What role does disease play in extinction?
Studies of the extinction rate over the past few centuries conclude that it is significantly greater than the background extinction rate, matching or exceeding that during the five previous mass extinctions (Barnosky et al., 2011; Dirzo et al., 2014; Ceballos et al., 2015). Understanding what is shaping extinctions, and how species respond to these pressures, is important if we are to conserve and protect populations under threat. One dramatic example of recent declines can be seen in North American bats. Several species have experienced severe mortality rates since the introduction of a new disease, white-nose syndrome (WNS) in 2006 (USFWS, 2016).
Although sometimes listed as a key causal factor in extinctions (Wilcove et al., 1998), less than 4% of species extinctions have been attributed to disease (Smith et al., 2006). Even in these cases, it is never listed as the single driver of extinction. However, the impact of disease on a population appears to increase when the species is threatened by other factors; or disease can lead to the population becoming endangered due to other factors once driven below a threshold (Heard et al., 2013).
There is some evidence that emerging infectious diseases (EIDs) are appearing at an unusually high rate and are a major cause of concern (Harvell et al., 1999), possibly as a result of interactions with climate change (Epstein, 2001). Further, as global travel for tourism and trade increases, parasites and pathogens are introduced into new areas. The impacts of this have been documented extensively – from extinctions and declines (Smith et al., 2006; Skerrat et al., 2007; Frick et al., 2010) to disruptions in communities and ecosystems (Holdo et al., 2009). Whether populations can adapt to and survive EIDs may be one key factor in shaping the future of biodiversity.
Fungal diseases can be especially damaging, and are responsible for 65% of pathogen-driven extinctions. Plant fungal diseases have long been recognised as a significant threat to food security, and recently pathogenic fungi have also been acknowledged as a danger to wildlife populations (Fisher et al., 2012).
Pathogenic fungi affecting wildlife populations include Nosema species, which cause colony collapse disorder in honeybees, and Fusarium solani, which causes hatch failure and juvenile mortality in loggerhead turtles (Fisher et al., 2012).
One extreme and high profile case of animals coming under threat from a fungal disease is chytridiomycosis, a dermal infection of amphibians caused by the fungus Batrachochtrium dendrobatidis (Bd) (Berger et al., 1998). Bd is thought to have spread globally from Africa, where the African clawed frog, Xenopus laevis, is not susceptible to chytridiomycosis (Weldon et al., 2004). As the infection has spread it has caused mass die-offs in a range of amphibian species around the globe (Skerrat et al., 2007).
This is similar to the spread of WNS in bats in North America. Understanding this disease is crucial both because of the massive global changes to biodiversity currently occurring and because of the key roles bats play in ecosystems. Bats make up about 20% mammals alive today (Tudge, 2000), and are the only mammals capable of powered flight (Hunter, 2007). The service provided to agriculture by bat control of insect pests is estimated to be worth $3.7 billion annually in the USA (Boyles et al., 2011). Therefore loss of bat species could prove devastating not just because of their intrinsic value but also due to direct economic effects.
What is white-nose syndrome?
WNS is a disease of insectivorous cave-hibernating bats caused by the fungus Pseudogymnoascus destructans (Pd) (previously Geomyces destructans) (Gargas et al., 2009; Lorch et al., 2011). It has many parallels with chytridiomycosis. It infects the skin of a vertebrate host and most likely was spread from a region where it is endemic into a new region with naïve hosts, leading to mass mortality. Both fungi can infect a large range of host species.
WNS was first seen in 2006 in Howe’s Cave, New York (Blehert et al., 2009) and has since spread cross much of the USA and Canada (Figure 1). It is named for the distinctive white fungal growth on the exposed epidermal surfaces of infected bats (Figure 2). Since 2006 it has resulted in the deaths of over six million bats in North America (USFWS, 2016). At some hibernacula (hibernating caves) 90-100% mortality rates have been observed (Frick et al., 2010).
