Quantifying host species contributions to pathogen transmission in a multihost community: the case of chytrid fungus in amphibian communities

Many pathogens of global health and conservation concern infect multiple host species. Ebola is a classic example, circulating naturally within a 'reservoir' host community, and with the potential to jump across to another host species with devastating effect. Clearly it is vital to understand how such pathogens are maintained in their host communities, and which species play a major role in spreading those pathogens. However, obtaining that understanding is notoriously hard. We have recently developed a mathematical approach to measure a host community's ability to maintain a pathogen, and identify 'key hosts' that drive pathogen spread. Importantly, unlike previous methods, this approach can be parameterised using relatively coarse-grained, easily-collected data (standard measures of host abundance and infection occurrence). We will provide the first rigorous test of this model, applying it to a natural 'multihost'-pathogen system of major conservation concern: chytrid fungus ('Bd') in amphibian communities. Bd is a major cause of amphibian declines worldwide, but we don't understand how it spreads through or is maintained by amphibian communities. We will apply our mathematical approach to historical and new data, and use it to identify those key hosts, and predict the effect of removing them. Crucially, we will then directly test those predictions by carrying out species removal experiments. Overall this will provide a rigorous test of our mathematical tool, show how host communities affect pathogen spread in general, and provide specific guidelines for the management of Bd in particular.

Chytridiomycosis affects wildlife communities globally and has devastated amphibian communities on four continents. There are major efforts to prevent its spread and manage its impact, but in every case the focus is on single host species conservation. This is clearly inadequate, as in every system where conservation efforts are underway the single host exists within a host community where potential reservoir species have been identified. By developing the analytical tool that allows the identification of key reservoir species and predicts the impact of removal, we hope to shift efforts to conserve single species threatened by chytridiomycosis to efforts to manage amphibian communities so that they may coexist with infection. Many relevant amphibian communities are found in structured habitats where management strategies that our method would recommend are possible. For example, across Europe and in North America threatened host populations are predominantly within high elevation communities where breeding habitat is spatially segregated and reproduction is broadly temporally predictable for component species. The fact that our study site is the index site where the first case of Bd was recorded in Europe - and where it is still continuing to circulate - is an example of this and makes our study particularly relevant, and important, for those working on the management of this disease. This group of stakeholders is not restricted to academics: field management of amphibian communities falls under the remit of government agencies, park managers, private landowners and NGOs committed to wildlife conservation. As well, current efforts to conserve amphibians threatened by chytridiomycosis overwhelmingly include a captive assurance component, where species threatened with extinction due to Bd are kept and bred in captivity until such time the threat has been mitigated. Because of limited infrastructure to support these efforts, amphibians are housed in mixed species facilities, where despite the best efforts at maintaining biosecurity, outbreaks of disease do occur. Our analytical approach can be applied to these settings, and can be used to guide how the various stakeholder facilities share responsibilities for captive assurance and structure their collections to minimize the possibility of outbreaks.
More broadly, our work will also be highly relevant to policy makers and those involved in the practical management of multihost diseases in general. Given that most emerging diseases involve multiple host species, it is clearly vital to understand the processes by which diseases circulate within and impact upon natural host communities. The majority of infectious diseases that are affecting livestock, crops and humans spill over from wildlife hosts and are maintained in wildlife communities. Host species responsible for spill over events are frequently identified but unlikely to be eliminated from contacts with susceptible spill over hosts (eg camels and MERS). What is instead needed is a toolkit that allows the identification of host community structures that pose the greatest risk for propagating spill over events. By using our system to develop and validate our generalised theoretical framework, this study has the potential to produce an invaluable tool for quantifying host contributions to transmission that can easily be applied by managers to other contexts and communities. As such, our study should influence the efforts of those who work to mitigate a wide range of multi-host diseases, by providing a validated framework with which to quantify the processes that underpin pathogen circulation and impact within natural communities.