Multidrug-Resistant Enteropathogenic Escherichia coli Isolated from Diarrhoeic Calves, Milk, and Workers in Dairy Farms: A Potential Public Health Risk

Enteropathogenic Escherichia coli (EPEC) is a leading cause of diarrhoeagenic diseases in humans and cattle worldwide. The emergence of multidrug-resistant (MDR) EPEC from cattle sources is a public health concern. A total of 240 samples (75 diarrhoeic calves, 150 milk samples, and 15 workers) were examined for prevalence of EPEC in three dairy farms in Egypt. Antimicrobial resistance (AMR) traits were determined by antibiogram and polymerase chain reaction (PCR) detection of β-lactamase-encoding genes, plasmid-mediated quinolone resistance genes, and carbapenemase-encoding genes. The genetic relatedness of the isolates was assessed using repetitive extragenic palindromic sequence-based PCR (REP-PCR). EPEC isolates were detected in 22.7% (17/75) of diarrhoeic calves, 5.3% (8/150) of milk samples, and 20% (3/15) of worker samples. The detected serovars were O26 (5%), O111 (3.3%), O124 (1.6%), O126 (0.8%), and O55 (0.8%). AMR-EPEC (harbouring any AMR gene) was detected in 9.2% of samples. Among isolates, blaTEM was the most detected gene (39.3%), followed by blaSHV (32.1%) and blaCTX-M-1 (25%). The qnrA, qnrB, and qnrS genes were detected in 21.4%, 10.7%, and 7.1% of isolates, respectively. The blaVIM gene was detected in 14.3% of isolates. All EPEC (100%) isolates were MDR. High resistance rates were reported for ampicillin (100%), tetracycline (89.3%), cefazolin (71%), and ciprofloxacin (64.3%). Three O26 isolates and two O111 isolates showed the highest multiple-antibiotic resistance (MAR) indices (0.85–0.92); these isolates harboured blaSHV-12 and blaCTX-M-15 genes, respectively. REP-PCR genotyping showed high genetic diversity of EPEC, although isolates belonging to the same serotype or farm were clustered together. Two worker isolates (O111 and O26) showed high genetic similarity (80–95%) with diarrhoeic calf isolates of matched serotypes/farms. This may highlight potential inter-species transmission within the farm. This study highlights the potential high risk of cattle (especially diarrhoeic calves) as disseminators of MDR-EPEC and/or their AMR genes in the study area. Prohibition of non-prescribed use of antibiotics in dairy farms in Egypt is strongly warranted.


