|Home | About | Journals | Submit | Contact Us | Français|
Polymyxin usage is common in Singapore, where carbapenem resistance is prevalent (1). Not surprisingly, polymyxin resistance has been detected in our setting (2). Until recently, polymyxin resistance has been considered rare and largely due to chromosomal mutations. Here, we report the first detection of plasmid-mediated polymyxin resistance in blaKPC-2-producing Enterobacteriaceae in Singapore.
Polymyxin B resistance, which was interpreted based on the Clinical and Laboratory Standards Institute's non-Enterobacteriaceae breakpoints (MICs ≥ 4 μg/ml), was detected in 63 of 321 (20%) clinical carbapenem-resistant Enterobacteriaceae isolates collected between 2008 and 2015 at our institution (2, 3). Using PCR and sequencing, we detected mcr-1 in three of the 63 (5%) resistant isolates (4). These isolates were resistant to all tested antibiotics except amikacin, tigecycline, and levofloxacin (EC702 only). Multiple resistance genes, including blaKPC-2, were detected in addition to mcr-1 (Table 1).
Escherichia coli EC249 and EC362 were recovered from two local patients admitted to different wards in a public tertiary care hospital on the same day in 2013, while Enterobacter aerogenes EC702 was recovered from a local patient admitted in 2014 (Table 1). All three patients had previous hospital admissions and exposure to broad-spectrum antibiotics, with no recent travel history. None of the patients had prior receipt of polymyxin B. Only E. coli EC362 was associated with clinical infection. It is interesting that these patients were colonized/infected with other Enterobacteriaceae isolates before or after the isolation of the index mcr-1-positive isolates (Table 1). Interestingly, these isolates were all blaKPC-2 positive. mcr-1 was absent in all other isolates from these patients, including those which were polymyxin B resistant.
The whole-genome sequences of the three isolates were analyzed (ENA sequence accession no. PRJEB14213). Comparison with complete mcr-1 plasmid sequences in GenBank indicated that the mcr-1-harboring contig from EC249 shared similarity (99.9% identity, 87% query coverage) with plasmid pHNSHP45 (GenBank accession no. KP347127), which was first reported in China. The mcr-1-harboring contig from EC702 shared similarity (100% identity, 99% query coverage) with pmcr1-IncX4 (GenBank accession no. KU761327), which was also isolated from a clinical specimen from China (5). There was, however, limited similarity of the mcr-1-harboring contig from EC362 with known mcr-1 plasmid sequences in GenBank. Analysis of the mcr-1-neighboring regions revealed that while the mcr-1–pap2 element was present in all three contigs, the ISApl1 mobile element was found only in EC362. As suggested in several studies, ISApl1 may not be always associated with mcr-1 (4, 6). The nikB gene was present in EC249. mcr-1 and blaKPC-2 were located on distinct unordered contigs for all three isolates. Further plasmid characterization, such as through long-read single-molecule real-time sequencing, is required to determine if mcr-1 and blaKPC-2 are located on the same plasmids, in addition to determining the diversity of the mcr-1-harboring plasmid backbones of the isolates in this study.
Successful transfer of mcr-1 was observed in the conjugation assays with E. coli J53 Azr for all three isolates. The transconjugants were positive for mcr-1 and blaKPC-2 (confirmed with PCR and sequencing). While the transconjugants had polymyxin B MICs similar to those for the donor isolates, the transconjugants' carbapenem MICs were mostly lower (2 to 8 μg/ml).
Comparative genomics suggests that genetic heterogeneity is present in the two E. coli strains isolated in the same period. The multilocus sequence type (MLST) profiles of the two E. coli isolates were similar—EC249 (ST2006) and EC362 (ST224) are single-locus variants, differing only in fumC (fumC-371 versus fumC-4) (7). However, variant calling by GATK revealed 3,963 single nucleotide polymorphisms (SNPs) between the two isolates, suggesting that they were not identical. The isolates in our study were of different sequence types than those detected previously.
In conclusion, we report the first cases of mcr-1 in clinical isolates in Singapore. While polymyxin resistance in local carbapenem-resistant Enterobacteriaceae isolates is high, it appears that plasmid-mediated resistance is rare and sporadic. Cocarriage of highly transmissible resistance determinants such as mcr-1 and various carbapenemase genes (blaKPC, blaNDM, blaVIM, and blaOXA-48) in human infections is increasingly being reported globally in regions such as Europe, the Americas, China, and South Africa (5, 8,–14). Close attention is warranted for these multidrug-resistant infections, for which there are few therapeutic options remaining.
We declare no competing interests.
We thank Hsu Li Yang for providing valuable comments and the members of the Carbapenemase Producing Enterobacteriaceae in Singapore (CaPES) group for contributing the whole-genome sequence of EC702.