| Summary: | Urbanization is characterized by rapid intensification of agriculture, socioeconomic change, and ecological fragmentation, which can have profound impacts on the distributional ecology of host populations and epidemiology of infectious disease within them. In this thesis, results from a large-scale field study conducted in Nairobi, Kenya are used to explore how anthropogenic and ecological changes associated with urbanisation influence the structure of sympatric wildlife, livestock and human host populations, and dictate bacterial epidemiology in wildlife hosts. As likely points of contact (and thus parasite transmission) between vertebrate wildlife, livestock, and humans, household 'interfaces' were chosen as sampling units. The ecological and sociological status of households was characterised through ecological surveys, questionnaire data and geospatial mapping, and faecal samples were collected from wildlife occurring within the household compound, and livestock and human inhabitants. Escherichia coli was isolated from faecal samples, and characterised both phenotypically (through antimicrobial sensitivity testing) and genetically (through whole genome sequencing). In the first part of this thesis I consider the influence of urban land-use change on the structure of host populations at household interfaces. Using unsupervised machine learning I describe variation in the host composition of wildlife-livestockhuman interfaces and, through multivariate regression analysis, demonstrate that citywide variation in ecological and anthropogenic factors (such as biotic habitat diversity and wealth) drive structural changes in wildlife, livestock and human populations across the urban landscape of Nairobi. Utilising commensal E. coli as an exemplar organism, I proceed to explore epidemiological connectivity between wildlife, livestock and humans at household interfaces, and link epidemiological processes in urban wildlife to their drivers across the urban landscape. Firstly, I explore the epidemiology of clinically relevant antimicrobial resistant (AMR)-E. coli in urban wildlife in Nairobi. Comparing E. coli isolates in wildlife to livestock, humans and the environment, I find that E. coli isolated from wildlife have a lower diversity of resistance phenotypes, and are thus an unlikely source of AMR. At household interfaces, I find evidence of AMR-E. coli exchange between rodents/seed-eating birds, and cattle and humans, and demonstrate that transmission is facilitated through anthropogenic resource provision in households. Next, utilising high resolution sequencing data, I explore the response of microbial communities in wildlife hosts to urban land-use change. Specifically, I test the hypothesis that communities of bacterial mobile genetic elements (MGEs) are deterministically structured, according to changes in host community structure. I show that the diversity of genes encoding virulence and AMR in avian-borne E. coli is determined by variation in the distribution and density of birds, livestock and humans at household interfaces, and that this varies along gradients of urbanisation. To conclude, I relate the findings in this thesis across multiple scales, linking the influence of abiotic factors such as habitat alteration and socioeconomics to host community structure at household interfaces and the epidemiology of wildlife-borne E. coli. Using this framework, I suggest future directions for research on urban disease emergence, and discuss implications of my findings for public health and urban planning.
|
|---|
