Detection and characterization of ESBL-producing Enterobacteriaceae from the gut of subsistence farmers, their livestock, and the surrounding environment in rural Nepal

The increasing trend of gut colonization by extended-spectrum β-lactamase (ESBL) producing Enterobacterales has been observed in conventional farm animals and their owners. Still, such colonization among domesticated organically fed livestock has not been well studied. This study aimed to determine the gut colonization rate of ESBL-producing Enterobacteriaceae and carbapenemase-producing Enterobacteriaceae (CPE) among rural subsistence farming communities of the Kaski district in Nepal. Rectal swabs collected by systematic random sampling from 128 households of subsistence farming communities were screened for ESBL-producing Enterobacteriaceae and CPE by phenotypic and molecular methods. A total of 357 (57%) ESBL-producing Enterobacteriaceae isolates were obtained from 626 specimens, which included 97 ESBL-producing Enterobacteriaceae (75.8%) from 128 adult humans, 101 (79.5%) from 127 of their children, 51 (47.7%) from 107 cattle, 26 (51%) from 51 goats, 30 (34.9%) from 86 poultry and 52 (42%) from 127 environmental samples. No CPE was isolated from any of the samples. blaCTX-M-15 was the most predominant gene found in animal (86.8%) and human (80.5%) isolates. Out of 308 Escherichia coli isolates, 16 human and two poultry isolates were positive for ST131 and were of clade C. Among non-cephalosporin antibiotics, the resistance rates were observed slightly higher in tetracycline and ciprofloxacin among all study subjects. This is the first one-health study in Nepal, demonstrating the high rate of CTX-M-15 type ESBL-producing Enterobacteriaceae among gut flora of subsistence-based farming communities. Gut colonization by E. coli ST131 clade C among healthy farmers and poultry birds is a consequential public health concern.

www.nature.com/scientificreports/ Antibiotic susceptibility testing. All the 394 ESBL-producing Enterobacteriaceae isolates were tested for susceptibility to various antibiotics. The antibiotic resistance profile of ESBL-producing Enterobacteriaceae isolates of each group is presented in Fig. 1, supplementary file 7: Table S4. Besides cephalosporin antibiotics, the highest rate of resistance was found against amoxicillin/clavulanic acid (63.1% of adult, 28.2% of cattle ESBL-producing Enterobacteriaceae isolates) followed by tetracycline (47% of the goat isolates, 25% of the environmental isolates) and ciprofloxacin (34.5% of the goat isolates, 14.9% of cattle isolates). Among animals, the poultry isolates showed the highest resistance to tested antibiotics, and the resistance rate of these poultry isolates towards nitrofurantoin was found to be maximum. All the ESBL isolates were susceptible to tigecycline, imipenem, meropenem, and ertapenem (Figs. 1, 4).
Multi-drug resistance pattern. As shown in Fig. 2 and supplementary file 1: Fig. S1, the highest multidrug resistance (non-susceptibility to at least one drug in three or more antibiotic categories) level was observed among the isolates of goat (61.5%) followed by poultry (60%). Adult (56.2%), child (55%), and environmental (52.3%) isolates had shown an almost similar level of multi-drug resistance. The lowest MDR percentage was found among the strains of cattle (39.4%). However, 7.7% of the goat, 3.1% of the environment, and 1.8% of child ESBL-producing Enterobacteriaceae isolates showed resistance to six antibiotics classes.

Discussion
Antibiotic resistance is recognized as a complex public health challenge because health risks arise from the resistant bacteria and their mobile genetic element transfer among humans, animals, and the environment through multiple interfaces with subsequent global dissemination 16,17 . Therefore, an international movement called One-Health is developed for an integrative approach to work in a sustainable way to inscribe the health risks at the human-animal-ecosystem interlinks 18 . Especially in LMICs, the emergence and re-emergence of disease by resistant bacteria are mainly due to poor sanitation, close interactions with livestock, climate change, easy access and irrational use of antibiotics, human behavior changes, unhygienic food preparation, and consumption practices 19 . The prevalence of ESBL-producing Enterobacteriaceae and CRE gut colonization in healthy humans and animals and their surrounding environment have been described previously in many regions of the world, particularly in developed countries 15 . Most of these studies were on commercial farm animals and their owners, which demonstraedt the high prevalence of AMR in conventional farming. However, such studies on organic feed livestock were sparse. This is the first one-health study investigating the occurrence of ESBL-producing Enterobacteriaceae and CPE in the gut of subsistence farming communities of rural areas of western Nepal. About 85 percent of the 26 million people of Nepal live in rural areas and are engaged in farming. The majority of farming is subsistent in nature, and the crop is mostly integrated with livestock (cattle, goat, pig, and poultry). The selling of livestock products in the local market is one of the crucial income sources of farm households. The subsistence-based agricultural community often includes local breeds' of organically raised farm animals and birds. The feed used usually is of homegrown cereals and grasses. This livestock is rarely exposed to antibiotics via food additives or treatments. In this study, a high proportion (57%) of ESBL-producing Enterobacteriaceae  Besides, selective pressures of antimicrobials present in healthcare and community settings, and movement of such genes by mobile genetic elements responsible for intracellular (insertion sequence elements, transposons, and integrons) and intercellular (plasmids and integrative conjugative elements) mobility can facilitate the emergence and dissemination of AMR 20 .
