Companion Animals as Potential Reservoirs of Antibiotic Resistant Diarrheagenic Escherichia coli in Shandong, China

Antibiotic resistance genes of Escherichia coli (E. coli) from companion animals were still poorly understood. Here, we investigated the extended-spectrum β-lactamases (ESBLs) resistance genes of E. coli from companion animals in Shandong, China. A total of 79 isolates (80.6%) were recovered from 98 healthy or diarrheal companion animals in 2021, among which ESBLs-producing isolates accounted for 43.0% (34/79), and more than half of ESBL E. coli (ESBL-EC) strains (n = 19) were isolated from healthy companion animals. Diarrheagenic E. coli isolates (45.6%, n = 36) were represented by enterotoxigenic (ETEC) (32.9%), enteropathogenic (EPEC) (10.1%) and enteroinvasive (EIEC) (2.6%), 20 isolates of which were from healthy pets. Among tested antibiotics, resistance to tetracycline (64.6%) was the most commonly observed, followed by doxycycline (59.5%) and ampicillin (53.2%). Notably, all isolates were susceptible to meropenem. The multidrug-resistant (MDR) rate was 49.4%, 20 isolates of which were ESBLs producers; moreover, 23.4%, 16.4% of ESBL-EC strains were resistant to 5 or more, 7 or more antibiotics, respectively. Among the 5 β-lactamase resistance genes, the most frequent gene was blaCTX-M (60.76%), followed by blaSHV (40.51%). The plasmid-mediated quinolone resistance (PMQR) gene aac(6’)-Ib-cr was detected in 35 isolates. Additionally, ESBL-associated genes (i.e., blaCTX-M, blaSHV) were found in 76.5% ESBL-EC strains, with six isolates carrying blaCTX-M and blaSHV. The marker gene of high-pathogenicity island gene irp2 (encoding iron capture systems) was the most frequency virulence gene. Our results showed that ESBL-EC were widespread in healthy or diarrhea companion animals, especially healthy pets, which may be a potential reservoir of antibiotic resistance, therefore, enhancing a risk to public and animal health.


Introduction
Escherichia coli (E. coli), a member of the Enterobacteriaceae family, widely distributes in healthy humans and animals [1]. Most E. coli strains colonize harmlessly in the intestines and rarely cause disease in healthy individuals; however, some pathogenic E. coli strains may lead to diarrhea in both humans and companion animals [2,3]. According to the virulence characteristics and clinical symptoms of diarrheagenic Escherichia coli (DEC), they have been further divided into five categories [4]: enteroinvasive (EIEC), enterotoxigenic (ETEC), enteropathogenic (EPEC), enterohemorrhagic (Shiga toxin-producing) (STEC), and enteroaggregative (EAEC). EPEC and ETEC strains have been reported from healthy or diarrhea companion animals [2,5,6]. EIEC are studied mainly in human infections [7,8], but studies in companion animals are limited.
The emergence of antimicrobial resistance in bacteria is an ongoing severe public health problem and results in 700,000 deaths annually [9]. Antibiotic misuse has led to the increasing emergence of multidrug-resistant (MDR) and pan-resistant strains in humans, companion animals, water and food animals [10]. Companion animals are kept for company, entertainment, or compassion for humans. In China, 55.03 million dogs and 44.12 million cats were kept in cities in 2019. The close contact between companion animals and humans enhances the risk of bacterial and even antimicrobial-resistant bacterial transmission across animals or humans by horizontal transfer and clonal spread [11]. Previous studies have shown that antimicrobial-resistant E. coli are frequently isolated from dogs and cats [12,13].
The Gram-negative bacterium E. coli is a common member of the microbiota of the lower intestine of mammals and, to a lesser extent, birds. E. coli is also considered to be the most important reservoir of resistance genes that may be accountable for treatment failures in both human and veterinary medicine [14,15]. In addition, most antibiotic classes used to treat colibacillosis are shared between veterinary medicine and human medicine in the community, regardless of whether these are first-generation antibiotics or critically important antibiotics [16]. Over the past two decades, there has been a significant number of infections caused by E. coli carrying extended-spectrum β-lactamases (ESBLs) [17]. ESBLs are becoming more common because this phenotype is being selected for by the use of and exposure to β-lactams [18]. Infection with ESBL E. coli (ESBL-EC) is increasingly associated with overt infections in humans and companion animals worldwide [19,20]. ESBL-EC has been identified in the faeces of healthy cats and dogs in a number of studies [21,22].
Virulence factors are usually expressed proteins encoded by genes located in the chromosome or in plasmids. Morato et al. [23] showed that Escherichia coli of companion animals is an important source of zoonotic infection. Therefore, further investigations are needed regarding the distribution of virulence genes of E. coli isolates from companion animals and their capacity as reservoirs of antimicrobial resistance genes.
Currently, national monitoring programs solely focus on the prevalence of bacteria in food animals, while there is little investigation of the risk of antibiotic resistance switching between humans and companion animals [24]. The prevalence of ESBL-EC in companion animals has been brought to the attention of the scientific community [25,26]; thus, this study aims to estimate antimicrobial resistance in E. coli from companion animals in Shandong, China.

