Chytridiomycosis, a disease of amphibians caused by the fungal pathogen Batrachochytrium dendrobatidis (Bd), is an ideal system for studying how temperature influences the host-pathogen relationship because both the host and the pathogen are ectothermic. In this dissertation, I employ a three-pronged approach to investigate the relationship between temperature and the prevalence, severity, and outcome of Bd infections in Northern cricket frogs (Acris crepitans) using both laboratory experiments and field studies. In Chapter One, I conducted a Bd exposure study to determine whether constant temperatures affect Bd pathogenesis differently in live hosts than in culture. Exposure frogs were inoculated weekly with 100 million zoospores of a virulent strain of Bd and frogs were incubated at one of six temperatures: 11, 14, 17, 20, 23, or 26°C. Infection prevalence and pathogen load in Bd-exposed frogs were highest at 14°C while survival was lowest at 11˚C. These results demonstrate that Bd pathogenicity peaks at different temperatures in live hosts than Bd growth peaks at in culture. In Chapter Two, I combine a field study that assesses the effect of microhabitat choice, seasonality and latitude on Bd prevalence and severity in wild amphibian populations with a laboratory study that investigates how temperatures that fluctuate during diurnal and nocturnal cycles influence Bd pathogenicity. Across all states, I found that year was the only significant predictor of infection prevalence, but probability of infection was predicted by year, month, and latitude, and pathogen load was predicted by month and body temperature. In my experimental study, I found infection prevalence and pathogen load to be highest, and survival lowest, in the Louisiana winter group (7˚-17˚C). In my third chapter, I used a biophysical model, Niche Mapper, to predict probability of survival and infection prevalence in Northern cricket frogs by modelling host body temperature. I found that the probability of an infected frog surviving thirty days in May was highest in Southern Louisiana and lowest in Northern Michigan. These studies demonstrate that temperature is an important influence on infection outcome in live hosts. / 1 / Julia Madeleine Sonn

Despite numerous studies, little is known about the influential factors impacting treponemal disease susceptibility. Yet, forms of treponemal disease, specifically yaws, are a major source of morbidity in the developing world. Bioarchaeological materials may assist with deficit by testing full ranges of disease expression. This study explores the relationship between treponemal disease ecology, specifically yaws, and susceptibility through six southeastern United States archaeological samples. Results show early life health experiences have a relationship to later life disease expression. However, influences of climate may only be seen at a cross-regional level. Finally, although nutrition, as evidenced by subsistence strategy, may play a role, within this study, frequencies of treponemal disease are higher in smaller, kinship based groups, rather than large scale societies due to patterns of disease epidemiology. These results should assist with the eradication campaign of yaws announced by the World Health (WHO) organization by 2020.

The consistent seasonal signal that characterizes annual influenza epidemics has long suggested a causal link between the physical environment and the transmission of influenza. Yet, despite considerable interest--dating as far back as Hippocrates--the environmental factors that facilitate the seasonal spread of influenza remain unclear. Historically, significant study of influenza seasonality was based almost exclusively on temperate regions,.due to a lack of high-quality influenza data in low-latitudes. In turn, although numerous hypotheses have been forwarded to explain the seasonal nature of influenza in temperate regions, few acknowledge the seasonal patterns in lower latitudes.This dissertation examines the scientific evidence for the seasonal mechanisms that potentially explain the complex seasonal patterns of influenza disease activity across the latitudinal gradient extending from temperate to tropical regions. I identified seasonal climatic variables that are potentially responsible for influenza seasonality from observational, experimental, ecological and anecdotal studies. I then used a global database of influenza seasonality to assess the consistency of relationships between influenza seasonality and the seasonality of relevant climatic variables. I determined that no single climatic variable is consistently correlated with seasonal influenza activity across temperate, subtropical and tropical regions.However, I did find a significant U-shaped relationship between specific humidity and influenza epidemics globally with epidemics becoming increasingly likely as specific humidity increases or decreases from approximately 12 g/kg. Further, I examined the temporal and spatial variation of influenza activity and specific humidity during the 2009 A/H1N1 pandemic across Mexico, which spans temperate, subtropical and tropical regions. I show that specific humidity may have modified the progression of three distinct waves of infection during the pandemic. These patterns are in agreement with the U-shaped relationship between specific humidity and seasonal influenza epidemics observed at a global scale. In all, this is the first time that relationships between climate and influenza (both seasonal and pandemic) activity have been successfully synthesized into a single parsimonious model across temperate, subtropical and tropical regions.

