Authors

A Fenton²; S M Withenshaw²; A Pedersen¹;  ¹ Institute of Evolutionary Biology and Centre for Immunology, Infection and Evolution, University of Edinburgh, UK;  ² Institute of Integrative Biology, University of Liverpool, UK

Discussion

Vector-borne pathogens (VBPs) are a major source of disease, suffering and economic loss. The epidemiology of VBPs can be complex as transmission arises from interactions between the pathogen, and potentially many vector and definitive host species. Furthermore, while many vector and pathogen species can infect multiple host species in the host community, their abilities to infect those different hosts can vary between the pathogen and vectors. Given these complexities, well-informed theory is needed to understand VBP dynamics in multihost-multivector-multipathogen systems. However, data on how different host species contribute to pathogen transmission are needed to inform these models, but are often lacking. Most studies of VBP dynamics in natural systems rely on purely observational data, which can reveal patterns of pathogen &hypen; vector &hypen; host co-associations, but cannot show definitively the roles that different species play in pathogen transmission. Hence, mathematical models of VBP transmission in multi-host communities remain poorly informed.
We present results from a large-scale experiment conducted on a natural multihost-multipathogen-multivector system, which we use to inform a mathematical model to infer rates of within- and between-species transmission. Our study focuses on woodland small mammal communities dominated by wood mice (Apodemus sylvaticus) and bank voles (Myodes glareolus). Each species hosts multiple flea species, which vector a range of apparently host-specialist and host-generalist pathogens (primarily Trypanosoma and Bartonella). We used targeted treatments with the insecticide Fipronil to interfere with vector (and hence VBP) transmission from one or other host species, and monitored changes in VBP prevalence in the drug-target and non-target host species. We then used these data to fit epidemiological models to infer differential rates of host use by the fleas, and the consequences for within-and between-species transmission. We show that cross-species transmission is much rarer than suggested by observational data, such that each host species has a relatively host-specific flea population, and even when vector sharing does occur, host-pathogen compatibility barriers further restrict cross-species transmission. Overall this combination of experimental perturbation and tailored mathematical models provides a powerful means to infer pathways of within- and between host species transmission that would not be possible with observational data alone