Concordance of Genotypic-Phenotypic Antimicrobial Resistance
The concordance of genotypic and phenotypic resistance is summarized in Table 3. We explored the correlation of genotypic-phenotypic antimicrobial resistance of the isolates. Interestingly, a comparatively stronger correlation was found between ceftriaxone and blaCTX-M among the strains (p = 0.010). Additionally, a significant correlation was also observed between phenotypic cephalexin resistance and blaOXA gene among isolates (p = 0.007). On the other hand, no statistically significant difference was observed between gentamicin phenotype and aaC4 resistance gene among strains (p > 0.05). On the other hand, all 61 tetracyclines resistant isolates did not harbor tetA and tetB genes encoding tetracycline resistance. Table 3. Correlation matrix between resistance phenotype and genotype.

Antibiotics
Characteristics of Strains

Concordance of Genotypic-Phenotypic Antimicrobial Resistance
The concordance of genotypic and phenotypic resistance is summarized in Table 3. We explored the correlation of genotypic-phenotypic antimicrobial resistance of the isolates. Interestingly, a comparatively stronger correlation was found between ceftriaxone and bla CTX-M among the strains (p = 0.010). Additionally, a significant correlation was also observed between phenotypic cephalexin resistance and bla OXA gene among isolates (p = 0.007). On the other hand, no statistically significant difference was observed between gentamicin phenotype and aaC4 resistance gene among strains (p > 0.05). On the other hand, all 61 tetracyclines resistant isolates did not harbor tetA and tetB genes encoding tetracycline resistance.  (6 )

Discussion
The relationship between companion animals and humans has become increasingly close in recent decades, leading to a high risk of zoonotic transmission of bacteria between pets and humans [27][28][29]. In this study, companion animals were evaluated as a potential source in the transmission of pathogenic bacteria.

