PhD Thesis defense by Pedro Miguela Villoldo at the VISAVET Centre of the Complutense University of Madrid

October 2^nd, 2023

Colistin is a cationic polypeptide whose use as antibiotic was withdrawn since the 1970s because of its toxicity, but was reintroduced later in human medicine to treat infections caused by multi-resistant Gram-negative bacteria, being categorised as a "last resort antimicrobial". Colistin has been used in veterinary medicine since its discovery for the treatment of infections in animals, especially post-weaning diarrhoea caused by E. coli in piglets and calves and colibacillosis in poultry.

Traditionally, colistin was linked exclusively to chromosome mutations in the two-component systems genes pmrA/B and phoP/Q. However, in 2015, a colistin resistance gene, called mcr-1, was described for the first time on a plasmid. Since colistin was one of the few alternatives available for the treatment of multidrug-resistant Gram negative bacteria infections and the spread of this plasmid gene could increase exponentially due to horizontal transfer, the European Medicines Agency (EMA) recommended in 2016 a reduction in the use of colistin in animals.

This doctoral thesis, entitled "Identification and characterisation of colistin resistant bacteria from different environments in Spain. Persistence and possible dissemination assessment", has focused on the study of colistin resistant enterobacteria from different phenotypic and molecular approaches.
In this thesis, isolation and phenotypic characterisation of colistin resistant bacteria and molecular detection of the mcr family genes associated with colistin resistance were carried out. Colistin resistant Enterobacteriaceae were isolated from animals, food of animal origin and water collected from an urban effluent treatment plant. Phenotypic analysis revealed moderate levels of resistance, with Minimum Inhibitory Concentration (MIC) values ranging between 4 and 8 mg/L. Most isolates also showed resistance to other antimicrobials. The detection of mcr genes showed a significantly higher frequency of the mcr-1 gene compared to other members of the mcr family. Other variants, such as mcr-4 and, occasionally, mcr-2 and mcr-3, were also detected, albeit in a lower proportion. The mcr genes showed a different distribution according to the bacterial species in which they were found. The majority of E. coli isolates harboured mcr genes, being the mcr-1 the most frequent variant. The mcr-2 and mcr-3 genes were only detected in this species. However, the presence of mcr genes was not usual in bacteria such as Salmonella enterica or Klebsiella pneumoniae which had mcr in less than half of the isolates, in spite of being phenotypically resistant to colistin, or in species of the genus Enterobacter spp., in which no isolates showed the presence of mcr genes.

This study revealed that the majority of mcr-1 genes were located on IncX4 and IncHI2 conjugative plasmids, in all cases appearing with the pap2 gene in their genetic environment, while mcr-4 genes were only located on small non-mobilizable ColE10 plasmids. From a timeline perspective, it could be observed that mcr genes were mostly located in small conjugative plasmids (IncX4) from 2006 to 2017, while plasmids carrying larger mcr genes (IncHI2) started to become more frequent from 2010, reaching similar values to those of IncX4 by 2018. However, from this year, another change characterised by an increase in the number of isolates with mcr genes located in the bacterial chromosome was observed, reaching larger numbers than those located in plasmids.

The effect of freezing biological samples on the colistin resistant Enterobacteriaceae population and the mcr-1 gene, as well as on the techniques used, bacterial culture and molecular detection by qPCR respectively, were also analysed. Freezing samples of caecal contents for six months at -40 oC reduced the viability and diversity of 84% of colistin-resistant Enterobacteriaceae, while qPCR results did not vary after treatment. These data suggest that techniques based on DNA detection are the best option for retrospective studies of colistin resistance in frozen stored samples and can be complemented by traditional culture-based methods.

The evolution of the mcr-1 gene over time was assessed by quantitative PCR directly on samples of caecal contents. The results obtained showed an upward progression of mcr-1 levels from 2004 to 2015, after which the quantity of the gene began to decrease up to 2021, when gene quantification values similar to those observed in 2008 were obtained. This decrease matched the reduction in the use of colistin in animals, recommended by the EMA and the Agencia Española de Medicamentos y Productos Sanitarios (AEMPS), demonstrating the relationship between polymyxin sales and colistin resistance associated with the mcr-1 gene.

Finally, the stability and dissemination of colistin resistant strains carrying the mcr-1 gene was evaluated by an in vivo study, to determine the effect that the administration of this antimicrobial could have on the selection of resistant strains, using Salmonella enterica as mcr-1-carrying bacteria and broiler chickens as animal model. The animals were divided in four groups, whereby two groups (G3 and G4) were challenged with the bacterial strain and two groups (G1 and G2) remained unchallenged. In addition, two groups were treated with colistin for seven days (G2 and G4), thus keeping one group free of both treatment and exposure to the mcr-1 – carrying bacteria (G1). The results showed a higher quantification of mcr-1 in the groups of animals inoculated with the bacteria, reaching maximum levels earlier (day 14) in the colistin treated group (G4) than in the untreated group (G3). Furthermore, mcr-1 levels were equal for both groups on day 21 of the study, albeit Salmonella spp. levels were reduced, indicating that the maintenance of mcr-1 levels could be due to horizontal transfer phenomena between intestinal bacteria.