Pet and Stray Dogs as Reservoirs of Antimicrobial-Resistant Escherichia coli

The close contact between dogs and humans creates the best bridge for interspecies transmission of antimicrobial-resistant bacteria. The surveillance of its resistance including the detection of extended-spectrum beta-lactamases (ESBLs) in Escherichia coli as indicator bacteria is an important tool to control the use of antimicrobials. The aim of this research was to evaluate the E. coli resistance in strains by phenotypic methods, isolated from pet and stray dogs of La Plata city, Argentina. Faecal samples were collected using rectal swabs from 50 dogs with owners (home dogs = HD) and 50 homeless dogs (stray dogs = SD). They were cultured in 3 MacConkey agar plates, with and without antibiotics (ciprofloxacin and cefotaxime). 197 strains were isolated, of which only 95 strains were biochemically identified as E. coli, 46 strains were from HD, and 49 were from SD. Antimicrobial susceptibility was evaluated by the Kirby–Bauer disk diffusion method. The most prevalent resistance was for tetracycline, streptomycin, and ampicillin. In both groups, the level of resistance to 3rd generation cephalosporins was high, and there were multiresistant strains. There was a higher level of antimicrobial resistance in strains from SD compared to HD. There were 8% of strains suspected of being ESBLs among samples of HD and 36% of SD. One (2%) of the strains isolated from HD and 11 (22%) from SD were phenotypically confirmed as ESBL. Pets and stray dogs are a potential source of E. coli antibiotic resistance in Argentina; therefore, its surveillance must be guaranteed.


Introduction
e role of pets as one of the most important disseminators of antimicrobial-resistant bacteria has been underestimated for a long time because the general focus has been aimed at food-producing animals as the principal source of resistant bacteria. However, the close contact between pets and humans creates the best bridge for interspecies transmission of multidrug-resistant (MDR) bacteria [1].
Dogs and cats which live in the same house with their owners make contact with the same surfaces and objects, and these habits increase the chances of antimicrobial resistance dissemination. Furthermore, veterinary practice evolution and the increasing sense of social responsibility for the welfare and health of pets have enhanced their life expectancy, which has increased the number of geriatric patients that frequently need antimicrobial therapy, since they often suffer from chronic diseases or immunocompromising conditions [2].
Antimicrobials are frequently used for therapeutic and prophylactic purposes, not only in dogs that live in homes but also in stray dogs. In Argentina, people without professional knowledge frequently administrate antimicrobials to stray dogs because of the unrestricted antibiotic sale and the increasing tendency to protect animals [3]. e inadequate use and abusive prescription of antibiotics, the application of subtherapeutic doses, the irregularity in the administration of antimicrobials, and the change in the social role of dogs among the community nowadays are important risk factors for antimicrobial resistance selection and the transference of bacteria with gene resistance determinants. Several studies have proved that antimicrobialresistant bacteria could be transferred from dogs to humans and from humans to dogs [4][5][6].
In recent years, there has been an important increment in the number of companion animals in Argentina. According to a recent study of the Department of Health and Animal Protection of the Government of Buenos Aires city, there are between 800000 and one million dogs and cats, with and without owners. is means that there is one pet for every 3 humans [7,8].
Several resistance microorganisms have been isolated from healthy and sick pets, such as methicillin-resistant Staphylococcus aureus (MRSA), methicillin resistance Staphylococcus pseudintermedius (MRSP), multidrug-resistance Gram-negative bacteria, and extended-spectrum betalactamase (ESBL)/AmpC-producing Enterobacteriaceae [9,10]. e great increase in strains carrying ESBLs in humans, animals, and their surrounding environments is a huge problem worldwide. Companion animals would have a possible role as reservoirs for ESBLs [11].
ese enzymes have hydrolytic activity against penicillins, third-generation cephalosporins (ceftazidime or cefotaxime), and aztreonam, but not the cephamycins (cefoxitin) or carbapenems, and are inhibited by beta-lactamase inhibitors as clavulanic acid [15]. Beta-lactams are possibly the antimicrobials most widely used not only in human medicine but also in animals due to the safety, antimicrobial spectrum, availability, and pharmacokinetic and pharmacodynamic properties [16]. First-generation cephalosporins and amoxicillin with clavulanic acid and penicillin with an aminoglycoside are prescribed routinely in daily veterinary clinic practice. irdgeneration cephalosporins represent one of a few therapeutic options to treat bacterial infections of difficult resolution; therefore, they are critically important drugs.
Enterobacteriaceae are the main producers of ESBL, particularly Klebsiella pneumoniae and E. coli [17]. Among the intestinal microbiome, E. coli has a role as an indicator of resistance because it is one of the most widespread groups of intestinal bacteria. e surveillance of its resistance mechanisms is an important tool in the control of nonprudent use of antimicrobials. It allows us to know which antimicrobials are selected for antimicrobial resistance, avoiding public health risks, reducing therapeutic failures and economic losses among producers [5].
Furthermore, ESBL strains often carry antimicrobial resistance genes for other antibiotics, such as fluoroquinolones, aminoglycosides, tetracyclines, sulfamethoxazole-trimethoprim, and chloramphenicol. Consequently, there is a potential risk of cross-resistance mediated by plasmids, and there is a great concern because of the increasing emergence of different phenotypes of resistance such as multidrug resistance (MDR: when at least one agent in three or more families of antibiotics is not sensitive), extreme resistance (XDR) when they are not sensitive to at least one agent in all categories of antibacterial families except one or two of them), and pan-resistance (PDR: when they are not sensitive to any of the agents in all the antimicrobial categories) [18,19]. e objectives of this research were to detect resistance patterns and in vitro ESBLs producing E. coli strains by phenotypic methods, isolated from dogs with owners and stray dogs of La Plata city, Buenos Aires, Argentina.