Discussion
The relationship between companion animals and humans has become increasingly close in recent decades, leading to a high risk of zoonotic transmission of bacteria between pets and humans [27][28][29]. In this study, companion animals were evaluated as a potential source in the transmission of pathogenic bacteria.
Specifically, a total of 79 E. coli isolates were confirmed from 98 companion animals (80.6%). Companion animals have been reported as a pivotal transmission reservoir for DEC [30,31]. The findings showed that 36 of 79 strains belonged to DEC, and ETEC was found to be mostly associated with diarrhea companion animals. In previous study, Zahraei et al. [32] found EPEC isolates from 113 non-diarrheic and diarrheic animals. Similar findings were also observed in our study ( Table 1), implying that, whether healthy or diarrheic, companion animals may act as a potential reservoir of EPEC. Compared to old animals, young companion animals are usually more susceptible to ETEC and EPEC infection [4]; however, 52.8% (19/36) of pets investigated in our study were older than 12 months. Accordingly, determining the virulence factor, and antibacterial resistance and evaluating the risk of potential transmission to people of companion animals E. coli strains are very crucial.
There is concern that antimicrobial resistance in E. coli harbored by companion animals can be transmitted from one host to another even by low bacterial numbers [33,34]. Previous studies point out that the close relationship of humans and their companion animals provide opportunities for sharing strains [35,36]. In this study, most of the isolates were resistant to TET (64.6%), DOX (59.5%) and AM (53.2%), which was similar to reports of E. coli isolates from other studies [37]. In contrast, other recent studies of E. coli isolates from dogs and cats reported resistance to a wide range of antibiotics, including quinolones and extended-spectrum beta-lactamases [38,39]; these high resistance rates are due to the wide use of antibiotics in companion animals. In addition, all the isolates were susceptible to MEM, and the restricted use of carbapenems may be a factor affecting the associated high antimicrobial sensitivity. Carbapenems should be administered only in cases of multidrug-resistant bacterial infection; their restricted use is probably a factor influencing the associated high antimicrobial sensitivity [40,41]. In the current study, the overall MDR frequency of 49.4% was similar to that in studies from the United States (52.0%) and Poland (66.8%) [42,43]. In contrast, the result in this study was much higher than that (28.0%) from healthy dogs from Canada [44]. The high MDR rates observed in the current study indicated that currently available antimicrobial treatment options for E. coli infections in companion animals are limited, and it is highly recommended that measures should be taken to control the potential risk from companion animals, such as reducing and standardizing antibiotic usage in clinics.
ESBLs are mainly associated with E. coli, which often show resistance to multiple antibiotics [45]. In this study, we found that 43.0% (34/79) of E. coli isolates from companion animals produced ESBLs; this result is significantly higher than that reported in Brazil, in which the prevalence of ESBL-EC in companion animals was 8.1% [46]. Most ESBL-EC strains (n = 20, 20/34) exhibited the MDR phenotype; this finding may be mirroring the result of irrational use of these antibiotics in veterinary clinics that eventually might cause high selection pressure of resistant bacteria. In Japan, the percentage of ESBL-EC in companion animals was 21.3%, and most of the isolates were sampled from urinary tract infections [47], and it is a common clinical diagnosis that urinary tract infection due to E. coli strains. In Europe and the United States, the proportions of ESBL carriers in companion animals varied according to the country and the sampling, but remained under 5% [25,48,49]. Data on the prevalence of ESBL-EC in different studies were difficult to compare based on differences in regions, sample types, and animal health status.
As reported previously, bla CTX-M genes are the most frequent ESBL-encoding genes identified in both humans and animals [50,51]. In this study, the prevalence of bla CTX-M was 60.8%, which was much higher than that in other studies in which the prevalence of bla CTX-M among bacteria derived from pets in different countries (including China) ranged from 10 to 21% [38,52,53]. Furthermore, higher resistance to quinolone antibiotics was observed in ESBL-EC strains, which may be because plasmids containing bla CTX-M often carry genes that confer resistance to other antibiotic families [54]. Therefore, the risk of zoonotic transmission of ESBL-EC is likely to be high, and the E. coli isolates from companion animals are reservoirs of ESBLs. In addition, we did not find bla TEM in this study, which was different from the E. coli from food-producing animals in which bla TEM was the most commonly identified β-lactamase gene [55].
PMQR is a threat to veterinary clinical therapy. To date, at least three types of quinolone resistance determinants, including the qnr family (qnrA, qnrB, qnrS, qnrC, aac (6')-Ib-cr and qnrD), and quinolone efflux pumps (oqxA), have been extensively reported [56]. In this study, we found a high prevalence of aac (6 )-Ib-cr among the E. coli isolates, and similar findings have been confirmed among the E. coli isolates from companion animals in Australia [57]. In addition, we detected the qnrS gene, which was the most prevalent PMQR gene among ESBL-producing Enterobacter spp. isolates from humans in China [58] but was not previously detected among isolates from companion animals in Australia [57]; these findings may suggest that the qnrS gene is locally spread among companion animals and humans in China.
In our study, 84.4% of non-ESBL producers carried at least 1 or more β-lactamase genes. In addition, 18 isolates of non-ESBL producers were resistant to cephalosporin antibiotics. The difference in phenotypic-genotypic cephalosporin resistance, a likely molecular mechanism was the presence of other AmpC β-lactamase genes. tetA and tetB genes were not detected in not detected in all tetracyclines resistance isolates, possibly because strains contain some other resistance genes, or involve other resistance mechanisms [59].
The high-pathogenetic island marker gene, irp2 mediates the iron-uptake system of highly pathogenic strains and was associated with E. coli virulence [60]. Several reports showed that irp2 was detected in pathogenic E. coli strains from humans [61,62], we observed a very high frequency of irp2 gene (87.3%). Additionally, eaeA (intimin) observed in DEC were found to be more pathogenic in humans and are involved in zoonotic transmission [63], a similar result was found in our study. Therefore, our study indicates that strains harbored irp2 may bring a risk for human and companion animals health. On the other hand, easA gene was mainly detected from diarrhea companion animals (Table S1), this finding showed that easA positive strains may are associated with animal diarrhea mechanism.

