Zoonotic Fecal Pathogens and Antimicrobial Resistance in Canadian Petting Zoos

This study aimed to better understand the potential public health risk associated with zoonotic pathogens in agricultural fairs and petting zoos in Canada. Prevalence of Salmonella, Shiga toxin-producing Escherichia coli (STEC) O157:H7, and top six non-O157 STEC serogroups in feces (n = 88), hide/feather (n = 36), and hand rail samples (n = 46) was assessed, as well as distributions of antimicrobial resistant (AMR) broad and extended-spectrum β-lactamase (ESBL)-producing E. coli. Prevalence of methicillin-resistant Staphylococcus aureus (MRSA) in pig nasal swabs (n = 4), and Campylobacter, Cryptosporidium, and Giardia in feces was also assessed. Neither Salmonella nor MRSA were detected. Campylobacter spp. were isolated from 32% of fecal samples. Cryptosporidium and Giardia were detected in 2% and 15% of fecal samples, respectively. Only one fecal sample was positive for STEC O157, whereas 22% were positive for non-O157 STEC. Multi-drug resistance (MDR) to antibiotics classified as critically and highly important in human medicine was proportionally greatest in E. coli from cattle feces. The β-lactamase-producing E. coli from pig, horse/donkey feces, and hand rail samples, as well as the STEC E. coli from handrail swabs were MDR. The diversity and prevalence of zoonotic pathogens and AMR bacteria detected within agricultural fairs and petting zoos emphasize the importance of hygienic practices and sanitization with respect to reducing associated zoonotic risks.


Introduction
Agricultural fairs and petting zoos provide the public with the opportunity to interact and learn about farm animals. However, farm animals also present a risk for transmission of zoonotic pathogens [1]. Illnesses and outbreaks involving Shiga toxin-producing Escherichia coli (STEC) O157:H7 and other non-O157 serogroups, Campylobacter spp., Salmonella spp., Cryptosporidium spp., and to a lesser degree, Listeria monocytogenes and Yersinia enterocolitica have been reported in fair and petting zoo visitors [2,3]. It has been estimated that 14% of all illnesses caused by these pathogens arises from direct animal contact [3].
Mississauga, ON, Canada) and Rappaport-Vassiliadis (RV; EMD Millipore, Etobicoke, ON, Canada) broths and incubated at 42 °C for 20-24 h. Tetrathionate enrichments were streaked onto Xylose Lysine Tergitol 4 agar (XLT-4; BD Canada, Mississauga, ON, Canada) and RV enrichments were transferred to modified semisolid RV (MSRV) agar plates and incubated at 37 °C for 24 h. A loop of MSRV culture from the edge of opaque zones was streaked onto modified Brilliant Green agar (MBGA; Hardy Diagnostics, Santa Maria, CA, USA) and incubated overnight at 37 °C for 24 h. Up to three colonies per XLT-4 and MBGA plate were sub-cultured to fresh XLT-4 plates and incubated at 37 °C for 24 h. Presumptive Salmonella isolates were tested for agglutination with Poly A-I + Vi antiserum (BD Canada, Mississauga, ON, Canada).

STEC Screening
Feces and sponge swab washings were enriched for O157 and the top 6 non-O157 STEC in EC broth (EMD Millipore Etobicoke, ON, Canada) at 37 °C for 6 h. Cultures were centrifuged (10 min at 6000× g) and DNA was extracted using a NucleoSpin Tissue kit (Macherey-Nagel, Düren, Germany) according to manufacturer's instructions.

STEC Screening
Feces and sponge swab washings were enriched for O157 and the top 6 non-O157 STEC in EC broth (EMD Millipore Etobicoke, ON, Canada) at 37 • C for 6 h. Cultures were centrifuged (10 min at 6000× g) and DNA was extracted using a NucleoSpin Tissue kit (Macherey-Nagel, Düren, Germany) according to manufacturer's instructions.

Generic E. coli Screening
Feces and pre-enriched swabs were streaked onto MAC agar and incubated overnight at 37 • C. Presumptive E. coli was streaked onto Tryptic Soy Agar (TSA) and incubated overnight at 37 • C, followed by testing for indole production (Ricca Chemical Co., Arlington, TX, USA). Lactose-fermenting and indole-producing isolates were screened via PCR for the presence of the E. coli-specific uidA gene [16,17].