Materials and Methods
2.1. Bacterial Strains. 50 faecal samples of dogs living in houses with owners (home dogs � HD) and 50 samples of homeless dogs (stray dogs � SD) were collected from October to December 2016 in La Plata city, Buenos Aires, Argentina. Dogs with owners were clinically healthy, adults, and raised in family homes. ey did not receive antimicrobial treatments in the last 3 months. For samples taken in a veterinary clinic, only those animals that come for routine health control or to reinforce vaccines by the health plan were selected. Stray dog faecal samples were collected in different areas of La Plata city following the spiral sampling method [20]. La Plata city is characterized by its particular design because it is a grid, a perfect square with two diagonals that cross the complete city from east to west and from north to south. Consequently, stray dogs were sampled beginning from the peripheral area to the centre of the square. Only nonaggressive stray dogs were chosen for sampling since no sedation method was used. Animals were restrained by two operators, and a third person took the rectal faecal sample using a sterile cotton swab. e protocol was carried out according to the Guide for the Care and Use of Agricultural Animals in Agricultural Research and Teaching (Federation of Animal Science Societies-FASS-) [21] and was approved by the Experimental Ethics Committee of the Faculty of Veterinary Science, UNLP, Argentina (47.3.15 J).
In the laboratory, each cotton swab was used to inoculate each sample onto 3 agar plates: MacConkey agar (Difco, Becton Dickinson, USA) supplemented with 2 mg/L cefotaxime (plate A), MacConkey agar supplemented with 0.05 mg/L ciprofloxacin (plate B) to select possible resistant strains present in the samples, and the third plate without antimicrobial (plate C). Cefotaxime and ciprofloxacin were obtained from Sigma-Aldrich, MO, USA. After overnight incubation at 37°C, 2-3 colonies with the phenotypic characteristics of E. coli were selected, Gram stained, and typified by biochemical tests (Urease production, Catalase test, Motility, Voges Proskauer, Indole production, Carbohydrate fermentation tests, Methyl red, and Citrate utilization). E. coli ATCC 25922® (American Type Culture Collection, USA) was used as quality control [22,23].

Susceptibility Testing.
e strains that grew in plates A and/or B were identified biochemically as E. coli and tested for antimicrobial susceptibility by the standard Kirby-Bauer disk diffusion method. But when there were no colonies on those plates (A or B) but in C, strains developed in this last plate without antibiotic were isolated to study. Twenty antimicrobials (Becton Dickinson, USA) were selected: amoxicillin/clavulanic acid (20 μg/10 μg), ampicillin (10 μg
For the ESBL confirmation, the Double-Disc Synergy Test was applied according to CLSI 2013; discs containing ceftazidime and cefpodoxime were put next to a disc with amoxicillin plus clavulanic acid (20 mm centre to centre). e positive result is indicated when the inhibition zones around any of the cephalosporin discs were augmented in the direction of the disc containing clavulanic acid. E. coli ATCC 25922® was used for quality control [22,23].

