Modeling Chagas disease vector infection prevalence: incorporating life history characteristics and community composition

Authors: Carolina Cabrera, Nicole L. Gottdenker

Abstract Multihost vector-borne pathogens play an important role in human and veterinary public health worldwide, and understanding factors that drive their transmission is critical to the development of vector-borne disease prevention and control. Two potentially important drivers of multihost vector-born pathogen transmission are 1) the community composition of reservoir host species that come in contact with the vector in a particular habitat, and 2) the life history characteristics of reservoir hosts. One of the most important multihost vector-borne pathogens in the Americas, infecting over 10 million people, is the protozoan parasite Trypanosoma cruzi, the cause of Chagas disease in humans. T. cruzi circulates between wild and domestic animal reservoirs and humans, and is transmitted by a triatomine vector. The objective of this study is to develop a mathematical model that attempts to incorporate biological realities of Trypanosoma cruzi transmission between reservoir hosts and a triatomine vector. Specifically, we evaluate the Chagas disease system in Panama, consisting of a wide range of mammalian reservoir hosts and the main vector Rhodnius pallescens. We link a deterministic SI model for pathogen transmission in the vector with an SI model that describes host community transmission, incorporating host community structure and host life history characteristics, as well as hosts that have been previously infected with T. cruzi, but have developed partial immunity and are less competent reservoirs. Using field and molecular blood meal data, and values from the literature, we calculate a reservoir potential index for the different habitats within this Chagas disease system and evaluate the degree to which changes in reservoir community structure and life history characteristics impact vector infection prevalence.

Download (PDF, 483KB)

Effects of temperature on larval immunity

2013 student Kaela Caballero worked with Professor Sonia Altizer and graduate student Alexa Fritzsche to investigate the effect of temperature on immunity to parasites in monarch butterfly larvae.



Recent studies indicate that environmental factors, particularly temperature, can affect the outcome of host-parasite interactions, with implications for predicting pathogen responses and parasite development to future climate change. In this study, we experimentally tested how ambient temperature affects both immunity and parasite infection focusing on monarch butterflies (Danaus plexippus) and a common protozoan parasite, Ophryocystis elektroscirrha (OE). Monarchs become infected as larvae when they eat milkweed plants covered in OE. Each year, populations of monarchs in eastern North America migrate from as far north as Canada to overwintering colonies in Mexico. Monarchs encounter a wide range of temperatures over the course of migration. For this reason, monarchs are an exemplary system for identifying relationships between temperature and immune defense. In this experiment, we explore the response of two measures of innate immune defense – phenoloxidase (PO) activity and the concentration of haemocytes in monarch larvae reared across a gradient of four temperatures: 23o C, 26oC, 29oC, 32oC. These temperatures represent a range normally experienced by monarchs during the summer months. Both PO activity and haemocytes are relevant indicators of immune competence and pathogen defense in a number of insect species. At the second instar stage, we placed 25 larvae into each of the four temperature treatments. At the fifth instar stage, we weighed and bled larvae for immune assays. We also measured the outcome of parasite infection for monarch larvae experimentally inoculated with OE and reared under the same temperature treatments. Warmer temperatures are known to accelerate growth and development in a linear fashion for a variety of insect species, but previous studies have indicated that rising temperatures are not positively correlated with improved immune competence. We hypothesized that both haemocyte concentrations and PO activity would be highest at intermediate temperatures, and would decrease with both cool and very warm temperatures, as has been shown for some other insect species. We also predicted that the severity of OE infections would be lowest at intermediate temperatures, corresponding to greater resistance of the monarchs. Our initial results reveal that haemocytes were greatest at 29oC, but decrease at both warmer and cooler temperatures implicating that some components of immunity perform best at intermediate temperatures.

Download (PDF, 1.5MB)

A within-host mathematical model for the treatment of active and dormant Mycobacterium tuberculosis

2013 student Alex Becker worked with Professor Andreas Handel to develop a mathematical model of drug resistance in tuberculosis.