Healthy subjects of the community are a vital reservoir for ESBLs, and global surveillance studies showed that the colonization rate was found to be increasing by approximately > 5% annually 7,21 . The rates are higher in the West Pacific (46%), Southeast Asia (22%), African (22%), and the eastern Mediterranean (15%) 21 . Worryingly, high rates of > 50% of ESBL-producing E. coli and Klebsiella spp. have been reported in some regions of Asia, Africa, and Latin America 22 . As stated in our previous study, the high rate of ESBL colonization (53.4%) with predominant genotype CTX-M-15 was found in healthy subjects of the rural community in Western Nepal 23 . Still, in the present study, it is striking to note that 77.6%, 43.9%, and 48% ESBL-producing Enterobacteriaceae with the predominant bla CTX-M gene were observed among the gut of healthy humans, their reared animals, and the environment, respectively. Such a high percentage of ESBL-producing Enterobacteriaceae colonization in community settings underlines the risk to human and animal health due to disseminating resistant bacteria or their resistance genes via foodborne transmission or environmental routes, such as farm waste. These ESBLproducing Enterobacteriaceae isolates were susceptible to tigecycline and carbapenem only. No carbapenemresistant or carbapenemase-producing Enterobacteriaceae was isolated in our study. In China, 2.4% of healthy people harbored CRE as enteric microbiota 24 . Similarly, Li et al. (2019) recently reported the presence of New Delhi Metallo-beta-lactamase type carbapenem-resistant E. coli (NDM-EC) among humans and backyard animals. Their study was the first to report the direct transmission of NDM-EC between humans and animals 25 .
In Nepal, the most common antibiotics used in animal sectors are tetracyclines, sulfa drugs, macrolides, polymyxins, bacitracin, nitrofurans, quinolones, and aminoglycosides, whereas chloramphenicol is the least antibiotic consumed in the veterinary sector. The rate of cotrimoxazole-resistant E. coli isolates from buffalo meat www.nature.com/scientificreports/ was reported as high as 79.6%, whereas chloramphenicol-resistant E. coli isolates from the milk sample were only 26% 26 . On the other hand, cotrimoxazole, amoxicillin, ciprofloxacin, chloramphenicol, nalidixic acid, gentamicin, and cephalosporins are commonly used in human health sectors 27 . In this study, higher than 50% [goat (61.5%) followed by poultry (60%), adult (56.2%), child (55%) and environmental (52.3%)] of ESBL-producing Enterobacteriaceae isolated from each subject showed an almost similar pattern of MDR with the highest resistance towards amoxicillin/clavulanic acid followed by tetracycline and ciprofloxacin as demonstrated in Fig. 2. The healthy children in this study had a similar fecal carriage rate of MDR isolates, as in a previous study reported from Bangladesh 28 . However, the rate of MDR in healthy adult humans was comparatively higher than that of studies previously conducted in Nepal 29 . The MDR bacterial colonization in animal and environmental isolates are comparatively lower in this study than that of a survey carried out in Nigeria 30 and India 31 . A single study determining the rate of MDR in all domains of one-health like environment, healthy humans and animals is limited. In this study, the ESBL-producing Enterobacteriaceae isolates of humans showed the highest (63.1%) resistance towards amoxicillin/clavulanic acid, and goat isolates showed 47% resistance towards tetracycline. The environmental strains showed 40.5% resistance towards nitrofurantoin. The highest resistance rate towards ciprofloxacin was observed in the ESBL-producing Enterobacteriaceae of goat isolates (34.5%). The rest of the antibiotic resistance rates by all three one-health domains showed less than 30%. The occurrence of combined   www.nature.com/scientificreports/ resistance towards ampicillin, tetracyclines, and cotrimoxazole has become common in ESBL isolates as the genes encoding resistance to these antibiotics are located on the same plasmid 32 . Farmers in rural areas of Nepal are still following traditional farming practices and are less likely to implement proper management and hygiene practices 26 . The irrational use of antibiotics to reduce mortality due to infection in cattle, pigs, and poultry without prior consultation of a veterinarian 26 is customary and widespread among conventional farming in Nepal. Appropriate and safe use of antibiotics in livestock is rarely practiced 25 . Subsistent farmers commonly use untreated manure as fertilizer or for bedding and irrigation of agricultural lands, thus discharging antimicrobial residues, resistant bacteria, and genes into the environment. Besides, frequent contact with a wide variety of animals without standard disinfection and hygiene procedures may lead to MDR-strains transmission to humans 33 .
