High Fecal Prevalence of mcr-Positive Escherichia coli in Veal Calves at Slaughter in France

The aim of this study was to determine the percentage of healthy veal calves carrying mcr-positive E. coli strains at the time of slaughter in France. Fecal samples were selectively screened for mcr-positive E. coli isolates using media supplemented with colistin. Screening for mcr genes was also carried out in E. coli isolates resistant to critically important antimicrobials used in human medicine recovered from the same fecal samples. Overall, 28 (16.5%) out of the 170 veal calves tested carried mcr-positive E. coli. As some calves carried several non-redundant mcr-positive strains, 41 mcr-positive E. coli were recovered. Thirty-one and seven strains were positive for mcr-1 and mcr-3 genes, respectively, while no strain was positive for the mcr-2 gene. Co-carriage of mcr-1 and mcr-3 was identified in three strains. All mcr-positive E. coli isolates, except one, were multidrug-resistant, with 56.1% being ciprofloxacin-resistant and 31.7% harboring blaCTX-M genes. All mcr-3-positive E. coli carried blaCTX-M genes, mainly blaCTX-M-55. This study highlights the high prevalence of mcr-positive E. coli strains in feces of veal calves at the time of slaughter. It also points out the multidrug (including ciprofloxacin) resistance of such strains and the co-occurrence of mcr-3 genes with blaCTX-M-55 genes.


Introduction
Colistin is currently considered as the last-resort antibiotic for human infections due to multidrug-resistant Enterobacteriaceae, notably those producing extended spectrum β-lactamases (ESBLs) and carbapenemases [1]. Colistin has also been used in veterinary medicine, notably, as a first-intention antibiotic for treating gastrointestinal infections in food-producing animals [2]. The future usefulness of this last-resort antibiotic was challenged by the first description of a plasmid-mediated colistin resistance gene (mcr-1) [3], especially as other plasmid-mediated colistin resistance genes (mcr-2 and mcr-3) were also identified [4,5]. Up to the present, seven additional mcr genes (mcr-4 to mcr-10) have been identified [6], with mcr-1 remaining the most prevalent [7]. The most common plasmids carrying the mcr-1 gene, which are IncI2, IncHI2 and IncX4, have been shown to be highly transferable and to carry other resistance genes [8][9][10]. They contributed to the widespread distribution of the mcr-1 gene worldwide [6].
As food-producing animals have been identified as a possible source of transmission of colistin-resistant bacteria to humans [6], the monitoring of such strains in the food chain is required. The detection of mcr-positive E. coli in food-producing animals has been recently reviewed [11]. This review indicated that the comparison of epidemiological studies, and their reported prevalence, is often challenging due to varying study design, methodologies and reporting. The first prevalence data were mainly obtained in a diseased context and through PCR screenings of ESBL-producing E. coli, and only a few reports focused on the detection of mcr genes in E. coli isolated from healthy animals [12]. Moreover, as only a few attempts were made to selectively isolate mcr-positive E. coli, the true prevalence of such strains in the commensal intestinal microbiota of healthy animals slaughtered for human consumption remains poorly described. This is especially true in the veal calf sector where only a few prevalence studies were conducted [11]. The monitoring of colistin resistance should be carried out in veal calves as they are frequently exposed to antibiotics, including colistin [2,13].
The objective of this study was therefore to evaluate the intestinal carriage of E. coli isolates positive for mcr-1, mcr-2 and mcr-3 by veal calves at the time of slaughter in France. A total of 170 fecal samples were selectively screened for mcr-positive E. coli isolates on colistin-supplemented media. Moreover, as co-detection of mcr genes with ESBL genes has been reported [12], ESBL-producing E. coli were also isolated from the same 170 fecal samples and screened for mcr genes. The occurrence of such genes in E. coli resistant to other critically important antimicrobials (CIA) for human medicine: ciprofloxacin and carbapenem was also investigated.