Introduction
Enteropathogenic Escherichia coli (EPEC) strains are leading etiological agents of diarrhoeagenic diseases in infants [1]. EPEC isolates have also been implicated in several diarrhoeic outbreaks among adults and AIDS patients worldwide [2,3]. In cattle, EPEC has been isolated from diarrhoeic calves and healthy carriers in several countries [4][5][6]. The EPEC pathotype is defined by the acquisition of an attaching and effacing (eae) gene with a lack of Shiga toxin-producing (stx1 and stx2) genes [1]. The eae gene encodes intimin, an outer-membrane protein responsible for attaching bacteria to the intestinal epithelium and effacing (destroying) the intestinal villi (A/E lesion). EPEC isolates that lack the EPEC adherence factor (EAF) plasmid that encodes the bundle-forming pilus (BFP) are atypical EPEC (aEPEC) [3]. aEPEC has mostly been isolated from animals, although has recently emerged as a human diarrhoeagenic pathogen [2,3].
Cattle and their milk are important reservoirs of E. coli-carrying AMR genes [11][12][13]; multidrug resistance (MDR) selection in this host is driven by the excess, non-prescribed use of antibiotics for the prevention or treatment of bacterial infections of economic importance, such as neonatal calf diarrhoea [11,12]. MDR E. coli carried by cattle can pass to humans through contact with live carrier animals or through consumption of their food products [14,15].
The extended-spectrum β-lactamase (ESBL)-producing genes, plasmid-mediated quinolone resistance (PMQR) genes, and carbapenemase-producing (CR) genes are leading contributors to resistance against β-lactam, fluoroquinolone, and carbapenem antibiotics, respectively. β-lactams and fluoroquinolones are the most commonly used antibiotics in clinical human medicine and veterinary practices worldwide [7]. Carbapenems are among the last-resort antibiotics for the treatment of MDR pathogens in humans. This class of antibiotics is prohibited for treatment of animals in several countries, yet there are several reports of emerging resistance against these antibiotics in isolates from cattle sources [13,16], posing a major public health threat.
In Egypt, the role of dairy cattle in the spread of EPEC and the prevalence of antimicrobial resistance in isolates of this pathotype is largely unknown. Therefore, the aim of this study aims are (1) to investigate the distribution frequency of EPEC isolates from diarrhoeic calves, milk, and workers in dairy farms in Egypt, (2) to record the antibiogram of isolates and their associated AMR genetic determinants, and (3) to assess the genetic relatedness between isolates using repetitive extragenic palindromic sequence-based PCR (REP-PCR) for evidence of potential infection pathways within farms.
Most ESBL and PMQR genes were predominant in isolates from diarrhoeic calves (Figures 1 and 2). However, no significant difference was observed in the rate of AMR genes acquired between isolates from different sources (p > 0.05). CR genes were not found in isolates from milk, and PMQR genes were not detected in any worker samples (Figures 1 and 2 (Figures 1 and 2). However, no significant difference was observed in the rate of AMR genes acquired between isolates from different sources (p > 0.05). CR genes were not found in isolates from milk, and PMQR genes were not detected in any worker samples ( Figures  1 and 2 and Table 3).
We sequenced the bla CTX-M-1 gene of two O111 isolates (ID 2 from diarrhoeic calves and ID 3 from a worker) and three bla SHV genes of three isolates of O26 (ID 13 and 18 from diarrhoeic calves and ID 17 from a worker). Analysis of bla CTX-M-1 of ID 2 (MW721313) and ID 3 (MW721311) isolates showed 99.8% and 100% similarity, respectively, to bla CTX-M-15 of several E. coli isolates in the GenBank database (including CP081589, CP075059, CP059120, and CP048916) ( Figure S1). The difference in the bla CTX-M-1 gene of ID 2 isolate is attributed to a single-nucleotide mutation (G to T at position 21); however, this point mutation did not alter the coded amino acid (valine) ( Figure S2). The sequenced bla SHV gene of the three isolates (ID 13, MW721318; ID 18, MW721319; and ID 17, MW721320) showed 100% similarity with bla SHV-12 from several E. coli isolates in the GenBank database (for example, MH460799, CP046002, and CP048293).
REP-PCR-based genotyping analysis of the 28 EPEC isolates is shown in Figure 1. The band patterns ranged from three to nine bands, with a size range of 150 to 2100 bp. The dendrogram map classified the EPEC isolates into two branches (BI and BII), six clades (C1-6), and 27 REP genotypes (RTs), with a discrimination index of 0.997. The EPEC isolates showed high genetic diversity, although isolates that belonged to the same serotype were clustered together. Unlike serotype, the source of the isolate did not discriminate between the isolates. The eight isolates of the O111 serovar belonging to the same farm (F1) displayed seven RTs with 80-95% genetic similarity. Six O111 isolates carried combined bla TEM /bla CTX-M-1 genes; two of these isolates (two diarrhoeic calves: ID 1 and 2) carried two additional PMQR genes (bla TEM /bla CTX-M-1 /qnrA/qnrS genotype) and showed identical RT (R1; Figure 1). One worker isolate (ID 3) showed high genetic similarity (>80%) with two isolates from diarrhoeic calves (ID 1 and 2), and the three isolates shared the same clade (C1). Twelve O26 isolates belonging to C3 displayed eleven RTs (R8-19) with 75-95% similarity. Nine O26 isolates from the same farm (F2) contained bla SHV ; of these, four isolates in one cluster (three diarrhoeic calves and one worker) carried an additional bla VIM gene (bla SHV /bla VIM genotype).

