Antimicrobial resistance profiles of Escherichia coli and prevalence of extended‐spectrum beta‐lactamase‐producing Enterobacteriaceae in calves from organic and conventional dairy farms in Switzerland

Abstract This study compared the antimicrobial resistance (AMR) among commensal Escherichia coli in the fecal microbiota of young calves raised on organic and on conventional dairy farms in Switzerland. Further, fecal carriage of extended‐spectrum beta‐lactamase (ESBL) producing Enterobacteriaceae was assessed for calves from both farming systems. Where possible, data on antimicrobial usage (AMU) were obtained. Antimicrobial susceptibility testing was performed on a total of 71 isolates using the disk diffusion method. ESBL producers were characterized by polymerase chain reaction‐based multilocus sequence typing and sequencing of the bla ESBL genes. Organically raised calves were significantly more likely to harbor E. coli that showed AMR to ampicillin (odds ratio [OR]: 2.78, 95% confidence interval [CI]: 1.02–7.61, p = 0.046), streptomycin (OR: 3.22, 95% CI: 1.17–8.92, p = 0.046), kanamycin (OR: 11.3, 95% CI: 2.94–43.50, p < 0.001), and tetracycline (OR: 3.25, 95% CI: 1.13–9.31, p = 0.028). Calves with reported AMU were significantly more likely to harbor E. coli with resistance to ampicillin (OR: 3.91, 95% CI: 1.03–14.85, p = 0.045), streptomycin (OR: 4.35, 95% CI: 1.13–16.7, p = 0.045), and kanamycin (OR: 8.69, 95% CI: 2.01–37.7, p = 0.004). ESBL‐producing Enterobacteriaceae (18 E. coli and 3 Citrobacter braakii) were detected exclusively among samples from conventionally farmed calves (OR: infinity [∞], 95% CI: 2.3–∞, p < 0.0013). The observations from this study suggest that AMR is highly prevalent among commensal E. coli in young dairy calves, irrespective of the farm management system, with proportions of certain resistance phenotypes higher among organic calves. By contrast, the occurrence of ESBL producers among young dairy calves may be linked to factors associated with conventional farming.


| INTRODUCTION
Antimicrobial resistance (AMR) is a major public health care issue that is increasingly reflected in veterinary medicine. Also in veterinary medicine, adequate antimicrobial treatment of bacterial infections is necessary to promote the health and welfare of animals. However, it is widely acknowledged that the use, overuse, and misuse of antimicrobial agents promote the emergence of resistant bacteria and the dissemination of AMR genes. In food-producing animals, the use of antimicrobials comes with the risk of spreading AMR at the animal/ human interface, through the food chain, or by contamination of the farm environment (Carattoli, 2008;Nüesch-Inderbinen & Stephan, 2016). Switzerland, like other European countries (World Health Organization, 2011), has developed and implemented a national strategy on antibiotic resistance (StAR) to address the threat of AMR affecting the human and animal health sectors and the environment (Swiss Federal Council, 2015). Data on the prevalence of AMR in zoonotic and indicator bacteria isolated from humans, livestock, and food are published regularly in the Swiss Antibiotic Resistance Report and Veterinary Office, 2020). In the EU regulation for organic dairy herds, antimicrobial therapy is restricted to three treatments per individual cow and year (Council Regulation [EC] No 834/2007 andEC No 889/2008). In Switzerland, the Swiss Ordinance on Organic Farming (SR 910.18) allows antimicrobial agents to be prescribed by a veterinarian only if homeopathic or phytotherapeutic products failed to prevent suffering or distress to the animal. Furthermore, the ordinance requires twice the legal withdrawal period for organically produced foodstuffs from treated animals. Animals that receive more than three courses of antimicrobial treatments within 1 year may no longer be classified as organically farmed. However, it is unclear whether this set of regulations results in a lower prevalence of AMR in organic dairy calves and there are currently no data that compare AMR among dairy calves reared in organic and conventional productions systems in Switzerland. Therefore, it was the aim of this study to assess the prevalence of AMR E. coli and the occurrence of ESBL-producing Enterobacteriaceae in young dairy calves on their birth farms and to evaluate any differences between calves from organically and conventionally managed dairy farms.