The main drivers of AMR development in humans are self-medication, over-prescription, under-prescription, syndromic management approaches, and irrational prescription of powerful antibiotics for a speedy cure, lack of well-equipped hospitals and clinics 1,23,26 . In our study, rectal samples of healthy subjects and environmental samples were collected after the disastrous earthquake struck in Nepal in 2015. The alarming rate of ESBL-producing Enterobacteriaceae (57%) in this study might be influenced by post-earthquake scenario. There was a dramatic increase in water-borne diseases and the destruction of irrigation and drainage canals, releasing more pathogenic strains into the environment. Despite considering the factors influenced by a natural disaster, the percentage of healthy subjects carrying ESBL-producing Enterobacteriaceae strains in our community was comparable to that of our neighboring countries; 68% in India 34 , up to 52% in Pakistan 35 , and 50% in China 25 .
The global picture of variant CTX-M is complex, but it is evident that bla CTX-M-15 has increased over the years and is dominant in most regions. The fecal carriage rate of CTX-M-15 is prevailing in Asia, whereas group 9 variants, especially CTX-M-14 in China, South East Asia, South Korea, Japan, and Spain, and CTX-M-2 in South America, remain significant 36 . CTX-M-15 type ESBL-producing Enterobacteriaceae reservoirs of environment and livestock regularly exchange clones, and mobile genetic elements with the human reservoir by transposition or transduction, leading to clonal and epidemic plasmid spread 9,36 . In this study, CTX-M-15 is found to be the dominant gene in the environment (100%), cattle (100%), poultry (100%), goat (84.6%), human adult (95.2%), and child (97.9%). In fact, the river, soil, and nearby drinking water sources of Nepal are contaminated with high levels of fluoroquinolones through wastewater effluents of pharmaceutical industries 26 . The pollution of the environment from human and animal waste, low sanitation standards, and contaminated drinking water increases the cycling of CTX-M possessing ESBL-producing Enterobacteriaceae between humans and the environment 36 . It can be hypothesized that reduced access to the lavatory in addition to high population density, high migration rate (remittance serves 25% of GDP, highest among South Asian countries; https ://nra.org.np/nra_news/remit tance -keepi ng-econo my-afloa t/) can make Nepal one of the epicenters of dissemination of ESBL-producing Enterobacteriaceae carrying bla CTX-M-15 . The situation was similar in India, where ESBL-producing Enterobacteriaceae carriage rates (> 68%) were the highest in the world. Such a tremendous rate warrants the dissemination of gut microbes carrying genes such as mcr-1 and bla NDM-1 34,36 . The persistence of resistance in commensal E. coli is a significant marker for the selective pressure enforced by antibiotic use and subsequent resistance predicted in pathogens 37  ) and E. coli (11.4%). One of the prominent human AMR high-risk clones includes E. coli ST131 with bla CTX-M types 20 . High-risk clone of ESBL-producing E. coli ST131 is dominating globally. Clade C is the most common global clade among clinical ST131 and is associated with fluoroquinolone resistance 9 . Sherchan et al. 38 revealed that > 90% of clinical ESBL-producing E. coli isolates in Nepal were CTX-M-15 positive, and more than half possessed ST131. ESBL-producing E. coli carrying ST131 isolates of dogs and cats were first reported from a Portuguese study 39 . The E. coli ST131 is seldom responsible for nosocomial outbreaks 9 . The transmission mode of E. coli ST131 in the community setting is currently unknown 9 . Therefore, investigations regarding the roles of environmental reservoirs, companion animals, and direct or indirect person-to-person transmission are epidemiologically pertinent in the community transmission of ST131 9 . We found that 9.2% of humans (adults 9, children 7) and 5.7% of poultry ESBL-producing E. coli isolates carried ST131 clade C with bla CTX-M-15 . The low prevalence of ST131 in non-human isolates suggests that ST131 might be originated from human sources and transmitted directly or indirectly among humans 9,40 . The study conducted by Ewers et al. 41 found that many human and companion animals' clinical ST 131 E. coli shared similar virulence, resistance, plasmid content, and Pulsed-field Gel Electrophoresis (PFGE) profile. The emergence of E. coli ST131 is due to the widespread use of fluoroquinolones and oxyimino-cephalosporins. Compared to other Extra-intestinal Pathogenic E. coli (ExPEC) clones or different ST131 clades, E. coli ST 131 clade C is found to be inherently more adaptable in the environment, even without antimicrobial selection pressure. Moreover, a single high-risk clone of E. coli ST131 clade C plays a significant role in the worldwide distribution of ESBL-producing bacteria 9,20,37 .
