Carriage of colistin-resistant Gram-negative bacteria in children from communities in Cape Town (Tuberculosis child multidrug-resistant preventive therapy trial sub-study)

Colistin is a last-resort antibiotic against multidrug-resistant, Gram-negative bacteria. Colistin resistance has been described in the clinical settings in South Africa. However, information on carriage of these bacteria in communities is limited. This study investigated gastrointestinal carriage of colistin-resistant Escherichia coli and Klebsiella spp. and mcr genes in children from communities in Cape Town. Colistin-resistant E. coli was isolated from two participants (4%, 2/50), and mcr-1-mcr-9 genes were not detected. Gastrointestinal carriage of colistin-resistant Enterobacterales was rare; however, continuous extensive surveillance is necessary to determine the extent of carriage and its contribution to resistance observed in clinical settings.


Introduction
Antibiotic resistance is threatening public health globally, and colistin remains one of the last-resort antibiotics for treating infections because of carbapenem-resistant Enterobacterales. Colistin resistance is increasingly being reported in Enterobacterales both worldwide and in South Africa, which is of great concern. 1,2,3,4,5, 6,7,8,9 The predominant mechanism of colistin resistance in Enterobacterales involves changes to the phosphate groups of lipid A by adding phosphoethanolamine (PEtN) and/or 4-amino-4-deoxy-L-arabinose (L-Ara4N), resulting in reduced anionic charge of the lipopolysaccharide (LPS). Mutations in the two-component regulatory systems, PhoPQ and PmrAB, and inactivation of the mgrB gene are the most common causes of colistin resistance. The mobile colistin resistance (mcr-1) gene was first reported in 2016 in China. 1 The mcr gene encodes a PEtN transferase enzyme, which transfers a PEtN to lipid A, thereby conferring resistance to colistin. Additional plasmid-mediated mcr genes, mcr-2-mcr-10 and different variants thereof, have since been detected, which also confer colistin resistance. 1,2 mcr genes have been mostly isolated from Enterobacterales, especially Escherichia coli and Klebsiella pneumoniae, which are responsible for both nosocomial and community-acquired (CA) infections, including sepsis, pneumonia, urinary tract infections and intra-abdominal infections 10,11 ; these organisms are also common gut commensal organisms. Numerous studies have detected the mcr gene in isolates from clinical, community and environmental settings. 1,2,3,4,5,6 The mcr-1 gene has been identified in clinical isolates from at least 11 hospitals across South Africa, 7,8,9 as well as in E. coli isolates from outpatients in Gauteng (n = 4) and the Western Cape (n = 1). 8 The mcr-1 gene has also been detected in E. coli isolated from broiler chickens, a pig and final effluents from wastewater treatment plants in South Africa. 3,4,5,6 However, there are limited data regarding the carriage of colistin-resistant organisms and mcr genes in healthy individuals from the community. The presence of mcr genes in the community could serve as a reservoir for colistin resistance in clinical settings, leading to difficult and more expensive treatment for human bacterial infections, increased hospitalisation, extended hospital stays and sometimes death. Gastrointestinal carriage of plasmid-mediated colistin resistance poses additional threat Colistin is a last-resort antibiotic against multidrug-resistant, Gram-negative bacteria. Colistin resistance has been described in the clinical settings in South Africa. However, information on carriage of these bacteria in communities is limited. This study investigated gastrointestinal carriage of colistin-resistant Escherichia coli and Klebsiella spp. and mcr genes in children from communities in Cape Town. Colistin-resistant E. coli was isolated from two participants (4%, 2/50), and mcr-1-mcr-9 genes were not detected. Gastrointestinal carriage of colistinresistant Enterobacterales was rare; however, continuous extensive surveillance is necessary to determine the extent of carriage and its contribution to resistance observed in clinical settings.

Carriage of colistin-resistant Gram-negative bacteria in children from communities in Cape Town (Tuberculosis child multidrug-resistant preventive therapy trial sub-study)
Read online:

Sample collection
Stool specimens were collected by nurses from children who were enrolled in the trial before commencing study treatment using a standard operating procedure. Six scoops of formed stool or three scoops of loose stool were collected using a 2.5 millilitres (mL) spoon (Lasec, South Africa) and mixed several times. Samples were delivered in a cooler box with an ice pack (2°C -8°C) to the National Health Laboratory Services (NHLS) microbiology laboratory at Tygerberg Hospital and were stored at -80°C.

