Food animals as reservoirs and potential sources of multidrug-resistant diarrheagenic E. coli pathotypes: Focus on intensive pig farming in South Africa

Background Diarrheagenic E. coli (DEC) strains are a major cause of diarrheal diseases in both developed and developing countries. Healthy asymptomatic animals may be reservoirs of zoonotic DEC, which may enter the food chain via the weak points in hygiene practices. Aim We investigated the prevalence of DEC along the pig production continuum from farm-to-fork. Methods A total of 417 samples were collected from specific points along the pig production system, that is, farm, transport, abattoir and food. E. coli was isolated and enumerated using Colilert. Ten isolates from each Quanti-tray were selected randomly and phenotypically identified using eosin methylene blue agar selective media. Real-time polymerase chain reaction (PCR) was used to confirm the species and to classify them into the various diarrheagenic pathotypes. Antimicrobial susceptibility was determined against a panel of 20 antibiotics using the Kirby-Bauer disk diffusion method and EUCAST guideline. Results The final sample size consisted of 1044 isolates, of which 45.40% (474/1044) were DEC and 73% (762/1044) were multidrug-resistant. Enteroinvasive E. coli (EIEC) was the most predominant DEC at all the sampling sites. Conclusion The presence of DEC in food animal production environments and food of animal origin could serve as reservoirs for transmitting these bacteria to humans, especially in occupationally exposed workers and via food. Adherence to good hygienic practices along the pig production continuum is essential for mitigating the risk of transmission and infection, and ensuring food safety.

Diarrheagenic E. coli (DEC) strains are considered major causes of diarrheal diseases in developed and developing countries (Aijuka et al. 2018;Estrada-Garcia & Navarro-Garcia 2012). Foodborne diseases resulting from the consumption of food contaminated by DEC have been recognised amongst the most challenging health issues worldwide (Galli et al. 2016).
Members of the DEC group are classified into six pathotypes, that is, enteroaggregative E. coli (EAEC), enteropathogenic E. coli (EPEC), enterotoxigenic E. coli (ETEC), enteroinvasive E. coli (EIEC), diffusely aggregative E. coli (DAEC) and enterohemorrhagic/Shiga toxin-producing E. coli (EHIEC/ STEC) (Acosta et al. 2016). The EPECis associated with infantile and persistent diarrhea. It is mainly detected by the presence of the eae gene (Ochoa & Contreras 2011). The ETEC is an important cause of diarrhea in children and travellers (Isidean et al. 2011). It is characterised by the presence of a heat-stable (ST) and/or heat-labile (LT) enterotoxin gene, encoding colonisation factors and toxin production (Gomes et al. 2016). This pathotype remains a problem for humans, pigs and calves. It causes diarrhea in neonatal and recently weaned piglets, which is considered one of the most important diseases affecting pig farming economically (Melkebeek, Goddeeris & Cox 2013). The EIEC causes dysentery in humans, is closely related to Shigella (Hosseini Nave et al. 2016), and can be distinguished from other E. coli by the detection of the ipaH gene (Van den Beld & Reubsaet 2012). The EAEC has been associated with persistent diarrhea in children, travellers and humans with immunodeficiency virus infections; it induces chronic inflammation in the absence of dysentery (Okhuysen & Dupont 2010). The EAEC is characterised by the transcriptional activator encoding aggR gene (Wang et al. 2017). The EHEC/STEC infections range from mild to severe, complicated, bloody diarrhea and haemolytic uremic syndrome (Friesema et al. 2011). This pathotype produces one or more types of Shiga toxin (Stx1 and/or Stx2) (Dias et al. 2016). The DAEC pathotypes consist of a heterogeneous group of organisms with variable virulence factors that may play a role in causing sporadic diarrheal illnesses, particularly in paediatric patients (Wang et al. 2017), and are detected by the presence of the daaD gene (Riveros et al. 2017).
Apart from causing infections, E. coli pathotypes have been reported globally for their resistance to numerous antibiotics, including those used in clinical medicine. The escalation of antibiotic resistance in pathogenic bacteria is a major global public health concern (Fernandez, Bert & Nicolas-Chanoine 2016) because it has led to prolonged illness and higher treatment failure rates (Van den Honert, Gouws & Hoffman 2018). This concern is further exacerbated by the emergence of multidrug-resistant (MDR) bacteria, resulting in reduced treatment options for the infections they cause (Lammie & Hughes 2016). The MDR E. coli has been increasingly isolated from livestock and animal products (Kallau et al. 2018). This increase has been attributed to increased antibiotic use and suboptimal biosecurity programmes on farms (Mbelle et al. 2019).
Human exposure to MDR DEC could lead to disease outbreaks with severe adverse public health consequences. Knowledge of the prevalence and distribution of these MDR pathotypes within the pig production continuum is therefore imperative. We investigated the prevalence of MDR DEC pathotypes in a pig production continuum using a farm-tofork approach. Such information could help identify areas needing attention across the continuum, to prevent the spread of these infectious agents to humans through both meat products and environmental exposure.

