The rates and routes of transmission of multidrug resistant Klebsiella clones and genes into the clinic from environmental sources.

Recent years have seen a dramatic global increase of strains of bacteria that pose a high risk in health care settings due to broad resistance to a wide range of antibiotics.

Our research project will concentrate on Klebsiella pneumoniae (Klebsiella), a leading cause of multidrug resistant hospital-acquired infections globally, particularly among neonates, the elderly and immunosuppressed patients, with reported mortality rates upwards of 50%. It is also responsible for an increasing burden due to community-acquired invasive infections, which can lead to pyogenic liver abscess, pneumonia, and meningitis. In addition to causing disease, Klebsiella is an ecological generalist and can be carried asymptomatically in the intestinal tract, skin, nose, and throat of healthy individuals, and can thrive in a range of arthropod, bird or plant hosts (including crops and trees) and environmental niches such as water and soil. It can also cause invasive disease in several animal species, and it is a common cause of mastitis in dairy herds.

In order to control the spread of Klebsiella through targeted surveillance and intervention policies it is necessary to identify the sources of emergent community and health-care associated infection from the interlinked and varied niches encompassing "the environment". Environmental prevalence data provides a critical baseline, but is not always sufficient to inform on public health policy. For instance, the observation that gulls frequently harbour Klebsiella is, by itself, insufficient to guide policy on managing gull populations, as direct transmission to humans remains very difficult to demonstrate. However, if, say, water were to be identified as a major Klebsiella reservoir with frequent transmission to and from the clinic, this might in turn impact on waste water treatment procedures. Similarly, if cockroaches are found to be associated with Klebsiella this should lead to increased pest control both in hospitals and in the community.

To address this, we will take samples in order to identify Klebsiella from multiple clinical, community, agricultural, veterinary and environmental settings. Then we will use a novel approach based on whole genome sequencing to provide a much more efficient means to understand ecological adaptation, the distribution of resistance and virulence genes in the bacteria, and to identify key environmental settings from which these Klebsiella bacterial clones emerge. This approach will involve sequencing multiple colonies of bacteria simultaneously in order to detect a wider sample of genes present within a given sample. We will build a comprehensive, data-driven, model of adaptation, diversification, gene flow and transmission dynamics between all relevant ecological and clinical compartments. This will then pave the way for the development of targeted disease management strategies.

Klebsiella pneumoniae (Kp) is a leading cause of multidrug resistant hospital-acquired infections globally, and is responsible for an increasing public health burden in the community. In order to control the spread of Kp through targeted surveillance and intervention policies it is necessary to identify the sources of emergent community and health-care associated infection from the interlinked and varied niches encompassing "the environment". To address this, we will sample from multiple clinical, community, agricultural, veterinary and environmental settings in and around a single town, Pavia, in Northern Italy, and supplement these data with matched samples from France and elsewhere. We will use whole genome sequencing of community (mixed-colony) samples to assay accessory gene abundance and distribution. This contrasts with the more common approach based on phylogenetic analysis of single colonies, which would be of limited utility over broad environmental scales due to the complexity of transmission chains, environmental dormancy, and high rates of recombination.

This gene-centric approach provides a much more efficient means to understand ecological adaptation, the distribution of resistance and virulence genes, and to identify key environmental reservoirs from which clinical clones emerge. A key deliverable of this project will be the establishment of a pan-genome database ('pangenomium') that will integrate with both existing Kp genome community resources established by project partners (BIGSdb-kp, and wgsa.net). We will also develop mathematical models to further explore gene flow between environmental compartments, and will use the Oxford Nanopore Minion TM platform to characterise AMR plasmids, and understand their distribution and transfer, using hybrid assemblies.

Klebsiella pneumoniae (Kp) is a highly diverse pathogen that is ubiquitous in the environment. This high level of variation will seriously hinder attempts to infer transmission dynamics between different environmental niches (including animal hosts) using approaches based on SNP variation within the core genome. This is due to factors such as the complex fitness landscape of the wider environment, the high level of genomic diversity (even within single hosts), high rates of co-colonisation, environmental dormancy, recombination, complex transmission chains and the simultaneous transmission of multiple variants. These issues would lead to inevitable gaps in sampling and statistical uncertainties stemming from the chance sequencing of one colony over another for any given sample.

This project will address these difficulties by providing a novel framework for understanding transmission on the basis of the distribution of accessory genes, rather than core genome variation, and in doing so will be of direct benefit to medical microbiologists and bacterial genomicists working on Kp, as well as the broader microbial genomics community. Our approach will be to construct a database of all known genes in this species, then used mixed colony sequencing to assay the distribution of these genes across multiple ecological niches within and around a single town (Pavia, northern Italy). This will enable us to identify niche specific genes, which will then be used as markers to determine the principal routes by which Kp moves through the environment and into healthcare settings. The identification of specific environmental settings enriched with particular AMR or virulence genes will inform on community risk assessment and policy guidelines. The project will involve the development of novel statistical and bioinformatics approaches and mathematical models, and all pipelines and tools will be readily accessible to the broader community via MRC-CLIMB. The resultant database and analysis tools will provide the means for users to infer the likely environmental source of query genomes by automated mining for the presence of our niche-specific marker genes. This will provide critical knowledge for future source attribution studies of Kp infections and to prioritize interventions to limit the spread of MDR-Kp from its natural reservoirs to human carriage and infections. Our team includes the developers of both Kp-BIGsdb and wgsa.net genomic platforms, and working to common standards between these two systems will be of further broad utility to the community at large. We will also develop and utilise our hybrid assembly methods to combine long read data generated by the Oxford Nanopore Minion with short read Illumina data which can be used for the complete assembly of plasmids (Bayliss et al, Gigascience 2017). In this way we will provide critical data concerning plasmid composition and mobility; understanding the genetic bases of (in)compatibilities between resistance genes, mobile elements and host chromosomes will play a key role in mitigating their spread. In addition to shedding light on plasmid transfer and evolution in Kp, our project will also provide important methodological advances for studying accessory elements on population scales.

In sum, this project will define, at what we believe is an unprecedented scale, the distribution of Kp lineages and genes in various environments. The project will also contribute to strengthen the bioinformatics infrastructures (wgsa.net, BIGSdb-Kp) used by the global microbiology and surveillance communities to communicate on the epidemiology and population biology of Kp. These advances will prove critical in the ongoing battle to mitigate the global spread of this serious pathogen.