SEARCH

SEARCH BY CITATION

Keywords:

  • Amphibian chytrid fungus;
  • Batrachochytrium dendrobatidis;
  • Infection Threshold;
  • Host Species

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Study methods
  5. Data analysis
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

Introduced pathogens are increasingly being implicated in population declines and their effects are difficult to manage. In the absence of methods to eradicate pathogens acting as threatening processes, intervention before population decline is necessary. Such an intervention requires an ability to predict when population declines will occur, and therefore, an understanding of when exposure will lead to infection, disease, death and population decline. This study investigates when pathogen exposure leads to disease for the amphibian chytrid fungus Batrachochytrium dendrobatidis, which has been implicated as a causal agent in the global amphibian decline. Susceptibility studies were conducted on two anuran species, the green and golden bell frog Litoria aurea and the striped marsh frog Limnodynastes peronii, when exposed to the fungus as either tadpoles or juveniles. Host species was found to significantly affect the outcome of exposure, with infection loads in L. aurea increasing over time and resulting in significantly lower survival rates than unexposed. By comparison, infection loads in L. peronii remained the same or decreased over time following the initial infection, and survival rates were no different whether exposed to B. dendrobatidis or not. These outcomes were independent of the life stage at exposure. Individuals with higher infection loads were not found to have lower survival rates; rather, an infection load threshold was identified where individuals with infection loads that crossed this threshold had high likelihoods of showing terminal signs of chytridiomycosis. Therefore, host species determined whether infection load crossed this threshold and the crossing of the threshold determined the incidence of disease and survival. The quantification of infection load thresholds for survival, along with the time it takes to reach them, will enable infection loads in wild populations to be related to the likelihood of disease and is the first step in the understanding and prediction of when exposure will result in population decline.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Study methods
  5. Data analysis
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

Pathogens of wildlife were historically believed to have little impact on host populations (Hassell et al., 1982; May, 1988; Scott, 1988; Daszak & Cunningham, 1998) but epidemiological studies (Keymer, 1981; Anderson & Crombie, 1984; Scott & Anderson, 1984; Scott, 1987) and theoretical modelling (Anderson & May, 1978, 1980) conducted throughout the 1970s and 1980s have shown that disease can suppress host population growth rate. In situations where transmission is density dependent, disease can cause cyclical fluctuations in population size and can play an important role in the regulation of host population dynamics and community structure (Scott, 1988; Thompkins & Begon, 1999). In recent decades, there has been an increased frequency of emerging infectious diseases of wildlife, which has resulted in rapid population declines and extinctions (McCallum & Dobson, 1995; Daszak, 1999a,b). This emergence has primarily been attributed to anthropogenically induced changes in the ecology of the host and/or pathogen. The loss and fragmentation of habitat have resulted in higher host densities and transmission rates, while environmental modifications have increased habitat suitability for pathogen growth and survival (Daszak, 1999a,b). The introduction of novel pathogens into naïve host populations through spill over from sympatric domesticated animals or the translocation of pathogens into new geographic ranges have also resulted in the emergence of infectious diseases (Daszak, 1999a,b). One of the most dramatic disease-induced losses of biodiversity in recent history has been the global amphibian decline.

Over 30% of the world's recognized amphibian species are currently classified by the IUCN as either extinct or threatened with extinction (Stuart et al., 2004; IUCN, 2010). Four hundred and thirty-five amphibian species are classified as rapidly declining and the cause(s) of decline in 207 of these species could not be attributed to habitat alteration/reduction or overexploitation (Stuart et al., 2004). Instead, these species declines may be due to the introduction of a novel pathogen, the amphibian chytrid fungus Batrachochytrium dendrobatidis (Skerratt et al., 2007). The conservation measures available for species in decline due to infectious disease are limited and emphasis must be placed on understanding disease dynamics within a population. The progression of pathogen exposure to infection, disease, death and population decline can be influenced by a range of factors. Identifying these factors and their effects under different situations will allow the outcome of pathogen exposure to be predicted so that informed decisions about the vulnerability of populations to extinction can be made and management actions can be prioritized. The first step in this process is to understand when exposure to a pathogen leads to the development of disease.

Batrachochytrium dendrobatidis infects keratinized epithelium in the outer epidermal layers of post-metamorphic amphibians (Berger et al., 1998). Infection can result in the fatal disease chytridiomycosis through an impairment of the skin's osmoregulatory function causing a reduction in electrolyte levels and circulatory collapse (Voyles et al., 2007, 2009). Batrachochytrium dendrobatidis also infects the keratinized mouthparts of tadpoles, which has been hypothesized to impede feeding ability, causing slower development rates (Parris & Baud, 2004). Mortality of tadpoles following exposure to low doses of B. dendrobatidis in the absence of detectable infection suggests that infection may be prevented or inhibited, but at a cost that reduces the probability of survival (Garner et al., 2009). Batrachochytrium dendrobatidis has a broad host range (Speare & Berger, 2004) but does not cause disease or death in every species that it infects. The survival rate of individuals exposed to B. dendrobatidis differ with host species (Parker et al., 2002; Davidson et al., 2003; Daszak, 2004; Blaustein et al., 2005; Woodhams et al., 2007) and life-history stage at exposure (Lamirande & Nichols, 2002; Carey et al., 2006) and research has focused on identifying the host traits causing these differences (Rollins-Smith et al., 2002b; Harris et al., 2006; Rowley & Alford, 2007; Woodhams et al., 2007). However, the effects of host species and life stage on the steps leading from exposure to disease have not been investigated. This study investigates the relationships between host species and life stage on the likelihood of becoming infected, infection load over time and the development of disease in tadpoles and juveniles of two anuran species with differing decline histories.

