Prevalence and mechanisms of fluoroquinolone-resistant Escherichia coli among sheltered companion animals

To better understand the prevalence of fluoroquinolone-resistant Escherichia coli among sheltered companion animals, we conducted a screening study of 38 dogs and 78 cats and investigated the resistance mechanisms and characteristics of the isolates. Fluoroquinolone-resistant E. coli was detected in 18 dogs (47.4 %) and 14 cats (17.9 %). The isolates carried one to four mutations in the gyrA, parC and parE genes of the quinolone resistance-determining region, and the number of mutations was proportional to the MIC for ciprofloxacin. For plasmid-mediated quinolone resistance, aac-(6′)-Ib-cr was detected in nine isolates, qnrS in five isolates and qnrB in one isolate. A relationship between the presence of these genes and MIC for ciprofloxacin was not apparent. Statistical analysis indicated that fluoroquinolone-resistant E. coli was widely distributed among sheltered companion animals with various attributes. This may relate to the wide dissemination of fluoroquinolone resistance among humans and other animals in Japan.


INTRODUCTION
Fluoroquinolones are broad-spectrum and effective bactericidal antimicrobials that have been approved for use in the treatment of various bacterial infections. Recently, fluoroquinolone resistance has increased in the clinical field nationwide [1]. The Japan Nosocomial Infections Surveillance (JANIS) reported that 60.2 and 59.0 % of Escherichia coli isolates from inpatient specimens obtained in 2015 and 2016, respectively, were resistant to levofloxacin [2]. The main fluoroquinolone resistance mechanism is the chromosomal mutation of quinolone resistance-determining region (QRDR) in DNA gyrase and topoisomerase IV. Another mechanism, plasmidmediated quinolone resistance (PMQR), was also found to contribute to quinolone resistance [3][4][5].
We have previously reported the prevalence of cephalosporinresistant Enterobacteriaceae among sheltered companion animals; 14.6 % of dogs and 11.0 % of cats harboured cephalosporin-resistant bacteria, and 33 % of these isolates were also resistant to fluoroquinolone. Based on the distribution of several types of antimicrobial-resistant bacteria among animals with a wide range of individual attributes, we concluded that companion animals may play a bridging role in the circulation of antimicrobial-resistant bacteria from humans and other origins [6]. Therefore, surveillance of sheltered companion animals may be useful in determining the prevalence of antimicrobial-resistant bacteria.
Several studies have demonstrated the presence and characteristics of fluoroquinolone-resistant E. coli among companion animals [7][8][9][10][11]. However, most of them included results from clinical isolates obtained from animal hospitals and did not aim to demonstrate the prevalence of fluoroquinoloneresistant bacteria among a wide range of companion animals.
Based on the above background, we first established a means to understand the prevalence of fluoroquinolone-resistant E. coli among diverse dogs and cats, and then screened a number of sheltered animals and investigated the resistance mechanisms and genetic characteristics of the isolates.

MeTHODS
Faecal samples or rectal swabs were collected between April 2016 and January 2017 from 38 dogs and 78 cats brought to Osaka Municipal Animal Care and Control Center. The sampling and classifying of animals by individual attributes were done as in our previous report [6]. Samples were inoculated on MacConkey agar (Nissui Pharmaceutical), supplemented with ciprofloxacin at 4 mg l −1 , and incubated at 37 °C for 20 h. Subculturing was performed on up to five E. coli-like colonies. Bacterial species were identified by biochemical tests with an API 20E identification kit (bioMérieux). E. coli O-serotyping, E. coli phylogenetic grouping and PFGE typing were carried out as in our previous report [6]. MICs were determined via a Dry Plate EIKEN (Eiken Chemical) and ETEST (bioMérieux) according to the manufacturers' instructions. The results were interpreted according to the Clinical and Laboratory Standards Institute (CLSI) guidelines [12]. Mutations in QRDR genes were confirmed by PCR amplification and sequencing [13]. PMQR genes were screened by PCR [14][15][16]. The β-lactamaseproducing phenotype and β-lactamase genes were detected as in our previous report [6].
When several isolates were cultured from the same animal, they were designated as the same clone if they exhibited the same PFGE patterns. The correlation between the prevalence of fluoroquinolone-resistant E. coli and individual attributes was estimated using Fisher's exact test. A P value of <0.05 was considered significant.

ReSUlTS
The correlation between individual animal attributes (sex, age group, presence/absence of an owner, health status and enrofloxacin medication treatment) and the prevalence of fluoroquinolone-resistant E. coli is presented in Table 1.
Fluoroquinolone-resistant E. coli was detected in 18 out of 38 dogs (47.4 %) and 14 out of 78 cats (17.9 %). Statistical analysis showed that prevalence among dogs was significantly higher than among cats (P=0.0017), and prevalence in males was significantly higher than in females among both dogs (P=0.0064) and cats (P=0.0088). No significant correlations were found relating to the presence/absence of an owner or treatment with enrofloxacin medication in the animal shelter.