Broad-and Extended-Spectrum β-Lactamase E. coli Screening
Feces and swab washings were added to EC broth containing cefotaxime (2 µg/mL) and incubated overnight at 37 • C. Enrichments were streaked onto MAC plates containing ceftriaxone (1 µg/mL) and incubated overnight at 37 • C. Typical E. coli colonies were streaked onto TSA containing ampicillin (32 µg/mL) and incubated overnight at 37 • C. Indole tests were completed as described above.

Cryptosporidium and Giardia Screening
Cryptosporidium oocysts were extracted from fecal samples using the PowerFecal ® extraction kit (Qiagen, Toronto, ON, Canada). Briefly, 0.25 g of feces were suspended in buffers C1 and C2, followed by five freeze-thaw cycles, alternating between a dry ice/ethanol bath and 95 • C heating block. The remainder of the extraction followed manufacturer's instructions. Cryptosporidium spp. were detected in the extracts by amplification of 18S rRNA as previously described [23] using Maxima Hot Start PCR Master Mix (Thermo Fisher Scientific, Toronto, ON, Canada) and final primer concentrations of 500 nM. Reactions were visualized on 1% agarose gels and 18S bands were extracted with the QIAquick gel extraction kit (Qiagen, Toronto, ON, Canada). Sanger sequencing was performed on 18S bands (Macrogen, Seoul, Korea). Genotype/species of Cryptosporidium-positive samples were identified based on the bioinformatic methods of Ruecker et al. [24].
A 292 bp fragment of the 16S small-subunit ribosomal RNA (SSU rRNA) was used to detect Giardia intestinalis and identification of genetic assemblages using endpoint PCR as described by Hopkins et al. [25] along with modifications as described by Smith et al. [26].

MRSA Screening
Alongside positive (S. aureus ATCC 43300) and negative (S. aureus ATCC 49775) controls, nasal swabs were transferred to SA broth (10 g tryptone/L, 75 g sodium chloride/L, 10 g mannitol/L and 2.5 g of yeast extract/L) and incubated at 35 • C for 24 h, then streaked onto MRSA CHROMagar plates (Dalynn Biologicals, Calgary, AB, Canada) and incubated at 35 • C for 24 h. Presumptive S. aureus colonies were transferred to 300 µL TSB and incubated at 35 • C for 24 h. Cultures were streaked onto MRSA CHROMagar prior to confirmation using the StaphTEX Blue latex agglutination test (Hardy Diagnostics, Santa Maria, CA, USA).

Antimicrobial Resistance Testing
A total of 583 E. coli isolates (425 generic, 99 putative-STEC, 59 ESBL-producing) and a control strain (E. coli ATCC 25922) were streaked onto TSA plates and incubated overnight at 37 • C. The direct colony method was used to prepare isolates for disk-diffusion testing according to CLSI M02-A12 [27]. The antibiotic panel (Sensi-disks, BD Canada) consisted of one or two representatives from eight classes ( Figure 1). Disk-diffusion plate readings were completed using the BIOMIC V 3 software version 7.1.1.2014 (Giles Scientific, Santa Barbara, CA, USA). Resistance was defined using established standards [28]. Multi-drug resistance (MDR) was defined as resistance ≥ 3 different classes of antibiotics [29].

E. coli Isolate Phylogroups
Phylogroup determination based on PCR results was completed for all E. coli isolates as described by Clermont et al. [30]. Briefly, an E. coli isolate was assigned to one of 8 phylogroups based on presence or absence of the genes arpA, chuA, yjaA, and TspE4.C2.

Statistical Analysis
Prevalence of STEC or Campylobacter spp. among sample types collected and animal species were compared with generalized linear mixed models (Proc Glimmix, SAS 9.3, SAS Institute Inc., Ottawa, ON, Canada) using a binomial distribution. Model adjusted means (back-transformed to original scale) were reported and deemed significant at p < 0.05.

Survey
Survey responses are summarized in Table 1. The survey had a completion rate of 24% (187 completed questionnaires). Visitors were allowed food within the animal contact area at 51% of petting zoos and 71% of livestock pavilions at fairs. Reportedly, 31% and 23% of livestock pavilions and petting zoos, respectively, never sanitize any surfaces. Furthermore, nearly 30% of petting zoos and livestock pavilions did not provide hand-washing facilities or hand sanitizer. Only 2% of fairs reported an outbreak within the last 10 years; with the majority (75%) identifying E. coli as the causative agent.
STEC O157 isolation frequency was lowest overall, isolated only from feces (0.6% of samples), while non-O157 STEC were isolated from all sources (22%), including hides/feathers and wooden hand rails (Table 2). Neither Salmonella nor MRSA were isolated from any samples. Cryptosporidium and Giardia were detected in 2% and 15% of fecal samples, respectively. Campylobacter spp. were the most prevalent and were isolated from 30% of fecal samples.