| Fecal sampling
During September 2020, a total of 24 officially registered organic, and 30 conventional dairy farms were visited throughout four cantons in the northwest region of Switzerland. Visits were conducted with the approval of the farmers. Only calves that were born on the respective farm were included in the study. Where available, data on antimicrobial usage (AMU) was recorded. In the current study, AMU included antimicrobial treatment of the calves or feeding of discard milk from cows treated with antibiotics.
The age of the calves ranged from 2 to 120 days. To ensure a noninvasive procedure, fresh feces were collected from pen floors and animal enclosures using plastic bags. A total of 196 samples were collected and placed in cooler boxes for transport and stored at −20°C until processing.
Before microbiological analysis, the samples were thawed at 4°C overnight. A sterile cotton swab of each sample was placed in a sterile blender bag (Seward), homogenized at a 1:10 ratio in Enterobacteriaceae enrichment (EE) broth (BD), and incubated at 37°C for 24 h.

| Identification of bla ESBL genes
DNA of ESBL-producers was extracted using a standard heat lysis protocol. Screening for bla TEM and bla SHV was carried out using primers described previously (Pitout et al., 1998). Screening for bla CTX-M alleles belonging to CTX-M groups 1, 2, 8, 9, and 25 was performed as described by Woodford et al. (2006). Amplicons for sequencing bla CTX-M genes were generated using primers described previously (Geser et al., 2012). Synthesis of primers and DNA custom sequen- nlm.nih.gov/blast/) was used.
2.6 | Phylogenetic analysis and multilocus sequence typing (MLST) of ESBL producing E. coli The distribution of phylogenetic groups among the ESBL-producing E. coli was determined by polymerase chain reaction (PCR) targeting the genes chuA, yjaA, arpA, and TspE4.C2, as described by Clermont et al. (2013). Isolates were thereby assigned to one of the eight phylogenetic groups A, B1, B2, C, D, E, F (E. coli sensu stricto), or Escherichia clade I.
For MLST of E. coli isolates, internal fragments of the seven housekeeping genes (adk, fumC, gyrB, icd, mdh, purA, and recA) were amplified by PCR as described by Wirth et al. (2006). The amplification products were custom sequenced and sequence types (STs) were determined using the E. coli MLST database website (https:// enterobase.warwick.ac.uk).

| Statistical analysis
For the descriptive statistical analysis, 95% binomial confidence intervals (CI) were obtained for the proportions of antibiotic resistance and ESBL-producers with the command BinomCI specifying the type "Jeffreys" in the package DescTools (Signorell et al., 2020), using R (R Core Team, 2021). To assess if antibiotic resistance is significantly associated with farming type or AMU, generalized linear mixed models were applied for all antibiotics where at least 10% of the isolates were resistant. The analysis was performed with the packages "geepack" (Halekoh et al., 2006) "lme4" (Bates et al., 2015), and "MASS" (Venables & Ripley, 2002). Model selection was based on likelihood ratio tests with the package "lmtest" (Hothorn et al., 2018).
To account for potential within-herd clustering, farms were included as random effects.
The prevalence of ESBL-producers at the farm level was compared with the two-tailed Fisher's exact test. AMU was noted for 17 calves. Seven calves were from organic farms and 10 from conventional farms. Among the seven calves from organically managed farms, five had been fed discard milk from cows treated with antibiotics for mastitis or other illnesses, one had a history of treatment with tetracycline, and one had received an aminoglycoside. Of the 10 calves from conventional farms, two had been fed discard milk, four had a history of treatment with tetracycline, three had received a penicillin-streptomycin combination, and one calf had florfenicol.