Detection of mcr-Positive E. coli Isolates
A specific detection of mcr-positive E. coli isolates in media supplemented with colistin was performed. For 24 out of the 170 calf fecal samples, at least one E. coli was isolated from colistin-supplemented RAPID'E. coli 2 agar plates which had been inoculated with colistin-supplemented BGBB turbid broths. A total of 27 presumptive colistinresistant E. coli isolates were recovered, as 2 non-redundant isolates were obtained from 3 calves as determined by ERIC-PCR typing (data not shown). PCR screening of the mcr-1, mcr-2 or mcr-3 gene showed that 23 out of the 27 colistin-resistant E. coli isolates were mcr-positive (identified as "COL" in Table 1). These originated from 21 calves. Meanwhile, PCR screening of the mcr-1, mcr-2 and mcr-3 gene was also performed in collections of CIA-resistant E. coli isolated from the same 170 fecal calf samples. From these, 110 yielded ciprofloxacin-resistant E. coli isolates. ERIC-PCR typing led to the recovery of 172 non-redundant ciprofloxacin-resistant E. coli isolates. Twelve mcr-positive E. coli isolates were identified amongst these strains (identified as "CIP" in Table 1). At least 1 ESBLproducing E. coli from 56 fecal samples was isolated. ERIC-PCR typing led to the recovery of 69 non-redundant ESBL-producing E. coli isolates, among which 9 were positive for mcr genes (identified as "CTX" in Table 1). No E. coli isolate was recovered from ChromoID CARBA and OXA-48 media. In all, 44 mcr-positive E. coli isolates were detected, including 23 isolated from media supplemented with colistin (COL), 12 corresponding to ciprofloxacin-resistant E. coli (CIP) and 9 corresponding to ESBL-producing E. coli (CTX) ( Table 1).
Thirty-one and seven E. coli isolates were positive for mcr-1 and mcr-3 genes, respectively. The remaining three isolates showed a co-occurrence of the two genes ( Table 1). All mcr-positive E. coli isolates, except for one strain (B2-8-COL-1), were multidrug-resistant (MDR), and 56.1% were resistant to ciprofloxacin. None of the mcr-positive E. coli isolates was positive for plasmid-mediated quinolone resistance (PMQR) genes. ESBL was produced by 13 mcr-positive E. coli (31.7%), and bla CTX-M genes were detected in all these strains. Co-carriage of bla CTX-M-1 or bla CTX-M-14 and mcr-1 was observed in three strains. bla CTX-M-55 was identified in nine mcr-3-positive E. coli strains, associated or not with the mcr-1 gene; bla CTX-M-14 was identified in one mcr-3-positive E. coli strain.

Prevalence of mcr-Positive E. coli Calf Ahedders at Alaughter
Overall, the prevalence of mcr-positive E. coli in fecal samples from healthy calves was 16.5% (28/170) ( Table 2). More precisely, it was 12.9% (22/170) and 3.5% (6/170) for calf carriers of mcr-1and mcr-3-positive E. coli isolates, respectively. Positive calves came from 10 out of the 32 farms providing calves that were tested at slaughter. Information about the use of colistin could be obtained for 25 farms with only 3 of them reporting the use of colistin. Positive calves originated from two of these farms ( Table 2). PFGE analysis showed that mcr-positive E. coli strains from calves originating from the same farm commonly shared PFGE types, suggesting circulation of strains or clones within a farm and potential transmission between animals. By contrast, PFGE types were distinct between farms, except for PFGE type 1 that was identified in farms 13 and 41 (Table 1).