Screening for STEC and Related Genes
A viable putative-STEC isolate was obtained from 23% of samples. Following PCR detection of stx1 and stx2 toxin genes, 66 isolates were confirmed to be STEC. Serogroups O26 and O103 were observed at most petting zoos and were the most frequent overall (Table 3). Resistance to antibiotics, albeit to only one class, was most common for O26 isolates. Resistance was more likely for isolates of serogroup O26 than O103 (p < 0.05). The O26 isolates showing resistance to kanamycin and streptomycin were primarily isolated from petting zoo A, while O26 isolates resistant to tetracycline originated primarily from petting zoo F. Three MDR serotypes (O45:H21, O45:H19, O103:H2) from either feather or hand rail swabs, were also collected from petting zoos A and F. All O26 and O45 isolates contained stx1, while other virulence markers were variable across serogroups. All O145 isolates contained eae and stx1 regardless of sample type or petting zoo. All O103:H38 isolates possessed ehxA, eae, and stx1 genes regardless of host, sample type, phylogenetic group, or petting zoo. Has there been an outbreak or problem with a pathogen in the past 10 years? (n = 164) Yes 2 What was the pathogen of concern? (among those that reported a problem or outbreak) (n = 4)

Question Answer % Fairs
Is there a children's play area directly adjacent? (L n = 130; P n = 115) Yes 18 30 Is there a food or beverage area directly adjacent? (L n = 131; P n = 116) Yes 16 2 Are there rules regarding visitors bringing food or drink into the animal area? (L n = 125; P n = 117) Not allowed 29 49 Allowed, but no touching animals 7 5 Allowed, not for feeding animals 25 25

Antimicrobial Resistance among Generic E. coli, Broad-and Spectrum β-Lastamase E. coli, and STEC
Antibiotic-resistant generic E. coli isolates were observed most often in pig (83%) and cattle (56%) feces, as well as feather swabs (66%; Table 5). The majority of AMR E. coli isolates from animal feces and hide/feather swabs (n = 69) were resistant to only one class of antibiotics. Six isolates were resistant to 3 or 4 classes. Isolates resistant ≥ 5 antibiotic classes were only obtained from cattle feces. Half of the AMR generic E. coli isolated from wooden hand rails were MDR. Conversely, AMR E. coli collected from metal hand rails were only resistant to tetracycline.
Of the 66 STEC isolates collected in this study, 38% were resistant to antimicrobials ( Table 5). All AMR STEC obtained from wooden hand rails were MDR, while those from feces and hides/feathers were resistant to only one class of antibiotic. One isolate from chicken feathers was the exception as it was resistant to five antibiotic classes.

Survey
To our knowledge, this is the first survey of hygienic practices employed in Canadian livestock fairs and petting zoos. Several investigations from the U.S. have identified contact with animals and their feces, poor hand-washing compliance, and lack of or inadequately equipped hand-washing facilities as factors that precipitate disease outbreaks [31,32]. Restrictions on food and drink in animal areas, frequent and immediate removal of animal feces, provision of hand-washing and hand-sanitizing facilities, and more frequent surface disinfection are practices that could also reduce health risks in Canadian petting zoos and fairs. The U.S. National Association of State Public Health Veterinarians released a report of recommendations to prevent disease associated with animals in public settings [33]. A similar document outlining standard operating, cleaning, and sanitation procedures from a Canadian authority would provide a beneficial resource for agricultural fair coordinators and operators of public animal venues in Canada. However, without official policies and enforcement in place, compliance with recommendations remains voluntary.

