Identification of bovine and porcine colistin-resistant mcr1-positive Escherichia coli.

OBJECTIVE
Polymyxins, especially colistin, have been used for years in veterinary medicine and were rediscovered a few years ago as last resort antibiotics in human medicine against multi-resistant Gram negative bacterial pathogens. For years, only chromosome-mediated resistance to colistin was identified as a consequence of mutation(s) in lipid A-encoding genes. Recently, however, a plasmid-located gene (mcr1) was identified in Gram-negative enterobacteria and has since been found by PCR in several, but not all, bovine, human, porcine and poultry colistin-resistant Escherichia coli (Liu YY et al. Lancet Infect Dis, 2016, 16(2), 161-168; Nordmann P and Poirel L. Clin Microbiol Infect, 2016, 22, 398-400 ; Schwarz S and Johnson AP. J Antimicrob Chemother, 2016, in press, doi: 10.1093/jac/dkw274).
The purpose of this study was to compare phenotypic and genetic for the detection of resistance to colistin and of the mcr1 gene in a collection of Escherichia coli isolated from different animal species and from humans.

METHODS
More than 3000 E. coli isolates from cattle, pigs, dogs, cats, horses, rabbits, chickens ducks and humans were tested for resistance to colistin by growing them on agar plates with 1g/ml of colistin. The Minimal Inhibitory Concentrations (MIC) of and the presence of the mcr1 gene in all growing isolates were determined using the E test® and colony hybridization assay with a mcr1 specific gene probe, respectively. The probe-positive isolates were further tested with the mcr1 gene specific PCR.

RESULTS
A total of 410 E. coli isolated grew on 1g/ml colistin-containing agar plates. The majority of isolates grew well, but several grew sparsely with only few isolated colonies. As determined by the E test®, MIC of 273 isolates (67%) was 1g/ml of colistin and higher; conversely, MIC of 137 isolates (33%) was lower than 1g/ml of colistin.
Of those 410 E. coli isolates, 34 from pigs and bovines (9% of isolates growing on colistin-containing agar plates; 25% of isolates with MIC higher than 1g/ml) hybridized with the mcr1 gene-derived probe: 5 from pigs and 11 from bovines gave black spots (including five from the same calf), while 18 from pigs and one from bovine gave grey spots. All but one pig isolate had a MIC between 1.5 and 16 g/ml of colistin.
Fifteen “black spot” probe-positive isolates tested positive with the mcr1 gene specific PCR as did 3 porcine “grey spot” probe-positive isolates, while the remaining 16 isolates repeatedly tested negative even after lowering the annealing temperature.

CONCLUSION
This study confirms that (i) the results of phenotypic assays for the detection of colistin resistance can not be always trusted; (ii) the mcr1 gene is not the only one mechanism of resistance to colistin; (iii) mcr1 variants may exist that can not be detected by the classical PCR.
Phenotypic assays like growth on colistin-containing agar plates can still represent a first base screening assay, although the MIC determination using the E test® confirms a >1g/ml MIC for only 2 out of 3 growing isolates.
Presence of mcr1 gene and putative variants (like the most recently described mcr2 gene; Xavier BB et al., Eurosurveillance, 21, 7 July 2016) in all probe-positive isolates will be confirmed after Whole Genome Sequencing that will also allow comparing the mcr1-positive plasmids and isolates from pigs and cattle to similar human E. coli isolates. Further studies should also be performed to identify the colistin resistance mechanism in mec-negative isolates.