Discussion
A significant number of studies on colistin resistance concentrated on pigs and chickens and highlighted high prevalence in these sectors. Fewer studies were conducted in cattle, and they tended to show low prevalence in dairy cattle [11]. The aim of our study was to accurately determine the prevalence of carriage of mcr-positive E. coli isolates by healthy veal calves at the time of slaughter. To do so, we isolated presumptive colistinresistant E. coli by inoculating BGBB with Durham tubes supplemented with 2 µg/mL of colistin. This isolation procedure was completed by parallel screening for the presence of mcr genes in ciprofloxacin-resistant and ESBL-producing E. coli recovered from the same fecal samples. Indeed, it has been shown that plasmids carrying mcr genes also carried genes encoding resistance to various antibiotics such as cephalosporins, carbapenems and fluoroquinolones [14][15][16]. This multifaceted approach improved the detection of mcr-positive E. coli isolates from fecal samples.
We identified that the prevalence of mcr-positive E. coli in fecal samples from healthy calves was 16.5%, with 12.9% and 3.5% of the calves carrying mcr-1-positive and mcr-3positive E. coli isolates, respectively. No mcr-2-positive isolate was detected here, in agreement with the detection of this gene from animals reported only in China and Belgium [4,17]. In the USA, Meinersmann et al. did not detect any mcr-positive Enterobacteriaceae when 1077 cecal contents from cattle were tested for presumptive colistin-resistant E. coli [18]. A pan-European study from 2008 to 2014 did not report any isolation of mcr-1-positive E. coli isolates when 3101 cattle were tested [19]. Reviewing the prevalence of mcr genes in cattle, Shen et al. reported that the prevalence of mcr-1 in bacteria was consistently low, from 0.5% to 0.8% in China, 2.6% in Egypt and from 0.1% to 3.3% in Europe [11]. The discrepancies between these studies and ours might partly result from the fact that mcr-positive E. coli form part of the sub-dominant fecal microbiota. The majority of prevalence studies screened for only a single E. coli isolate per animal, consequently, mcr-positive strains present in the intestinal microbiota of the animal could remain undetected and are named the 'phantom resistome' [20]. In our study, the use of colistin-supplemented media after an enrichment step in the presence of colistin likely increased the identification of mcr-positive E. coli isolates, which is in agreement with previous reports showing the importance of an enrichment step with colistin to detect mcr-positive E. coli [21,22]. Screening performed on ESBL-producing or ciprofloxacin-resistant E. coli also contributed here to the identification of higher proportions of mcr-positive E. coli isolates in cattle and also shown by others in France. Haenni and colleagues tested 1398 ESBL producers from diseased veal calves and detected 15.0% and 2.6% of mcr-1and mcr-3-positive E. coli isolates, respectively [23,24]. These higher proportions are consistent with the fact that these diseased calves might have been treated with antibiotics, and especially colistin.
Our results confirmed that mcr-positive isolates were resistant to other antibiotics [10,25,26]. Except for one isolate, all mcr-positive strains were MDR and 56.1% of them were resistant to ciprofloxacin. Regarding resistance to cephalosporins, only 15.8% of mcr-1-positive E. coli had bla CTX-M genes, whereas all mcr-3-positive E. coli carried such genes. The co-occurrence of mcr-3 genes and bla CTX-M 55 was the most frequently observed gene combination. Although the CTX-M 55 subtype is very common in Asia and, until recently, seemed to be limited to this continent [27], the emergence of bla CTX-M 55 associated with mcr-1 and/or mcr-3 genes was reported in France in cattle and in a pediatric infection, suggesting that CTX-M-55-producing E. coli could be vectors of mcr-3 dissemination outside Asia [23,28].
Calves harboring mcr-positive E. coli came from ten farms, from which only two reported the use of colistin in the batch from which the calves sampled at the slaughterhouses originated. Out of the 32 farms, another farm reported the use of colistin, but no calf harboring mcr-positive E. coli was detected at the time of slaughter. Since the year 2017 when the sampling was performed, the use of colistin in food animal production has drastically decreased in France and several countries around the world, in order to limit the spread of mcr genes [29,30]. The occurrence and distribution of mcr genes was monitored after the withdrawal of colistin at pig and chicken farms. Studies showed different results, and the association between colistin withdrawal and control of the spread of mcr genes was not evidenced [25,[31][32][33]. It should be stressed, however, that in the absence of colistin use, treatment with other antibiotics might co-select mcr genes and favor their spread. Indeed, eight out the ten farms with mcr-positive isolates did not report the use of colistin.
In conclusion, this study revealed a high prevalence of fecal carriage of mcr-positive E. coli by healthy veal calves at the time of slaughter. It also pointed out the multidrugresistance of such strains. Although the use of colistin has drastically decreased in foodproducing animals in recent years, it should be kept in mind that the use of other antibiotics could lead to the selection of mcr-positive isolates. The monitoring of such strains in the food chain is still required.

Study Population and Sampling
Fecal samples were collected from 170 calves aged from 6 to 8 months, originating from 32 calf fattening units and intended for meat production. Feces were collected after evisceration at the time of slaughter in five French slaughterhouses (identified as A, B1, B2, B3 and RA) during 8 sampling campaigns from March to September 2017 (Table 2). A survey was conducted at the farms to identify their use of colistin in the batch from which the calves sampled at the slaughterhouses originated.

Collections of Presumptive Colistin-and CIA-Resistant E. coli
Feces were diluted (10 g in 90 mL) in modified tryptone soya broth (mTSB). In order to detect presumptive colistin-resistant E. coli, fecal samples were diluted 1:10 in brilliant green bile broth (BGBB) with Durham tubes and supplemented with 2 µg/mL of colistin. Duplicate BGBB tubes were incubated at 44 • C for 24 h. The positive BGBB tubes were then plated onto RAPID'E. coli 2 agar (Bio-Rad ® ) with 2 µg/mL of colistin (COL). At the same time, CIA-resistant E. coli were detected by plating fecal suspensions in duplicate onto: (i) RAPID'E. coli 2 agar (Bio-Rad ® ) supplemented with 1 µg/mL of cefotaxime (CTX), (ii) RAPID'E. coli 2 agar (Bio-Rad ® ) supplemented with 1 µg/mL of ciprofloxacin (CIP), (iii) ChromID CARBA agar and (iv) ChromID OXA-48 agar (bioMérieux ® , Marcy-l'Étoile, France). All plates were incubated overnight at 44 • C. Up to 8 E. coli colonies were collected per positive plate, confirmed for species using the indole production test and typed using the Enterobacterial Repetitive Intergenic Consensus (ERIC) -PCR fingerprinting assay with the ERIC 2 primer [34]. The DNA fingerprints of the isolates collected from each animal were visually compared to identify non-redundant resistant strains in each positive calf that was submitted for further characterization.

PFGE Typing
E. coli strains were typed using PFGE on XbaI-digested DNA following the CDCstandard operating procedure [45]. Analysis was conducted using BioNumerics software. A Dice similarity coefficient with a UPGMA dendrogram was generated based on 1% tolerance windows and 1% optimization. A cutoff line at 95% was considered to identify genetically related isolates.