This study is also the first to investigate whether B. dendrobatidis could have been a causal agent in the decline of the green and golden bell frog Litoria aurea, an endangered Australian anuran. Litoria aurea is a large basking tree frog that was once common throughout eastern NSW and north-eastern Victoria but has declined from over 90% of its range since the 1970s (White & Pyke, 1996). This species is now considered endangered in NSW under the Threatened Species Conservation Act 1995 and vulnerable nationally under the Environmental Protection and Biodiversity Conservation Act 1999. Batrachochytrium dendrobatidis has been implicated as a contributing factor in this range contraction and has been associated with population declines and disappearances (Penman et al., 2008; Stockwell et al., 2008). The striped marsh frog Limnodynastes peronii is a common ground-dwelling species that co-occurs with L. aurea throughout much of its range but has not undergone a decline (Pyke & White, 1996). Despite the different decline histories of these two species, they are similar in their ecologies, both being generalist species that are opportunistic lentic breeders (Pyke & White, 2001; Hamer, Lane & Mahony, 2007). Identifying differences in the outcome of exposure to B. dendrobatidis in these species at different life stages will contribute to understanding when exposure leads to disease and will provide the building blocks for predicting population declines.

Study methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Study methods
  5. Data analysis
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

Two experiments were conducted simultaneously, one that exposed tadpoles (Gosner stage 34–37; Gosner, 1960) to B. dendrobatidis and one that exposed juveniles 2–3 weeks post-metamorphosis. Experimental animals were bred in captivity using four adult breeding pairs (two L. aurea clutches produced 4 weeks apart and two L. peronii clutches produced 5 weeks apart) making individuals of each species in each experiment full siblings. Full siblings were utilized in each experiment because of constraints on the timing of breeding. However, this meant that both species and clutch effects were confounded. Despite the use of different cohorts for different experiments, the results obtained across cohorts were consistent (see ‘Results’), implying that the differences measured represent species-level differences.

Both experiments consisted of 38 replicate tanks and species were paired, either as tadpoles (Experiment A) or juveniles (Experiment B), in each tank. Each experimental tank was a 15 L plastic aquarium with a gravel substrate sloping into 2 L of 20% Holtfreters solution as a water source (Wright & Whitaker, 2001) and three pieces of autoclaved eucalyptus bark for shelter. Tanks were positioned randomly on shelves within a constant temperature room that was maintained at 22 °C under an automated 12 h light/12 h dark regime. Water changes were conducted fortnightly by the removal and replacement of 1 L of 20% Holtfreters solution. Waste water was disinfected through the addition of bleach to total 2% sodium hypochlorite (Johnson et al., 2003). Tadpoles in Experiment A were fed one trout pellet twice a week until metamorphosis. Once one tadpole in a tank metamorphosed one trout pellet was added to the tank once a week and four small crickets were added to the tank twice a week. Once both tadpoles in a tank had metamorphosed, eight small crickets were added to the tank twice a week. In Experiment B, the two juveniles in each tank were provided with eight small crickets twice a week. Strict hygiene protocols were followed when conducting water changes or feeding.

Water in half of the tanks in each experiment was inoculated with 2 mL of B. dendrobatidis suspension (strain: Gibbo River-Llesueuri-00-LB-1) obtained by flooding actively growing TGhL agar plates (Johnson, 2003) with sterile distilled water. The remaining 19 tanks in each experiment were negative controls and were inoculated with a sham suspension obtained by flooding sterile TGhL agar plates with sterile distilled water. Following inoculation, animals in both experiments were observed daily to monitor for signs of chytridiomycosis and humane endpoints were determined for both tadpoles and post-metamorphic life stages. To monitor body condition in tadpoles, the width of the body across the abdomen was observed relative to the width behind the eyes as this decreases with food limitation (M. P. Stockwell, pers. obs.). If the abdominal width was observed to be less than the width of the body behind the eyes, the tadpole was removed from the water and mouthparts were inspected for signs of degradation. Where more than 75% of the tooth rows and jaw sheaths were missing, the animal was euthanized for ethical reasons by immersion in 0.4% tricaine methanesulphonate (MS-222) solution.

In post-metamorphic individuals, the signs of chytridiomycosis that were monitored for included lethargy and poor righting reflex. Lethargy was identified as an absence of hunting behaviour when crickets were provided or unresponsiveness to the opening of the tank and lifting of bark during inspections (healthy animals will adopt a cringing posture, back into a corner or hop when disturbed). Where lethargic behaviour was observed, the strength of the individuals righting reflex was determined. For individuals with terminal chytridiomycosis, a slow righting reflex is indicative that death will occur within the next 48 h (Berger et al., 2005). Healthy L. aurea juveniles will right themselves in 1–4 s for up to four turns and L. peronii juveniles will right themselves in 1–2 s for up to four turns (M. P. Stockwell, pers. obs.). Therefore, where the animal was unable to right itself or where the time taken to right itself was greater than double the maximum turning time for that species (i.e. 8 s for L. aurea and 4 s for L. peronii), the animal was observed closely for the next 12 h and the righting reflex was tested after 6 and then 12 h. If there was no change or worsening of the rate of righting, the animal was euthanized for ethical reasons, by immersion in 0.4% tricaine methanesulphonate (MS-222) solution.

The infection status of animals in each experiment was determined by swabbing at 2 weeks, 3 and 6 months after inoculation. Swabbing methods differed for tadpoles and post-metamorphic animals because of the distribution of keratin, and therefore, the site of B. dendrobatidis infections in the two life stages (Berger et al., 1998; Marantelli et al., 2004). Tadpoles were swabbed by gently wiping the mouth parts (jaw sheaths and tooth rows) 10 times each with sterile fine-tipped swabs. Metamorphs and juveniles were swabbed by wiping the left and right sides of the animal's ventral surface and the inner and outer thighs eight times each, as well as the ventral surface of the fore and hind feet twice. In Experiment A, the metamorphosis of tadpoles during the experiment and the different swabbing techniques made it impossible to compare tadpole and metamorph infection levels. Therefore, data for tadpoles and metamorphs were analysed separately.