Abstract. With the ongoing problem of Multiple Drug Resistant and Extensively Drug Resistant Tuberculosis (TB), treatment strategies are being reassessed in hopes of making therapy more effective as well as shorter. We use a compartmental within host mathematical model and a random sampling method to simulate the effect of TB treatment on active and latent bacterial populations during full course therapy. We assess conventional treatment protocols before addressing the questions posed by the TB Modeling and Analysis Consortium: what are the best first therapy regimens, replacing rifampin with rifabutin (a rifampin like drug with a longer half life), and what are the ideal pharmacokinetic parameters of a new drug that could replace rifampin, shorten TB treatment, and clear latent TB. We use our model to show that a rifabutin regimen can considerably improve treatment success, up to 99%, in a shorter period of time, with average clearance in 50 days compared to 70-90 days. Additionally, we confirm that a novel drug with a slower decay and longer efficiency curve could shorten treatment even more, with average clearance in 37 days. Our model and results add another perspective and set of predictions to TB treatment and drug combinations.

Download (PDF, 183KB)


Parasite species richness in hoofed mammals

Student Rachel Mercaldo, who participated in summer 2013, chose a mapping study for her project. She worked with mentors Patrick Stephens and John Gittleman to map ungulate parasite richness worldwide. The abstract from her summer project is below.

Abstract. The effect of host traits, individual parasite biology, and phylogeny on parasite species richness (PSR) was studied in ungulates.  To correct for differences in sampling effort among host species, in addition to using residuals of PSR on sampling effort, three non-parametric estimators of PSR were employed for analyses of the 22 best studied species.  Data for a wide variety of host traits were obtained from a published species level data base of mammalian traits (PanTHERIA) and tested for correlations with PSR. Data on the biological traits of parasites were obtained from a variety of literature sources and used to determine if the correlates of PSR are influenced by parasite type (e.g., bacterial, viral) or parasite transmission mode. To account for the influence of phylogeny, multivariate analyses were conducted using phylogenetic GLM. Of the 27 host traits considered, only geographical range area and litter size were found to be consistent correlates of variation in PSR.  The type of parasite was also significant, with factors influencing PSR varying for different parasite types, but not transmission mode.  Analyses also showed a strong influence of the latitude and longitude where host species occur on PSR, as well as the temperature, rainfall, and human population density within a host species range.  Future work will focus using spatially explicit models to determine whether the latter results reflect spatial autocorrelation or the influence of environmental variables on PSR.

Download (PDF, 1.99MB)


When worms compete, who wins?

In summer 2013, student Jessica Ramadin studied interactions among gastrointestinal worms of African buffalo together with mentors Sarah Budischak and Vanessa Ezenwa. Here’s the abstract from her summer project.

Abstract. Competition occurs when organisms must vie for limited resources and is often evidenced by a reduction in body size, decrease in reproductive success, change in resource use, or decline in numbers. Competition can occur within a species (intraspecific competition) or between species (interspecific competition) and plays a large role in the formation of natural communities. In this study, I investigated intraspecific and interspefic competition among gastrointestinal helminth parasites co-infecting African buffalo. Response to competition was measured by worm body length and fecundity. Specifically, I measured the length of 10 female and 10 male Cooperia from each of 10 African buffalo using digital photography and a dissecting microscope. Female length measurements and fecundity were first correlated to determine if there was a significant relationship between the two variables. Although there was no correlation between female length and fecundity, the relationship seemed to be influenced by buffalo host identity. To test for the presence of intraspecific competition, I compared worm body length and fecundity to the total number of Cooperia present in each individual host. Results indicated that both worm body length and fecundity were not affected by intraspecific competition. To test for the presence of interspecific competition, I compared worm body length and fecundity to the total number of non-Cooperia worms present in each individual host. Results indicated that interspecific competition did not affect fecundity, but competition and gender interacted to affect worm length. Moreover, this effect seemed to be driven by the presence of Parabronema worms. Overall, these results shed light on the complex interactions between parasite populations and how these interactions may influence disease transmission and severity.

Download (PDF, 1.6MB)


Influence of larval aquatic habitat on the growth and development of Aedes aegypti mosquitoes

2013 student Victoria Knight worked with Professor Michael Strand and graduate students Kerri Coon and Kevin Vogel to investigate the effect of larval habitat on mosquito gut microbial communities.

