Presence of mcr-1-positive Enterobacteriaceae in retail chicken meat but not in humans in the Netherlands since 2009

Citation for published version (APA): Kluytmans-van den Bergh, M. F., Huizinga, P., Bonten, M. J., Bos, M., De Bruyne, K., Friedrich, A. W., Rossen, J. W., Savelkoul, P. H., & Kluytmans, J. A. (2016). Presence of mcr-1-positive Enterobacteriaceae in retail chicken meat but not in humans in the Netherlands since 2009. Eurosurveillance, 21(9), 12-18. https://doi.org/10.2807/1560-7917.ES.2016.21.9.30149


Introduction
The worldwide emergence of extended-spectrum betalactamases (ESBL) and carbapenemases has limited the available treatment options for infections with Gram-negative bacteria [1].Colistin is considered to be an antibiotic of last resort for the treatment of infections with carbapenem-resistant bacteria, and its use in humans is increasing [1].
In November 2015, the presence of a plasmid-mediated colistin-resistance gene, mcr-1, in Enterobacteriaceae from food animals, food and patients in China was reported [2].The mcr-1 gene was detected in 21% of Escherichia coli isolates cultured from pigs at slaughter and in 15% of E. coli isolates cultured from retail meat between 2011 and 2014.In addition, the mcr-1 gene was present in 1.4% of E. coli isolates and 0.7% of Klebsiella pneumoniae isolates from clinical cultures from patients in two Chinese hospitals in 2014.Directly following this publication, the mcr-1 gene was reported to be present in 0.2% of ESBL-and AmpC-producing E. coli isolates from human bloodstream infections, and in 2% of E. coli isolates cultured from imported chicken meat in Denmark since 2012 [3].Hereafter, several reports have documented the global presence of the mcr-1 gene in Enterobacteriaceae cultured from humans, food animals and food [4][5][6][7][8][9][10][11][12][13].
Traditionally, colistin resistance was thought to be mediated by chromosomal mutations only, and to spread exclusively via clonal transmission of resistant isolates [14].The emergence of plasmid-mediated

Table 1
Enterobacteriaceae isolates from retail chicken meat, rectal samples, clinical cultures and outbreaks by year of culture, type of isolate, and colistin-resistance, analysed by whole genome sequencing for the presence of the mcr-1 gene, the Netherlands, 2009-2015 (n = 2,471)

Hospitalised patients, rectal samples
The retail chicken meat surveys in 2009 and 2014 were accompanied by hospital-wide prevalence surveys in patients who were admitted to four hospitals in the region where the chicken meat was bought [15,16] used to perform whole genome SNP (wgSNP) analysis (BioNumerics v7.6 beta, Applied Maths).

Antimicrobial susceptibility testing
Isolates for which antimicrobial susceptibility data were available were screened for the presence of colistin resistance.Susceptibility testing of the three mcr-1-positive E. coli isolates was performed using Vitek2 (bioMérieux, France) and Etest (bioMérieux, France).The breakpoint tables of the European Committee on Antimicrobial Susceptibility Testing (EUCAST) were used for the interpretation of minimum inhibitory concentrations (MICs) [26].Isolates with a colistin MIC > 2 mg/L were considered colistin-resistant.

Results
An overview of the 2,471 Enterobacteriaceae isolates from retail chicken meat, rectal samples, clinical cultures and outbreaks is presented in Table 1.Colistin resistance was found in two (1.6%) of 122 chicken meat-derived ESBL-E isolates, in 14 (1.1%) of 1,247 isolates from ESBL-E rectal carriers, and in 15 (1.8%) of 813 ESBL-E isolates from clinical cultures.The mcr-1 gene was detected in three (1.5%) of 196 chicken meatderived ESBL-producing E. coli isolates, one cultured in 2009 and two in 2014.For all three isolates, the mcr-1 sequence showed 100% similarity to the gene reported in China [2].None of the 2,275 human isolates harboured the mcr-1 gene.
Table 2 shows the general and molecular characteristics of the three mcr-1-positive E. coli isolates.
The isolate that was cultured in 2009 had sequence type ST2079, was CTX-M-1-positive and harboured 17 acquired resistance genes.Both isolates from 2014 had sequence type ST117, were SHV-12-positive and harboured five acquired resistance genes.Although these two isolates were cultured from different meat samples of non-Dutch origin, the meat samples had the same lot number and were bought in the same supermarket on the same day.Plasmid replicons were identified in all three isolates, eight in the isolate from 2009 and two in both isolates from 2014.However, none of the plasmid replicons could be linked to the mcr-1 gene.
Antimicrobial susceptibilities for the three mcr-1-positive E. coli isolates are shown in Table 3.All three isolates were colistin-resistant (MIC > 2 mg/L).The isolate from 2009 tested colistin-susceptible by Vitek2 (MIC = 2 mg/L), but resistant by Etest (MIC = 3 mg/L).wgMLST analysis showed that the two isolates from 2014 differed by only three (0.07%) of 4,243 shared loci, whereas the isolate from 2009 differed by 3,606 (95.1%) of 3,791 shared loci (Table 4).The two isolates from 2014 differed by only eight SNPs in wgSNP analysis.