<p>In the Anthropocene, amphibians must not only cope with natural stressors but also a suite of human-made stressors that have been experienced relatively recently within their evolutionary history. Because it has become increasingly common for natural and anthropogenic stressors to co-occur in aquatic ecosystems, the study of their separate and combined effects on ecosystems and their component species is increasingly necessary. This is especially important for amphibians, which have experienced global declines and can be highly sensitive to both natural and anthropogenic stressors. Pesticides and parasites are two commonly co-occurring stressors that can have complex individual and synergistic detrimental effects in amphibian populations. Here, I conducted three studies to broadly assess the separate and interactive effects of pesticides and parasites on amphibians. More specifically, I explored: 1) the underlying physiological mechanism allowing amphibians to induce increased tolerance to a pesticide within a single generation, 2) the effects of exposure timing to two functionally similar cryptic parasite species on infection outcomes, and 3) population-level variation in susceptibility to parasites and whether prior exposure to pesticides influenced the outcome of host-parasite interactions. First, to test the hypothesis that induced pesticide tolerance is the result of a generalized stress response, I exposed tadpoles to an anthropogenic stressor (sublethal pesticide dose), a natural stressor (cues from a caged predator), or a simulated stressor via exogenous exposure to the stress hormone corticosterone (CORT). I then exposed the larvae to a lethal carbaryl treatment to assess how the stressor exposures influenced survival. I found that prior exposure to exogenous CORT and predator cues induced tolerance to a lethal concentration of carbaryl, providing evidence that pesticide tolerance can be induced by a generalized stress response both in the presence and absence (exogenous CORT) of specific cues. Second, I explored how the timing of host exposure to two co-occurring cryptic echinostome species influences infection outcomes. I found that echinostome infection success in larval anurans can differ significantly based on the parasite species makeup, density, and exposure timing. I also found that priority effects can occur even between functionally similar cryptic species, with an early exposure to <i>Echinoparyphium </i>lineage 3 reducing the infection success of <i>Echinostoma trivolvis </i>three days later. Finally, I assessed the influence of pesticide exposure on host-parasite interactions and population-level variation in these responses. This was accomplished by exposing wood frog larvae from eight populations to one of two treatments (a sublethal carbaryl concentration or a pesticide-free control) followed by controlled parasite exposures to either echinostome trematodes or ranavirus. Then, I assessed how pesticide exposure influenced infection loads, infection prevalence, and survival in each population. I found significant population-level variation in infection outcomes. Interestingly, however, I found no significant effects of pesticide exposure on disease outcomes. Together, these three studies demonstrate the wide-ranging and surprising outcomes that can result from interactions among and between natural and anthropogenic stressors.</p>

