Characterisation of colistin resistance in Gram-negative microbiota of pregnant women and neonates in Nigeria

Abstract

A mobile colistin resistance gene mcr was first reported in 2016 in China and has since been found with increasing prevalence across South-East Asia. Here we survey the presence of mcr genes in 4907 rectal swabs from mothers and neonates from three hospital sites across Nigeria; a country with limited availability or history of colistin use clinically. Forty mother and seven neonatal swabs carried mcr genes in a range of bacterial species: 46 Enterobacter spp. and single isolates of; Shigella, E. coli and Klebsiella quasipneumoniae. Ninety percent of the genes were mcr-10 (n = 45) we also found mcr-1 (n = 3) and mcr-9 (n = 1). While the prevalence during this collection (2015-2016) was low, the widespread diversity of mcr-gene type and range of bacterial species in this sentinel population sampling is concerning. It suggests that agricultural colistin use was likely encouraging sustainment of mcr-positive isolates in the community and implementation of medical colistin use will rapidly select and expand resistant isolates.

Fig. 1. An infographic map indicating the location of sampling, number of samples per type, and prevalence of mcr.

a Shows a map of Nigeria highlighting the location of the three hospital sites NK Kano Nigeria: NN: National Hospital Abuja, Nigeria, NW: WUSE hospital Abuja, Nigeria, with (b) embedded pie charts showing the prevalence of mcr and the split proportion and number of which were maternal rectal samples (MR) and neonatal rectal swabs (BR). NN and NW are both in Abuja and are 8 km apart, NN to NK is 350 km. The map of Nigeria was created using ggplot2 and map_data in R.

Fig. 2. Bacterial species distribution per hospital site, sample type and mcr variant.

A sunburst diagram delineating the mcr gene in the context of the sample type (MR – mother rectal, BR – baby rectal), hospital site (National Hospital Abuja NN, Wuse District Hospital Abuja NW, Murtala Mohammad Specialist Hospital Kano NK) and bacterial species. EAS Enterobacter asburiae, ECL Enterobacter cloacae, EKO Enterobacter kobei, ERO Enterobacter roggenkampii, ESP Enterobacter sp., ECO Escherichia coli, KQI Klebsiella quasipneumoniae, SHI Shigella sp. Forty-nine bacterial isolates carrying an mcr gene are represented in this sunburst diagram.

Fig. 3. Antimicrobial resistance of bacterial isolates purified from mother and neonatal samples.

a Stacked bar graph showing the MIC profile percentages of mcr positive isolates (n = 41) and mcr negative isolates (n = 83) screened against colistin, ampicillin, gentamicin and meropenem. Blue denotes isolates sensitive to both colistin and ampicillin, gentamicin or meropenem CS XS. Yellow denotes isolates sensitive to colistin and resistant to ampicillin, gentamicin or meropenem CS XR. Orange denotes isolates resistant to colistin and sensitive to either ampicillin, gentamicin or meropenem CR XS. Pink denotes isolates resistant to both colistin and ampicillin, gentamicin or meropenem CR XR. b A stacked bar graph showing the percentage profile of isolates that were resistant to how many 0–4 of the antibiotics tested colistin and ampicillin, gentamicin or meropenem. Split by mcr positive isolates (n = 41) and mcr negative isolates (n = 83). 0 (white) shows no resistance and 4 (black) shows resistance to all 4 tested antibiotics. c A heatmap highlighting the skipped well phenomena in Enterobacter spp., showing that the bacteria cease growing at low concentrations (no growth depicted as yellow) of colistin but later return at higher concentrations (depicted in orange), this also indicates this is not an effect restricted to mcr positive isolates as this was observed in both mcr positive and mcr negative isolates.