Pd is thought to be a novel pathogen from Europe (Warnecke et al., 2012), where it is present but does not cause massive die-offs (Puechmaille et al., 2011; Zukal et al., 2016) indicating bats have co-evolved with the fungus. Asian bats also have much greater resistance to WNS than North American bats (Hoyt et al., 2016). so it is also possible the fungus originated in Asia and has spread through Europe and to America. While it hasn’t been proven, it is likely that tourists or cavers introduced fungal spores from Europe or Asia to America. It is a psychrophilic (cold-loving) fungus that grows between 0˚C and 19.8˚C, and optimally between 12.5˚C and 15.8˚C (Verant et al., 2012). This means that bats are vulnerable when their body temperature drops during winter torpor (Blehert et al., 2009).
The fungus can survive in abiotic reservoirs, outside of a living host, and therefore persist in caves even while they are uninhabited by bats. It has been detected in soil and viable spores have been found in the walls of some hibernacula (Hoyt et al., 2015). This is another similarity with Bd, which can survive in soil and water (Johnson & Speare, 2005). Having an environmental reservoir makes it significantly more likely that a fungus will lead to extinctions (de Castro & Bolker, 2005). The host-generality of Pd may also increase its persistence, as it is protected against declines in a specific host species population, and species that are tolerant (can carry the infection without mortality) may spread it further (de Castro & Bolker, 2005).
What have the effects of WNS been?
What are the effects on individual bats?
There have been several investigations into how Pd causes mortality in bats (Cryan et al., 2010; Lorch et al., 2011; Willis et al., 2011; Reeder et al., 2012; Warnecke et al., 2012). Bats hibernate at roughly the same temperatures required for optimal growth of Pd (Blehert et al., 2009). The immune system is also naturally downregulated in hibernation, providing little resistance to infection (Bouma et al., 2010).
Experimental evidence suggests that Pd must be passed by physical contact from bat to bat, or from the environment to a bat (Lorch et al., 2011). Dense clustering to save energy or reduce water loss during hibernation may increase the rate of infection (Cryan et al., 2010).
Unlike Bd, Pd does not just grow on the skin but can erode and digest the skin, invading tissues (Figure 3). Although called white-nose syndrome, the most damaging impacts of WNS are from fungal growth on and in the wings. Wing membranes are used for gas exchange, blood pressure regulation, thermoregulation and to help to maintain water balance. During torpor, bats are particularly vulnerable to dehydration (Cryan et al., 2010).
Hibernating bats occasionally awake from torpor and return to a euthermic state. Infection appears to cause dehydration and therefore increase the frequency of waking. These arousals are energetically expensive, as they must raise their temperature by producing metabolic heat; so increased arousal causes bats to deplete their energy reserves too soon before the end of winter. This leads to them dying of starvation (Cryan et al., 2010; Willis et al., 2011; Warnecke et al., 2012).
What have the effects on populations been?
Common name Latin binomial name
Little brown bat Myotis lucifugus
Indiana bat Myotis sodalis
Gray bat Myotis griscescens
Northern long-eared bat Myotis septentrionalis
Eastern small-footed bat Myotis leibii
Tricoloured bat Perimyotis subflavus
Big brown bat Eptesicus fuscus
The characteristic lesions of WNS (Fig. 2; Fig. 3) have been recorded on 7 species of bat in North America (Table 1). All affected species co-occur in the same hibernacula but have had vastly different population responses to infection (Langwig et al., 2016). WNS has been responsible for a population crash in the little brown bat, Myotis lucifugus, previously the most abundant North American bat species (Frick et al., 2010). Local extinctions have been seen in all species in Table 1 except the big brown bat, Eptesicus fuscus. The Northern long-eared bat, M. septentrionalis, has suffered the highest rate of local extinctions, and is now extinct at 69% of previous known hibernacula (Frick et al., 2015).
Pd has been detected on five species of North American bat without diagnostic symptoms (USFWS, 2015). This raises the concern that these species are tolerant of the infection and may spread it further to vulnerable species.