There are rules and regulations regarding the judicious use of antibiotics and the infection control system in Nepal. However, the problem lies in implementation 42 . Such non-compliance activities and guidelines have expedited the further emergence and spread of resistant microbes 43 . To restrict AMR risk and mitigate its effect on human and animal health, a multidisciplinary approach involving public-private collaborators and government agencies is required. It is crucial to enforce initiatives such as immunization, public literacy, better hygiene standards, antibiotic stewardship, genomic surveillance program, decreased use of antibiotics in agriculture and livestock, and good husbandry practices. Moreover, proper waste segregation, disposal system, handling, transport, and medical waste treatment should be practised to mitigate the direct effects on human health and the environment. Thus, public health and environmental professionals, health care practitioners, and veterinarians www.nature.com/scientificreports/ should regard the ' one-health approach' as a professional imperative for the shared interests of health promotion and global tackling of antimicrobial resistance.

Limitations
Although the study is foremost in Nepal, there are some limitations. The study lacks an extensive investigation of the different sequence types and the clonal diversity of ESBL-producing Enterobacteriaceae isolates. Similarly, carbapenemase genes were not amplified by genotype based methods. Moreover, all PCR products were not subjected to sequencing to affirm the ESBL variants. Therefore, intensive one-health surveillance, molecular sorting, and genomics-based research will help in understanding the significance of MDR gut flora in the dissemination of AMR and its interdependence in human and animal infections.

Conclusion
This is the first one-health study in western Nepal, determining the high rate of CTX-M-15 type ESBL-producing Enterobacteriaceae among gut flora of subsistence-based farming communities. Gut colonization by E. coli ST131 clade C among healthy farmers and poultry birds is a significant public health problem. The wide dissemination of ESBL-producing Enterobacteriaceae among organically raised livestock and owners, including their children, may contribute to the shortfalls in infection control practices and public health management in western Nepal. Further implementation of AMR mitigation strategies (the framework to identify, prioritize and implement) across the one-health spectrum is essential to combat and unravel the complexities associated with the emergence, evolution, and dissemination of antimicrobial resistance in subsistence-based farm settings.

Materials and methods
Study design, site, and enrollment of participants. During the study period, rural areas of districts in Nepal were subdivided into federal entities called Village Development Committees (VDCs), and the urban or metropolitan regions were divided into wards. Nepal was divided into 14 zones, and the study area Kaski district was under the Gandaki zone (Fig. 5). The Kaski district was divided into 47 VDCs and two metropolitan areas (Pokhara and Lekhnath), with a total population of 492,098 (2011 census). The population in each VDC was around 3,000-12,000. Based on statistical power calculations and the expected farming communities in the Kaski district, our goal was to obtain 125 subsistence farming families. All 47 VDCs were numbered according to alphabetic order; 23 (49%) VDCs were randomly selected applying the formula = RANDBETWEEN (1, 47). The random numbers procreated were 1,2,4,15,40,27,16,33,31,5,22,17,37,26,45,34,23,9,41,39,8,13, and 28.