Fig. 4. Genetic context of mcr-1-positive bacterial isolates.

a A comparative analysis using Mauve of two mcr-1.1 containing IncHI2A plasmids from the same sample; E. coli (pNN-MR49-1, 293,168 bp) ST1266, and a K. quasipneumonaie (pNN-MR49-2, 294,497 bp) compared to two similar non-mcr carrying plasmids accession number NZ_CP069683.1 (from Poland) and NZ_MT929286.1 (from Czech Republic). b The genetic context of the mcr-1.1 Shigella spp. (pNW-MR1609) ST484 IncX4 plasmid relative to Klebsiella pneumoniae IncX4 plasmid (accession number CP041105) from Thailand, where the mcr−1.1 gene was the only major difference. c Despite the very different plasmid backbones for our 3 mcr-containing plasmids they all share an identical 2602 bp without evidence of IS element or transposase presence.

Fig. 5. Two copies of mcr-9 (black arrows) were found in isolate NN-MR-803, one on a plasmid and one inserted into the chromosome.

A previous publication by Tyson et al., had defined three separate cassettes defined by IS elements IS903B and IS1R (blue arrows). The mcr-9 gene carried on the plasmid was completely homologous to the “type A” configuration in their Fig. 3, including the qseB-like and qseC-like genes (green arrows) required for induction of colistin resistance. However, while the NN-MR-803 genomic copy of mcr-9 had similarity to Tyson et al.’s Fig. 3C, an additional intervening ISEsp1 insertion was observed (ISfinder identified genes shown as purple arrows). However, an identical mcr-9 configuration was found on the chromosome of GenBank accession number CP012999 Enterobacter sp. Isolate E20. Conserved cupin fold metalloprotein, WbuC is shown as a grey arrow and other open reading frames are shown as yellow arrows (including a transposase overlapping ISEsp1 from GenBank annotation). Large discrepancies between genomic and plasmid environments surrounding mcr-9 indicates independent acquisition of these from separate sources.

Fig. 6. Twelve isolates were found to carry mcr-10 (black arrows), nine on plasmids and three on chromosomal (chrom) insertions.

In all occurrences the resistance gene was confined to a conserved 3378 bp region that included an adjacent tyrosine-type integrase (tyr-Int or In; brown arrows), defining the smallest unit transposon. Analysing the adjacent environment using ISfinder these unit transposons were located on inserts of various lengths bordered by a variety of IS elements (purple arrows). These could largely be classified as (A) minimal: 5132–5902 bp; B small: 14,539-17,400 bp (although NK-BR1052 was truncated by contig assembly); C intermediate: 25,862–30,548 bp (which includes a region of 20,215 bp homologous) to a Enterobacter roggenkampii plasmid (GenBank Accession number CP116250) containing a cassette of tellurite resistance genes (green arrows); and (D) complex regions of >50,000 bp, two of which include the Enterobacter roggenkampii plasmid tellurite resistance cassette, as well as genes from the tra family of conjugal elements (light blue arrows). Two of these isolates have >20,000 bp regions with high homology to a tra gene family cassette carried by an unnamed plasmid in Enterobacter sp. Z1 (Genbank accession number CP099720). No consistent pattern of boundary IS elements could be discerned to account for a larger conserved mobile element. Yellow arrows indicate open-reading frames that could not be ascribed resistance or mobility but were interspersed between other genes of interest. A full annotation of these regions is not included to highlight areas of greater interest.

Fig. 7. A core genome characterization of 130 Enterobacter spp.

isolates using Panaroo to generate a core genome alignment and IQtree to build the phylogeny. The colours in the circles indicate the sample type (MR, mother rectal or BR, baby rectal). The colour of the label which represents the isolate identification code (leaf) denotes the hospital location (NN, NW & NK). The date displayed is the month of the rectal swab collection. The sequence type (ST) of the isolate is denoted in the numerical text following the isolate date. The colour strip represents colistin sensitivity/resistance with red denoting resistance, yellow sensitivity, and grey indicates data not available. Followed by a heatmap of resistance genes, beginning with the presence of mcr gene, and ARGs were grouped per antibiotic class.