Discussion
The overall prevalence of EPEC in this study was 11.7%. The highest rate of EPEC was detected in diarrhoeic calves (22.7%). Lower prevalence rates (4.5-9.7%) were reported in Egypt [6], India [4], and Turkey [5]; however, EPEC was recovered at a higher rate (38%) from diarrhoeic calves in Belgium [17]. In contrast, EPEC was not a common cause of calf diarrhoea in dairy farms in Sweden [12]. In this study, EPEC was recovered from 5.3% of milk samples from healthy dairy cattle. In agreement with these results, low rates of EPEC (0.9-4.5%) were reported in milk worldwide [18][19][20]; however, dairy cattle were reported to be faecal carriers with relatively high rates (31-36%) of EPEC [21,22]. In addition, Lambertini et al. [21] suggested that milk seems play a smaller role than played by cattle faeces in environmental and interspecies dissemination of pathogenic E. coli in dairy farms. In support of this suggestion, our findings showed that EPEC isolates were 5.5 times more likely to be associated with diarrhoeic calves than with milk (OR = 5.5, p = 0.001). In this study, EPEC was detected in 20% of workers' stool samples. In agreement with these results, dairy farm workers were previously reported as carriers of pathogenic E. coli [14,23]. This may highlight the critical role of farm workers in infection dynamics within dairy farms.
In this investigation, all EPEC isolates were atypical (aEPEC; lacking the bfp gene), which is in agreement with another report from China [20]. Four of the five detected aEPEC serovars (O26, O111, O126, and O55) belonged to the classical EPEC (typical and atypical) serovars defined by the World Health Organization [1]. O124, a non-classical aEPEC serovar, was detected in 1.7% of samples. Similarly, non-classical aEPEC serovars were detected worldwide, especially from cattle sources [3]. O26 was the predominant serovar (detected in 5% of samples and 42.9% of EPEC isolates), which is in line with several reports worldwide [4,17]. The O26, O111, and O55 serovars were associated with several outbreaks of aEPEC diarrhoea cases among children and adults worldwide [1][2][3], which highlights the public health risk of these serovars.
AMR-EPEC isolates with at least one antimicrobial resistance gene were detected in 11.7% of samples. These isolates were more likely to be associated with diarrhoeic calves than with milk samples from healthy cattle (18.7% vs. 4%; OR = 5.5, p = 0.001).
Similarly, ESBL-EPEC (17.3% vs. 4%) and PMQR-EPEC (8% vs. 0.7%) were more likely to be associated with diarrhoeic calves than with milk (p = 0.002-0.02). These findings are consistent with those reported by de Verdier et al. [12], suggesting a greater association of antibiotic resistance with clinical diarrhoeic calves than healthy calves. Overuse and misuse of antibiotics for prophylaxis or therapy in young calves may have contributed to our findings, as previously described [11,12]. Additionally, the development of resistance among E. coli isolates from calves may be attributed to drinking milk containing antibiotic residues from treated cows on the same farm [24].
ESPL-EPEC isolates were detected in 13.3% of samples from dairy farm workers. This result is comparable with that of another report (12.5%) from Germany [14]. Unlike diarrhoeic calf samples, there was no difference between worker and milk samples in terms of the prevalence of ESPL-EPEC isolates (p = 0.1). This suggest that diarrhoeic cases may contribute more to the dissemination of ESPL-EPEC within the farm compared to other sources. In contrast, workers may contribute more than or equally to diarrhoeic calves in the spread of CR-EPEC, highlighting their important role as a source of these pathogens (OR = 1.7, p = 0.7).
In this study, the frequency of individual AMR genes ranged from 7.1% to 39.3% among isolates from different sources. This is in agreement with other reports (10.4-34.3%) in Egypt and elsewhere [25].
The bla TEM , bla SHV , and bla CTX-M-1 ESBL genes were detected at high rates (25-39.3%) among EPEC isolates from various sources. Similarly, these ESBL genes were detected at variable rates in cattle and workers on dairy farms worldwide [10,14,25]. None of the detected isolates in this study harboured bla OXA1 ; this gene is usually detected at low rates (1-3%) in E. coli isolates from dairy farms in Egypt [13,25].