Study design, sample collection and enumeration of E. coli
The study was a longitudinal study conducted over 18 weeks (September 2018 -January 2019) covering the farm-to-fork pig production continuum. It included sampling on the farm, transport system (truck) and attached abattoir, using the guidelines of the World Health Organization Advisory Group on Integrated Surveillance of Antimicrobial Resistance (WHO-AGISAR) (WHO 2017).
A total of 417 samples were collected and processed as previously described (Abdalla et al. 2021). Samples consisted of faeces, litter and slurry (farm), swabs of transport vehicles, and ceacal, carcass swabs and final meat cut swabs (abattoir). Briefly, the defined substrate ColilertTM-18 system from IDEXX (IDEXX Laboratories (Pty) Ltd., Johannesburg, South Africa) was used to detect and quantify E. coli according to manufacturer instructions. All processed samples were incubated for 18-24 h at 37 °C and examined under UV light for fluorescence.

E. coli confirmation and detection of pathotypes
Pure E. coli strains were obtained from fluorescent Quanti-Tray wells as previously described (Abia, Ubomba-Jaswa & Momba 2015). Ten isolates were randomly selected from each of the 417 samples and phenotypically identified on eosin methylene blue agar (HiMedia Laboratories Pvt. Ltd., Mumbai, India). The isolates were further streaked on nutrient agar (Neogen, Lansing, Michigan, United States) and incubated at 37 °C for 24 h. Deoxyribonucleic acid (DNA) was extracted from the isolates using the boiling method (Amoako et al. 2019). Real-time polymerase chain reactions were used to confirm E. coli and determine the various DEC pathotypes targeting specific genes (Appendix 1). The reaction mixtures and thermal cycling conditions were as previously described (Abia et al. 2015), except for the master mix where the Luna® Universal qPCR Master Mix (New England Biolabs, Ipswich, Massachusetts, United States) was used in this study. After the final extension step, a melt curve was generated and analysed as previously described (Molechan et al. 2019). All reactions were performed on a QuantStudioTM 5 (ThermoFischer Scientific, Bedford, MA, USA). The DNA from reference E. coli strains was used as positive controls (Table 1-A1), whilst the reaction mixture with no DNA (replaced with nuclease-free water) was used as a no template control. All controls were obtained in-house from the Antimicrobial Research Unit microbial bank.

Antibiotic susceptibility testing
The antibiotic susceptibility profiles of the confirmed diarrheagenic isolates were determined against 20 antibiotics using the disk diffusion method on Muller Hinton Agar (Oxoid, Basingstoke, Hampshire, England) as previously described (Abdalla et al. 2021). The E. coli ATCC® 25922 was used for quality control. The diameters of the zones of inhibition were measured and interpreted according to the European Committee on Antimicrobial Susceptibility testing breakpoints (EUCAST 2017). The Clinical and Laboratory Standards Institute guidelines (CLSI) were used (CLSI 2017) for antibiotics that did not have published breakpoints in the EUCAST guidelines. All the antibiotic discs were purchased from Oxoid (Basingstoke, Hampshire, England). Isolates showing resistance to ≥ 1 agent in > 3 distinct antibiotic classes were considered as MDR (Amoako et al. 2016). A multiple-antibiotic resistance index (MARI) was calculated for these isolates to ascertain whether these isolates originated from high antibiotic use environments (Abdalla et al. 2021).

Statistical analysis
The data were analysed using the Statistical Package for the Social Science (SPSS) version 26 (IBM Corporation, Armonk, New York, United States). Descriptive statistics were used to describe the frequency of DEC along the pork production chain. The statistical significance of the differences in counts and DEC prevalence between different sources was determined using the one-way analysis of variance (ANOVA) with Tukey's Honestly Significant Difference (HSD) post-hoc test. A p-value < 0.05 was considered statistically significant.