The study area of the Kaski district is demonstrated in Fig. 5. Houses were randomly selected from the 2008 voter registration list provided by the Nepal Government Electoral Commission. The recruitment of subsistence farmers' houses was made by visiting the residences in the selected VDCs who fulfilled the inclusion criteria. During the study period (May 2016 to December 2018), a rectal swab specimen was collected from healthy subjects of 128 houses involved in subsistence farming who volunteered to participate in the study. On average, five houses were selected per VDC. The subsistence farming communities having one adult and one child and a minimum of one livestock were involved in specimen collection. Organic fed pasture-raised animals and their owners and children without a history of antibiotic consumption or hospitalization during the last three months at the time of sample collection were included in the study. The exclusion criteria included; participants who refused to provide samples and consent, a history of hospitalization (< 3 months), antibiotic consumption (< 3 months), family members working in hospitals, and the presence of unhealthy animals. The selected residences, where either owner or livestock was not available during the visit for sample collection, were also excluded.
Specimen collection and questionnaire. The rectal swab from human, livestock and environmental specimens. Sterile cotton swab (HiMedia Laboratories, India) pre-moistened with sterile normal saline was inserted into 1-1.5 inches deep into the rectum and gently rotated for a few seconds. After specimen collection, the swabs were inoculated in a tube containing 0.1% peptone water (HiCulture Transport Swab, HiMedia, India). A cloacal swab from chickens was taken by inserting a swab into the vent and by gently swabbing the mucosal wall till the swab was stained with fecal material. For environmental specimens, the swab specimens were collected from the drainage or sewage area near each selected house. All samples were transported to the Microbiology Laboratory of Manipal Teaching Hospital in specimen transport containers with ice packs. Based on the distance between the laboratory and the place of a visit, the samples were processed within a maximum of 8 h of collection. Detailed questionnaires on demographic and husbandry practices were included. Antimicrobial susceptibility testing. The antibiotic susceptibility profiles of the ESBL-positive isolates were determined using a panel of antibiotics of human and veterinary clinical pertinence by Kirby-Bauer disk diffusion method as per CLSI guidelines 48 and interpreted based on CLSI 2016 and 2017 breakpoints 49 . FDA breakpoint (http://www.acces sdata .fda.gov/drugs atfda _docs/label /2009/02182 1s016 lbl.pdf) was used for the interpretation of tigecycline. One or two representatives from various classes of antibiotics (MASTDISCS, U.K) were included: co-amoxiclav (beta-lactam combination agents), cefotaxime and ceftazidime (3rd generation cephalosporins), cefoxitin (2nd generation cephalosporins), tetracycline (tetracyclines), amikacin (aminoglycosides), nalidixic acid, ciprofloxacin (fluoroquinolones), nitrofurantoin (nitrofurans), tigecycline (glycylcycline) and imipenem, meropenem and ertapenem (carbapenems). E. coli ATCC 25922 was used as quality control strain.

Detection of ESBL-producing
Genotypic screening of ESBL phenotypes. Phenotypically ESBL-positive isolates were screened for ESBL encoding genes by multiplex PCR. As per Dallenne C et al. 50 , two multiplex PCR were assayed to detect the presence of β-lactamase genes bla TEM /bla SHV /bla OXA-1 and bla CTX-M, including phylogenetic groups 1, 2, and 9.
The primers (Sigma-Aldrich) used, and the size of the expected DNA products for each enzyme group are shown in supplementary file 5: Screening for E. coli ST131 clades. Sequence type-131 clonal lineage and ST131 clades (A, B, and C) of all ESBL-positive E. coli isolates were screened by multiplex PCR as described by Matsumura et al. 52 . After running at 100 V for one hour on 2% agarose gel containing ethidium bromide, amplicons were visualized. A size marker of 100 bp DNA ladder (Eurofins Scientific, India) was used. For positive and negative controls, known E. coli ST131 and non-ST131 E. coli were used, respectively. The size of the expected DNA products and the primers used in this study for each enzyme group are listed in supplementary file 5: Table S3.
Ethics approval and consent to participate. The research proposal was approved by the Institutional Ethics Committee, Manipal Teaching Hospital, Pokhara, Nepal. Reference number: MEMG/IRC/GA/1269/2015. The study was conducted in compliance with the latest version of the Declaration of Helsinki. All methods were carried out in accordance with relevant guidelines and regulations. This study was carried out in compliance with the ARRIVE guidelines approved by MCOMS, Nepal and MAHE, India.
Statistical analysis. Significant differences in the prevalence and high degree of antibiotic resistance between different populations were assessed by the Epi-info and SPSS software. GraphPad Prism Version 8.1.2 (227) was used to develop a heat map showing an antibiotic resistance profile and ESBL genes.