We lack a clear understanding of the spread and persistence of plague at both the within population and landscape levels.  Plague is a highly virulent, vector-borne disease caused by the bacterium Yersinia pestis.  Within populations of natural rodent hosts, plague dynamics exhibit both enzootic persistence and epizootic die offs depending on the host species.  At the landscape level, plague persists even when epizootic hosts predominate.  Well-accepted hypotheses suggested that plague transmission within populations occurs mainly via blocked flea transmission, where infectious vectors develop a blockage of the proventriculous formed by Y. pestis that leads to regurgitation of infectious bacteria as the flea feeds.  At the landscape level, dominant hypotheses suggested that plague persistence was due to persistence within enzootic hosts with occasional spillover to epizootic hosts.  We have used a modeling framework that incorporates information about vectors to test these long standing hypotheses against viable alternatives.  First, we developed well-parameterized and validated differential equations model of plague spread within a prairie dog town (an epizootic host) and used sensitivity analysis to investigate the relative importance of transmission from blocked flea vectors, from aerosolized droplets of infectious hosts, and from a short-term infectious reservoir.  Second, we developed a well-parameterized and validated model of plague spread and persistence at the landscape level within a metapopulation of prairie dog towns and used model selection to test the role of prairie dogs and their fleas versus alternative vector hosts in spreading plague at the landscape level.

Results/Conclusions

Our results strongly suggest that blocked flea transmission cannot drive the prairie dog epizootics observed in the field and support the suggestion that transmission from a short-lived infectious reservoir drives prairie dog epizootics.  One possible form of reservoir transmission is “early-phase” transmission by unblocked fleas.  Recent empirical work found “early-phase” transmission rates can be as high as blocked transmission rates, and preliminary results from models that incorporate early-phase transmission explicitly suggest that it may be an important mechanism for initiating epizootics.  At the landscape level, we found that plague could persist in a metapopulation system of prairie dog hosts and their fleas if plague-resistant alternative vector hosts are present, suggesting that enzootic maintenance hosts are not required for landscape level persistence of plague.  Both sets of modeling results underscore the importance of explicitly considering vector biology in hypotheses and models of vector-borne diseases.