In this study, sequence analysis showed that the bla CTX-M-15 and bla SHV-12 genes were detected in O111 isolates (one diarrhoeic calf and one worker) and O26 isolates (two diarrhoeic calves and one worker), respectively. In agreement with these results, Dandachi et al. [11] reported that bla CTX-M-1 (including bla CTX-M-15 ) and bla SHV-12 are the most prevalent ESBL genes in isolates from animals in the Mediterranean Basin (including Egypt). MDR E. coli isolates that harbour bla CTX-M-15 and/or bla SHV-12 genes have been associated with several clinical outbreaks among humans and cattle worldwide [10,11,25]. Additionally, both EPEC O111 and O26 isolates were found to carry multiple AMR genes, and they showed the highest MAR indices (O111 also carried bla TEM1 and PMQR genes; O26 also carried bla VIM ). This highlights the emerging risk of these isolates with respect to both cattle and public health in the study area.
In the current study, PMQR genes were detected in diarrhoeic calves and milk isolates (12.5-35.3%). A comparable rate (36.8%) was reported in diarrhoeic calf isolates in another study in Egypt [25]; however, Chen et al. [9] failed to detect any PMQR genes in cattle isolates in China. In contrast to animal isolates, none of the human isolates contained PMQR genes. In China, 8% of human diarrhoeic isolates harboured PMQR genes [9].
In this study, CR genes were detected in 14.3% of isolates (three diarrhoeic calves and one worker isolate). Despite the lack of carbapenem use in veterinary practice, an emerging increase in CR genes acquired among E. coli isolates from cattle has been reportedworldwide [13,16]. In Egypt, despite very few reports, E. coli isolates carrying CR genes were previously reported in faeces of healthy cattle [26], in milk, in faeces of diarrhoeic cattle, and in the stool samples of diarrhoeic humans [13]. bla VIM gene was the only detected CR gene in this study. Similarly, this gene was reported in cattle E. coli isolates from Egypt [13] and elsewhere [16].
Interestingly, Elmonir et al. [13] detected O26 isolates harbouring bla VIM combined with ESBL genes (ESBL/CR genotype) in milk and diarrhoeic (calves and human) cases in another study in the same research area (Kafrelsheikh governorate). This suggests clonal dissemination and the possible emergence of this CR-O26 serovar in the study area.
Around two-thirds (64.3%) of the isolates were resistant to CIP. A similarly high resistance rate (41.3%) was recorded for fluoroquinolones in Turkey [5]; however, a much lower rate (1.5%) was reported in China [20]. The lowest resistance rates were reported for CN (7.1%) and IPM (14.3%). No resistance to CN and IPM was previously reported [6,20]; however, resistance to carbapenems among cattle isolates was reported in previous studies in Egypt [13,26]. There was a concordance between the acquisition of AMR genes and the expression of phenotypic resistance to corresponding antibiotics (β-lactam, carbapenem, and fluoroquinolones) in all EPEC isolates in this study. However, few isolates lacking any of the investigated AMR genes showed phenotypic resistance to β-lactam and/or fluoroquinolones (ID 6, 7, and 20), which may be attributed to other AMR genes or mechanisms that were not studied in this research.
Variance in the rates of AMR genes acquisition/phenotypic resistance in EPEC isolates recovered in this study, as well as previous studies conducted elsewhere, may be linked to various factors, including ecological factors (environmental conditions), host factors (previous antibiotic therapy), genetic factors (virulence or mobile genetic elements), or technical factors (sampling and detection techniques).
REP-PCR-based genotyping analysis of the 28 EPEC isolates proved the high genetic diversity (6 Cs and 27 RTs), regardless of the sample sources, although isolates of the same serotype were clustered together. This is consistent with previous reports in Egypt [13] and elsewhere [27]. Interestingly, isolates of the same serotype that shared the same farm showed higher genetic similarity than isolates of this serotype in another farm (for example, O26 isolates in Farm 2 displayed 85-95% similarity, although they shared only 75% similarity with O26 isolates from Farm 3). This farm (location)-associated genetic similarity was previously reported worldwide [15,28]. These findings prove vertical and horizontal clonal dissemination of EPEC serovars between different sources within the farm, as previously reported [14,15,28].
Importantly, O111 and O26 worker isolates (ID 3 and 17, respectively) showed high genetic similarity (80-95%) with matched serotypes of diarrhoeic calves (O111 ID 2 and O26 ID 18, respectively). This may highlight potential interspecies zoonotic transmission of these pathogens between workers and calves within the farm, as reported in other studies [14,23]. This means that dairy farm workers are at constant occupational risk and may also play a role in reverse transmission of MDR E. coli isolates and/or their AMR genes within the farm [14,15]. Furthermore, Silvestro et al. [23] provided genetic evidence of infection transmission from the dairy workers to their families; therefore, they may also pose a health risk to their contacts.

