High Occurrence of Shiga Toxin-Producing Escherichia coli in Raw Meat-Based Diets for Companion Animals—A Public Health Issue

Feeding pets raw meat-based diets (RMBDs) is becoming increasingly popular but comes with a risk of pathogenic bacteria, including Shiga toxin-producing Escherichia coli (STEC). In humans, STEC may cause gastrointestinal illnesses, including diarrhea, hemorrhagic colitis (HC), and the hemolytic uremic syndrome (HUS). The aim of this study was to evaluate commercially available RMBDs with regard to the occurrence of STEC. Of 59 RMBD samples, 59% tested positive by real-time PCR for the presence of Shiga toxin genes stx1 and/or stx2. STECs were recovered from 41% of the 59 samples, and strains were subjected to serotyping and virulence gene profiling, using whole genome sequencing (WGS)-based methods. Of 28 strains, 29% carried stx2a or stx2d, which are linked to STEC with high pathogenic potential. Twenty different serotypes were identified, including STEC O26:H11, O91:H10, O91:H14, O145:H28, O146:H21, and O146:H28, which are within the most common non-O157 serogroups associated with human STEC-related illnesses worldwide. Considering the low infectious dose and potential severity of disease manifestations, the high occurrence of STEC in RMBDs poses an important health risk for persons handling raw pet food and persons with close contact to pets fed on RMBDs, and is of concern in the field of public health.


Introduction
Feeding companion animals raw meat has become increasingly popular among cat and dog owners aiming to provide their pets with a natural and healthy diet [1,2]. Raw meat-based diets (RMBDs), also known as Biologically Appropriate Raw Food (BARF), include uncooked raw muscle meats, organ meats, and meaty bones of livestock or wild animals, and are mostly based on the by-products of animals slaughtered for human consumption [3,4]. Since RMBDs are not cooked or pasteurized, concerns have been raised regarding bacterial contamination and the possible transmission of pathogens to pets and humans [5][6][7]. Enterobacteriaceae are the most frequently recovered bacteria from commercially available RMBDs, with a high proportion of sampled RMBDs failing to meet the microbiological standards set out by EC regulation no.1069/2009 in the EU for animal by-products intended for pet food, or the threshold levels for raw human meat products which apply in North America [2,[7][8][9]. Of particular concern, Shiga toxin-producing Escherichia coli (STEC) were identified in 4% of commercially available RMBDs in the US [10], and contaminated RMBDs have been associated with an outbreak of human STEC infections in the UK [11].
STEC belonging to various serotypes within the O157, O26, O103, O91, and O145 serogroups constitute the main STEC associated with human infections in the EU and in Switzerland and are considered a major concern to human health in Europe [19,21].
Worldwide, STEC causes an estimated 2.8 million acute illnesses and 3890 HUS cases annually, representing a major public health issue [22]. Although frequently linked to foodborne outbreaks, a majority of STEC infections remain sporadic and are significantly associated with consuming undercooked or raw meat, person-to-person transmission, or contact with animals or their environment [23,24]. Although studies demonstrating the occurrence of STEC in RMBDs are rare, they raise the question on the safety of raw pet food and the level of pathogenic potential of STEC occurring in RMBDs [2,10,11].
The aim of this study was to assess the occurrence of STEC isolated from commercially available raw pet food in Switzerland and to characterize the strains by using whole genome analyses.

Sample Collection
During September 2018 and May 2020, a total of 59 RMBD products were purchased from ten different suppliers (designated A-J) either on site in pet-food stores or from certified Swiss RMBD producing enterprises, or in online stores of suppliers located in Switzerland and Germany. The products were purchased frozen or shipped frozen to the laboratory and stored until analysis, according to the recommendations of the suppliers.
The tested products contained either pure muscle or pure organ meat, mixed muscle and organ meat products, or meat supplemented with plant ingredients. Details are listed in Supplementary Table S1.

Screening for Stx Genes
Each sample (10g) was enriched at a 1:10 ratio in Enterobacteriaceae enrichment (EE) broth (Becton, Dickinson, Heidelberg, Germany) for 24 h at 37 • C. One loopful of each of the enrichment cultures was cultured on sheep blood agar (Difco™ Columbia Blood Agar Base EH; Becton Dickinson AG, Allschwil, Switzerland), using the streak-plate method. The resulting colonies were washed off with 2 mL 0.85% NaCl. Samples were then screened by real-time PCR for stx1 and stx2, using the Assurance GDS ® for Shiga Toxin Genes (Bio Control Systems, Bellevue, WA, USA).