Abstract  Mosquitoes host diverse bacterial communities that vary depending on many factors, including environment. The composition of bacteria present in the gut of adult mosquitoes is determined by the microbial community present in their larval aquatic habitat. Previous work indicates that the presence of bacteria in the larval aquatic environment is required for normal mosquito development. However, the degree to which larval habitat influences mosquito fitness is not well understood. Water samples were collected from natural breeding habitats around Athens, GA. Aedes aegypti larvae were reared in these water samples in the lab and larval development time, pupation rate, and wing length were used as proxies for fitness of emerging adult mosquitoes. Statistical analyses identified water source as a significant factor affecting larval development and wing length in adults. Filter sterilizing water samples resulted in no larval development, which was rescued by the addition of bacteria. Our results show that adult fitness varies with respect to larval aquatic environment, and that these differences may be explained by variation in the microbial communities present in different locations. This study supports an essential role for bacteria in larval growth and development, and suggests that differences in microbial communities may alter aquatic habitat quality and mosquito production in the field.

Host factors associated with blood parasite infections in aquatic turtles in Georgia

Student Candace Cooper worked with Professor Michael Yabsley during summer 2013 studying Haemogregarine parasites in turtles. This is the abstract from her summer project.

Abstract. Haemogregarines are common intraerythrocytic protozoan parasites of reptiles. Of the four genera of haemogregarines that infect reptiles, Haemogregarina is the most common genus reported from aquatic turtles. The life cycle of these parasites includes leeches as definitive hosts and vectors and aquatic turtles as the intermediate hosts. Previous research has suggested that turtle behavior may play a role in risk of infection and parasitemia with Haemogregarina spp. The objective of this study was to better understand what factors (e.g. basking habits, seasonality, etc.) affect both parasitemia and prevalence of infection. Additionally, to determine if multiple Haemogregarina spp. infect aquatic turtles, we genetically characterized a subset of samples. We hypothesized that Haemogregarina infection prevalence and parasitemia levels would differ between basking and non-basking turtle species and with season as both factors would alter contact rates with leeches. Six different aquatic turtle species were trapped with baited hoop nets in eight locations in Clarke County, Georgia. Giemsa-stained blood smears were analyzed to determine prevalence of haemogregarines and parasitemias based on number of infected cells per 7,000 cells examined. To date, significantly higher prevalence were noted for non-basking species (n=103, 88%) compared with basking species (n=286, 43%). Similarly, non-basking species (0.41%) had significantly higher parasitemias compared with basking species (0.016%). Among the four most commonly sampled species, the non-basking musk turtles (Sternotherus odoratus and S. minor) (0.51%) had higher parasitemias (p=0.0001) compared with the non-basking common snappers (Chelydra serpentina) (0.147%). The parasitemias of the three basking species were different with sliders (Trachemys scripta) (0.022%) having significantly higher parasitemia (p=0.028) than painted turtles (Chrysemys picta) (0.012%) and river cooters (Pseudemys concinna) (0.007%). No differences in prevalence or parasitemia were noted between genders. These data support the importance of turtle behavior (basking or non-basking) in prevalence and parasitemia levels. Lower prevalence and parasitemias in basking species could be attributed to a reduction in leech exposure, an increase in the host immune response, biology of different haemogregarine species, or another unknown factor. To address Haemogregarina species diversity, a segment of the 18s rRNA region (~600bp) was amplified and sequenced. Sequences were highly similar but three groups, possibly different species of Haemogregarina, were identified. Each group contained a diversity of turtle species (including basking and non-basking species) and five turtles showed evidence of infection with two different strains. Additional research is needed to better understand the ecology of these understudied parasites of turtles.

Download (PDF, 446KB)

The Perfect Storm:  Factors that lead to increased transmission and resistance emergence of heartworm in the United States

2013 student Paige Miller worked with Dr. Andrew Park to investigate dynamics of canine heartworm in the United States.

  Abstract  Heartworm disease has been observed all over the world but is distributed heterogeneously and hotspots are thought to be promoted by factors such as climate, pet and owner demographics, and percent of canid population given medication. Recent concerning evidence has suggested the establishment of drug resistant worm in some areas of the United States. Because only one class of drugs exists to treat heartworm, resistance presents a large problem. Two of the first models for heartworm disease dynamics and drug resistance emergence, one deterministic and one stochastic, were developed in order to identify factors that could lead to higher rates of transmission or faster rates of resistant allele increases. Factors analyzed included vector abundance, treatment coverage, and fitness benefit of mutation. It was found that areas with high mosquito abundance and low treatment coverage are more likely to suffer from higher worm burdens in general. Speed of resistance emergence and probability of resistant worm invasion depend on the fitness cost of mutation in heartworms and founding size of resistant worm population. Collectively, these models help to identify key factors and regions that are associated with successful and rapid establishment of drug resistant heartworm populations.

Download (PDF, 3.87MB)