Results
Bacterial growth was observed on the MacConkey agar for all 50 stool samples. A total of 70 isolates, 55 E. coli from 25 participants and 15 K. pneumoniae from six participants, were obtained on the MacConkey agar with colistin and vancomycin. Escherichia coli and/or K. pneumoniae were obtained from the colistin containing media of 30 participants. More than one E. coli or Klebsiella spp. colony was isolated from some of the participants because of differences in colony morphology, and both E. coli and K. pneumoniae were isolated from one participant. For 20 of the participants, no E. coli or Klebsiella spp. were obtained from the colistin containing media. Colistin resistance was confirmed by BMD in only 3% (2/70) of the isolates, both of which were E. coli (MIC = 4 mg/L). The colistin MICs of the remaining isolates were in the range of 0.125 mg/L -0.5 mg/L. The two colistinresistant E. coli were from two separate individuals, and therefore, the carriage rates of colistin-resistant E. coli and K. pneumoniae were 4% (2/50) and 0% (0/50), respectively.
The plasmid-mediated mcr-1 to mcr-9 colistin resistance genes were not detected in either the colistin-resistant E. coli isolates or the DNA extracted directly from the stool samples.

Discussion
Colistin is regarded as one of the last-resort antibiotics, and therefore, any emerging resistance outside of hospital settings is of great concern. In South Africa, there is a lack of information regarding the carriage of colistin resistance and mechanisms of resistance in children and in the community, in general. In another Cape Town study, nasopharyngeal colonisation with resistant pathogens was reported in human immunodeficiency virus (HIV)-positive children, with extended-spectrum beta-lactamase production in 50% of Enterobacteriaceae 15 ; this poses a risk of becoming infected by these resistant organisms. The TB CHAMP trial http://www.sajid.co.za Open Access provided an opportunity to obtain data on gastrointestinal carriage of colistin resistance in healthy children from communities in Cape Town.
This study showed that carriage of colistin-resistant organisms was uncommon in healthy children enrolled from local communities, with 4% of participants carrying colistinresistant E. coli. The majority (68/70) of the colonies obtained on the colistin-containing media were not colistin-resistant based on BMD. This could be related to the poor diffusion of colistin in agar as previously reported 16 or could be influenced by the amount of stool cultured, or the abundance of bacteria in the stool sample. However, resistance was confirmed by BMD, which is considered the only reliable colistin susceptibility method. 17 A previous study has selected colistin-resistant Gram-negative bacteria (GNB) with a higher (4 mg/L) colistin concentration in agar 18 ; however, this approach has the potential to miss isolates with a borderlineresistant MIC (2 mg/L -4 mg/L). Development of a costeffective and user-friendly commercial BMD panel for colistin susceptibility testing or direct mcr screening could be valuable for surveillance purposes.
None of the mcr-1-mcr-9 genes were detected in any of the isolates or stool samples, suggesting that although mcr genes have been detected in clinical settings in South Africa, 7,9 these genes are not widely disseminated in these communities. The two colistin-resistant E. coli isolates could have chromosomal mutations in pmrA, pmrB, phoP, phoQ, eptB or mgrB genes, which lead to L-Ara4N and/or PEtN modification of lipid A in the bacterial outer membrane resulting in colistin resistance. 19,20,21,22,23 In addition, other mcr genes (mcr-10) or potentially undescribed, plasmidmediated colistin resistance genes could be responsible for colistin resistance.
To the best of our knowledge, this is the first study to investigate the carriage of colistin resistance in a South African community setting. Colistin resistance in communities, especially the dissemination of mcr genes, could lead to infections with these resistant organisms, which could further spread to hospital settings. Although the sample size was small, our data suggest that community carriage of colistin-resistant organisms in children is uncommon. This also reassures that colistin can still be used in this community as a last-resort antibiotic when necessary. Ongoing surveillance is needed to better understand and estimate the prevalence of colistin resistance in the community.

Conclusion
We report a low carriage rate of colistin-resistant Enterobacterales and no plasmid-mediated colistin resistance genes in children sampled from the community level. Colistin is the last-resort antibiotic for the treatment of multidrugresistant GNB, and therefore, there is a need to understand the prevalence of colistin resistance in community settings. Ongoing surveillance is required in additional community settings in order to determine the extent of carriage and its contribution to resistance observed in clinical settings.