Statistical Analysis.
Chi-square test was used to determine the significance of differences in resistance prevalence between stray dogs and pets. A value of p ≤ 0.05 was considered significant.

Results and Discussion
197 strains were isolated (77 from HD and 120 from SD) from the 100 faecal samples. Only 95 strains biochemically identified as E. coli (46 from HD and 49 from SD) were studied; every strain was individual and becomes from one of the 3 plates mentioned above (A, B, or C). From a few samples, we obtained isolations biochemically identified as E. coli from both plates A and B at the same time, but we studied them individually. ey were chosen as follows: from samples of HD, 3 strains of E. coli grew on plate A (with cefotaxime), 19 grew on plate B (with ciprofloxacin), and the rest of the strains (n � 24) only developed on agar without antimicrobials; on the other hand, from samples of SD, 19 strains of E. coli developed on plate A and 16 on plate B; the rest (n � 14) grew in the plates without antimicrobials (plate C).
ose strains that had grown on agar with cefotaxime (plate A) or with ciprofloxacin (plate B) were selected to carry out antibiograms. Strains of E. coli that were inhibited in plates A and B but developed in plate C without antimicrobial were the third group studied, including susceptibility test. erefore, a total of 46 strains of dogs with owner and 49 without owner were selected for the study. e most common resistance observed was to tetracycline (70%) in SD-derived strains and to nalidixic acid (38.3%) in HD-derived strains. Resistance to streptomycin (46%) and ampicillin (44%) was the second high in SD strains. For HD, there were 25.5% of resistance to ampicillin and 19.1% to sulfamethoxazole-trimethoprim. In both groups, the level of resistance to 3rd generation cephalosporins was high (ceftriaxone: 30% SD-21.3% HD and cefotaxime: 28% for SD). ere was a higher level of antimicrobial resistance in strains from SD compared to HD. No E. coli from either group showed resistance to imipenem or nitrofurantoin.
ere were multiresistant strains in both groups, and their resistance profile includes one, two, three, and more than four antimicrobials (see Figure 1 and Table 1).
ere were 8% of strains suspected of being ESBL among samples of HD and 36% of SD. One of the suspected ESBL strains from HD (2%) and 11 (22%) of SD were confirmed. e most prevalent phenotypes of resistance detected among the E. coli isolates recovered from HD were nalidixic acid (NAL � 5 isolates) and nalidixic acid with tetracycline (NAL + TET � two isolates). e multiresistance occurrence was considerably higher in SD strains, with 32.7% of MDR profiles to more than 4 antimicrobials. e most prevalent phenotypes detected for SD isolates were tetracycline (TET � 8 isolates), and as HD strains, nalidixic acid with tetracycline (NAL + TET � three isolates).
All the strains that showed resistance to 4 antimicrobials or more in both groups (HD and SD strains) belong to isolates that had previously developed in plates with antimicrobial (ciprofloxacin and cefotaxime) (see Tables 2 and  3).
In our study, samples were obtained from faeces of animals with and without owners, but they were not clinical samples because the principal aim of our work was to know the prevalence of antimicrobial resistance in E. coli as indicator bacteria in La Plata city, in dogs without recent antimicrobial therapy. Most authors have published results obtained from clinical samples such as urine, urinary tract, intestinal tract, and ears [24][25][26]. is makes it difficult to compare results since there is an expected difference in the prevalence of resistance between populations of sick and healthy dogs.
Resistance to antimicrobials in the present study appeared to be higher than those previously reported in other studies done with healthy dogs. e highest level of resistance obtained was for tetracycline in SD-derived strains but not in pets (p < 0.0001). Ampicillin was the second most prevalent resistance observed (SD � 44%; International Journal of Microbiology HD � 25.5%); however, it was not quite significant (p � 0.8888). Resistance to nalidixic acid was common in both groups of animals (HD � 38.3%; SD � 44%), considered not significant (p � 0.6798). Stray dog isolates also showed high resistance against streptomycin (46%), but it was low (17%) in pets (p < 0.0001). Resistance to ampicillin and tetracycline is the most frequently reported in most studies made with healthy dogs, in different countries [27][28][29]. e resistance percentages obtained for the antimicrobials mentioned above in pet strains (HD) are similar to those reported in a study of Wedley et al. [27], from healthy dogs living in a semirural community in Cheshire, UK (24% for ampicillin and 19.7% for tetracycline). In another study carried out in dogs visiting veterinarians from the UK, they  found a higher