| Antimicrobial susceptibility of E. coli isolates
Using Rapid'E. coli agar, E. coli was recovered from all 196 fecal samples. For further analysis, a total of 71 isolates were selected for AST. Thereof, 54 originated from calves without AMU (24 E. coli from organically farmed calves and 30 E. coli from conventionally farmed calves), representing one isolate per farm. In addition, 17 E. coli from calves with a history of AMU were included (seven isolates from four organic farms, and 10 isolates from six conventional dairies).
Of the 18 ESBL-producing E. coli, four isolates from three different farms harbored bla CTX-M-1 , four isolates originating from the same farm harbored bla CTX-M-3 , two E. coli from one farm contained bla CTX-M-14 , and eight isolates from four farms carried bla CTX-M-15 (Table 2). Fifteen (83%) of the ESBL-producing E. coli were MDR (Table 2).
Phylogenetic classification allocated 13 E. coli to phylogenetic Group A and four to phylogenetic Group B1, both of which typically contain commensal strains. One strain belonged to phylogenetic Group C (Table 2).
AMU was known in three cases of calves harboring CTX-M-15producing E. coli, all of which had received penicillin-streptomycin and were from the same farm (Table 2). The three C. braakii harbored bla CTX-M-1 and were isolated from calves from one farm (Table 2).    There is however no data that assess AMR among E. coli from dairy calves at the farm level or studies that compare AMR prevalence in different production systems in Switzerland.
In this study, for some antimicrobials such as ampicillin, tetra- isolates from calves on organic dairy farms compared to conventionally produced animals (Sato et al., 2005;Sjöström et al., 2020;Wilhelm et al., 2009). In view of organic farming in other livestock sectors, it is worth mentioning that on organic broiler farms, AMR commensal E. coli are still common, albeit less frequent than on conventional farms (Musa et al., 2020;Pesciaroli et al., 2020).
However, organic production regulations, AMU, as well as study designs and methodologies vary between studies and should be considered when comparing data (Wilhelm et al., 2009 that in addition to exposure to antimicrobials, environmental and host factors may be associated with the selection of AMR in young calves (Haley et al., 2020). Drivers that select for AMR in young dairy calves are currently poorly understood and warrant further investigation (Haley & Van Kessel, 2021). Our data further suggest that calves with AMU from both farming types are more likely to carry ampicillin-, streptomycin-, and kanamycin-resistant E. coli than calves without AMU.
In contrast to AMR commensal E. coli, ESBL-producing Enterobacteriaceae were detected exclusively among calves reared on conventionally managed dairy farms.
With a prevalence of 33% at the farm level, our data correspond to the herd prevalence of 32% reported for Swiss fattening calves at slaughter (Federal Food Safety and Veterinary Office, 2019). The similar prevalences among the two age groups suggest that ESBL producers may be maintained at a high level in the microbiota of calves during the fattening period. The absence of ESBL producers among organically reared calves confirms recent data from a comparable study from Sweden that examined the prevalence of ESBL producing E. coli in calves from organic and conventional dairy farms (Sjöström et al., 2020), but the contrast with a previous study from the Netherlands which reported that 11% of organic dairy herds were positive for ESBL producers (Santman-Berends et al., 2017).
However, while the study by Sjöström et al. (2020) as well as the present study used healthy young calves as an indicator for AMR and ESBL producers in organic and conventional dairy herds, Santman-  Notably, our findings differ from results obtained from studies on poultry farms, where ESBL-producers show a high prevalence both in conventional and organic poultry farming (Saliu et al., 2017).
In both poultry farming systems, vertical transmission of plasmids within the production pyramid rather than the clonal spread of certain lineages account for the occurrence and long-term maintenance of ESBL-producers among broilers (van Hoek et al., 2018;Zurfluh et al., 2014).
In this study, the most frequently observed ESBL variants were CTX-M-15 which is the most important ESBL in human medicine (Cantón et al., 2012), and CTX-M-1 which represents the pre- Finally, C. braakii harboring bla CTX-M-1 was found in three calves on the same farm. This species infrequently causes infections in humans, and reports on ESBL-producing C. braakii remain rare (Liu et al., 2020). CTX-M-producing C. braakii have been identified in pork in China, in fish in Tanzania (Moremi et al., 2016), and in raw milk in Germany (Odenthal et al., 2016). Thus, our data provide further evidence for the occurrence of ESBL-producing C. braakii in the dairy farm environment.
This study has some limitations. The first limitation is the small number of farms and sampling occasions, and some differences between conventional and organically reared calves may have remained undetected. Second, data on AMU in the present study were communicated verbally by farm owners or staff and may have been subject to unprecise or non-objective reporting. The number of calves with a history of AMU was very small and consequently, the results should be interpreted with caution. Finally, we did not account for additional factors that may have an impact on AMR on dairy farms, for example, by including environmental samples, and the observational nature of our study design leaves the possibility of containing confounders.
Nevertheless, the findings of this study offer new useful information with regard to developing farming management strategies that aim to mitigate the occurrence of ESBL-producing Enterobacteriaceae on dairy farms and their dissemination to the environment and the food chain.

| CONCLUSIONS
AMR was shown to be prevalent among commensal E. coli from young calves from both organic and from conventional dairy farms, with particular resistance phenotypes occurring more frequently in E. coli from organic calves. In both groups, AMR E. coli occurred in calves with and without AMU. The occurrence of ESBL-producing Enterobacteriaceae was found to be significantly associated with conventionally managed farming systems. Further research on environmental and host factors that promote AMR in commensal E. coli of young dairy calves is required. Factors associated with conventional dairy farming that may co-select for the maintenance of ESBL producers should be identified.

ACKNOWLEDGMENTS
We thank the participating farmers for their cooperation in sample and data collection.

CONFLICTS OF INTEREST
None declared.

ETHICS STATEMENT
This study was performed following the Swiss Animal Welfare Act (SR 455) and the involved procedures were noninvasive and are ethically accepted by the FSVO. Participation in this study was voluntary, and the farmers approved of the purpose and methods of this study. All information was treated anonymously.