Prevalence of O157 and Non-O157 STEC
Due to the low infective dose of STEC O157:H7 and the severity of illness that can develop, particularly in young children, it is the pathogen of utmost concern for petting zoos [2,5]. Previous studies reported a high prevalence of STEC O157 at county fairs in the U.S. (75-97%) with cattle being the primary host (11% host prevalence), followed by sheep and goats (3.6%), and swine (1.2%) [34]. In the current study, STEC O157 was isolated once from goat feces. Cattle were not common at the petting zoos sampled in this study; hence, only nine composite cattle fecal samples were collected compared to 36 from sheep and goats.
Pathogenic E. coli O157 can survive in the environment outside the animal host for weeks to months [4], and viable E. coli O157 have been cultured from fairgrounds and petting zoos months after outbreaks have occurred [35,36]. Hand rails can become contaminated at fairs and petting zoos and have been implicated as sources of E. coli O157 following outbreaks [36,37]. Although E. coli O157 was not isolated from hand rail swabs in the current study, non-O157 STEC were, even from hand rails that were reportedly sanitized daily.
To our knowledge, this is the first study to examine the prevalence of the top 6 non-O157 STEC serogroups in animals at fairs and petting zoos. Most studies of non-O157 STEC report prevalence in food animals, of which cattle, small ruminants, pigs, and poultry are important sources [5,38]. However, non-O157 STEC are widespread and have been isolated from numerous non-food domestic and wild animals [5,32]. According to previous studies, llamas, horses, and donkeys are not considered important STEC reservoirs [4,39]. In the current study, STEC were isolated from llama feces from the same petting zoo on two separate occasions, and from horse and donkey feces collected from two different petting zoos ( Table 2). The relatively high prevalence of STEC in fecal samples from these animals may reflect animal-to-animal transmission, considering the proximity of housing within petting zoos and fairs [40]. This possibility is supported by the isolation of STEC from multiple animal species at the same petting zoos that possessed the same serotype, virulence, and AMR phenotypes.
Research of STEC loads on hides is primarily focused on cattle since hides represent a major source of carcass contamination during meat processing [5]. No research comparable to the current study for STEC prevalence on the hides, fur, or feathers of other animal species could be identified. Considering that contact with animals at petting zoos is primarily from active contact with hides and feather, findings of the current study emphasize the importance of supplying hand-washing facilities within petting zoo areas.

Prevalence of Salmonella, MRSA, Campylobacter, Cryptosporidium and Giardia
Research suggests that the prevalence of Salmonella is lower in Canada and the northern U.S. than the southern U.S. [41,42] and that management practices may have a profound effect on the transmission and persistence of salmonellae within an integrated production system [43]. Animals sampled within this study were not a part of the production system, coupled with the small sample size, the lack of Salmonella in this study was perhaps not surprising.
Pigs are a reservoir for S. aureus, and MRSA have been isolated from pigs at an agricultural fair in the U.S. [44]. Though pigs at fairs in Canada have not previously been tested, other studies indicate MRSA are common in pigs on Canadian production farms [45]. Since pigs were not frequently encountered at the petting zoos visited in this study resulting in a small sample size, it is perhaps not surprising that none of the nasal swabs were positive for MRSA.
Compared to the other pathogens in this study, Campylobacter is thought to cause the highest number of human illnesses due to transmission from animals [3]. Healthy poultry, cattle, swine, and sheep are considered important reservoirs for this organism [20,46]. We isolated Campylobacter spp. from sheep and goats, pigs, horses and donkeys, and birds. Isolation of Campylobacter from healthy horses has been reported, but much less frequently than for other host species [47]. A similar study isolated Campylobacter spp. from cattle, sheep, goats, and chickens at a petting zoo in California [11]. In contrast to the high proportion of Campylobacter-positive samples found in our study, a meta-analysis estimate of Campylobacter prevalence among petting zoo animals was determined to be 6.5% [48]. The meta-analysis was derived from only seven studies, as there is a lack of research on Campylobacter at fairs and petting zoos in Canada. Furthermore, non-jejuni/coli species are of uncertain significance to human health [46]. Additional work is required to fully assess the risk of Campylobacter infection from petting zoos and fairs in Canada.
Cryptosporidium was only detected in sheep and goat feces, with the Cryptosporidium bovis identified in goat/sheep fecal samples occasionally being associated with human illness [49]. A zoonotic assemblage of Giardia intestinalis (assemblage A) was detected in cattle, horses, and donkeys. Neither pathogen was detected in feces of petting zoo animals in a similar U.S. study [11].

