PS 6-82 - Scaling up the stress-induced susceptibility hypothesis: an experimental test in the Ranavirus-wood frog system

Abstract: The stress-induced susceptibility hypothesis posits that chronic exposure to stressors—physiologically challenging conditions—is immunosuppressive, making infections more likely or more severe. Stressors can lead to homeostatic imbalance, can reduce or divert energetic or material resources need to mount immune responses, or can lead to direct immunosuppression (e.g., via the effects of glucocorticoids). The assumption is that these individual-level effects scale up to population-level outcomes; that is, stressors are thought to make epidemics more likely and more severe, too. However, there are few empirical tests of this assumption, and reason to doubt its general validity (e.g., if stressors make infections more lethal, this leads to a shorter window of time in which to transmit the infection, and presumably smaller epidemics).

We tested the effects of elevated salinity and elevated temperature, factorially, on the likelihood, speed, and severity of Ranavirus (Rv) epidemics initiated by introducing a pair of Rv-infected tadpoles into each of 96 replicate populations of larval wood frogs (Lithobates sylvaticus) in mesocosms (32 Rv-exposed + 8 controls per treatment). Saline waters, such as those cause by the runoff of deicing salts into roadside ponds, are energetic stressors of amphibian larvae and cause more intense glucocorticoid responses to Rv infections as well as more intense and lethal infections. We thus expected populations in the elevated salinity treatments would be more likely to experience an epidemic, given Rv introduction, and for those epidemics to be more lethal. We also expected that populations in the elevated temperature treatments (3°C above ambient, achieved with aquarium heaters regulated by Arduino microcontrollers) would be faster given that viral replication, as well as host immune responses, are temperature dependent.

We found that neither the likelihood or severity of the replicate epidemics varied noticeably by treatment: epidemics were observed in all but two of the 96 mesocosms with Rv introduced and mortality averaged over 95% in all treatments. Elevated temperatures, however, did speed up epidemics by several days. While the speed and severity of the epidemics likely reflect experimental conditions (e.g., necessarily small and simple habitats), making it difficult to directly extrapolate to wild populations, our results strongly suggest that worse outcomes among individual animals do not translate simply or clearly to worse population-level outcomes.