Sampling Size Determination
The required sample size was determined based on a confidence interval (CI) of 95%, an E. coli expected prevalence of 50% and an accepted error of 10%, giving a value of N will be 98 samples in this study according to a previous report [64].

Bacterial Isolates
Rectal and oronasal swabs were collected from 65 dogs and 33 cats attended in different veterinary clinics in the city of Tai'an, Shandong during a 3-month period (January to March 2021). Specimens were refrigerated until processing. The health status and age of the companion animals are shown in Table S1. Ethical approval was not required for the study because the sampling process did not harm the animals.
The swab samples were tested for the presence of E. coli as previously described [65]. The swab samples were streaked onto MacConkey (Hope, Qingdao, China) agar plates and incubated overnight at 37 • C. After overnight incubation, the phenotypic characteristics of E. coli were selected, and isolates were subcultured onto MacConkey agar plates. The subcultured colonies were confirmed as E. coli by matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) with a Vitek-MS (bioMerieux, Marcy-Etoile, France). Isolates were inoculated into 5 mL of lysogeny broth (LB) and incubated at 37 • C with shaking for 18 h. A 1 mL aliquot of this suspension was combined with 0.5 mL of sterile glycerol and stored at −80 • C. Additionally, Escherichia coli ATCC 25,922 was kept in the Laboratory of Veterinary Public Health, Shandong Agricultural University.

Determination of Serotypes, and Genomic DNA Extraction
According to the manufacturer's instructions (Tianrun, Ningbo, China), using polyvalent and monovalent agglutination sera to serotype. Briefly, commercially available antisera were blended with a E. coli suspension on a slide, and then the serotype was determined in one minute.
The isolates were cultivated on 2 mL of LB and incubated at 37 • C with shaking for 10 h. Genomic DNA was extracted by using the Genomic DNA Purification Kit (Tiangen, Beijing, China), and DNA templates were stored at −20 • C until use.

Statistical Analysis
The statistical analyses were performed using SPSS software version 21.0 (IBM Corp., Armonk, NY, USA), employing the chi-square test. A p value of < 0.05 was considered to be statistically significant.

Conclusions
Overall, our results highlight the seriousness of the antibiotic resistance problem among E. coli isolates from whether healthy or diarrhea companion animals in Shandong, China, which may provide a significant reference for pet clinical veterinarians, public health agencies and other researchers. Our findings also emphasize that the presence of ESBL-producing E. coli among companion animals needs further continued surveillance. Therefore, it is necessary to establish national standards for the rational use of antibiotics in companion animals.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/antibiotics11060828/s1, Table S1: Characteristics of the E. coli collected from companion animals.

Conflicts of Interest:
The authors declare no conflict of interest.