There are extremely variable mortality rates in different populations and species affected by WNS. Fungal load appears to be the best predictor of disease impact, with some evidence of a ‘threshold fungal load’ above which the probability of mortality dramatically increases (Langwig et al., 2016). Fungal load appears to be greater in species that hibernate in warmer microclimates. This may have been an adaptive behaviour, decreasing the energy requirements for arousal from torpor throughout winter, before the emergence of WNS, but now have become maladaptive as it increases the fungal growth rate (Langwig et al., 2016).
M. lucifugus suffers from a high degree of evaporative water loss. Bats that are less susceptible to dehydration are less or more variably affected by infection with Pd (Willis et al., 2011). Those with larger bodies or that hibernate in colder and drier hibernacula are more likely to survive WNS (Hayman et al., 2016).
WNS may also have indirect impacts on species that are not affected, or are less affected. Even where it does not lead to extinction, WNS may cause changes in ecology. Data from Fort Drum in New York show that following the emergence of WNS and decline in M. lucifugus activity, spatial and temporal niche partitioning in the bat community was reduced (Jachowski et al., 2014). Colony size of hibernating bats has decreased 10-fold since the emergence of WNS (Frick et al., 2015). They now more closely resemble the colony sizes of related European species, leading to speculation that WNS did, at one time, cause high mortality in European bats.
E. fuscus was originally thought to be vulnerable to WNS due to a 41% population decline found by Turner et al. (2011). However further research has found evidence that E. fuscus is resistant to WNS and these declines may be due to other factors (Frank et al., 2014). They seem to be one of the cave-hibernating bat species most likely to survive WNS and may even become more abundant due to competitive release. In 2009-2010, after the arrival of WNS in Virginia there was a 17% increase in capture rates of E. fuscus (Francl et al., 2012).
The post-WNS success of E. fuscus also raises the question of whether affected species may find themselves outcompeted by resistant species in the future, even if small populations do manage to survive and develop resistance. It may be important to study which ecological niches are or aren’t filled by these species, to know whether unaffected bat species can act as ecological analogues for those that have become locally extinct.
Can these effects be mitigated? How?
Could bats co-evolve with Pd?
The fact that WNS is present in Europe but does not cause massive die-offs suggests that European bat species have coevolved with Pd, preventing it from being lethal. This provides some hope that if North American bats can survive alongside Pd for long enough, they may be able to evolve resistance or tolerance to WNS. Resistance refers to mechanisms that decrease fungal load, while tolerance describes the ability to survive with the pathogen without inhibiting its growth or spread (Roy & Kirchner, 2000).
There is some evidence that bats can survive WNS (Dobony et al., 2011; Meteyer et al., 2011). There are records of M. lucifugus being able to heal the wing damage caused by WNS and still successfully reproduce. However, it has also been shown that WNS infection has impacts on the timing and success of reproduction (Francl et al., 2012). Recovering from WNS is likely to be extremely energetically costly and so have fitness repercussions for those infected.
There is evidence that colonies of M. lucifugus that are persisting following infection with Pd have lower fungal loads (Figure 4) that may be attributed to having evolved some resistance (Langwig et al., 2016b). This resistance does not appear to be from reducing the initial pathogen growth rate, but from growth coming to a stop later in the winter. This may mean there is less selective pressure on Pd to evolve in response (Langwig et al., 2016b). Persisting colonies of M. lucifugus appeared to stabilise 4 years after WNS detection, although at just 2-20% of their original size (Langwig et al., 2012). Genotyping could reveal whether there are differences between those individuals that survive Pd and those that die, or between persisting colonies and epidemic colonies, and therefore whether genes for resistance could spread through the population.
