Drivers of variation in species impacts for a multi-host fungal disease of bats

Abstract

Disease can play an important role in structuring species communities because the effects of disease vary among hosts; some species are driven towards extinction, while others suffer relatively little impact. Why disease impacts vary among host species remains poorly understood for most multi-host pathogens, and factors allowing less-susceptible species to persist could be useful in conserving highly affected species. White-nose syndrome (WNS), an emerging fungal disease of bats, has decimated some species while sympatric and closely related species have experienced little effect. We analysed data on infection prevalence, fungal loads and environmental factors to determine how variation in infection among sympatric host species influenced the severity of WNS population impacts. Intense transmission resulted in almost uniformly high prevalence in all species. By contrast, fungal loads varied over 3 orders of magnitude among species, and explained 98% of the variation among species in disease impacts. Fungal loads increased with hibernating roosting temperatures, with bats roosting at warmer temperatures having higher fungal loads and suffering greater WNS impacts. We also found evidence of a threshold fungal load, above which the probability of mortality may increase sharply, and this threshold was similar for multiple species. This study demonstrates how differences in behavioural traits among species-in this case microclimate preferences-that may have been previously adaptive can be deleterious after the introduction of a new pathogen. Management to reduce pathogen loads rather than exposure may be an effective way of reducing disease impact and preventing species extinctions.This article is part of the themed issue 'Tackling emerging fungal threats to animal health, food security and ecosystem resilience'.

Keywords: Geomyces destructans; Myotis lucifugus; emerging infectious disease; multi-host pathogen; white-nose syndrome; wildlife disease.

Figure 1.

The distribution of white-nose syndrome (hatched) as of March 2016, and distribution maps for six hibernating bat species (colours) in the USA and Canada. Inset: Counts of hibernating bats at sites from 1979 to 2011, with per cent declines in the first year of WNS detection indicated in the upper left of the graph.

Figure 2.

Impact of white-nose syndrome, measured as the change in population growth rates (λ) following the detection of WNS (based on data and analyses in [12]) and infection prevalence (a,c) and intensity (b,d) of P. destructans for six bat species. Error bars show ±s.e. of the predicted mean. Lines show significant relationships, and the dashed line in (a) indicates that this relationship was significant only in a univariate regression, but not significant in a model with early or late fungal loads. Equations in the figure report results from phylogenetic regression in the format slope ± s.e. of the slope plus the intercept. Late hibernation prevalence was not a significant predictor of WNS impacts (intercept: −0.22, slope: 0.92 ± 0.5, p = 0.14, R² = 0.45).

Figure 3.

Estimated P. destructans loads in late hibernation and hibernation roosting temperature for six bat species. Species' average load and roosting temperature were restricted to individual bats in which both roost temperatures and load were collected (n = 419). Statistics in the figure report results from phylogenetic regression in the format slope ± s.e. of the slope plus the intercept.

Figure 4.

Distributions and skewness of fungal loads on bats in early (solid line) and late hibernation (dashed line). Early and late hibernation skewness differed significantly for little browns (p = 0.03), tri-colored (p < 0.0001) and big brown bats (p = 0.004). Skewness differences were compared by calculating skewness of simulated load distributions from 9999 draws from the grouped winter load data by species, and then comparing the bootstrapped skew differences with our observed skew differences to assess significance.

Figure 5.

Change in P. destructans fungal loads over time plotted against fungal loads at the beginning of hibernation for five species of bats in 19 colonies (where early hibernation prevalence was greater than 50%). Statistical results are shown in electronic supplementary material, table S3.