If an animal was euthanized between the scheduled swabbing time points, it was swabbed immediately after death, along with the other individual sharing the same tank to allow comparisons between species over identical time intervals. Similarly, when the individuals in one tank were swabbed, so too were those in a randomly selected tank from the other group (i.e. control or treatment) to allow comparisons of infection load between groups over equivalent time intervals. Additional swabs were also streaked 10 times through the water of each tank of Experiment A 2 weeks after inoculation and before the first water change. Quantitative PCR (qPCR) values derived from these swabs were used to set the background environmental level of B. dendrobatidis in the water. Twice the maximum background level of B. dendrobatidis detected in any tank from Experiment A was considered to be the confirmed infection level, below which a positive qPCR from any experimental animal was a false positive.

Extraction and quantification of B. dendrobatidis on swabs were performed following standard protocols for a qPCR Taqman assay (Boyle et al., 2004) using a Rotor Gene 6000 real time DNA amplification system (Corbett Life Science, Sydney, Australia). Each swab was analysed in triplicate and where amplification of B. dendrobatidis DNA occurred in all three replicates, the number of genomic equivalents (GE) detected at a standardized cycle threshold was summarized as the geometric mean. Where an equivocal result occurred with amplification in less than three of the replicates and the negative template control revealed no contamination, the sample was considered positive for the presence of B. dendrobatidis and the geometric mean of all three replicates was calculated. The zero values were included in this calculation as it was assumed to be the result of a low quantity of DNA in the sample. Sampling probability dictates that if the density on the original swab was low, the probability of it being aliquoted into all three replicates would also be low (Hyatt et al., 2007).

Where amplification did not occur in any of the replicates, the sample was considered negative for the presence of B. dendrobatidis, provided the PCR reaction was not inhibited. To detect inhibition within the reactions, internal positive controls were included in one replicate of each sample and in the negative B. dendrobatidis template control. Following qPCR, the cycle number 5 that crossed a threshold set midway up the amplification curve were compared. If the sample crossed the threshold more than five cycles after the negative B. dendrobatidis template control, the sample was considered inhibited. Where inhibition was detected, a 1/100 dilution of the originally extracted DNA was prepared to dilute inhibitory agents and the reaction was repeated. In samples determined to be positive, the geometric mean number of GE was multiplied by 10 to account for a dilution step in the extraction process, or was multiplied by 100 if a 1/100 dilution was conducted to remove inhibition. This mean number of GE reflects a relative measure of the infection load in each individual, provided that it was greater than the environmental levels detected by swabbing the water column.

Data analysis

  1. Top of page
  2. Abstract
  3. Introduction
  4. Study methods
  5. Data analysis
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

All data analyses were conducted separately for each experiment. The prevalence of B. dendrobatidis infection in each species was calculated as the proportion of individuals found to be positive. For each individual, the infection status data were binary (infected or not infected) and thus were compared between species using binary logistic regression. The infection load data were normalized with logarithmic transformations and its relationship with host species and swabbing event (time) was investigated using linear mixed models. Models were constructed with species nested within tank included as a random effect and swabbing event included as a repeated measure. The effect of B. dendrobatidis exposure on survival rate was determined for each species using log rank Mantel–Cox tests in Kaplan–Meier survival analysis (Hosmer & Lemeshow, 1999). The relationship between the number of days until death and infection load at swabbing event 2 was investigated using Cox regression models for censored survival data (Hosmer & Lemeshow, 1999).

Receiver operating characteristic (ROC) curves were used to investigate whether individuals exposed to B. dendrobatidis could be classified into two groups, those that showed signs of chytridiomycosis and were euthanized and those that did not show signs and survived, based on their infection load at swabbing event 2. ROC curves plot the proportion of true positives (i.e. the proportion of individuals correctly classified as being euthanized based on infection load, also called the sensitivity) against the proportion of false positives (i.e. the proportion of individuals incorrectly classified as being euthanized based on infection load, also called 1−specificity where specificity is the proportion of true negatives) as the discrimination threshold (i.e. the infection load threshold) changes (Zou, O'Malley & Mauri, 2007). The area under the ROC curve (AUC) was used as a measure of how dissimilar the two groups were in their infection loads, where an AUC close to 1 represents a high level of dissimilarity. The infection load data at swabbing event 2 were used in Cox regression models and ROC curves as 100% of individuals exposed to B. dendrobatidis were infected at that time point. Infection load at swabbing event 3 was not used as all infected L. aurea had been euthanized (see ‘Results’).

The inclusion of equivocal results from qPCR as positives for the presence of B. dendrobatidis meant that low levels of contamination, even when the negative controls were negative, may have been included in analyses as false positives. In addition, the requirement of infection loads to be above the environmental level may have excluded some true infections, resulting in false negatives. To ensure that both false positives and false negatives did not alter the interpretation of the results, all data analyses were repeated with equivocal samples considered negative and samples with infection loads below the environmental level included as positives.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Study methods
  5. Data analysis
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

Three of the 19 samples of treatment tank water tested positive for the presence of B. dendrobatidis (0.14, 0.26 and 0.91 GE). Therefore, the confirmed infection level was set at 1.82 GE. Mean GE scores for seven frog samples were <1.82 and were treated as negative. All of these samples were taken during the first and third swabbing events and hence binary logistic regression and linear mixed models were repeated with these samples included as positives for prevalence and infection load comparisons, respectively (results follow). Equivocal results were obtained for four samples, all of which had scores below 1.82 GE and were included in the previous reanalysis. Therefore, a separate reanalysis of equivocal results was not required. No samples collected in the second swabbing event were found to have infection loads below the confirmed infection level, and hence a repeat of the ROC analysis was not required.