Ethical considerations
Ethical approval was obtained from the Animal Research Ethics Committee (AREC 073/016PD) and the Biomedical Research Ethics Committee (BCA444/16) of the University of KwaZulu-Natal. A Section 20A permit (12/11/1/5) was further obtained from the South African National Department of Agriculture, Forestry and Fisheries.

Mean E. coli count per sampling site
The E. coli was isolated from all the samples collected in this study. The mean E. coli concentrations per sampling site are shown in Figure 1. The highest mean E. coli count was recorded in the faecal samples (1.59 × 10 6 MPN/100 mL), whilst the lowest was recorded in truck samples. There was an overall statistically significant difference ( p = 0.002) between the E. coli counts from the different sampling points (Table 2-A1). The post-hoc analysis revealed a statistically significant difference in the overall E. coli count between the farm and truck ( p = 0.045), and the farm and abattoir ( p = 0.004). However, no statistically significant difference was observed between the truck and abattoir E. coli counts ( p = 0.183).

Identification of E. coli pathotypes
A total of 1044 isolates were confirmed as E. coli using realtime PCR. Pathotyping using pathotype-specific genes showed that 45.40% (474) of the isolates carried at least one of the virulence genes tested (Table 1). The highest percentage of DEC was isolated on the farm, whilst the lowest was found at the transport (truck). The EIEC at 19% was the most prevalent pathotype, whilst EHEC was the least prevalent (0.1%).
Like with the abundance of E. coli across the continuum, there was an overall statistically significant difference ( p = 0.000) between the prevalence of the DEC pathotypes from the different sampling points (Table 3-A1). Similarly, Tukey's HSD post-hoc analysis of the DEC pathotypes prevalence was statistically significantly different between the farm and truck ( p = 0.000), farm and abattoir (p = 0.039) and truck and abattoir ( p = 0.006) ( Table-3A1).

Antibiotic susceptibility profiles
Only 1% (5 isolates) of the total DEC was susceptible to all the antibiotics been tested.
Overall, the highest resistance was against tetracycline 90.5% (429), whilst all the isolates were susceptible to meropenem ( Figure 2); the percentage resistance to each antibiotic differed by source ( Figure 3).

Discussion
This study investigated the prevalence of DEC pathotypes and their antibiotic resistance profiles in intensive pig farming in uMgungundlovu District, KwaZulu-Natal, South Africa, using the farm-to-fork approach. The E. coli was recovered from all the samples along the pig production continuum. Of the total number of 1044 E. coli isolates constituting the sample size across the continuum, 45% were DEC, with EIEC being the most prevalent pathotype. Also, 99% of the isolates were resistant to at least one of the antibiotics tested, 73.84% were MDR, most recording a MARI greater than 0.2.

Enumeration of E. coli across the continuum
The enumertion results reported in this study are lower than those previously reported in an earlier study in South Africa (King & Schmidt 2017) and Bulgaria (Petkov et al. 2006). However, the reported results in the current study were higher than those reported in a study involving 13 pigs farms in Australia (Chinivasagam et al. 2004). Despite the discrepancies between our study and the previous study, the E. coli abundance in the faeces observed in these studies was expected, as the human and animal intestines serve as reservoirs (Jafari, Aslani & Bouzari 2012). Unlike most studies that focus on farms or slaughterhouses, the present study used a farm-to-fork approach, ensuring adequate evaluation of the presence and abundance of E. coli along the entire continuum. Thus, although E. coli was statistically significantly most abundant in the farm, substantial numbers were also recorded at the abattoir, whilst the lowest E. coli counts were found in the truck samples. The presence of E. coli at the abattoir revealed that although the scalding and singeing method has a significant impact in lowering the E. coli   (2002). Whilst the abundance of E. coli in the faecal samples may not be considered a significant threat to human health, the considerable numbers in the meat portions could represent a meat safety challenge that needs to be addressed by ensuring that more efficient hygienic practices are applied to prevent packaging of contaminated meat. It has been demonstrated that meat protein is a potential source of bacterial foodborne disease because of its high protein content, water activity and approximately neutral pH, allowing the proliferation of bacteria (Jaja, Green & Muchenje 2018).