Discussion
In our study, covering the period 2009 to 2015, we detected the recently described plasmid-mediated colistin resistance gene, mcr-1, in three ESBL-producing E. coli isolates from retail chicken meat samples obtained from Dutch supermarkets in 2009 and 2014.
All three mcr-1-positive isolates were colistin-resistant, and two of them were genetically closely related.No mcr-1-positive isolates were detected in a large collection of Enterobacteriaceae isolates of human origin that were collected during the same time period and included isolates of four outbreaks with colistin-resistant Enterobacteriaceae.
The observed 1.5% prevalence of mcr-1-positive isolates is comparable with the reported 2% (5/255) prevalence in imported chicken meat in Denmark, and is lower than the 15% (78/523) prevalence in retail meat in China [2,3].This lower prevalence may be related to the relatively low rates of polymyxin use in livestock in Europe.In 2014, polymyxins constituted only 0.4% (0.34 defined daily dose animal (DDDA)/animal year) of all antibiotics used in broilers in the Netherlands, with a decreasing trend over the last few years [27].
It is noteworthy that the observed 1.5% prevalence of mcr-1-positive isolates in ESBL-E isolates from retail chicken meat in this study is similar to the 1.5% phenotypic colistin resistance that was found in E. coli isolates cultured from Dutch retail chicken meat in 2014 [27].Unfortunately, no data are currently available on the resistance mechanisms involved in this phenotypic colistin resistance.
The genetic identity between the two mcr-1-positive isolates that were obtained from the same batch of meat samples most likely represents batch contamination from a common source.
The mcr-1-positive isolates in this study belong to different sequence types as compared with those that were found to be related to the mcr-1 gene in the Chinese and Danish study [2,3].E. coli ST2097 is uncommon in humans, but has been reported once before in a study on ESBL-producing bacteria in flies from broiler farms in the Netherlands [28].E. coli ST117, on the other hand, is common in both poultry and humans [16,29].The detection of the mcr-1 gene in isolates that belong to different sequence types illustrates the potential for horizontal transfer of this resistance gene.Although all chicken meat samples were bought in Dutch supermarkets, the labelling of the samples did not provide any clue with respect to the country where animals were raised.Available data on the origin of the chicken meat were limited to the producing country for the samples from 2014 (non-Dutch, European), for the 2009 isolate this information was not available.A non-European origin of the mcr-1-positive meat samples can, therefore, neither be confirmed, nor excluded.
The absence of the mcr-1 gene in human isolates of various origins is in accordance with observations in previous studies that the presence of the mcr-1 gene in clinical isolates is still rare.In China, 1.4% (13/902) of clinical E. coli isolates and 0.7% (3/420) of clinical K. pneumoniae isolates were mcr-1-positive, and in Denmark, only 0.2% (1/417) of ESBL-and AmpCproducing E. coli isolates from bloodstream infections [2,3].This absence of the mcr-1 gene in current Dutch collections of human Enterobacteriaceae may in part be due to the low use of colistin and its analogues, the polymyxins, in humans in the Netherlands.In 2014, polymyxins constituted less than 0.1% (0.01 defined daily dose (DDD)/1,000 inhabitant-days) of all systemic antimicrobials used in primary care and ca 0.3% (0.2 DDD/100 patient-days) of systemic antimicrobials used in the hospital setting [30].
Short-read sequence data are not optimal for the assembly of plasmid sequences, which are known to contain multiple repetitive elements.This may explain why the analysis of our sequence data did not reveal a link between the mcr-1 gene and the plasmid replicons identified.
Although the prevalence of mcr-1-positive isolates in meat samples was low, the presence of this colistin resistance gene in food represents a potential public health threat, as it is located on mobile genetic elements that have the potential to spread horizontally to other bacteria.With the increase in carbapenem resistance, the use of colistin is increasing and, herewith, the selective pressure for the spread of mcr-1 genecontaining plasmids.As colistin has become one of the last resort antibiotic options to treat severe infections with Gram-negative bacteria, the continued monitoring of colistin resistance and its underlying resistance mechanisms is important, not only in humans, but also in food production animals and food.The emergence of plasmid-mediated colistin resistance underpins the recent proposal of veterinary experts to reconsider the use of colistin and its analogues in food production animals [31].
In conclusion, the plasmid-mediated colistin resistance gene mcr-1 was detected in three ESBL-producing E. coli isolates that had been cultured from retail chicken meat from Dutch supermarkets in 2009 and 2014.Two isolates were obtained from the same batch of meat samples, which most likely represents contamination from a common source.The mcr-1 gene was not present in a large collection of human isolates collected between 2009 and 2015 in the Netherlands.These findings indicate that mcr-1-based colistin resistance currently poses no threat to healthcare in the Netherlands, but requires continued monitoring of colistin resistance and its underlying mechanisms in humans, livestock and food.
Two E. coli isolates and one K. pneumoniae isolate were not available for whole genome sequencing.
c d Retail chicken meat Two ESBL-producing Enterobacteriaceae (ESBL-E) prevalence surveys in Dutch retail chicken meat were performed in 2009 and in 2014 [15,16].A total number of 196 ESBL-E isolates were obtained, 74 isolates from 71 ESBL-E-positive meat samples in 2009 (89 samples cultured), and 122 isolates from 86 ESBL-E-positive meat samples in 2014 (101 meat samples cultured).