Experiment A: tadpole experiment

None of the tadpoles showed signs of poor body condition and all survived to metamorphosis. However, within 3 months of metamorphosis, three L. peronii and four L. aurea from control tanks along with the three L. peronii and 19 L. aurea from treatment tanks showed poor righting reflexes and were subsequently euthanized. Negative controls survived significantly better than exposed L. aurea (χ2=36.43, P<0.001; Fig. 1), an effect not detected in L. peronii (χ2=0.20, P=0.65; Fig. 1). All animals in the control tanks not exposed to B. dendrobatidis tested negative for infection at all swabbing events. In treatment tanks, 100 of L. peronii and 89.5% of L. aurea were infected at the first swabbing event and this difference between species was not significant. By the second swabbing event, 100% of both species were infected and by the third swabbing event all L. aurea had been euthanized. By comparison, only three L. peronii from treatment tanks had been euthanized and the prevalence of infection in remaining individuals had dropped to 12.5%, a significant decrease compared with swabbing event 2 (χ2=35.05, P<0.001). At swabbing event 3, the GE of one L. peronii sample was below the confirmed infection level of 1.82 GE. Inclusion of this sample as a positive changed the prevalence of infection in this species at this time point to 18.8% which still constituted a significant decrease compared to swabbing event 2 (χ2=17.43, P<0.001).

image

Figure 1.  Survival curves for Experiment A in which Litoria aurea and Limnodynastes peronii tadpoles were inoculated with sterile sham suspensions of zero Batrachochytrium dendrobatidis zoospores (controls) or suspensions of c. 20 million B. dendrobatidis zoospores (treatments).

Download figure to PowerPoint

By the first swabbing event, three L. aurea tadpoles had metamorphosed compared with 16 L. peronii. The remaining L. aurea tadpoles were found to have significantly higher infection loads than the remaining L. peronii tadpoles (F=5.41, P=0.03; Fig. 2). By the second swabbing event, all tadpoles had metamorphosed and a significant difference in the infection load was found between species (F=33.91, P<0.001) and time (F=4.93, P=0.035). Litoria aurea was again found to have higher infection loads than L. peronii at swabbing event 2 but no comparative data were available for swabbing event 3 because all 19 L. aurea had been euthanized. By comparison, only three L. peronii in treatment tanks were euthanized and infection loads were found to be lower at swabbing event 3 than 2 (Fig. 2). The inclusion of the L. peronii sample with an infection load below the confirmed infection level in analysis did not affect the outcome with infection loads still significantly different between species (F=127.95, P<0.001) and time (F=10.44, P=0.005).

image

Figure 2.  The mean number of genomic equivalents detected (±se) in experimentally infected Litoria aurea and Limnodynastes peronii tadpoles from Experiment A before and after metamorphosis, over three swabbing events. Swabbing event 1 occurred 2 weeks after inoculation while swabbing events 2 and 3 occurred 3 and 6 months after metamorphosis, respectively, or at the death of an individual or the death of an individual in the same tank. Sample sizes at each swabbing event are shown next to data points. No data were available for L. aurea at swabbing event 3 due to 100% mortality. Infection loads for tadpoles and metamorphs are not comparable because of differences in infection distribution and swabbing techniques.

Download figure to PowerPoint

Days to death post-metamorphosis was significantly and positively related to GE measured from swabs taken at swabbing event 2 (Wald=30.03, P<0.001). The area under the ROC curve was found to be 0.88 (±0.05) indicating a high level of dissimilarity in infection load between individuals that were euthanized and those that were not (Fig. 3). The most sensitive and specific infection load threshold was 15 GE. Individuals with infection loads greater than this at swabbing event 2 had a 77% chance of showing terminal signs of chytridiomycosis and being euthanized, while individuals with infection loads less than this threshold had a 90% chance of surviving to the end of the experiment (Fig. 3).

image

Figure 3.  Relationship between the number of Batrachochytrium dendrobatidis genomic equivalents detected on Litoria aurea and Limnodynastes peronii individuals from Experiment A at swabbing event 2, that were exposed as tadpoles. The dotted line shows the most sensitive and specific infection load threshold of 15 genomic equivalents.

Download figure to PowerPoint

Experiment B: juvenile experiment

None of the L. aurea and L. peronii juveniles in control tanks showed signs of chytridiomycosis and all survived to the end of the experiment. However, within 3 months of inoculation two L. peronii and all 19 L. aurea in treatment tanks showed poor righting reflexes and were subsequently euthanized. Negative controls survived significantly better than exposed L. aurea (χ2=43.48, P<0.001; Fig. 4), an effect not detected in L. peronii (χ2=1.26, P=0.26; Fig. 4). All animals in the control tanks not exposed to B. dendrobatidis tested negative for infection at all swabbing events. In treatment tanks, 78.9% of L. peronii and 84.2% of L. aurea were infected at the first swabbing event and this difference between species was not significant (χ2=0.03, P=0.87). By the second swabbing event, 100% of both species were infected and by the third swabbing event all L. aurea had been euthanized. By comparison, only two L. peronii from treatment tanks had been euthanized and the prevalence of infection in remaining individuals had dropped to 29%, a significant decrease compared with swabbing event 2 (χ2=25.23, P<0.001). At swabbing event 1, the GE of two L. peronii and three L. aurea was below the confirmed infection level of 1.82 GE. Inclusion of these samples as positives changed the prevalence of infection in these species at this time point to 89.5 and 100%, respectively, and were still not found to be significantly different (χ2=2.55, P=0.11). At swabbing event 3, the GE of one L. peronii sample was below the confirmed infection level and inclusion of this sample as a positive changed the prevalence to 35.3% which still constituted a significant decrease compared with swabbing event 2 (χ2=22.24, P<0.001).

image

Figure 4.  Survival curves for Experiment B in which Litoria aurea and Limnodynastes peronii juveniles were inoculated with sterile sham suspensions of zero Batrachochytrium dendrobatidis zoospores (controls) or suspensions of c. 20 million B. dendrobatidis zoospores (treatments).

