Multidrug-Resistant Escherichia coli in Bovine Animals, Europe

Of 150 Escherichia coli strains we cultured from specimens taken from cattle in Europe, 3 had elevated MICs against colistin. We assessed all 3 strains for the presence of the plasmid-mediated mcr-1 gene and identified 1 isolate as mcr-1–positive and co-resistant to β-lactam, florfenicol, and fluoroquinolone antimicrobial compounds.

T he dissemination of mobile genetic elements containing antimicrobial resistance genes and the emergence of carbapenem β-lactamases (e.g., Klebsiella pneumoniae carbapenemase-2 and New Delhi metallo-β-lactamase-1) have narrowed the chemotherapeutic options available to clinicians (1,2). Treatment of infections associated with carbapenem-resistant Enterobacteriaceae requires the use of polymyxin B and polymyxin E (colistin). These cationic peptides are considered to be the last line of defense for infections in humans.
Colistin is a drug with a bactericidal action that targets the lipid A component of the lipopolysaccharide structure located in the outer wall of some gram-negative bacteria. Consequently, the drug exhibits a broad spectrum of activity against Enterobacteriaceae (3). Despite its use in animal production in certain countries, rates of resistance to colistin have so far remained low in animals and humans (3,4). Polymyxin resistance can develop after modification of the lipid A component in the lipopolysaccharide structure through mechanisms that are chromosomally mediated and result in a reduction in the affinity for these cationic peptides (5,6). In a recent report, Liu et al. (7) described the first known case of plasmid-mediated colistin resistance involving the mcr-1 gene coding for a phospoethanolamine transferase-like enzyme.
Considering the importance of colistin in the control of multidrug-resistant (MDR) nosocomial human infections caused by gram-negative bacteria and the use of this drug in veterinary medicine, the identification of the mcr-1 gene in food-producing animals is of major public health importance. The objective of our study was to retrospectively investigate a large collection of E. coli cultured from cattle that had suspected enteric or mastitic infections.

The Study
During 2004-2010, we cultured 150 E. coli strains from fecal samples collected from cattle with suspected enteric infection or milk-aliquots collected from cattle with suspected mastitis in France and Germany. We conducted antimicrobial susceptibility testing by using disk diffusion against a panel of 17 compounds consisting of penicillin G, amoxicillin, and amoxicillin/clavulanic acid; cephalothin, cefoxitin, cefotaxime, and cefepime; ertapenem, meropenem, and imipenem; marbofloxacin, ciprofloxacin, and nalidixic acid; gentamicin; tetracycline; florfenicol; and trimethoprim/sulfamethoxazole. We interpreted results according to the criteria of the Clinical and Laboratory Standards Institute where appropriate (8,9).
A subset of these E. coli (n = 45) were classified as MDR and expressed resistance to >3 drug classes. We determined plasmid profiles and PCR-based replicon types as described previously (10,11) and detected plasmids ranging in size from 2 to 200 kbp. Our PCR-based replicon type analysis identified several incompatibility (Inc) types, including IncX4 in E. coli strain 11-1896 and the previously reported IncHI2 type in E. coli strain 29957 (Table). We then determined the MICs of these 45 MDR isolates for colistin by using broth microdilution. Three of 45 demonstrated MICs >2 mg/L, which we interpreted as being colistin resistant based on breakpoint tables of the European Committee on Antimicrobial Susceptibility Testing (12). We identified these isolates as E. coli 22134 O9:H9 U/ST10, E. coli 11-1896 O9:H12 U/ST58, and E. coli 29957 O101:H9 A or C/ST167 (Table). All were additionally resistant to >2 drug classes, including aminoglycosides, aminopenicillins, cephalosporins, fluoroquinolones, phenicols, tetracyclines, and trimethoprim and sulfonamides. One of the 3 isolates (E. coli 29957) was resistant to all of the antimicrobial compounds tested, including β-lactams, florfenicol, and fluoroquinolone compounds (Table). In addition, PCR results indicated that this isolate was positive for the presence of the mcr-1 gene (online Technical Appendix Figure, panel A, http://wwwnc. cdc.gov/EID/article/22/9/16-0140-Techapp1.pdf) (7). We conducted whole-genome sequencing of 3 isolates with increased MICs for colistin by using the Nextera XT DNA Library Preparation Kit and the Illumina MiSeq platform (Illumina, Inc., San Diego, CA, USA) to produce 300-bp paired end reads (v3 chemistry). We assembled these data de novo using SPAdes version 3.6.2 (http://bioinf.spbau.ru/spades) and then generated queries by using the PlasmidFinder 1.3 (https://cge.cbs.dtu.dk/services/ plasmidfinder) and ResFinder 2.1 (http://cge.cbs.dtu.dk/ services/resfinder) databases to identify plasmid replicon types and antibiotic resistance genes using BLAST+ (https://blast.ncbi.nlm.nih.gov/Blast.cgi). Several antibiotic resistant genotypes, including some that were acquired, were identified. Four of these genotypes occurred in E. coli 22134 isolates, 8 in E. coli  isolates, and 21 in E. coli 22957 isolates (Table). These isolates harbored genes or mutations that confirmed the phenotypes detected in most of the suspected cases of infection in the cattle in our study. We also identified plasmid replicons in all 3 isolates, including 3 types in E. coli 22134; 5 in E. coli , and 7 in E. coli 29957. We did not detect the mcr-1 gene in E. coli 22134 or E. coli 11-1896; however, we identified several nonsynonymous amino acid substitutions in genes previously shown to be associated with colistin resistance, including pmrA and pmrB. We also identified phoP and eptB in E. coli . Similarly, we identified the mcr-1 gene in E. coli 29957 (a feature that was previously confirmed by PCR) and 1 nonsynonymous substitution in pmrB. The mcr-1 gene was located in a 4,752-bp contig, which when used to query the cur-rent databases matched an identical region containing a transposase gene, a phosphoethanolamine transferase gene (the mcr-1 encoding gene), a hypothetical protein/ phosphoesterase gene, and another transposase. The mcr-1 gene was 100% similar at the nucleotide level to that reported in China and was found to be located distal to the same insertion sequence element ISApl1 that mapped to the IncHI2 type plasmid pHNSHP45 (online Technical Appendix Figure, panel B) (7).