Table 2
Characteristics of the mcr-1-positive Escherichia coli isolates from retail chicken meat, the Netherlands, 2009-2015 Isolate 14M009386 and 14M009387 were cultured from different meat samples with the same lot number.
[20]16]09, ESBL-E rectal carriage was detected in 45 (5.1%) of 876 patients, who carried 50 ESBL-E isolates.Two repeated prevalence surveys in one of the four hospitals in 2013 and 2014, yielded 63 ESBL-E isolates obtained from 63 (5.9%) ESBL-E carriers among 1,065 patients cultured[17].A multi-centre cluster-randomised study comparing contact isolation strategies for known ESBL-E carriers was performed in 14 Dutch hospitals between 2011 and 2014 (SoM study)[18].All consecutive adult patients with a routine clinical culture with ESBL-E were placed on contact precautions and enrolled in the study (= index patient).Ward-basedESBL-E prevalence surveys were performed one week after enrolment of the index patient.Perianal swabs were obtained from 10,691 patients and identified 992 (9.3%) ESBL-E carriers, from whom 1,134 ESBL-E isolates were cultured.Hospitalised patients, clinical culturesIn 2009, 2013 and 2014, all consecutive ESBL-E isolates from blood cultures were prospectively collected in the four hospitals that participated in the ESBL-E rectal carriage prevalence surveys[15,16].A total number of 102 ESBL-isolates from blood cultures were obtained, 25 isolates from 23 patients with an ESBL-E-positive blood culture in 2009, and 77 isolates from 76 patients in 2013 and 2014.Three isolates that were collected in 2014 were not available for whole genome sequencing.In the SoM study, a total number of 711 clinical ESBL-E isolates were obtained from 654 ESBL-E-positive patients.isolates);(iii)an outbreak of colistin-resistant E. cloacae in an intensive care unit between 2009 and 2014 (86 isolates); (iv) an outbreak of colistin-resistant KPCproducing K. pneumoniae in a nursing home in 2012 (10 isolates)[20]; (v) an outbreak of colistin-resistant K. pneumoniae in patients after endoscopic retrograde cholangio-pancreaticography (ERCP) procedures in a 2014-2015 (50 isolates); and (vi) an outbreak of (intrinsic) colistin-resistant Serratia marcescens in a neonatal intensive care unit in 2014-2015 (26 isolates).

Whole genome sequencing and analysis of sequence data
[24]ated version 2.1, including the mcr-1 sequence) of the Center for Genomic Epidemiology (http://www.genomicepidemiology.org/)[22].For isolates from two outbreaks (colistin-resistant E. cloacae and ERCPrelated colistin-resistant K. pneumonia), the thresholds for sequence identity and coverage length were set to genomes of E. coli and Shigella spp.(BioNumerics v7.6beta, Applied Maths).Additionally, single nucleotide polymorphism (SNP) calling was performed by mapping the paired-end reads of isolate 14M009387 and isolate 213 to the de novo assembled genome of isolate 14M009386, using Bowtie 2.5.5[24]andSAMtools[25].Resulting Binary Alignment Maps (BAM) files were

Table 4
Whole genome multilocus sequence typing analysis and whole genome single nucleotide polymorphism analysis of mcr-1-positive isolates from retail chicken meat, the Netherlands, 2009-2015 a