Assessment of two selective agar media to isolate colistin-resistant bovine Escherichia coli: Correlation with minimal inhibitory concentration and presence of mcr genes

Highlights

• Of 158 bovine E. coli, 94 were colistin-resistant with a MIC >2 μg/ml and 64 susceptible with a MIC ≤2 μg/ml

• The positive predictive value of the two selective agar media studied was 0.98 for the bovine colistin-resistant E. coli

• But only one third of the bovine colistin-resistant E. coli were PCR-positive for mec-1 or mec-2 genes

• At the end, 94% of R E. coli at the disk diffusion assay were colistin-resistant and grew on the selective agar media

Abstract

The identification of colistin-resistant enterobacteria in veterinary medicine is impaired by the absence of first-line reliable phenotypic assay. The purpose of this study was to assess two selective agar media for the detection of colistin-resistant bovine pathogenic Escherichia coli. A total of 158 E. coli (46 R <resistant>, 96 I <intermediate> and 16 S <sensitive> at the disk diffusion assay) isolated between 2013 and 2018 from <3 month-old calves suffering enteritis or septicaemia, were (i) tested by the broth dilution assay to determine colistin Minimal Inhibitory Concentrations (MIC); (ii) streaked on CHROMID® Colistin_R and CHROMagar™ COL-APSE agar plates; (iii) submitted to a pentaplex PCR to identify the presence of mcr-1 to mcr-5 genes. Of the 92 E. coli growing on both agar media, 90 had a MIC > 2.0 μg/ml as had the 3 E. coli that grew only on the CHROMID® Colistin_R agar medium and one E. coli that grew on neither agar media. Therefore, the positive predictive values of the CHROMID® Colistin_R and CHROMagar™ COL-APSE agar media were both 0.98 whereas their negative predictive values were 0.98 and 0.94, respectively. Also noteworthy 43 of the 46 R isolates had a MIC > 2.0 μg/ml and grew on both selective media as did half of the 96 I isolates and only 1 of the S isolates. Conversely, only 30 of the 90 isolates that grew on both agar media and with a MIC > 2.0 μg/ml tested positive for the mcr-1 or mcr-2 genes with the pentaplex PCR. These two selective agar media can be used to reliably detect colistin-resistant E. coli. Positive growth was highly correlated with R results at the disk diffusion assay, but not with the presence of mcr genes.

Introduction

Polymyxins are bactericidal antibiotics produced by Paenibacillus polymyxa with activity against most Gram-negative bacteria, via an interaction with the lipid A moiety of the lipopolysaccharide (Dowling, 2013; Stansly and Schlosser, 1947). Polymyxin E, or colistin, has been used for decades in veterinary medicine more especially against Escherichia (E.) coli infections in farm animals (Dowling, 2013; Kempf et al., 2016). A few years ago, human medicine began to also use colistin against the carbapenemase-producing Enterobacteriaceae in hospitals (Caniaux et al., 2017). Unfortunately it was not long before some species such as Klebsiella (K.) pneumoniae, developed resistance due to mutations in chromosomal genes required for the synthesis of lipid A. The mechanism of resistance is a modification of the negative charge of lipid A that leads to a decreased interaction with the positively-charged polymyxins (Caniaux et al., 2017; Olaitan et al., 2014).

In November 2015, the first plasmid-located transferable mechanism of resistance to colistin was identified in a porcine isolate of E. coli in China (Liu et al., 2016). During the following months and years the mcr-1 (after “mobilized colistin resistance”) gene was identified by PCR in E. coli isolates from cattle, humans, piglets, poultry in several countries in different continents and in several other species of the Enterobacteriaceae family, including Salmonella (S.) enterica and K. pneumonia (Sun et al., 2018). Moreover other mcr genes have been since described: mcr-2 in a porcine E. coli isolate in Belgium (Xavier et al., 2016); mcr-3 in a porcine E. coli isolate in China (Yin et al., 2017); mcr-4 in a porcine S. enterica isolate in Italy (Carattoli et al., 2017); mcr-5 in a poultry S. enterica in Germany (Borowiak et al., 2017); mcr-6, originally named mcr-2.2, in a porcine Moraxella pleuranimalium-like isolate in Great Britain (AbuOun et al., 2017); mcr-7 in a poultry K. pneumoniae isolate in China (Yang et al., 2018); and mcr-8 in a porcine K. pneumoniae isolate in China (Wang et al., 2018).

Identifying acquired resistance of pathogenic bacteria in most routine diagnostic laboratories classically relies on the disk diffusion assay but is not suitable for polymyxins as these poorly diffuse in agar. Actually, the most reliable method for polymyxins is the determination of the Minimal Inhibitory Concentration (MIC) by the broth dilution assay but this procedure is not compatible with the workload of veterinary first-line diagnostic laboratories. Therefore, three selective media for the detection of Gram negative pathogens with acquired resistance to colistin have been commercialized and assessed in human medicine: SuperPolymyxin or Rapid Polymyxin NP (ELITech MICROBIO, Signes, France), ChromID® Colistin R (BioMérieux, Lyon, France) and CHROMagar™ COL-APSE (CHROMagar, Paris, France) (Abdul Momin et al., 2017; Girlich et al., 2018). As far as the authors know, only the SuperPolymyxin medium has been assessed in veterinary medicine for the detection of porcine colistin-resistant E. coli and K. pneumoniae (Kieffer et al., 2017).

Since E. coli is the most frequent pathogenic bacteria isolated from young calves with enteritis or septicaemia (Mainil and Fairbrother, 2014), the purpose of this study was to assess two of those selective agar media, ChromID® Colistin R and CHROMagar™ COL-APSE for the routine detection of colistin-resistant E. coli in a veterinary diagnostic laboratory by correlating the bacterial growth with the MIC and with the presence of mcr genes.