Behavior and disease interact bidirectionally and on multiple levels of host organization, and these interactions can have important consequences for population-level disease dynamics. I explored how behavior can both influence and respond to infectious disease in a banded mongoose population experiencing epidemics of tuberculosis (TB) caused by the bacterial pathogen Mycobacterium mungi in the M. tuberculosis complex (Alexander et al. 2010). Banded mongooses are highly social carnivores that live in troops of 5 to 65 individuals. Mycobacterium mungi appears to be primarily environmentally transmitted, but direct horizontal transmission cannot be ruled out. Approximately 10-20% of mongooses become diseased with TB each year in the study population in and around Chobe National Park, Botswana, and all mongooses with clinical signs of TB die within months. Characteristics of both banded mongooses and clinical TB provided a productive study system for exploring interactions between behavior and disease: first, free-living mongooses can be habituated and directly observed; second, the clinical signs of TB can be visually assessed non-invasively; and third, the mongooses' high sociality and egalitarianism provide a unique and ecologically relevant host social system for examining bidirectional interactions between behavior and infectious disease. I found that banded mongooses influenced and responded to disease through their behavior at both the individual and troop level, with possible implications for banded mongoose population and disease dynamics. Due to the environmental transmission of M. mungi, which appears to invade mongooses through breaks in the skin and nasal planum (Alexander et al. 2010), I focused on aggressive interactions as a potential risk factor for acquiring TB in this system. Troops with higher levels of aggression had more injuries, and at the individual level, injuries were a strong predictor of TB, suggesting that aggression may increase risk of disease by creating potential invasion sites for the pathogen. Troops were more aggressive when they foraged in garbage than when they foraged in other habitats, presumably due to the concentration of resources at this highly modified habitat. Overall, my results on how behavior can influence disease in this system suggest that anthropogenic supplementation of food, albeit inadvertent in this system, augments aggression levels in banded mongooses and may in turn lead to a higher incidence of TB. Second, I examined how behavior responds to disease in banded mongooses. Diseased individuals showed significantly lower activity and alertness, but intriguingly, did not show a reduction in overall social behaviors. Diseased individuals were less likely to disperse than healthy individuals, and healthy individuals with diseased troopmates may have been more likely to disperse than individuals without diseased troopmates. Despite this latter possible increase in dispersal in the presence of diseased conspecifics, diseased individuals were not avoided by their troopmates in daily social interactions. For example, diseased individuals were allogroomed at a higher than expected rate even though their reciprocation during allogrooming was approximately half that of healthy individuals. These interactions between behavior and disease have implications for banded mongoose troop and population dynamics, via changes in dispersal behavior and mortality, and can also affect disease dynamics, such as transmission rate. For example, changes to dispersal may affect the amount of inbreeding and outbreeding that occurs in this normally inbred species, and disease might be amplified in areas where aggression is increased by resource augmentation from humans. Additionally, the role that garbage plays in mongoose aggression suggests that humans may be inadvertently increasing disease incidence in this system, as well as in other taxa for which anthropogenic food augmentation may alter disease dynamics via changes in intraspecific aggression. This research sheds light on ways that behavior can influence and respond to disease that are often overlooked in disease ecology. / Ph. D.

Despite the fact that eight cases of Lyme disease were diagnosed in San Luis Obispo County between 2005-2013, the identity of wildlife hosts serving as sources for tick infection in this region remained unidentified. The primary cause of Lyme disease in the U.S. is the spirochetal bacterium Borrelia burgdorferi sensu stricto, and this agent had not been previously isolated from the region. Borrelia bissettii, a related species that has not been implicated as a common causative agent of Lyme disease, was isolated in small rodents inhabiting coastal scrub and chaparral habitats in a previous San Luis Obispo County study. However, B. burgdorferi was not detected. In northwestern California, B. burgdorferi has been primarily associated with high populations of the tick vector Ixodes pacificus in dense woodlands or hardwood-conifer habitats, particularly in the western gray squirrel reservoir host, Sciurus griseus. My study investigated the role of S. griseus and other associated rodents as potential reservoirs for B. burgdorferi in central coastal California woodland habitats. Rodents were live-trapped at four sites in San Luis Obispo County in oak and mixed woodland. Rodent ear samples were tested for B. burgdorferi genospecies by bacterial culture and PCR. Ticks were collected from captured rodents and surrounding environments and tested by PCR for the presence of Borrelia. Of 119 captured rodents, seven were positive for Borrelia infection (5.9%) and of these, six were positive for B. burgdorferi (5.0%). There were multiple infected rodent species that included two western gray squirrels, three deer mice (Peromyscus maniculatus), and one brush mouse (P. boylii). Borrelia spp. were not detected by PCR from the 81 ticks recovered from the environment and rodents. Here, for the first time, we verify the presence of B. burgdorferi sensu stricto in San Luis Obispo county rodents. However, in contrast to previous Northern California studies, the western gray squirrel may not be the primary reservoir host for B. burgdorferi in this region. Multiple rodent species in oak woodlands may be involved in spirochete maintenance in San Luis Obispo County.