Recovery of STEC
In the event of a stx-positive PCR result, one loopful each of the washed-off suspension was streaked onto on three to five STEC Chromagar plate (CHROMagar, Paris, France) and on three-to-five Brolacin agar plates (Bio-Rad, Hercules, CA, USA) to get single colonies. The plates were incubated overnight at 37 • C.
From each plate, 20-180 individual colonies were picked (mauve colonies on STEC Chromagar plates; yellow colonies on Brolacin Agar plates) and suspended in 0.5 mL 0.85% NaCl. The suspensions were pooled in groups of ten colonies to simplify the screening process. The pooled suspensions were screened for stx1 and stx2 genes by realtime PCR (LightCycler R 2.0 Instrument, Roche Diagnostics Corporation, Indianapolis, IN, USA), using the QuantiFast Multiplex PCR Kit (Qiagen, Hombrechtikon, Switzerland) according to the guidelines of the European Union Reference Laboratory [25]. In the event of a positive PCR result for stx1 or stx2, the pool was taken apart and the ten colonies were tested individually. From plates yielding more than one stx positive colony, one presumptive STEC isolate was randomly chosen for subsequent characterization.

DNA Extraction and Whole Genome Sequencing
The strains were grown on sheep blood agar at 37 • C overnight prior to DNA isolation, using the DNA blood and tissue kit (Qiagen, Hombrechtikon, Switzerland). The DNA libraries were prepared by using a Nextera DNA Flex Sample Preparation Kit (Illumina, San Diego, CA, USA). Whole-genome sequencing was performed on an Illumina MiniSeq Sequencer (Illumina, San Diego, CA, USA). The Illumina-reads files passed the standard quality checks, using the software package FastQC 0.11.7 (Babraham Bioinformatics, Cambridge, UK), and were assembled by using the Spades 3.14.1based software Shovill 1.0.4 [26,27], using default settings. The assembly was filtered, retaining contigs >500 bp and annotated by using the NCBI prokaryotic genome annotation pipeline [28]. Stx types were determined by an in silico PCR, using the perl script in_silico_pcr (https://github.com/egonozer/in_silico_pcr, accessed on 20 January 2021), using the option "-m, allow one mismatch" and primer sets described in the EU Reference Laboratory for E. coli manual for stx genes detection [29]. The O and H-types were identified by using SerotypeFinder 2.0 [30]. The virulence gene profiles and antimicrobial resistance genes were determined by using VirulenceFinder 2.0 [31] and Resistance Gene Identifier (RGI) 4.2.2 [32].
The sequence types (STs) of each strain were determined based on seven housekeeping genes, using the tool "MLST" (https://github.com/tseemann/mlst, accessed on 20 January 2021), using PubMLST as the database (https://pubmlst.org/, accessed on 20 January 2021) [33]. The isolates were compared by using core genome MLST (cgMLST) analyses comprising 2513 loci of E. coli, using the Ridom SeqSphereC+ software (version 5.1.0; Ridom GmbH, Münster, Germany). Minimum spanning tree (MST) were generated for visualization of strain relatedness, and the threshold for cluster identification was ≤10 alleles between a pair of isolates, according to the Ridom SeqSphereC+ software.

Real-Time Screening for Stx Genes and Isolation of STEC
By real-time PCR screening of enrichment cultures, stx1 and/or stx2 were detected in 35 (59%) of the 59 raw pet-food samples analyzed in this study. Thereof, the majority (n = 32) contained stx2 alone or in combination with stx1. The distribution of stx genes among the different types and categories of meat is shown in Table 1 and Figure 1. RMBDs containing stx genes were detected in products from nine of ten suppliers ( Figure 2). STEC was isolated from 24 of the 35 samples with presumptive presence of STEC, corresponding to a recovery rate of 69% and an overall STEC prevalence of 41%. Three samples (beef RMBD samples AT 15 and LS 01, and venison RMBD sample AT 11, respectively) contained two or more distinct STEC strains ( Table 2). A total of 28 STEC strains were retrieved. The types and categories of meat from which STEC-positive samples were recovered are shown in Table 1 and Figure 1.  Microorganisms 2021, 9, x FOR PEER REVIEW 5 of 14
The population structure of the strains was visualized by constructing a phylogenetic tree based on cgMLST. The isolates grouped according to serotypes and STs, but they were phylogenetically clearly distinct from each other, with ≥5 different alleles between each pair of neighboring isolates (Figure 2). The genomes of strains belonging to ST33, ST442, and ST641 were compared with the available genomes of corresponding STs present in the database of the Swiss National Reference Centre for Enteropathogenic Bacteria and Listeria (NENT) which collects all STEC strains from confirmed human cases nationwide and performs Illumina-based whole-genome sequencing. The cgMLST-based phylogenetic trees are shown in Figure 3 and details are available in Supplementary Table S3). None of the STEC strains belonging to ST33, ST442, or ST642 clustered with a strain in the database, thereby ruling out a direct match with any STEC of those STs reported from a case of human disease in Switzerland.   and Listeria (NENT) which collects all STEC strains from confirmed human cases nationwide and performs Illumina-based whole-genome sequencing. The cgMLST-based phylogenetic trees are shown in Figure 3 and details are available in supplementary Table S3). None of the STEC strains belonging to ST33, ST442, or ST642 clustered with a strain in the database, thereby ruling out a direct match with any STEC of those STs reported from a case of human disease in Switzerland.