AMR, Broad-and Extended-Spectrum β-Lactamase-E. coli
AMR E. coli were observed among all animal groups and sample types tested. Collectively, 22% of the generic E. coli isolates from this study were resistant to one or more antimicrobials. Other studies report varied rates of resistance in generic E. coli from different animal groups in Canada [50,51].
Very few isolates (1.3%) were resistant to the critically important 3rd generation cephalosporins or quinolones. No isolates were resistant to imipenem. Resistance to tetracycline was most common and has been widely reported in dairy and beef cattle, small ruminants, pigs, poultry, equines, and petting zoo animals [11,[51][52][53][54]. Resistance to ampicillin, streptomycin and kanamycin is also common in E. coli from animal feces [11,52,54].
Although MDR among generic E. coli and STEC was not common, MDR E. coli including STEC were isolated from wooden hand rails, possibly posing an even greater risk of transmission within the petting zoo than bacteria directly associated with animal feces. Although these isolates may not be pathogenic, they may be carriers of AMR genes that can be readily exchanged with other bacteria [50].
Since the late 1990s, ESBL-producing E. coli have emerged globally [10]. Early β-lactamase-producing E. coli possessed mainly TEM and SHV β-lactamases, while the significance of CTX-M-type enzymes has since increased and currently represents the most common type in humans [10]. Infection with β-lactamase-producing E. coli is of concern because of limited therapeutic options [10]. Only the most common broad-spectrum β-lactamase and ESBL genotypes were screened for in this study, suggesting the isolates with unidentified genotypes (>60%) may have been of less common β-lactamase types. The detection of broad-and extended-spectrum β-lactamase-producing E. coli in this study is to be considered a preliminary screening. Future studies must also include phenotypic confirmation of putative ESBL-producing E. coli isolates.

Phylogroups
The clonal population structure of E. coli has enabled assignation of isolates to phylogroups (A, B1, B2, C, D, E, F, and cryptic clades) [29,30] which vary in host association, presence of virulence factors, and persistence in the non-host environment [29]. Distinct ecological niches characterize the different phylogroups. For example, phylogroups A and B2 are predominant in humans, whereas phylogroups A and B1 are more common in animals [55].
Although representatives from every phylogroup were identified, the most common phylogroup was B1. These results are consistent with previous studies which have identified a higher proportion of B1 among cows, goats, and sheep [56]. Our results suggest that E. coli from phylogroups A, B1, and E are the most common in petting zoo animals and their environments. From a human health perspective, E. coli responsible for mild and chronic diarrhea are found throughout the phylogenetic tree, whereas E. coli that cause more severe pathologies are frequently restricted to phylogroups A, B1, and E [55]. Whether any of the isolates collected in the current study could cause mild or severe illness in humans remains unknown, but measures that reduce the risk of ingestion of these bacteria [2], as well as cleaning and sanitizing procedures to decrease environmental persistence undoubtedly have merit.

Conclusions
Effective hygienic preventative measures within fairs and petting zoos rely on assessing risks for human exposure to zoonotic pathogens. Despite the limited number of petting zoos and samples in this study, we have shown that some zoonotic pathogens are clearly present within petting zoo animals and their environments. Of concern are STEC and Campylobacter species based on their prevalence and/or potential to cause serious illness. In addition, a variety of virulence genes and AMR E. coli were present in a variety of animal species and on hand rails suggesting the potential for transfer of these elements to other E. coli and/or pathogens.
Certain inadequacies are evident within the operation of animal areas accessible to the public at fairs in Canada, such as the absence of hand-washing facilities at many venues. Sanitization of surfaces within petting zoos and livestock pavilions is largely sporadic, with some venues lacking sanitization procedures. These data should not deter the public from visiting agricultural fairs but should serve as reminder to event operators and visitors alike, of the importance of good hygiene and sanitation practices. Materials used to construct animal areas and pens should be non-porous, such as metal rather than wood, allowing more effective sanitization. Fair and petting zoo operators are encouraged to consult the recommendations outlined by the National Association of State Public Health Veterinarians. Ongoing efforts to reduce risks of human exposure to zoonotic pathogens can help to ensure children and adults continue to enjoy agricultural events and the educational opportunities they provide.
Further research is necessary to better understand the human health risk associated with fairs and petting zoos. Future studies should include identification of plasmids in E. coli strains to identify the dissemination potential and possible transfer of AMR genes to other E. coli strains and other members of Enterobacteriaceae. Multilocus Sequence Typing (MLST) of E. coli strains could be used to determine clonality and animal/human dissemination potential. Additionally, the inclusion of animal and human clinical isolates could be used to identify common characteristics between clinical isolates and those collected from petting zoos.