial flora of E. coli and Enterococcus spp. In addition to the fact that there were differences in resistance rates among diseased and healthy dogs and cats, there were similarities in the types of antimicrobials they showed resistance to (e.g., E. coli resistance to nalidixic acid and ampicillin; Enterococcus spp. resistance to tetracycline, erythromycin, and ciprofloxacin).
According to data collected by JVARM in 2016 and 2017 (18), E. coli isolates from cattle, pigs and broilers in slaughterhouses were most resistant to streptomycin (19.0-51.3%) and tetracycline (21.0-56.7%), while isolates from diseased livestock were highly resistant to tetracycline (54.5-87.3%) followed by streptomycin (38.9-74.5%), ampicillin (33.3-74.5%) and chloramphenicol (11.1-69.6%). Resistance to cefotaxime and ciprofloxacin was generally lower in E. coli  isolates from healthy than diseased livestock (0.0-5.7 and 2.9-8.9 vs. 0.0-12.0 and 11.1-28.5%, respectively). Notably, <30% of all E. coli isolates from diseased and healthy cattle, pigs and broilers, but >30% of those from diseased dogs and cats were resistant to cefotaxime and ciprofloxacin. In contrast, E. faecalis and E. faecium isolates collected from healthy pigs and broilers in 2017 were highly resistant to oxytetracycline (31.8-84.6%) and erythromycin (27.3-61.5%) (18). Kimura et al. (19), also found similar trends for Enterococcus spp. isolated from diseased companion animals in Japan (non-susceptibility rates against doxycycline and minocycline of 40-56% and erythromycin of 40-93%). Data collected by the Japan Nosocomial Infections Surveillance System (20) indicate that trends in E. coli in human resistance to penicillins, quinolones/fluoroquinolones, and cephalosporins are similar to those of dogs and cats, in addition to high resistance of Enterococcus spp. to tetracycline and erythromycin.
Although the trends in resistance of Enterococcus spp. concurred with those of all animal species considered in this study, there were notable differences in the AMR profiles of E. coli isolates from humans, dogs and cats vs. livestock, which may be due to differences in the antimicrobial classes used for treatment of infections.
Consistent with previous studies, most E. faecium and few E. faecalis isolates from diseased dogs and cats were resistant to ampicillin (21). However, interestingly, the resistance rates of E. faecium from healthy dogs and cats were significantly lower than from diseased dogs and cats (22.2 and 9.1 vs. 91.0 and 92.1%, respectively), even in each years. Jackson et al. reported that 47.4 and 51.6% of E. faecium isolates from healthy dogs and cats in the United States were resistant to penicillin (22), which are not exceptionally high resistance rates. Origins of the human E. faecium infection are broadly categorized as hospital or community-associated (23,24). Most cases of hospital-associated E. faecium are resistant to ampicillin while community-associated cases are generally susceptible to ampicillin, suggesting the possibility of similar types of E. faecium infection of dogs and cats.
As shown in Figures 1, 2, there were small annual fluctuations with regular, rather than random patterns. Despite collecting only 3 years of data, this monitoring system seemed to be useful to illustrate trends. Although there were differences in the resistant rates between diseased and healthy, resistant rates between animal species, dogs and cats, were similar in both E. coli and Enterococcus spp. isolates.
Few countries monitor AMR bacteria for dogs and cats. Swedish Veterinary Antimicrobial Resistance Monitoring (SVARM) program (4), refers to MICs of E. coli from clinical urine samples of dogs and cats. In the SVARM report, 13% of E. coli isolates from dogs and 16% from cats were resistant to ampicillin. In France, the national surveillance network for antimicrobial resistance in bacteria from diseased animals (RESAPATH) (5), collects data via the disc diffusion method of E. coli isolates from clinical samples of dogs and cats with various pathologies. According to the RESAPATH data, 70, 61, and 73% of E. coli isolates from kidney/urinary tract, Frontiers in Veterinary Science | www.frontiersin.org skin/soft tissue, and otitis samples, respectively, from dogs and 70% from cats with all pathologies were susceptible to amoxicillin. Furthermore, according to the national monitoring program data for antimicrobial resistance in the veterinary and food production sectors in Norway (NORM-VET) (6), 20.2 and 46.5% of E. coli isolates from urine of dogs with urinary tract infections and other infections, respectively, were resistant to ampicillin. It should be noted that the SVARM data includes ECOFFs as defined by the EUCAST, while the RESAPATH system uses the NF U47-107 standard of the Antibiogram Committee of the French Society of Microbiology, and the NORM-VET program uses the clinical breakpoints or ECOFFs.
Each of the cited surveillance systems adopted different breakpoints and there were differences in the numbers and types of tested antimicrobial agents as well as the sampling methods (passive sample collection method of the SVARM, RESAPATH and NORM-VET systems vs. an active sample collection method in the present study). Therefore, data comparisons were challenging. In our study, 54.4 and 64.1% of the E. coli isolates from diseased dogs and cats, respectively, were resistant to ampicillin, which were higher rates than reported in Sweden, France and Norway, where aggressive measures against AMR have been enacted.
In the present study, the resistance rates of E. coli and Enterococcus spp. isolates from diseased dogs and cats were significantly higher than those from healthy dogs and cats against most of the tested antimicrobials, indicating that use of antimicrobials could be selective pressure for resistant E. coli and Enterococcus spp.
The limitation of our study lies on difference of sample origins between diseased and healthy dogs and cats, urine/genital/ear sample origin vs. rectal swab origin due to collect enough numbers of strains. Also, there are lack of genetic data including serotype. However, our results are valuable to know AMR situation in dogs and cats in Japan and useful to consider AMR measures.
In conclusion, this is the huge step toward continued AMR monitoring of isolates from diseased and healthy dogs and cats. To the best of our knowledge, nationwide monitoring systems of AMR bacteria isolated from both diseased and healthy dogs and cats are rare. In 2020 (25), our group published guidelines for use of antimicrobials for companion animal veterinarians to avoid selection of AMR bacteria. The results of this study suggest that it is crucial to promote prudent use of antimicrobials in companion animals and to continue monitoring trends in AMR bacteria.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

AUTHOR CONTRIBUTIONS
YF, HS, YS, RA, MO, MM, SH, and MKa contributed to conception and design of the study. MU, MKu, and TS organized the sampling methods and approached to the related organizations. All authors contributed to manuscript revision, read, and approved the submitted version.