Conclusions
Plasmid-mediated colistin resistance identified in MDR bacteria of animal origin represents a serious risk to public health. Our data further support recent findings demonstrating that the mcr-1 gene is not just present in Asia but can also be found in some countries in Europe (e.g., the mcr-1 gene identified in an E. coli strain cultured from a food-producing animal in France in 2007) (Table). Other arrangements of the mcr-1 gene on plasmids can occur, such as that observed in the IncX4 type (13). Liu et al. (7) reported that plasmid pHNSHP45 exhibited an in vivo transfer rate between different E. coli strains (measured at 10 -1 to 10 -3 per recipient) (7), a feature that could contribute to the successful dissemination of the mcr-1 gene. Similarly, in our study, we can also confirm the transfer of the mcr-1 gene from E. coli 29957 via conjugation, albeit at a reduced frequency (data not shown). Especially concerning is the extensive resistance profile of E. coli 29957, a feature noted in other studies, which have indicated that colistin resistance might be co-selected after the use of cephalosporins and other compounds (14,15).
The mcr-1 gene has now been reported in food-producing animals and in humans located in different geographic regions. In several of these regions, the gene was linked to extended-spectrum β-lactam and florfenicol resistance in the same bacterial isolate (15). Because E. coli 29957 was identified in 2007, this finding cannot be considered a recent occurrence. Given the genetic mapping reported to date, selective pressure imposed after the administration of broadspectrum cephalosporins and other compounds might have the potential to co-select for colistin resistance and vice versa, thereby contributing to the dissemination of mcr-1 (15). Molecular epidemiologic studies are required to discover the origin and means of transmission of this gene as a first step in attempting to limit its dissemination, particularly among pathogenic bacteria that threaten human health.
Financial support for this study was kindly provided by Vétoquinol SA as part of the University College Dublin Foundation Newman Scholarship Program that funded M.M.
Mr. Brennan is a research scientist working at the University College Dublin Centre of Food Safety. His primary research interest is in microbiology, including antimicrobial resistance genes and their means of dissemination.