Download figure to PowerPoint

The infection load of individuals differed significantly between both species (F=20.49, P<0.001) and time (F=14.08, P<0.001). Litoria aurea had infection loads that increased over time and were higher than L. peronii infection loads at swabbing events 1 and 2 (Fig. 5). No comparative data were available for swabbing event 3 because all 19 L. aurea had been euthanized. By comparison, only two L. peronii were euthanized and infection loads in surviving individuals remained the same over time (Fig. 5). A significant interaction was found between the effects of species and time (F=12.29, P=0.002). A repeat analysis including the two L. peronii and three L. aurea samples from swabbing event 1 and the L. peronii sample from swabbing event 3 that were initially removed as their infection loads were below the confirmed infection level did not affect the outcome of analysis. Significant differences in infection load remained between species (F=20.68, P<0.001) and time (F=11.84, P<0.001) and there was still a significant interaction between the two (F=18.68, P<0.001).

image

Figure 5.  The mean number of genomic equivalents detected (±se) in experimentally infected Litoria aurea and Limnodynastes peronii juveniles from Experiment B over three swabbing events. Swabbing events 1, 2 and 3 occurred 2 weeks, 3 and 6 months after inoculation, respectively, or at the death of an individual or the death of an individual in the same tank. Sample sizes at each swabbing event are shown next to data points. No data were available for L. aurea at swabbing event 3 is due to 100% mortality.

Download figure to PowerPoint

Days to death post-metamorphosis was significantly and positively related to GE measured from swabs taken at swabbing event 2 (Wald=33.59, P<0.001). The area under the ROC curve was found to be 0.92 (±0.04) indicating a high level of dissimilarity in infection load between individuals that were euthanized and those that were not (Fig. 6). The most sensitive and specific infection load threshold was 32 GE. Individuals with infection loads greater than this at swabbing event 2 had an 86% chance of showing terminal signs of chytridiomycosis and being euthanized, while individuals with infection loads less than this threshold had a 95% chance of surviving to the end of the experiment (Fig. 6).

image

Figure 6.  Relationship between the number of Batrachochytrium dendrobatidis genomic equivalents detected on Litoria aurea and Limnodynastes peronii individuals in Experiment B at swabbing event 2, that were exposed as juveniles. The dotted line shows the most sensitive and specific infection load threshold of 32 genomic equivalents.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Study methods
  5. Data analysis
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

The development of disease following exposure to a pathogen requires the successful establishment of an infection and the replication of the pathogenic organism to a level that impairs biological function (McConnell, 2007). Litoria aurea and L. peronii were found to be equally susceptible to infection with B. dendrobatidis as all exposed tadpoles and juveniles tested positive for infection. However, the multiplication rate of B. dendrobatidis differed significantly between species. Mean GE were significantly higher in post-metamorphic L. aurea than L. peronii and increased over time, whereas mean GE in L. peronii either remained the same or decreased below the detectable limit. Given that both the environmental and exposure conditions were standardized in this experiment, these results suggest that post-metamorphic L. peronii possess an innate mechanism that inhibits the replication of B. dendrobatidis following an initial infection and that this mechanism is lacking in L. aurea.

There is increasing evidence that the innate immune responses of amphibians are responsible for species-specific differences in the outcome of B. dendrobatidis infection (Rollins-Smith et al., 2002a,b, 2003, 2006; Harris et al., 2006; Woodhams et al., 2006, 2007). The antimicrobial skin secretions of Limnodynastes tasmaniensis, a species closely related to L. peronii, inhibits B. dendrobatidis more effectively than those from two Litoria species (Woodhams et al., 2007). While not tested against B. dendrobatidis, skin secretions of L. aurea seem to lack antifungal activity (Apponyi et al., 2004). Differences in the efficacy of antimicrobial peptides produced by L. peronii and L. aurea may be responsible for the observed infection outcomes.

In addition to the inhibition of infection, the prevalence of B. dendrobatidis in L. peronii significantly decreased 6 months after exposure, suggesting that this species also possessess a means of clearing themselves of infection. The loss of B. dendrobatidis infections has been observed in other amphibian species (Davidson et al., 2003; Woodhams, Alford & Marantelli, 2003; Berger et al., 2004; Kriger & Hero, 2006) and has been linked to increasing temperature. Alternatively, the rapid and frequent sloughing of a salamander species has been suggested to reduce infection load and contribute to the loss of infection over time (Davidson et al., 2003). Given that temperature remained constant and within the thermal optima for B. dendrobatidis throughout this experiment, sloughing by L. peronii, occurring either in the absence of re-infection or at a greater rate than B. dendrobatidis maturation, may be responsible for the observed loss of infection.

No signs of lethargy or poor righting reflex were detected in L. peronii that exhibited lower mean GE scores. By comparison, all L. aurea exposed to B. dendrobatidis showed signs associated with chytridiomycosis which, under the assumption that death occurs within 48 h (Berger et al., 2005), would have resulted in 100% mortality. This result is consistent with the field observations for this species where over 90% of populations have disappeared (White & Pyke, 1996) and provides support for the role of this pathogen in the L. aurea decline. The survival of L. peronii with infection also echoes field observations as this species has not undergone a decline and persists throughout the former L. aurea range. Additionally, the findings that L. peronii carry infections asymptomatically for at least 6 months and that the tadpoles of both species survive with infection until metamorphosis suggest that they may be acting as reservoirs for B. dendrobatidis, exacerbating the L. aurea decline. Such reservoir hosts are required if a pathogen is to drive a more susceptible host species to extinction (McCallum & Dobson, 1995).

The association between higher infection loads and chytridiomycosis in this study is suggestive of a linear causal relationship. However, the relationship between infection load and survival time was counter-intuitive as individuals that survived for longer were found to have higher infection loads, indicating that infection load does not determine the timing of mortality. Rather, mortality appeared to be determined by the infection load crossing a threshold, beyond which survival time was variable as infection loads increased until death. Therefore, host species determined whether infection load increased over time and the probability of death was in turn dependent upon the infection load crossing a threshold value. An infection load threshold that determines a high probability of mortality is reflective of the number of pathogenic organisms infecting the host that impairs biological function and causes disease (McConnell, 2007). In the case of B. dendrobatidis infections, this threshold may be the point at which the inhibition of osmoregulation causes electrolyte concentrations to fall below the level required for normal cardiac function (Voyles et al., 2009).