Discussion
While recent years have seen a rise in popularity of feeding pets RMBDs, there is rising concern that this trend may come with the risk of exposure to zoonotic pathogens, including STECs. STECs constitute part of the flora of the gastrointestinal tract of a variety

Discussion
While recent years have seen a rise in popularity of feeding pets RMBDs, there is rising concern that this trend may come with the risk of exposure to zoonotic pathogens, including STECs. STECs constitute part of the flora of the gastrointestinal tract of a variety of healthy domestic and wild animals and may therefore contaminate meat during slaughter, evisceration, processing, and packing [1,2,34].
In this study, the presence of stx1 and stx2 genes was detected in 59% of the enrichment cultures, indicating that the overall contamination of STEC among RMBDs is high. In the majority (69%) of the stx-positive samples, the isolation of STEC strains confirmed the presence of the stx genes in viable bacterial cells. With an overall prevalence of 41%, the level of STEC contamination in the present study is considerably higher than what another study found previously, where STEC was isolated from 4% of raw pet food in the US [10]. It is also higher than the prevalence of 14% reported during an investigation of raw meat for dogs in the UK [11]. However, comparative data are still scarce, and discrepancies between results of different studies may be due to differences in the testing methodologies. Nevertheless, the present study provides evidence that the occurrence of STEC in RMBDs may currently be underestimated.
The pathogenic potential of STEC may vary according to the presence or absence of a variety of virulence factors. Thus, in addition to serotyping, comprehensive virulence gene profiling is important for risk assessment.
Notably, the serotype most frequently associated with HC and HUS, STEC O157:H7, was not detected in the present study, indicating that STEC O157:H7 may not be considered a suitable marker for STEC detection in RMBDs.
In our study, we found that 29% of the STEC isolated from RMBD harbored stx2a or stx2d, which are the stx subtypes that have the strongest association with HUS [14].
Other toxin genes, including ehxA and subA found in 57% and 61% of the STEC in this study, are considered important virulence markers for STEC pathogenesis and are frequently detected among human clinical isolates [21,35].
Despite the vast majority (93%) of the strains being negative for the eae gene, 25 of the 28 strains harbored iha, which is thought to contribute to pathogenicity of eae-negative STEC by facilitating attachment to intestinal cells [40]. Taken together, our data indicate that the STECs occurring in RMBDs have the potential to cause disease in humans.
By contrast, STEC-related illness appears to be rare in companion animals [41]. However, there are several studies that provide evidence for the intestinal carriage of STEC in dogs and cats, highlighting their potential epidemiological role as a source for human STEC infections [42][43][44][45]. Hence, it is possible that pets fed RMBDs contaminated with STEC could serve as asymptomatic shedders through their feces, transmitting the pathogen to humans and into the environment.
Interestingly, ten (36%) strains harbored one or more virulence factors, namely fyuA, kpsE, and papC, which are characteristic of extra-intestinal pathogenic E. coli (ExPEC), including uropathogenic E. coli (UPEC) [46][47][48]. STEC/ExPEC hybrid strains are rarely reported, but, nevertheless, they must be considered high-risk pathogens due to the possibility of a systemic infection in combination with gastrointestinal disease [49]. Furthermore, papC and fyuA are also prevalent among E. coli causing urinary tract infection (UTI) in cats and dogs [50,51]. Therefore, our data indicate that a subset of STEC present in RMBDs may have the potential to cause disease in pets as well as in humans.
Finally, in this study, six (21%) strains, including one STEC/ExPEC strain, carried two or more transmissible AMR genes, predominantly genes conferring resistance to aminoglycosides, which are antimicrobials categorized by the World Health Organization (WHO) as critically important in human medicine [52]. These findings are consistent with previous data that document the presence of AMR genes on mobile genetic elements in STEC [53]. AMR in STEC is worrisome because of the likelihood of horizontal transfer of resistance genes to other pathogens. In view of the ongoing global antimicrobial resistance problem, feeding RMBD to dogs that are undergoing antimicrobial treatment should be reconsidered in order to avoid selection and dissemination of AMR bacteria.

Conclusions
This study identified commercially available RMBDs as a potential source of STEC, including strains with serotypes, stx subtypes, and other virulence traits that are associated with human disease, such as BD, HC, and HUS. In view of the low infectious dose and potential severity of disease manifestations, the high occurrence of potentially harmful STEC in RMBDs represents a risk of infection for persons handling raw pet food and for persons with close contact to pets fed on RMBDs. Our data provide further evidence for the public health risks of raw feeding and highlight the importance of promoting awareness among veterinary and public-health agencies, RMBD suppliers, and pet owners, with the need to focus on safe and hygienic handling of RMBD to protect human and animal health.  Data Availability Statement: This whole-genome shotgun project was deposited at DDBJ/ENA/ GenBank under the accession numbers JAETXY000000000-JAEUYO000000000. The versions described in this paper are versions JAETXY000000000.1-JAEUYO000000000.1 (https://www.ncbi.nlm. nih.gov/nuccore, accessed on 1 July 2021). Raw sequence data are also available in the Sequence Read Archive (SRA) of the NCBI under BioProject no. PRJNA694525 (https://www.ncbi.nlm.nih.gov/sra/, accessed on 1 July 2021).