Extremely drug resistant E. coli and K. pneumoniae : novel approaches against their spread and related infections

Abstract

Antibiotic resistance in Enterobacteriaceae is a well-established hindrance and the spread of pathogenic high-risk clones harboring extended-spectrum -lactamases (ESBLs) and carbapenemases is posing a global health challenge, due to very less treatment alternatives. Particular lineages from the major representative species E. coli and K. pneumoniae are predominant in such perspective, like the notorious ST131 and the emerging ST410 and ST1193 in E. coli, and the infamous CG23 and CG258 in K. pneumoniae. Recent additions to the scenery are also K. pneumoniae ST307 and ST147. Nowadays, with the progresses in whole genome sequencing and the reduction in costs, in silico analysis is the gold standard for bacterial classification and detection of accessory genes, including resistance and virulence factors. Moreover, the development of user-friendly tools for the analysis eased the burden of dealing with the immense amount of data generated by the genomic sequencing and reduced the time. This machinery is very useful in the identification and surveillance of high-risk clones, as well as easily detecting novel mutations or genes involved in the resistance. Apart from characterization, which is very useful for a background knowledge, the problem remains in the limited treatment options available for antibiotic resistant bacteria. At the moment, few novel antimicrobials have been developed and the major solution given relies on antimicrobial combinations, such as -lactam--lactamase inhibitors, or revisited therapeutics, such as colistin and fosfomycin. Additionally, phage therapy resurfaced, with a strong interest in the killing activity of lytic phages for standalone treatment in cocktails or as a support in antibiotic treatment. Although few propositions to counteract the spread and the pathologies caused by E. coli and K. pneumoniae high-risk clones, an even lower number has been approved and put into routine clinical practice. Generally, the aim of the thesis was to investigate the clonal composition and presence of high-risk clones in a subset of ESBLs- and carbapenemases-producers Enterobacteriaceae and determine the validity of novel therapeutic approaches, particularly for the β-lactam-β-lactamase inhibitor combination aztreonam-avibactam (AZT-AVI) and K. pneumoniae lytic phages.

The overall number of isolates tested were 449, distributed into 364 E. coli and 85 K. pneumoniae, all of them being ESBLs-producing and carbapenem resistant. These strains belonged to diverse collections from a carbapenemase resistance surveillance in Oman (paper I; 35 E. coli), a cross-sectional study on ESBLs-producing Enterobacterales in Swedish travelers developing diarrhea (paper II; 77 E. coli and 7 K. pneumoniae), a prospective study in ESBLs-producing Enterobacterales causing urinary-tract infection (UTI) in a Swedish cohort (paper III; 223 E. coli and 12 K. pneumoniae), a collection of carbapenemase-producers and ST307 K. pneumoniae (paper IV; n=53) and recent isolates of carbapenemase-producers Enterobacteriaceae from Karolinska University Hospital (paper V; 29 E. coli and 19 K. pneumoniae). In paper IV, a library of 15 previously isolated phages was tested. The isolates belonging to paper I, II and III underwent WGS and in silico characterization for determining sequence types, resistance, and virulence genes; phylotyping was performed on paper I and II isolates. Genomic data about sequence types and resistance factors was already available for isolates in paper IV and V. Additionally, all strains except those in paper IV were phenotypically tested for antimicrobial susceptibility with disk diffusion, broth microdilution and agar dilution methods and the results interpreted by the EUCAST guidelines. In paper IV, phages were tested for host-range, efficacy in killing the indicator hosts, determination of preferrable amplification methods and expansion of range; they were also checked for morphology and challenged for pH stability. Phage DNA was sequenced with Nanopore and screened for undesirable genes with ResFinder, VFDB and Victors databases; comparative genomic analysis with BLAST and annotation databases was performed to determine the novelty of our phages and the similarity with previously detected ones.

Various STs were identified among the isolates, including high-risk clones. E.coli high-risk clone ST131 that carried CTX-M enzymes was predominantly detected (n=79), especially in UTI patients (n=72); other interesting STs were ST38 (n=29), ST69 (n=17), ST410 (n=8) and ST1193 (n=3). Except from isolates in paper IV, in which 12 strains were ST307 harboring CTX-M -lactamases, 6 were ST14, 5 were ST147 and 3 were ST258, K. pneumoniae showed polyclonality. Phylogroup determination in isolates from paper I and II defined phylogroup A as predominant (n=44), followed by phylogroup D (n=34) and B2 (n=12), the latter including ST131 and ST1193. From virulence analysis in paper I, II and III, the most common factor was the presence of adhesins; moreover, in paper II, the virulence patters revealed 45 isolates belonging to DEC pathotypes, majorly EAEC (n=26) and EIEC (n=17), 34 to ExPEC, and 24 that shared both. The most common resistance determinants were ESBLs CTX-M (n=291), OXA-1 (n=64) and TEM-1B (n=82), whereas NDM (n=40) and OXA-48-like (n=27) were the majority for carbapenemase-producers. Moreover, in paper I, II and III, aminoglycosides and fluoroquinolones resistance genes were detected in 223 and 138 isolates respectively. Genetic and phenotypic manifestation of resistance correlated in all isolates but 3 from paper I and 2 from paper III. In particular, the MICs of isolates in paper V to AZT-AVI were below or equal the putative resistance breakpoint of 8 mg/L, although 14 isolates were considered “less susceptible”, and 5 E. coli were able to produce resistant mutants at a frequency of 10-8 and above.

In paper IV, 14/15 phages showed activity, with a narrow range, towards the K. pneumoniae isolates tested. Specifically, three isolates belonging to ST307 (SIV 018, STO 037, DHS 004) and one belonging to ST334 (AKP-2) which were used as indicator strains. After amplification, expansion of host range was detected in KI_14 and KI_15. The threshold of 108 PFU/mL defined only 5 phages as productive. They were very stable to all the pH conditions but pH2. Genomic analysis and TEM images defined KI_1 and KI_5 as Siphoviridae, KI_14 and KI_15 as Myoviridae and KI_3 as Podoviridae. In silico investigation revealed similarity with other phage sequences from NCBI but not identity, and no resistance and virulence genes were detected.

Presence and risk of dissemination of high-risk clones with antibiotic resistance capacity against extended spectrum cephalosporins and carbapenems is disclosed within the E. coli and K. pneumoniae collection included in this work. Particularly for UTI patients, E. coli ST131 played an important role, being the predominant causative agent, and was detected at least once in the other papers. Additionally, the emergent high-risk lineages ST1193 and ST410 were identified; the latter carrying carbapenemases and generally demonstrating a decreased susceptibility to AZT-AVI that might develop into selection of resistance in vivo. The recent K. pneumoniae high-risk clone ST307 was found to be the main target of a previously identified phage library. Of the initial number, 5 novel phages displayed in vitro efficacy and stable replication within the hosts and were determined as free from detrimental genes, desirable features for in vivo studies.