The infection load threshold that causes disease will vary considerably between individuals due to differential effects of host susceptibility and pathogen virulence (McConnell, 2007; Beldomenico & Begon, 2010). For example, the pre-existence of electrolyte imbalances or cardiac abnormalities in infected individuals would be expected to lower the threshold at which infection causes disease, as would infection by a B. dendrobatidis strain that is more virulent relative to infection load. The results of this study also indicate that life stage at exposure to B. dendrobatidis may determine host susceptibility as the infection load threshold was found to be lower for individuals exposed as tadpoles than as juveniles. The time it takes for infection load to reach the threshold level (the incubation period) would also vary between individuals due to exposure doses, different rates of contact with infected individuals and subsequent secondary infection opportunities, as well as the suitability of both host tissue and the external environment for pathogen growth and multiplication (Nelson, Williams & Graham, 2001; Carey et al., 2006; McConnell, 2007; Beldomenico & Begon, 2010). However, understanding infection dynamics at the level of the individual is important for understanding the outcome of pathogen exposure on host populations.

Modelling of B. dendrobatidis infection in California's mountain yellow-legged frogs has found within-host infection load dynamics to be key determinants of population persistence or extinction (Briggs, Knapp & Vredenburg, 2010). Individuals in populations that persist with B. dendrobatidis were found to have low infection loads while those undergoing declines were considerably higher (Briggs et al., 2010). Within the same species complex, population declines have been associated with mean infection loads reaching a 10 000 zoospore equivalent threshold, following an infection load increase of 0.15 zoospores/day (Vredenburg et al., 2010). The generation of such population summary statistics, relating to individual infection thresholds, incubation periods and/or pathogen multiplication rate, would allow point estimates of infection loads (like those collected through swabbing and qPCR) to be related to disease dynamics and survivorship, something that is currently lacking in B. dendrobatidis surveillance programmes. These data, in combination with a knowledge of infection prevalence within a population and transmission dynamics, could then be used to model population viability and predict decline events before they are observed. In the absence of being able to eradicate the effects of a pathogen acting as a threatening process, the ability to initially identify species' predisposed to disease outbreaks and to then intervene before population decline, through the inhibition of individual infection loads or the creation of insurance populations, may be the only means of preventing further extinctions.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Study methods
  5. Data analysis
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