Sampling
Samples were collected from three dairy farms in the Kafrelsheikh district of the Kafrelsheikh governorate. The governorate is located in the northern region of the Nile delta of Egypt. We collected a total of 240 samples from diarrhoeic calves (75 calves), dairy cows (150 milk samples), and farm workers (15 workers) on three dairy farms during 2019. We collected rectal swabs samples from diarrhoeic calves (those with ≥3 loose faeces per day) ranging in age from 1 day to 4 months. We collected composite milk samples (150-200 mL per animal) from individual dairy cows that showed no signs of mastitis or ill

Isolation, Identification, and Serotyping of Enteropathogenic E. coli (EPEC)
We inoculated faecal calf samples and worker stool samples on MacConkey's agar (Oxoid, Hampshire, UK), and the plates were incubated at 37°C for 24 h. Lactose fermenter (pink in colour) colonies were then subcultured on eosin methylene blue agar (EMB; Oxoid) and were incubated under the same conditions. For milk samples, ten dilutions on Tryptone Soy broth (TSB; Oxoid) were incubated at 37°C for 6 h, and then an inoculum of each sample was cultured on MacConkey's agar, followed by EMB agar at 37°C for 24 h each. Suspected E. coli colonies on EMB (green metallic sheen in colour) were biochemically confirmed by API-20E (bioMérieux, Marcy-l'Etoile, France). All EPEC isolates were serotyped using polyvalent and monovalent O-antisera sets (Denka Seiken Co., Tokyo, Japan) according to the manufacturer's instructions.

Molecular Confirmation of EPEC
All E. coli isolates were examined for detection of eae, bfpA, stx1, and stx2 virulence genes. DNA of each isolate was extracted from an overnight TSB culture using a QIAamp DNA Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. Uniplex PCR was conducted for eae gene detection. The PCR mix contained 12.5 µL of EmeraldAmp MAX PCR master mix (Takara Bio, Kusatsu, Japan), 1 µL (20 pmol) of each primer, 5 µL of DNA template (~50 ng), and water up to a final volume of 25 µL. The PCR reaction was conducted in an Applied Biosystem 2720 thermal cycler (Applied Biosystem, Foster City, CA, USA) under the following cycling conditions: 94 • C for 7 min; 35 cycles of 94 • C for 1 min, 51 • C for 1 min, and 72 • C for 2 min; and a final extension at 72 • C for 10 min. bfpA, stx1, and stx2 genes were detected under the same PCR conditions as for eae. The primers and annealing temperature specific to each gene are illustrated in Supplementary Data (Table S1). E. coli O157:H7 Sakai (positive for stx1 and stx2 genes) and E. coli strain E2348/69 (positive for eae and bfpA genes) were used as positive controls in all PCR reactions. Isolates that were positive for eae gene and negative for stx1 and stx2 genes were defined as EPEC [1]. EPEC isolates that were negative for bfpA gene were considered atypical EPEC (aEPEC).

