Discussion
Our results demonstrate that Bd strains are typically more successful at infecting and replicating, and thus causing mortality, in local than novel hosts, supporting local adaptation or the hypothesis that pathogens are best adapted to take advantage of their local hosts (Lively & Jokela 1996; Kaltz & Shykoff 1998; Lively & Dybdahl 2000; Morran et al. 2011; Strauss et al. 2012; Lymbery et al. 2014; Urban et al. 2020; Johnson et al. 2021). However, the rise of catastrophic modern pandemics from pathogen pollution has led to speculation that introduced pathogens are especially devastating to naïve hosts (Carey et al. 1999; Cunningham et al. 2003; Anderson et al. 2004; Mastitskyet al. 2010). While we did find substantial variation among host-pathogen outcomes, there was no evidence that novel host-pathogen interactions generally cause especially high mortality. Thus, the major concern with pathogen pollution is not that naïve hosts are especially vulnerable but that increasing introductions of novel pathogens increases the chance that virulent pathogens will encounter highly vulnerable hosts. This perception that novel and introduced pathogens are especially deadly might be because of detection bias – most introduced pathogens might go undetected because low pathogen fitness in novel host populations reduces the likelihood of establishment (Torchinet al. 2003; Torchin & Mitchell 2004). However, there are >8,000 species of amphibians and a multitude of Bd strains and thus the full strength of local adaptation and the scope of variation in host-strain outcomes remains unclear in this system.
We found significant variation in mortality, infection success, and pathogen load among Bd strains (Figure 3 & S2), consistent with other studies showing variation in the traits of Bd strains within the global panzootic lineage (Lambertini et al. 2016; Becker et al.2017). Specifically, when averaging across host species, Bd from Louisiana was most deadly, most likely to cause infection, and produced the highest infection burdens (Figure 3 & S2). We also found significant variation in mortality, prevalence, and pathogen load among host species (Figures 3 & S2). Specifically, when averaging across Bd strains, toads from Arizona and Tennessee had the highest mortality and infection prevalence (Figure S2). Rapid mortality occurred in our experiment when highly susceptible hosts were exposed to especially virulent Bd strains. Specifically, toads from Arizona and Tennessee died an average of 29 days sooner when exposed to Bd from Louisiana or Ohio relative to any other novel host-pathogen interactions (Figures 3B & S3). Further, our meta-analysis revealed that Bufonoidea species (toads and relatives) were especially susceptible to Bd-induced mortality compared to other host taxa, confirming previous research (Scheeleet al. 2019a) (Figure 3 & S2).
By combining a continental-scale factorial experiment with a global-scale meta-analysis, we show that local host-pathogen interactions typically resulted in higher host mortality, greater infection success, and higher pathogen loads. However, we also found that novel host-pathogen interactions do not always result in low mortality, infection success, and pathogen loads. There was substantial variation in novel host-pathogen outcomes, including pairings of especially vulnerable hosts and especially deadly pathogen strains that resulted in extremely rapid mortality. In fact, the host- and strain-level effects accounted for more variation in our models than strain novelty. Therefore, frequent introduction of novel strains increases the risk of especially vulnerable hosts encountering especially deadly pathogens despite pathogens being better adapted to invade and replicate in local hosts. Therefore, we provide support for both the local adaptation and naïve host syndrome hypotheses, highlight how the two hypotheses are complementary rather than conflicting, and emphasize the need for greater integration of these hypotheses and their associated semi-disparate literature. Stochastic encounters between deadly pathogen strains and vulnerable hosts will continue to rise as human connectivity and encroachment on and degradation of wildlife habitat increase. Thus, it is imperative that policy makers mitigate the risk of pathogen spillover and dispersal (Altmann & Kolby 2017; Aguirreet al. 2021). Policies regarding surveillance, biosafety, and security of wildlife and livestock trade that utilize One Health approaches are especially promising for reducing risk (Cunninghamet al. 2017; Aguirre et al. 2021).