A genetic basis for resistance to WNS could come from the major histocompatability complex (MHC) or other immune system genes generating an immune response. There is evidence that an immune response is generated in response to infection with Pd, even in hibernation (Field et al, 2015). This includes an inflammatory response, although this may actually contribute to WNS mortality by causing arousals from torpor. However antibody-mediated immune response does not appear to be sufficient to explain resistance in persisting American bat populations or in European populations (Johnson et al., 2015). This could suggest that other factors such as behavioural responses are contributing to survival.
The MHC impacts the bacterial flora that grows on the skin, and so could result in differences in fungal suppression by probiotics. The United States Fish and Wildlife Service (USFWS) are currently funding studies into physiological characteristics of bats that have survived Pd infection over several years, and the role of skin secretions and proteins in resistance to Pd (Box 1). Hoyt et al. (2015b) found that six species of Pseudomonas bacteria that naturally occur on the skin of bats are effective at decreasing the growth of Pd in vitro, although it as yet unknown whether this could contribute to differential mortality in infected bats, or be used as a probiotic treatment in vivo.
Another possibility is that heritable differences in hormone levels and sensitivity could produce variation in the response to infection – for example, leptin is a hormone that signals satiation. Bats with less sensitivity to leptin should put on more mass before hibernation, and so be more likely to survive WNS (Willis & Wilcox, 2014). Hormones involved in water balance and in metabolic rate could also be important.
Another theory for why WNS doesn’t cause mass mortality in Europe is that environmental conditions outside hibernacula are different, i.e. if it is warmer outside then bats may be able to survive infection by leaving the cave (Flory et al., 2012). If this is true, then warmer conditions to the south could help bat populations to survive at the margins of the species range.
It is possible that behavioural differences can lead to resistance. E. fuscus, which appears to be resistant to some degree, is not always a cave hibernating species. Although it does co-exist with other affected species in WNS infected caves, it also utilises hibernacula other than caves (Whitaker & Gummer, 1992). It would be revealing to study whether E. fuscus are utilising alternative hibernacula at greater rates since the emergence of WNS and therefore avoiding infection due to behavioural plasticity. They are also more likely to hibernate alone or in small groups (Whitaker & Gummer, 1992) reducing the spread of infection.
HIgher proportions of both M. lucifugus and M. sodalis started roosting individually after WNS detection (Figure 5). M. sodalis has overall remained more social, and WNS mortality rates have ameliorated less in M. sodalis than M. lucifugus (Langwig et al., 2012). This further suggests that behavioural plasticity has a role in resistance.
Could anthropogenic interventions help to prevent extinctions?
Modelling suggests generalised culling of bat species is likely to be ineffective in preventing the spread of WNS (Hallam & McCraken, 2011). However, targeted culling of infected bats could be more effective, especially in combination with measures to reduce the carrying capacity of hibernacula for Pd (Meyer et al., 2016). A suggested mechanism for targeted culling is ‘poisoned thermal traps’ to attract infected bats when they arouse from torpor. Reducing the carrying capacity for Pd could be achieved by removing sediment in which it can grow. While this was effective in Meyer et al.’s model, they assume traps could target infected bats with complete accuracy. The plausibility of this method is far from guaranteed.
One idea that has received attention is localised artificial warming in hibernacula, creating ‘thermal refugia’ (Boyles & Willis, 2010). In much the same way as it is theorised warmer conditions outside hibernacula could help bats to survive, it has been suggested that if bats roused from torpor by WNS could migrate to a warmer area of the cave, it would lower the energy demands of arousal and allow them to survive the winter. Since cold temperatures are required for hibernation these ‘refugia’ would have to be able to exist without increasing the temperature in the hibernaculum as a whole (Boyles & Willis, 2011). Modelling of this system predicted a decrease in WNS mortality from 81.9% to 63.0% if frequency of arousal bouts is the factor increased by WNS infection, and even greater benefits if the length of arousal periods increases.
However, the work done by Langwig and colleagues (2016) suggests that warmer microclimates may increase fungal load and therefore increase mortality. They recommend instead that the access to warmer, more humid parts of hibernacula should be restricted. Further experimental work on the costs and benefits of increased temperature to WNS infected bats should be carried out before enacting either of these suggestions.