Molecular Detection of Antimicrobial Resistance Genes in EPEC Isolates
We screened all EPEC isolates using a uniplex PCR to detect each of the following antibiotic resistance genes: ESBL-encoding genes (bla TEM , bla CTX-M-1 , bla OXA-1 , and bla SHV ), carbapenemase-encoding genes (bla VIM , bla NDM-1 , and bla IMP ), and PMQR genes (qnrA, qnrB, and qnrS). The PCR mix and cycling conditions were the same as those used for eae, with the exception of the primer sets and annealing temperature specific to each gene (Table S1). The following reference strains were used for quality control: K. pneumoniae ATCC BAA-1705 (bla TEM and bla SHV ), E. coli NCTC 13353 (bla CTX-M-1 ), E. coli NCTC 13476 (bla IMP ), E. coli ATCC BAA-2469 (bla NDM-1 ), and K. pneumoniae NCTC 13439 (bla VIM-1 and qnrS). Additionally, an E. coli isolate positive for qnrA and qnrB was kindly provided by the Central Laboratory of the Faculty of Veterinary Medicine, Kafrelsheikh University, Kafr el-Sheikh, Egypt. The E. coli ATCC 25922 reference strain was used as a negative control for all PCR tests.

Sequencing of EPEC Isolates Positive for bla CTX-M-1 and bla SHV
We selected two O111 isolates positive for bla CTX-M-1 (ID 2 from diarrhoeic calves and ID 3 from a worker) and three O26 isolates positive for bla SHV (ID 13 and 18 from diarrhoeic calves and ID 17 from a worker). These isolates showed the highest MAR indices (0.77-0.92). The PCR products were purified using a QIAquick gel extraction kit (Qiagen, Valencia, CA, USA). PCR product sequencing in both directions was conducted in an Applied Biosystems 3130 genetic analyser (Applied Biosystems, Foster City, CA, USA) using a BigDye Terminator v3.1 cycle sequencing kit (Applied Biosystems) following the manufacturer's instructions. The nucleotide sequence identity for each gene was confirmed using the BLAST 2.2 program (National Center for Biotechnology Information, NCBI). The accession numbers of the sequenced bla CTX-M-1 gene are MW721313 (isolate ID 2) and MW721311 (isolate ID 3). The accession numbers of the sequenced bla SHV gene are MW721318 (isolate ID 13), MW721319 (isolate ID 18), and MW721320 (isolate ID 17).

Genotyping of EPEC Isolates
We genotyped the EPEC isolates by the REP-PCR method as previously described [30]. We constructed a REP-PCR-based dendrogram using the Dice coefficient and the unweighted pair group method with arithmetic mean in GelJ v.2.0 software [31]. The Simpson's discrimination index for REP-PCR genotyping was calculated according to Hunter and Gaston [32].

Statistical Analysis
The odds ratio and potential association between antibiotic resistance traits and the source of EPEC isolates were assessed using a univariate logistic regression model. Statistical analyses were conducted using SPSS v19 software (IBM, Armonk, NY, USA), with significance set at p ≤ 0.05.

Conclusions
In conclusion, this study showed a high rate of MDR-EPEC-carrying AMR genes from various sources (especially diarrhoeic calves) in dairy farms in Egypt. This highlights the potential public health risk of these pathogens or their AMR genes crossing to humans in contact with dairy cattle or milk consumers in Egypt. Restriction of unauthorised use of antibiotics in dairy farms, use of protective clothing, enhanced awareness among workers, and extended surveillance for MDR-EPEC at a national level are mandatory prevention measures. Research is currently underway to further characterize the MDR-EPEC from other hosts in a wide geographical area.

Informed Consent Statement:
For dairy farm worker samples, we explained the research details to all participating workers, and informed consent was obtained before sample collection.

Data Availability Statement:
The authors confirm that the data supporting the findings of this study are available within the article.