Zoonotic Escherichia coli is a potential Driver of Antimicrobial Resistance among Pastoralist Communities in Uganda: A Laboratory based Cross sectional Study


 Background Non-prescribed use of antimicrobials in Agriculture incurs a transfer risk of resistant pathogens to humans, complicating treatment. The aim of this study was to determine the potential of Zoonotic E. coli to serve as drivers of antimicrobial resistance among animals and humans in pastoralist communities in Kasese district, so as to protect the community.Materials and methods A laboratory based cross-sectional study was done using archived E. coli isolates previously obtained from humans in pastoralist communities of Kasese district, Uganda. Recovery of the isolates was done by conventional culture, and Identification by biochemical methods, serotyping and PCR. Antimicrobial resistance profiling was done by using Kirby Bauer disc diffusion method. The isolates were also screened for resistance mechanisms including Extended Spectrum β-lactamase (ESBL), Carbapenemase and AmpC production using disc diffusion based methods.Results The prevalence of Enterohemorrhagic E.coli (EHEC) was 16% (28/180). These EHEC isolates belonged to phylogroups; B 1 (94%, 26/28), B2 (3%, 1/28) and A (3%, 1/28). The most prevalent virulence gene was Stx1 (100%, 28/28) followed by Stx2e (94%, 26/28); none of the isolates was Stx2 positive. Highest resistance was seen to Cotrimoxazole (89%, 25/28), Tetracycline (71%, 20/28), Ampicillin (65%, 18/28) and Nitrofurantoin (28%, 8/28), these are the most commonly used antimicrobials in the agricultural sector in Uganda. Minimal resistance was observed to the antimicrobials that are commonly used in human medicine especially β-lactams, β-lactam+inhibitors and Carbapenems. Of the 28 zoonotic E. coli isolates, 17%, (5/28) were ESBL positive and among these 1 (3%, 1/28) was a Carbapenemase producer.Conclusion There is a high prevalence of highly pathogenic and resistant zoonotic E. coli among humans in pastoralist communities in Uganda. We suspect that these pathogens, along with their AMR genes, were acquired from animals because they largely contained the animal specific Vero toxin gene VT2e and majority belonged Pylo-group B 1 which has been documented as the most common EHEC phylo-group inhabiting domestic animals. We recommend that studies involving relatedness of drug resistant isolates from humans and animals should be conducted to ascertain the role of enterohemorrhagic E. coli in the zoonotic spread of antimicrobial resistance in pastoralist communities.


Introduction
Globally, the incidence of antimicrobial resistant pathogens is increasingly manifesting its self [1,2]. This has continued to make the treatment of infections caused by these pathogens very difficult and expensive, imparting a great risk of death [3]. Uncontrolled use of antimicrobials in farming is a major factor causing the emergence of resistant bacterial pathogens [4]. The emergent zoonotic strains may pass on to humans via the food chain [5][6][7]. At present, E. coli O157:H7 is one of major pathogens implicated in the zoonotic spread of antimicrobial resistance (AMR) globally [8]. In East Africa, and Uganda in particular, antimicrobial resistant E. coli with the potential for transmission from animals to humans are commonly implicated in human diarrheal diseases [9,10].
In Kasese district, western Uganda, farmers in and around Queen Elizabeth National park graze domestic animals within the Park. The interactions between domestic animals, wild animals and humans create a porous interface for zoonotic transfer of drug resistant pathogens [11]. This district has for long suffered outbreaks of antimicrobial resistant diarrheal diseases caused by zoonotic bacterial agents [12]. Though studies about these zoonotic pathogens have been done [9,13], little is known about the prevalence of antimicrobial resistant E.coli 0157:H7, and the potential role of this strain in the zoonotic spread of antimicrobial resistance among pastoralist communities in this area. 4 This study aimed to determine the prevalence, population structure, and antimicrobial resistance profiles of zoonotic E. coli, among bacterial isolates obtained from humans in pastoralist communities of Kasese district: to ascertain whether E. coli is a potential zoonotic driver of antimicrobial resistance in such settings, in order to guide AMR stewardship.

Study design
A Laboratory based cross-sectional study was conducted from January to August 2017 in the department of medical microbiology at Makerere University College of health Sciences (MakCHS).

Study population
One hundred and eighty (180) E. coli isolates archived in the medical microbiology laboratory (MakCHS), were used. The isolates had been obtained from patients of all ages and sexes, with fever and/or diarrhea, among pastoralist communities in Kasese district, Uganda.

Culture and identification of Zoonotic E. coli
The E. coli isolates were recovered from storage by culturing on blood agar followed by sub culturing on Sorbitol MacConkey agar. Upon inoculation onto the solid media, plates were incubated at 37 0 C for 18-24 hours. Presumptive identification was done using; colony characteristics, gram stain and Biochemical methods including citrate, urea, and 5 triple sugar iron and Sulphur indole motility. Further identification was done by Serotyping, and virulence genotyping using conventional PCR.
Serotyping: The translucent E. coli colonies from sorbitol MacConkey, following biochemical identification, were sub-cultured on Mueller Hinton agar and subjected to
Highest resistance was observed to Trimethoprim-sulphurmethoxazole followed by Tetracycline and then Ampicillin. Least resistance was observed in Ceftazidime, Imipenem and Nalidixic Acid. None of the Isolates was resistant to Cefepime, Ertapenem and Amikacin.