Impact Statement
Fluoroquinolones are broad-spectrum antimicrobials that are useful for the treatment of bacterial infections. However, fluoroquinolone resistance has increased in the clinical field. This paper first clarified that fluoroquinolone-resistant Escherichia coli was widely distributed among sheltered companion animals with a wide range of individual attributes. This finding is important with regard to the circulation of fluoroquinolone resistance among humans, among animals and in the environment. It may have considerable influence on the treatment of infectious disease in both animals and humans given that 46.9 % of isolates were classified as multidrug-resistant bacteria. In order to lower the overall level of fluoroquinolone resistance, it will be necessary to reduce fluoroquinolone use and continue surveillance not only in the human clinical field but also in the veterinary field.  a, MICs for ciprofloxacin, nalidixic acid and cefotaxim were confirmed with an E-test. b, Mutation in gyrB was not found. c, PMQR: qnrA [14], qnrB [14], qnrC [14], qnrD [16], qnrS [14], qepA [14], oqxA [15], oqxB [15] and aac(6')-Ib-cr [14] were tested. d, Prime symbols (', '', ''', '''' indicate >80 % similarity between band patterns. e, Negative for the shiga toxin gene. f, β-Lactamase phenotype and β-lactamase genes were not detected.  The resistance mechanisms and characteristics of 32 fluoroquinolone-resistant E. coli isolates are given in Table 2. All isolates showed MICs ≥256 µg ml −1 for nalidixic acid and demonstrated at least one mutation in the QRDR. Four isolates with MICs of 2-3 µg ml −1 for ciprofloxacin harboured a single mutation in gyrA. Fourteen isolates with MICs of 3-24 µg ml −1 for ciprofloxacin carried two or three mutations in the QRDR: the combination of a double mutation (except for isolate 37E-1) in gyrA and a single mutation in parC. Fourteen isolates with MICs ≥32 µg ml −1 for ciprofloxacin carried a total of four mutations in the QRDR: the combination of a double mutation in gyrA, a single mutation in parC and a single mutation in parE; or the combination of a double mutation in gyrA and a double mutation in parC. For PMQR, nine isolates harboured aac-(6′)-Ib-cr, five isolates harboured qnrS and one isolate harboured qnrB. The qnrS gene was always harboured simultaneously with aac-(6′)-Ib-cr. A relationship between the presence of PMQR and ciprofloxacin MIC values was not apparent. There were seven isolates resistant to cefotaxime and six isolates produced β-lactamase from the various β-lactamase gene(s). Extended-spectrum β-lactamase (ESBL) genes (CTX-M-14 or CTX-M-27) were harboured simultaneously with aac-(6′)-Ib-cr, while the DHA-1 gene was harboured with qnrB. For the O-serotype, most of the E. coli isolates were untyped, except for serotypes O1, O25 and O157, each from one isolate. For phylogenetic grouping, three isolates belonged to phylogenetic group B2, and 11 isolates belonged to group D, while the remaining isolates belonged either to group A or to group B1.
PFGE and a dendrogram analysis were performed to assess genetic relationships (Fig. 1). There were 17 distinct patterns with <80 % similarity from the 32 fluoroquinolone-resistant E. coli isolates. The similarity of band patterns ranged from 14.3 to 100 %. There were five PFGE types that indicated >80 % similarity, for a total of 20 isolates ( Table 2, Fig. 1).  [9]. In the present study, the use of enrofloxacin in the animal shelter was not related to possessing the bacteria, although a full medication history was not determined.

DISCUSSION
Co-resistance to cephalosporin was observed in 21.9 % (7/32) of fluoroquinolone-resistant isolates. These included four ESBL-producing and two AmpC-producing isolates. Moreover, 68.8 % (22/32) of fluoroquinolone-resistant isolates were also resistant to other categories of antimicrobials. Fluoroquinolone-resistant E. coli with co-resistance to other antimicrobials, especially cephalosporin, is prevalent nationwide [3,4], and has also been detected among clinical E. coli isolated from companion animals [10,17]. According to the international standard definition for multidrug-resistant (MDR) bacteria [18], 15 isolates (46.9 %) were classified as MDR with non-susceptibility to three or more antimicrobial categories. These findings may have great relevance for the treatment of infectious diseases in both animals and humans.
For QRDR, all isolates harboured a mutation in gyrA at Ser83, the most common mutation site. Highly resistant isolates carried a combination of mutations within gyrA and parC, and the number of mutations was proportional to the MIC for ciprofloxacin. In line with previous reports regarding fluoroquinoloneresistant E. coli from companion animals [7,10], mutations in QRDR genes play a significant role in mediating resistance to fluoroquinolone. For PMQR genes on transferable plasmids, which confer a lower level of resistance to fluoroquinolone than QRDR [4,5], aac-(6′)-Ib-cr was the most prevalent, followed by qnrS and qnrB. The predominance of the aac-(6′)-Ib-cr gene and qnr variant genes agrees with the previous reports of E. coli isolates recovered from companion animals in the USA [10] and Europe [7,8]. This trend has been confirmed in human clinical isolates as well [3]. Furthermore, many different PMQR genes have also been found in E. coli isolates from Chinese companion animals [11].
PFGE patterns were widely diverse, but some isolates showed the same or similar band patterns. Although these animals were from separate locations, cross-contamination may have occurred on admittance to the animal shelter or may also have been caused by the movement of staff or materials.
In conclusion, we found that fluoroquinolone-resistant E. coli was widely distributed among sheltered companion animals with a wide range of individual attributes. This may be related to the dissemination of fluoroquinolone resistance among Japanese human clinical E. coli [2], livestock animals [19] and different environments [20]. To lower the overall fluoroquinolone resistance, it will be necessary to reduce fluoroquinolone use and continue surveillance not only in the human clinical field but also in the veterinary field.

Consent for publication
This paper does not include details, images or videos relating to an individual person. Therefore, informed consent is not necessary.

Funding information
This work was partially supported by JSPS KAKENHI, Grant Number 16H07501.
Five reasons to publish your next article with a Microbiology Society journal