There has been research into using biocontrol to treat WNS. The bacterium Rhodococcus rhodochrous can inhibit the growth of some fungi, including Pd (Cornelison et al., 2014). 75 bats that were infected with WNS were treated using volatile organic compounds (VOCs) from R. rhodochrous and rereleased into the wild in Missouri in May 2015 (The Nature Conservancy, 2015). While possible under controlled conditions, it is difficult to know if this method could be used on a large scale in completely wild hibernacula. Biocontrol is very controversial and can be dangerous. This method may lead to extinctions of resident fungal species.
Obviously it is crucial to take extreme caution when attempting to manipulate wild populations, creating the need for modelling and experimental studies. There is also a concern that increasing survival rates of WNS could increase the spread of the disease and have an overall negative effect. There is ongoing research into a number of treatment and management options (Box 1), and hopefully significant progress will be made with some of these while it can still be helpful to North American bat populations.
Finally, population declines due to WNS should not be considered in isolation. There are other threats to bat species in North America, including land management (USFWS, 2007), deforestation (USFWS, 2007), and wind turbines (Arnett et al., 2008). Given the additional extinction risk posed by disease when it occurs in conjunction with other threats (Smith et al., 2006; Heard et al., 2013) we should focus on providing protection from other pressures alongside WNS.
Will WNS lead to the extinction of some North American bats?
Ecological and climate traits both appear to have a strong impact on the vulnerability of bats to WNS. The extreme variation in mortality rates suggests other factors such as behavioural plasticity and possibly even the skin microbiome may also have an affect. For this reason, further research is needed to truly understand the complexities of species response to WNS. However, it is possible to make some predictions.
Local extirpation of M. lucifugus due to WNS is anticipated by 2026. Although there has been evidence of resistance in surviving colonies of this species, even if the disease ameliorates with time from infection the chance of extinction was estimated to be greater than 90% by 2075 (Frick et al., 2010). Modelling predicts that a full recovery of the species is impossible (Russell et al., 2015). The temperature and relative humidity that M. lucifugus requires for water balance perfectly matches the growing conditions for Pd (Hayman et al., 2016), and this species suffers from a high degree of evaporative water loss during hibernation, making it extremely vulnerable. Genetic analyses show there is little barrier to gene flow across the species range, although western populations show less connectivity than eastern populations, which may alter the disease dynamics west of the Rocky Mountains (Vonhof et al., 2015).
Other species that appear to be at extreme risk of local extirpation are M. sodalis, which was already under threat before the emergence of WNS (Thogmartin et al., 2013), making it more vulnerable to extinction; and M. septentrionalis, which has been lost at 69% of its prior hibernacula already (Frick et al., 2015).
It seems very likely that at the very least WNS will cause regional extinctions of M. lucifugus, M. sodalis, and possibly some other Myotis species. If small populations of these, and other affected species, can survive at the warmer south margins of their range, or due to behavioural adaptations or anthropogenic interventions, then they may be able to evolve resistance and eventually co-evolve with Pd to the point where it does not cause mass mortality. They may then be able to recolonize colder, drier areas where they have become regionally extinct. However there is a possibility that unaffected or resistant bat species such as E. fuscus will become dominant in these regions.
More experimental work needs to be done on mitigation methods for WNS, in the hope that we can find safe and effective ways to treat or prevent infection in the wild in the near future. We should try to find the basis for resistance in European and Asian bats, as well as those colonies persisting in North America. This could be approached with both studies into genotyping and behavioural plasticity, since resistance may come from either or both of these elements.
Given the host-generality of Pd and its ability to survive and persist in the environment outside of a host, eradication seems very unlikely. Bat species will only survive in the long term if they are able to adapt to the fungus. This may lead to some long-term changes in the ecology of species, including colony size, sociality, and temperature of hibernation.