Sir, The first description of the plasmid-mediated mcr-1 gene in November 2015 proved its capacity for horizontal transfer and shed an unexpected light on the last-resort colistin antibiotic.1 Then, numerous studies demonstrated the worldwide dissemination of mcr-1 in Enterobacteriaceae, mainly in livestock, but also in humans. Variants of mcr-1 have been described, as have the new mcr-2, mcr-3, mcr-4 and mcr-5 genes. Whereas mcr-2 remains very rare, the recent discovery of mcr-3 in China, Malaysia, Thailand and the USA has been rapidly followed by case reports in Europe, mostly in humans and one in a healthy calf.2–5 Here, we describe the epidemic spread in diseased veal calves of mcr-3/blaCTX-M-55-positive ST744 Escherichia coli collected through the French national surveillance network for antimicrobial resistance (RESAPATH, www.resapath.anses.fr). A collection of 1398 ESBL-producing E. coli, continually sampled from the same peripheral laboratories between 2006 and 2016 as part of the routine surveillance activity of the RESAPATH network, was screened for the presence of mcr-3 using published primers.2 This gene was not detected during 2006–10, but 36 mcr-3-positive E. coli isolates have been identified since 2011 (Table S1, available as Supplementary data at JAC Online). WGS of all 36 isolates was determined by de novo assembly of 2 × 150 bp paired-end reads generated with Illumina sequencing technology (San Diego, CA, USA). Assemblies were performed with SPAdes6 and mappings with a Burrows–Wheeler aligner. Thirty-five isolates carried the prototypic mcr-3.2 gene (T488I variant of mcr-3),5 whereas the last one (no. 37922) presented a variant of mcr-3.2 with the coding sequence disrupted by an IS1-like sequence after nucleotide 174. Interestingly, apart from the first description,2 only the mcr-3.2 gene has been reported yet in E. coli, in a Danish patient and a calf at a slaughterhouse in Spain.4,5 No other mcr variant was found. MICs of colistin, as determined by broth microdilution, ranged from 2 to 4 mg/L for 33 of 36 isolates. Two isolates (nos 30457 and 37922) exhibited MIC values of <0.5 mg/L, probably due to the mcr-3.2 gene disruption in isolate no. 37922 (see above) and to an IS26 inserted 118 bp upstream of mcr-3.2 in isolate no. 30457. The third isolate (no. 41172) displayed an MIC of 1 mg/L, but no mutation was found in or upstream of the mcr-3.2 coding sequence. The reason for this low MIC thus remains unexplained. Besides mcr-3, 10–22 associated resistance genes (modal: 19) were detected, as characterized through a manually curated database derived from the Comprehensive Antibiotic Resistance Database7 and ResFinder database (https://cge.cbs.dtu.dk/services/ResFinder/). In particular, all 36 isolates carried blaCTX-M-55, an ESBL gene highly prevalent in Asia,8 but rare in Europe.9,10,blaCTX-M-55 has also been detected in most mcr-3-positive Enterobacteriaceae reported so far, suggesting an Asian origin for these isolates, possibly through travel or imported food.3 In addition, aacC2, floR, strA, strB, sul1 and sul2 were amongst the other most prevalent resistance genes in this collection (Figure S1). According to the Achtman's E. coli MLST scheme, the in silico analysis revealed considerable predominance of ST744 (29 of 36, 80.6%), which was mostly detected in very distant farms across France over 6 years (Figure 1a and Table S1). Seven other STs were also identified. Among these 29 ST744 isolates, 26 shared ≤20 SNPs among the 4739176 bp recombination-free core genomes, proving an epidemic spread (Figure 1b). The clonal spread of CTX-M-producing E. coli is quite rare in cattle, particularly in 3–5-week-old veal calves whose life does not exceed 5–6 months and who have been reared in different locations over 6 years (Figure 1a). ST744, though infrequent, has been sporadically reported in humans and animals including wild birds, and also in isolates harbouring the mcr-1 gene.11,12 (a) Geographical representation of the 36 mcr-3.2 isolates. All isolates originated from different individuals, mostly from different farms and from 10 different departments across France. Most animals were reared in north-west France, which is a major dairy cattle area. (b) SNP-based minimum spanning tree showing the clonal relatedness among the ST744 isolates. Relative date of isolation is expressed as day (d-). Branches are labelled by the pairwise number of SNP differences and the width of each branch is scaled from these values. Borders of leaves are coloured according to the geographical zone. Squares (a) and rectangles (b) indicate ST744 isolates diverging by ≤ 20 SNPs. Interestingly, another common feature in this collection is the detection using PlasmidFinder (https://cge.cbs.dtu.dk/services/PlasmidFinder/) of the F46:A-:B20 IncF plasmid subtype in all 36 isolates (Figure S2). Hybridization experiments on S1-PFGE gels further revealed the co-localization of mcr-3.2/blaCTX-M-55/IncF in 13 isolates (Table S1). This finding is consistent with the widespread incidence of the same ST744 E. coli clone. It also highlights the role of a plasmid backbone other than the previously reported IncHI2 in the spread of mcr-3. Moreover, IncF plasmids have been associated with the international spread of clinically relevant genes such as blaCTX-M-15, and may actively disseminate the mcr-3.2 gene. Indeed here, the mcr-3.2/blaCTX-M-55/IncF F46:A-:B20 IncF plasmid was also detected outside the ST744 lineage. In conclusion, we report the epidemic spread of mcr-3.2/blaCTX-M-55-carrying E. coli ST744 in cattle in France between 2011 and 2016. The source of this clonal spread remains unknown, but may have an ancient Asian origin, as suggested by the CTX-M-55 epidemiology worldwide. Therefore, we hypothesize its possible introduction in 2011 in France before its further epidemic dissemination in the French veal calf sector. Finally, as the mcr-3 spread was principally by a single ST744 E. coli clone, and our study was confined to ESBL producers, these findings do not reflect the true prevalence of mcr-3 in animals. Nonetheless, the mcr-3 proportion appeared lower than the mcr-1 one when comparing them with the same collection of 1398 ESBL producers (mcr-1: 210 of 1398, 15.0%; mcr-3: 36 of 1398, 2.6%).13,14 In all, our findings importantly document the wide dissemination of plasmid-mediated colistin-resistance genes, and recent reports of the mcr-4 and mcr-5 genes suggest an ongoing story.15,16 We thank all peripheral veterinary laboratories that are members of the RESAPATH network. This work was supported by the Agency for Food, Environmental and Occupational Health and Safety (ANSES), a grant from the Animal Health and Welfare (ANIHWA) ERA-Net project ANR-14-ANWA-0006-04 (France) and a grant from the Joint Programming Initiative on Antimicrobial Resistance (JPI-EC-AMR) project Trans-Comp-ESC-R JPIAMR2016-077 (France). None to declare. Table S1 and Figures S1 and S2 are available as Supplementary data at JAC Online.
No takes yet. Share an insight, caveat, or question.
Haenni et al. (2017) studied this question.