Prevalence of diarrheagenic E. coli pathotypes
The DEC pathotypes continue to be a major health threat globally because of the disease outbreak-causing potentials.
Thus, identifying all possible reservoirs of these pathotypes is essential to ensure public health. In the current study, the highest prevalence of DEC was found on the farm. This was logical because the samples isolated from the farm were predominantly faecal. Also, the number of samples isolated from the farm was much higher than those isolated from the truck and abattoir, increasing the chances of isolating the pathotypes at farm level. The prevalence of DEC in the truck after the loading of pigs was higher, indicating their possible faecal sources.
Although different DEC pathotypes, including EPEC, have been implicated in colibacillosis in pigs, ETEC is the most frequently isolated pathotype (García-Meniño et al. 2018).
The ETEC coli is a major cause of severe diarrhea in suckling and weaning animals, and is a cause of significant losses in the African pig industry (Kagambega et al. 2012;Kylla et al. 2019). The relatively low prevalence of EPEC and ETEC could be because this study was conducted on healthy animals, and no incidence of diseased pigs was recorded throughout the study.
Apart from EPEC and ETEC, the other DEC pathotypes were also recorded in the current study, albeit at low prevalences ( this dominant occurrence could not be explained, it is important to note that this pathotype is highly implicated in bacillary dysentery in humans, particularly in low-income countries (Pasqua et al. 2017). It has also been implicated in foodborne disease outbreaks even in highly industrialised countries with improved hygiene and sanitation, such as the United States (Venkitanarayanan & Doyle 2003). Therefore, its detection across the continuum, including substantially at the abattoir, calls for stringent implementation of hygienic protocols in intensive pig farm systems. We also detected EHEC (1%), including O157:H7 (3%), at the farm level. Despite their low prevalence, this percentage was still significant because, although most pathogenic E. coli are infective at high doses, EHEC requires only ten cells to be infective (Limayem & Martin 2014).

Antibiotic susceptibility profiles
The growing problem of antibiotic resistance has become a significant public health concern (Tanih et al. 2015). In South Africa, sulphonamides, tetracyclines, macrolides, ionophores, quinoxalines, glycolipids, polypeptides, streptogramins, oligosaccharides, phosphonic acid, nitroimidazoles and polymeric compounds are registered in various dosage for veterinary use and freely available to farmers for prophylaxis, metaphylaxis, growth promotion and infection treatment (Eagar & Naidoo 2017). This could explain the antibiotic resistance profiles observed, with most isolates showing resistance to tetracycline, chloramphenicol, ampicillin and trimethoprim-sulfamethoxazole ( Figure 2a). The incidence of drug-resistant DEC pathotypes revealed that pork might pose a public health risk. Although tigecycline and imipenem recorded low resistance rates in the current study, they are still alarming because they are considered last-resort antibiotics for human use. The high percentage of MDR isolates obtained in the current study suggest that these organisms were exposed to high antibiotic use environments. This is substantiated by the fact that most of the MDR isolates recorded MARIs greater than 0.2. A MARI greater or equal to 0.2 indicated potential high-risk as the organisms probably originate from environments where antibiotics are extensively used (Teshome et al. 2020).
Diarrhea caused by foodborne pathogens such as DEC pathotypes is an important cause of death, especially in children in low-and middle-income countries (Jaja et al. 2018). The rise in MDR DEC aggravates the situation because MDR strains are known to limit treatment options. In order to minimise the risk and prevent microbial contamination along the food production chain, good hygiene manufacturing practice (GHMP) and hazard analysis critical control points (HACCP) practices should be implemented, in addition to appropriate retail and consumer meat handling and processing (Galli et al. 2016). In South Africa, meat safety control is a shared responsibility between the Department of Agriculture, Land Reform and Rural Development (DALRRD) and the Department of Health (DoH). Approaches to meat safety control include farm-to-fork control with the publication of HACCP regulations.

Conclusion
The presence of MDR DEC pathotypes in animal food production environments and food of animal origin indicates that these animals serve as reservoirs and potential sources of these pathogenic organisms that could be transmitted to humans. This is especially in occupationally exposed workers and through consumption of undercooked pork. Adherence to good hygienic practices along the pig production continuum and thorough cooking are essential for mitigating the risk of transmission and infection and ensuring food safety.

Acknowledgements
We would like to acknowledge managers and workers at both the farm and abattoir for their helpful cooperation.

Disclaimer
Any opinion, finding and conclusion or recommendation expressed in this material are those of the authors and do not necessarily reflect the views of the organisations or agencies that provided support for the project, and the publisher. The funders had no role in the study design or the decision to submit the work for publication.