Doctor of Philosophy / Department of Biology / Samantha Wisely / Emerging infectious diseases play an increasingly critical role in many biological fields, including conservation biology and public health. Many emerging diseases originate in wildlife, most are caused by viruses, and often emergence is due to adaptation to and amplification in a new host, frequently in areas where ecological transformation is occurring. These emergence patterns suggest that the complex interactions among host, virus, and landscape drive disease emergence. Terrestrial rabies in striped skunks (Mephitis mephitis) in the central Great Plains is an excellent model system to investigate the interactions among the components of disease emergence: host ecology, pathogen properties, and landscape features. Striped skunks are not only numerous in the central Great Plains, they are also the reservoir for two genetically distinct rabies strains that co-occur in the region. Additionally, the landscape in the central Great Plains has undergone significant land use change over the last 70 years through increased urbanization and industrial agriculture practices. I used a combination of molecular and spatial techniques to investigate the interactions among host, pathogen, and landscape. Molecular epidemiology results indicated that rabies strains in the central Great Plains exhibit different epidemiological properties, while population genetic analyses indicated that striped skunks in the region are highly admixed and comprise a single population. Spatial analysis revealed that landscape features such as rivers are not a barrier to striped skunk dispersal, but differentially influence the movement of the two rabies strains. Because striped skunks are reservoirs for many diseases other than rabies and are ubiquitous throughout North America, I also examined the historical movements and distribution of striped skunks in North America using a phylogeographic approach. Results revealed that a combination of multiple Pleistocene dispersal events and Holocene admixture are responsible for the contemporary population structure of striped skunks in North America, and allowed me to place my regional-scale striped skunk rabies study into a larger biogeographic context. My results support the use of a holistic approach for studying emerging infectious diseases that includes studies of viral characteristics, host ecology and biogeography, and spatial features.

Unravelling the complexities of a disease with multiple wildlife host and multiple tsetse vector species is no easy task. After over a century of field observations, experimental studies, anecdotal evidence and conjecture, the role of wildlife in the transmission of trypanosomes is still unclear. Recently, however, frameworks used in the studies of other vector-borne diseases with wildlife reservoirs showed that not only is it possible to understand transmission, but that spatio-temporal predictions of human disease risk and targeted control are realistic aims, even in such complex systems. This thesis explores the epidemiology of human African trypanosomiasis (HAT) in the Serengeti-Mara ecosystem in Northern Tanzania, where recent cases in tourists have highlighted the disease as a public health and economic concern. Assessment of the prevalence of trypanosome infections in different wildlife species is the first step in investigating the relative importance of different species in disease transmission. Identification of trypanosomes relies on sensitive and specific diagnostic tests. Polymerase chain reaction (PCR) protocols based on interspecies differences in the length of the ribosomal internal transcribed spacer (ITS) regions have been widely used in livestock to identify multiple trypanosome species in one PCR reaction. This study represents the first assessment of these protocols on blood samples collected from wildlife. Clonal sequence analysis of PCR products revealed a large range of trypanosomes circulating in wildlife, including Trypanosoma congolense, Trypanosoma brucei, Trypanosoma simiae Tsavo, Trypanosoma godfreyi and Trypanosoma vivax. In addition sequences similar to known sequences, termed Trypanosoma simiae-like and T. vivax-like trypanosomes, may reflect further diversity. However, further characterisation is needed before ITS protocols can be used widely for epidemiological studies in wildlife. The prevalence of T. brucei s.l. and T. congolense varied widely between species. This variation was predominantly explained by taxonomic classification, suggesting intrinsic differences in response to trypanosomes. Trypanosoma brucei rhodesiense, the subspecies responsible for HAT, was identified in lion, hyaena and reedbuck. Age significantly affected the prevalence of T. congolense in lion and hyaena, with the highest prevalence in subadults. The lack of statistically significant differences in prevalence between animals sampled live or after death confirmed that post-mortem sampling provides a method for increasing sample sizes in wildlife studies. The complex relationship between tsetse density and prevalence of trypanosome infections illustrated the difficulties of assessing data from diverse ecosystems with many potential confounding factors. A cross-sectional study of Glossina swynnertoni and Glossina pallidipes, the main tsetse species in Serengeti, highlighted the difficulties of integrating the results of microscopy and PCR to generate meaningful measures of the prevalence of transmissible T. brucei infections for epidemiological studies. However, PCR results suggested that G. pallidipes may be more important as a vector of T. brucei s.l. than has been previously recognised. Spatial variation in both tsetse density and the prevalence of trypanosome infections suggests human disease risk is heterogeneous. The results of this study, along with relevant literature, are considered within the context of frameworks used for other vector-borne diseases and the implications for disease management discussed.