This study was supported by Port Waratah Coal Service through the Kooragang Wetland Rehabilitation Project. The authors would like to thank Don Barker and Annette Lynch for the Barker Scholarship and the Australian Animal Health Laboratories, especially Donna Boyle, for training in qPCR and provision of the B. dendrobatidis isolate. We also acknowledge Kim Colyvas and Trevor Moffiet for their statistical advice and to the two anonymous reviewers that commented on earlier versions of this paper. All work was conducted according to the Australian Government National Health and Medical Research Councils Code of Practice for the Care and Use of Animals for Scientific Purposes and under approval from the University of Newcastle Animal Care and Ethics Committee, project number 9891208.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Study methods
  5. Data analysis
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References
  • Anderson, R.M. & Crombie, J. (1984). Experimental studies of age-prevalence curves for Schistosoma mansoni infections in populations of Biomphalaria glabrata. Parasitology 89, 79106.
  • Anderson, R.M. & May, R.M. (1978). Regulation and stability of host-parasite population interactions: I. Regulatory processes. J. Anim. Ecol. 47, 219247.
  • Anderson, R.M. & May, R.M. (1980). Infectious diseases and population cycles of forest insects. Science 210, 658661.
  • Apponyi, M.A., Pukala, T.L., Brinkworth, C.S., Maseli, V.M., Bowie, J.H., Tyler, M.J., Booker, G.W., Wallace, J.C., Carver, J.A., Separovic, F., Doyle, J. & Lewellyn, L.E. (2004). Host-defence peptides of Australian anurans: structure, mechanism of action and evolutionary significance. Peptides 25, 10351054.
  • Beldomenico, P.M. & Begon, M. (2010). Disease spread, susceptibility and infection intensity: vicious circles? Trends Ecol. Evol. 25, 2127.
  • Berger, L., Speare, R., Daszak, P., Green, D.E., Cunningham, A.A., Goggin, C.L., Slocombe, R., Ragan, M.A., Hyatt, A.d., Mcdonald, K.R., Hines, H.B., Lips, K.R., Marantelli, G. & Parkes, H. (1998). Chytridiomycosis causes amphibian mortality associated with population declines in the rain forests of australia and central america. Proc. Natl. Acad. Sci. USA 95, 90319036.
  • Berger, L., Speare, R., Hines, H., Marantelli, G., Hyatt, A.D., Mcdonald, K.R., Skerratt, L.F., Oslen, V., Clarke, J.M., Gillespie, G., Mahony, M., Sheppard, N., Williams, C. & Tyler, M.J. (2004). Effect of season and temperature on mortality in amphibians due to chytridiomycosis. Aust. Vet. J. 82, 434440.
  • Berger, L, Marantelli, G., Skerratt, L. F. & Spear, R. (2005). Virulence of the amphibian chytrid fungus Batrachochytrium dendrobatidis varies with strain. Dis. Aquat. Org. 68, 4750.
  • Blaustein, A.R., Romansic, J.M., Scheessele, E.A., Ham, B.A., Pessier, A.P. & Longcore, J.E. (2005). Interspecific variation in susceptability of frog tadpoles to the pathogenic fungus Batrachochytrium dendrobatidis. Conserv. Biol. 19, 14601468.
  • Boyle, D.B., Boyle, D.B., Olsen, V., Morgan, J.A.T. & Hyatt, A.D. (2004). Rapid quantitative detection of chytridiomycosis (Batrachochytrium dendrobatidis) in amphibian samples using real-time taqman pcr assay. Dis. Aquat. Org. 60, 141148.
  • Briggs, C.J., Knapp, R.A. & Vredenburg, V.T. (2010). Enzootic and epizootic dynamics of the chytrid fungal pathogen of amphibians. PNAS 107, 96959700.
  • Carey, C., Bruzgul, J.E., Livo, L.J., Walling, M.L., Kuehl, K.A., Dixon, B.F., Pessier, A.P., Alford, R. & Rogers, K.B. (2006). Experimental exposures of boreal toads (Bufo boreas) to a pathogenic chytrid fungus (Batrachochytrium dendrobatidis). EcoHealth 3, 521.
  • Daszak, P. (1999a). Extinction by infection. Trends Ecol. Evol. 14, 279.
  • Daszak, P., Berger, L., Cunningham, A.A., Hyatt, A.D., Green, D.E. & Speare, R. (1999b). Emerging infectious diseases and amphibian popultion declines. Emerg. Infect. Dis. 5, 735748.
  • Daszak, P., Strieby, A., Cunningham, A., Longcore, J.E., Brown, C.C. & Porter, D. (2004). Experimental evidence that the bullfrog (Rana catesbeiana) is a potential carrier of chytridiomycosis, an emerging fungal disease of amphibians. Herpetol. J. 14, 201207.
  • Daszak, P. & Cunningham, A.A. (1998). Wildlife disease biology comes of age. Oryx 32, 238239.
  • Davidson, E.W., Parris, M.J., Collins, J.P., Longcore, J.E., Pessier, A.P. & Brunner, J. (2003). Pathogenicity and transmission of chytridiomycosis in tiger salamanders (Ambystoma tigrinum). Copeia 601607.
  • Garner, T.W.J., Walker, S., Bosch, J., Leech, S., Rowcliffe, J.M., Cunningham, A.A. & Fisher, M.C. (2009). Life history tradeoffs influence mortality associated with the amphibian pathogen Batrachochytrium dendrobatidis. Oikos 118, 783791.
  • Gosner, K.L. (1960). A simplified table for staging anuran embryos and larvae with notes on identification. Herpetologica 16, 183190.
  • Hamer, A.J., Lane, S.J. & Mahony, M.J. (2007). Life-history of an endangered amphibian challenges the declining species paradigm. Aust. J. Zool. 55, 7988.
  • Harris, R.N., James, T.Y., Lauer, A., Simon, M.A. & Patel, A. (2006). Amphibian pathogen Batrachochytrium dendrobatidis is inhibited by the cutaneous bacteria of amphibian species. EcoHealth 3, 5356.
  • Hassell, M.P., Anderson, R.M., Cohen, J.E., Cvjetanovic, B., Dobson, A.P., Gill, D.E., Holmes, J.C., May, R.M., Mckeown, T., Pereira, M.S. & Tyrell, D.A.J. (1982). Impact of infectious diseases on host populations. In Population biology of infectious diseases: 1536. Anderson, R.M. & May, R.M. (Eds). New York: Springer-Verlag.
  • Hosmer, D.W. & Lemeshow, S. (1999). Applied survival analysis: regression modeling of time to event data. New York: Wiley.
  • Hyatt, A.D., Boyle, D.G., Olsen, V., Boyle, D.B., Berger, L., Obendorf, D., Dalton, A., Kriger, K., Hero, M., Hines, H., Phillott, R., Campbell, R., Marantelli, G., Gleason, F. & Colling, A. (2007). Diagnostic assays and sampling protocols for the detection of Batrachochytrium dendrobatidis. Dis. Aquat. Org. 73, 175192.
  • IUCN. (2010). Summary statistics for globally threatened species. Available at http://www.iucnredlist.org/about/
  • Johnson, M. (2003). Media for in-vitro cultivation of Batrachochytrium dendrobatidis. Available at http://www.jcu.edu.au/school/phtm/PHTM/frogs/protocols/bd-media-protocol.pdf (Accessed 11 November 2006)
  • Johnson, M.L., Berger, L., Phillips, L. & Speare, R. (2003). Fungicidal effects of chemical disinfectants, uv light, desiccation and heat on the amphibian chytrid Batrachochytrium dendrobatidis. Dis. Aquat. Org. 57, 255260.