Discussion
We report a 16% prevalence of zoonotic E. coli in this study. Most of these zoonotic isolates harbored Vero toxin genes VT1 and VT2e. Of the 28 zoonotic E. coli isolates, 94% belonged to phylogroup B 1 , 3% belonged to phylogroup B 2 and 3% belonged to phylogroup A. Highest resistance in these isolates was seen to Cotrimoxazole followed by Tetracycline and Ampicillin. The most common mechanism underlying antimicrobial resistance was Extended Spectrum β-lactamases (ESBL).
The 16% prevalence of zoonotic E. coli reported in this study is higher than the prevalence reported in an earlier study conducted in Western Uganda by Majalija et al, 2008 which reported a prevalence of 8.5%. This rise may be partly attributed to the increasing number 9 of people settling and grazing animals in Queen Elizabeth National park. The domestic animals interact with wild animals and end up spreading such pathogens to humans.
The high prevalence of Vero toxin genes VT1 (100%) and VT2e (93%) observed in our study is in agreement with Omisakin et al, 2003, and reflects the high pathogenic nature of our zoonotic E. coli study isolates [15]. These Vero toxin genes mediate the virulence factors that enhance the organism's potential to cause disease [8,[16][17][18]. Vero toxin gene VT2e is animal specific and is common among isolates that have encountered the animal host [19]; hence its high prevalence in our study suggests a possible zoonosis.
Furthermore, Pylo-group B 1 , observed as the most prevalent phylo-type circulating among individuals in our study population has been documented as the most common enterohemorrhagic E. coli phylo-group inhabiting domestic animals [19,20]. This affirms our earlier assumption that these isolates are potential conduits for the zoonotic spread of antimicrobial resistance to humans who interact frequently with the animals (the one health concept).
Highest resistance was observed to; Trimethoprim-sulphurmethoxazole, Tetracycline, Ampicillin and Nitrofurantoin: these have been documented as the most commonly used antimicrobial agents in the agricultural sector in Uganda [21]. These finding are consistent with studies done elsewhere on similar isolates whose source was animals [22]. In addition, these studies also reported a possible transmission of the resistance in these isolates to humans. The implication of this is that the resistance seen in our zoonotic E.
coli study isolates may be of agricultural origin; however this remains speculative.
Our study successfully proved that 16% of the E. coli isolates obtained from humans in this pastoralist community are zoonotic and highly resistant to antibiotics, using 10 conventional culture and molecular methods. The zoonotic nature of these pathogens implies that they could have originated from animals from which they potentially transmit resistance to humans. We were unable to compare our study isolates with drug resistant isolates obtained from animals to certainly infer zoonotic transmission of antimicrobial resistance in this setting.
The high prevalence of resistant, enterohemorrhagic E.coli is of utmost significance because these bacteria have been reported as the most common causative agents of diarrheal disease, a major cause of death in humans especially children under five years [8,23]. The linkage of this serotype to zoonosis creates a risk for the development of MDR phenotypes when the already resistant zoonotic strains get exposed to new classes of antimicrobial agents while in the human host. The high antimicrobial resistance burden observed in this study setting is of global public health concern because the number of international tourists visiting Queen Elizabeth National park in this area continues to rise amidst the expanding local populations of animals and humans in this locality [24][25][26].
This global connectivity, which is now associated with the rapid worldwide spread of infectious agents and their resistance genes [24], implies that the AMR observed in our study may consequently spread to the rest of the world.

Conclusions
There is high prevalence of highly pathogenic and resistant zoonotic E. coli pathogens among humans in pastoralist communities in Uganda. We suspect that these were acquired from animals where the resistance observed in the human host may have originated. We therefore recommend that studies involving isolates from humans and animals should be conducted, and relatedness of the resistant isolates from the two host groups investigated, to confirm the role of enterohemorrhagic E. coli in the zoonotic 11 spread of antimicrobial resistance in this setting. A one health approach should be used to establish the drivers of MDR spread in pastoralist communities.

Consent for publication was approved by the School of Biomedical Sciences, Higher
Degrees Research and Ethics Committee (HDREC) of Makerere University.

Availability of data and material
Datasets generated and/or analyzed during this study are available from the corresponding author on reasonable request.

Competing interests
There were no competing interests in this study.

Authors' contributions
"AW Performed the laboratory work, data entry and analysis, and wrote the primary draft of the manuscript.
"JSI Supervised the laboratory work, advised on the data analysis plan and was a major contributor in writing the manuscript.
"HK provided some of the laboratory supplies and was a major contributor in writing the manuscript.
"CFN was a major contributor in writing and editing the manuscript.
"BBA Was the senior advisor and supervisor in the study, availed the necessary funds, was a major contributor in writing the manuscript and performed final editing of the manuscript.
All authors read and approved the final manuscript."