Influenza causes thousands of illnesses and deaths annually in the United States. In part, this is a product of rapid changes in influenza genetics, resulting in different variants than a previous season. Influenza virus traverses landscapes by infecting susceptible hosts, thus allowing seasonal influenza to move great distances due to the mobility of humans who occupy diverse natural, social, and built environments. Using H3N2 influenza viral sequences from Minnesota in the 2012-2013 influenza season we explored relationships between the diversity of influenza genetics and the environments in which humans live. Landscape genetic methods were used to test for relationships between genetic diversity of influenza viruses with different concepts of distance separating the viruses in time and space. Additional analyses were used to identify relationships between influenza genetic evolution and socio-environmental characteristics of Minnesota zip code tabulation areas (ZCTAs) where those viruses were isolated. Influenza hospitalization data in Minnesota ZCTAs was also analyzed with spatial and statistical methods to compare differences and similarities between environmental features driving influenza genetic evolution and influenza morbidity. Findings indicated a complex genetic landscape with few significant correlations between genetic distance and other distance concepts. Elderly populations and populations without health insurance were found to be drivers of H3 hospitalizations. The synthesis of information from theses analyses can be used to inform our overall understanding of influenza diffusion and will allow for more targeted and effective public health prevention strategies.

Doctor of Philosophy / Department of Biology / Jack F. Cully, Jr. / Black-tailed prairie dogs (Cynomys ludovicianus) are epizootic hosts for plague (Yersinia pestis); however, alternate enzootic hosts are important for the maintenance of the pathogen. We determined small rodents and prairie dog associations and quantified rodent and flea relationships in the presence and absence of prairie dog colonies and plague. We identified potential alternate hosts and flea vectors for the maintenance and transmission of plague in the prairie ecosystem. This is the first multi-year study to investigate associations between prairie dogs, rodents and fleas across the range of the black-tailed prairie dog. Few rodent species associated with black-tailed prairie dogs and were found to be highly abundant on colonies. Rodent species implicated in plague were present at study areas with and without plague. Peromyscus maniculatus and Onychomus leucogaster, two widely occurring species, were more abundant in areas with a recent history of plague. Flea community characteristics varied within each study area in the presence and absence of prairie dogs. Based on flea diversity on rodents, and the role of rodents and fleas in plague, we identified P. maniculatus and O. leucogaster and their associated fleas, Aetheca wagneri, Malareus telchinus, Orchopeas leucopus, Peromyscopsylla hesperomys, and Pleochaetis exilis to be important for the dynamics of sylvatic plague in our study areas. Peromyscus maniculatus and O. leucogaster were consistently infected with Bartonella spp., another blood parasite. Presence of prairie dog fleas on other rodents at both off and on prairie dog colonies suggests the potential for intra and interspecific transmission of fleas between rodent hosts, and between other small rodents and prairie dogs.