  • Keymer, A. (1981). Population dynamics of Hymenolepis diminuta in the intermediate host. J. Anim. Ecol. 50, 941950.
  • Kriger, K.M. & Hero, J.-M. (2006). Survivorship in wild frogs infected with chytridiomycosis. EcoHealth 3, 171177.
  • Lamirande, E.W. & Nichols, D.K. (2002). Effects of host age on susceptability to cutaneous chytridiomycosis in blue-and-yellow poison dart frogs (Dendrobates tinctorius). In Proceeding of the sixth international symposium on the pathology of reptiles and amphibians: 313. Mckinnell, R.G. & Carlson, D.L. (Eds). University of Minnesota, Printing Services, Minnesota.
  • Marantelli, G., Berger, L., Speare, R. & Keegan, L. (2004). Distribution of the amphibian chytrid Batrachochytrium dendrobatidis and keratin during tadpole development. Pac. Conserv. Biol. 10, 173179.
  • May, R.M. (1988). Conservation and disease. Conserv. Biol. 2, 2830.
  • Mccallum, H. & Dobson, A. (1995). Detecting disease and parasite threats to endangered species and ecosystems. Trends Ecol. Evol. 10, 190194.
  • Mcconnell, T.H. (2007). The nature of disease. Baltimore: Lippincott, Williams and Wilkins.
  • Nelson, K.E., Williams, C.M. & Graham, N.M.H. (2001). Infectious disease epidemiology: theory and practice. Maryland: Aspen Publishers.
  • Parker, J.M., Mikaelian, I., Hahn, N. & Diggs, H.E. (2002). Clinical diagnosis and treatment of epidermal chytridiomycosis in african clawed frogs (Xenopus tropicalis). Comp. Med. 52, 265268.
  • Parris, M.J. & Baud, D.R. (2004). Interactive effects of a heavy metal and chytridiomycosis on gray treefrog larvae (Hyla chrysoscelis). Copeia 344350.
  • Penman, T.D., Muir, G.W., Magerey, E.R. & Burns, E.L. (2008). Impact of a chytrid-related mortality event on a population of the green and golden bell frog Litoria aurea. Aust. Zool. 34, 314318.
  • Pyke, G.H. & White, A.W. (1996). Habitat requirements for the green and golden bell frog Litoria aurea (anura: Hylidae). Aust. Zool. 30, 224232.
  • Pyke, G.H. & White, A.W. (2001). A review of the biology of the green and golden bell frog Litoria aurea. Aust. Zool. 31, 563598.
  • Rollins-Smith, L., Carey, C., Conlon, J.M., Reinert, L.K., Doersam, J.K., Bergman, T., Silberring, J., Lankinen, H. & Wade, D. (2003). Activities of temporin family peptides against the chytrid fungus (Batracochytrium dendrobatidis) associated with global amphibian declines. Antimicrob. Agents Chemother. 47, 11571160.
  • Rollins-Smith, L.A., Carey, C., Longcore, J.E., Doersam, J.K., Boutte, A., Bruzal, J.E. & Conlon, J.M. (2002a). Activity of antimicrobial skin peptides from ranid frogs against Batrachochytrium dendrobatidis, the fungus associated with global amphibian declines. Dev. Comp. Immunol. 26, 471479.
  • Rollins-Smith, L.A., Doersam, J.K., Longcore, J.E., Taylor, S.K., Shamblin, J.C., Carey, C. & Zasloff, M.A. (2002b). Antimicrobial peptide defenses against pathogens associated with global amphibian declines. Dev. Comp. Immunol. 26, 6372.
  • Rollins-Smith, L.A., Woodhams, D.C., Reinert, L.K., Vredenburg, V.T., Briggs, C.J., Nielsen, P.F. & Conlon, J.M. (2006). Antimicrobial defenses of the mountain yellow-legged frog (Rana muscosa). Dev. Comp. Immunol. 30, 831842.
  • Rowley, J.J.L. & Alford, R.A. (2007). Behaviour of Australian rainforest stream frogs may affect the transmission of chytridiomycosis. Dis. Aquat. Org. 77, 19.
  • Scott, M.E. (1987). Regulation of mouse colony abundance by Heligmosomoides polygyrus. Parasitology 95, 111128.
  • Scott, M.E. (1988). The impact of infection and disease on animal populations: implications for conservation biology. Conserv. Biol. 2, 4056.
  • Scott, M.E. & Anderson, R.M. (1984). The population dynamics of Gyrodactylus bullatarudis (monogenea) within laboratory populations of the fish host Poecilia reticulata. Parasitology 89, 159194.
  • Skerratt, L.F., Berger, L., Speare, R., Cashins, S., Mcdonald, K.R., Phillott, A.D., Hines, H.B. & Kenyon, N. (2007). Spread of chytridiomycosis has caused the rapid global decline and extinction of frogs. EcoHealth 4, 125134.
  • Speare, R. & Berger, L. (2004). Global distribution of chytridiomycosis in amphibians. Available at http://www.jcu.edu.au/school/phtm/PHTM/frogs/chyglob.htm (Accessed 15 April 2005)
  • Stockwell, M.P., Clulow, S., Clulow, J. & Mahony, M. (2008). The impact of the amphibian chytrid fungus Batrachochytrium dendrobatidis on a green and golden bell frog Litoria aurea reintroduction program at the hunter wetlands centre australia in the hunter region of NSW. Aust. Zool. 34, 379386.
  • Stuart, S.N., Chanson, J.S., Cox, N.A., Young, B.E., Rodrigues, A.S.L., Fischman, D.L. & Walker, R.W. (2004). Status and trends of amphibian declines and extinctions worldwide. Science 306, 17831786.
  • Thompkins, D.M. & Begon, M. (1999). Parasites can regulate wildlife populations. Parasitol. Today 15, 311313.
  • Voyles, J., Berger, L., Young, S., Speare, R., Webb, R., Warner, J., Rudd, D., Campbell, R. & Skerratt, L.F. (2007). Electrolyte depletion and osmotic imbalance in amphibians with chytridiomycosis. Dis. Aquat. Org. 77, 113118.
  • Voyles, J., Young, S., Berger, L., Campbell, C., Voyles, W.F., Dinudom, A., Cook, D., Webb, R., Alford, R.A., Skerratt, L.F. & Speare, R. (2009). Pathogenesis of chytridiomycosis, a cause of catastrphic amphibian declines. Science 326, 582585.
  • Vredenburg, V.T., Knapp, R.A., Tunstall, T.S. & Briggs, C.J. (2010). Dynamics of an emerging disease drive large-scale amphibian population extinctions. PNAS 107, 96899694.
  • White, A.W. & Pyke, G.H. (1996). Distribution and conservation status of the green and golden bell frog Litoria aurea in new south wales. Aust. Zool. 30, 177189.
  • Woodhams, D.C., Rollins-Smith, L.A., Carey, C., Reinert, L., Tyler, M.J. & Alford, R.A. (2006). Population trends associated with skin peptide defenses against chytridiomycosis in Australian frogs. Oecologia 146, 531540.
  • Woodhams, D.C., Ardipraja, K., Alford, R.A., Marantelli, G., Reinert, L.K. & Rollins-Smith, L.A. (2007). Resistance to chytridiomycosis varies among amphibian species and is correlated with skin peptide defenses. Anim. Conserv. 10, 409417.
  • Woodhams, D.C., Alford, R.A. & Marantelli, G. (2003). Emerging disease of amphibians cured by elevated body temperature. Dis. Aquat. Org. 55, 6567.
  • Wright, K.M. & Whitaker, B.R. (2001). Pharmacotherapeutics. In Amphibian medicine and captive husbandry: 309319. Wright, K.M. & Whitaker, B.R. (Eds). Malabar: Krieger Publishing Company.
  • Zou, K.H., O'malley, A.J. & Mauri, L. (2007). Receiver operating characteristic analysis for evaluating diagnostic tests and predictive models. J. Am. Heart Assoc. 115, 645657.