level of resistance for both antibiotics; 37.2% of the isolates were resistant to ampicillin and 30% to tetracycline [28]. Nevertheless, these results are still lower than those reported for SD isolates in our work. e results reported by Costa et al. [30], in healthy pets from Portugal isolates, were even lower: 20.5% of resistance for tetracyclines and 7.7% for ampicillin. Interestingly, in some studies done with clinical isolates, in the E. coli isolated from urinary infections, septicaemia, and skin and soft tissue infections, the resistance to ampicillin and tetracycline was lower than that of our results [24,31]. Beta-lactams and tetracyclines are the antimicrobials more used in pets; the different levels of resistance reported in every work correlate with the frequency and magnitude of their use in different countries. Furthermore, our results are disturbing because they demonstrate that in La Plata it seems to be an abuse and/or misuse of broad-spectrum antimicrobials such as tetracyclines, as well as beta-lactams. e identification of ESBLs producing strains is very important to guide an adequate antimicrobial therapy. ere were 8% of strains suspected of being ESBLs among samples of HD and 36% from SD. One of the ESBL strains suspected from HD (2%) and 11 (22%) of SD were confirmed by phenotypic tests, but there was no genetic confirmation, which is a limitation in our study. Again, the results obtained for HD strains are similar to those published by Wedley et al. [27,28], but the percentage of ESBL in SD strains was considerably higher. e prevalence of multidrug resistance (resistance to more than 3 antimicrobials) was considerably higher in SD strains, compared to several studies published by other authors [27,28,[32][33][34]. In pet strains (HD), there was a similar percentage (28%) of MDR to those reported by Murphy et al. [34], in a study made on healthy dogs from Canada. Again, the findings in SD strains are a big concern. Antimicrobials can be acquired without a professional prescription in Argentina. is is one of the possible reasons that lead to the administration of them by anyone, even without being a veterinarian. Another possible source of antimicrobial resistance transmission is the garbage; stray dogs break the garbage bags and eat them. It is possible that bacteria contained in those bags carry antimicrobial resistance genes. On the other hand, household antibiotic waste contaminates the nonhome environment, selecting antimicrobial resistance in bacteria there, which then infect stray dogs and other humans. In Argentina, data are extremely limited for the amount of antimicrobial-resistant bacteria on pets and in the near-home environment. Furthermore, whilst antimicrobials are monitored in stuff entering the formal market, there is a significant informal market, and antimicrobial levels in the environment are not routinely measured. One potential response to the rising threat of antimicrobial resistance is regulation, and both international organizations-led by the WHO-and individual countries have sought to formulate, for example, new regulatory frameworks to address the use of antimicrobials, as well as the household waste management. e source of the resistance genes in those bacteria of stray dogs is still unknown, and further studies should be done to detect such environmental source(s).
One of the most prevalent phenotypes detected for SD and HD isolates was nalidixic acid with tetracycline (NAL + TET � three isolates). ose strains were NAL R CIP S ; it means that they were resistant to nalidixic acid and susceptible to ciprofloxacin.
is is a frequent finding in daily veterinary clinic laboratory. e mechanisms of resistance to quinolones are target alterations or efflux pump overexpression. e most relevant mechanism is the mutation of genes encoding quinolone targets (DNA gyrase and topoisomerase IV). e mutation on a point of the gyrA gene is frequently in E. coli clinical strains and has been associated with this phenotype NAL R CIP S . Bacteria that have this kind of mutation have shown to develop more frequently higher levels of resistance in the presence of quinolones [35].

Conclusions
It is very important to know the prevalence of antimicrobial resistance by surveillance studies not only in pets but also in homeless dogs. e unrestricted antibiotic sale, weak management of household waste, and the increasing tendency to protect animals in Argentina are important risk factors for antimicrobial resistance selection and the transference of gene resistance determinants among bacteria from pets, homeless dogs, and humans. e results of the present study indicate that similar studies in other cities of Argentina would be very useful for monitoring antimicrobial resistance at the national level.

Conflicts of Interest
All the authors declare that there are no conflicts of interest regarding the publication of this paper.