VectorBiTE 2018 working groups:
1. Assessing the variability of host-seeking behaviors affecting host-finding success among tick subpopulations.
Proposed activities/objectives:
The blacklegged tick (Ixodes scapularis) is the primary vector of Borrelia burgdorferi, the
causative agent of Lyme disease in the eastern and midwestern US, as well as the primary
vector of six other emerging zoonotic pathogens. According to previous work by a
member of this proposed group, the basic reproduction number (R 0 ) for at least some tick-borne
pathogens are highly sensitive to the probability that engorged larvae molt into nymphs and find a
host. Incorporating regional variation in nymph host-finding success into models, would allow
for more accurate predictions of local tick-borne pathogen persistence or emergence.
Host-finding depends not only on the abundance and composition of local host populations, but
also on the questing behavior of individual ticks. Common-garden experiments conducted by
members of this group and others have demonstrated heritable variation in questing propensity
between ticks from regions with differing climates. Any prediction of host-finding success
therefore requires us to consider the relative impacts both of host population characteristics and
of tick behavior. While models of host-finding success in relation to host populations have been
thoroughly investigated elsewhere, real-world measurements of these functions are
generally unable to separate host effects and climate effects. These models are therefore
limited in their generalizability to new environments.
To investigate potential determinants of nymph behavioral variation, we have developed a
dynamic state variable model of questing I. scapularis nymphs that takes into account several
tradeoffs associated with questing. This approach allows us to examine the theoretical impacts
of climate variables, lipid use rates, and host abundance on the fitness of questing nymphs –
and by extension concomitant impacts on transmission of the Lyme disease bacterium and other Ixodes-borne pathogens. We have previously applied this analysis both to locally adapted
(modeled as behaving optimally) regional subpopulations and to ticks that have been exposed
to new climates at a timescale that precludes full adaptation (modeled as behaving
suboptimally) (McClure and Diuk-Wasser, in preparation). Defining these behavioral changes
may have predictive power with regard to: i) variations in tick behavior throughout a questing
season ii) variations in tick behavior between regions with differing climates or host communities, and iii) changes in tick behavior and fitness following migration to a new region.
In line with the focus of VectorBite 2018 (i.e., heterogeneity of behavioral traits), we propose
using this optimal foraging model as a starting-point for separating the relative effects of host
population from climate-driven tick behavior, on tick host-finding success. We also intend that
this working group will generate an improved model that will serve as a hypothesis generator to
drive future empirical studies and future collaborations between the modelers and the field
biologists.
2. Heterogeneity in mosquito blood-feeding behavior, behavioral manipulation by malaria
parasites, and consequences for disease transmission and control
Proposed activities/objectives:
The degree of mosquito-host contact is a key predictor for the transmission intensity of
malaria. To obtain a successful blood meal, a female mosquito must balance the risk of
death due to host defensive behavior against the benefits of feeding on a host species that
maximize its fertility. Such trade-offs, along with other selective forces (e.g. host diversity,
density, and distribution in the insect environment) have likely shaped the level of host
specificity and the innate host preference of mosquitoes. However, many environmental
factors, acting in combination with the innate preferences, influence the final host selection.
These factors include host availability, host accessibility (e.g. use of bed-net) or mosquito
previous experience. As a result, there is an important heterogeneity of host preference /
choice both between and within species. Some populations or species of malaria vectors
display generalist or opportunistic feeding behaviors (i.e. propensity to feed on different
vertebrate species) whereas others show stronger tropism toward either human
(anthropophily/phagy behaviors) or animals (zoophily/phagy). Beside mosquito innate
preference and environmental influences, an increasing amount of empirical studies show
that malaria parasites can manipulate the feeding behavior of their mosquito vector in ways
that increase transmission. First, parasites can increase vector’s biting rate and motivation
towards feeding on a blood-source (“quantitative” manipulation). Second, recent evidence
indicates that the human malaria parasite Plasmodium falciparum can modify its vector host
choice, such that parasite transmission towards suitable hosts is enhanced and/or attraction
to unsuitable hosts is reduced (“qualitative” manipulation). Third, the parasites can modify
the vertebrate hosts’ odor profile to increase attractiveness to vectors.
In an epidemiological context, heterogeneity in mosquito feeding behaviors coupled
with the high host-specificity of malaria parasites suggests that increasing the availability
and accessibility of non-reservoir animals to divert mosquito bites away from suitable host
species (i.e. zooprophylaxis or dilution effect) is a possible sustainable strategy for malaria
control. Because of the complexity and the numerous sources acting on vector trophic
preference, it is crucial to be able to forecast what could be the net impact of increasing
abundance of non-reservoir animals on malaria transmission. Current mathematical models
of malaria transmission are generally centered on the human populations and ignore most of
the above cited environmental- and genetic-determinants of mosquito blood-feeding
behaviors. For example, most existing models assume that uninfected and infected vectors
have similar feeding behavior and do not consider host behavioral manipulation by
parasites.
We propose to use existing empirical data and build on previous models to evaluate
the overall effect of heterogeneity of mosquito blood-feeding behavior and parasite
manipulation on disease transmission. Because each of the factors influencing mosquito
blood-feeding may have contrasting impact on mosquito life-history traits, the main goal is
to develop, relying on what has been done for within-host model of chronic infections, an explicit within-mosquito model of parasite life-cycle. Therefore, through exploration of
different trade-off scenarios, it will be possible to investigate what could be the
epidemiological outcomes of the interactions among genetic and environmental factors
yielding to heterogeneity in mosquito blood-feeding behaviors, and then identify the
conditions under which dilution effect could be effective in reducing transmission
considering the possibility of parasite manipulation.
3.Vector Life History Trade-Offs and Their Consequences for Transmission
Proposed activities/objectives:
Insect vectors, parasites, and their vertebrate hosts associate in a world that is rapidly changing. While recent work has highlighted the effects of changes in land use, climate, and current disease interventions (e.g. drugs, bed nets, vaccines, insecticides) on disease transmission, the majority of this work has focused on effects on individual parasite or vector traits. This approach has begun to yield valuable insight. However, these parasite and vector traits do not operate in isolation. Instead, many of the most critical vector and parasite determinants of disease transmission interact with one another both in the context of the single organism and in their association in the transmission cycle. We argue that by incorporating these interactions in a life history framework will lead to a better prediction and management of vector-borne disease in the face of global change.
In addition to formatting the data for submission to VecDyn, we aim to use these data to incorporate life history trade-offs into a simple transmission model and explore how environmental variation impacts the nature and shape of these trade-offs. Preliminary work by Penny Lynch suggest that when we assume that there is a trade-off between immune investment (increased immune investment = decreased vector competence) and the daily probability of mortality, mechanistic models tend to over-estimate (when immune investment is low, vector competence high, and the daily probability of mortality is low) and underestimate Ro (when immune investment is high, vector competence low, and the daily probability of mortality high). However, data from other non-vector systems demonstrate that additional trade-offs could be important to incorporate, such as between current reproduction and future survival and between current immune investment and future reproduction, all of which can impact our predictions for both mosquito population dynamics and transmission of vector-borne diseases.
4. Assessing the power of rate summation to predict performance in a thermally fluctuating environment
Proposed activities/objectives:
Using a combination of empirical and modeling approaches, we want to assess the power of rate summation to predict how disease transmission will vary in a fluctuating environment (using temperature as an example) across a diversity of organisms, including insect vectors. Rate summation may vary in its capacity to predict trait performance in fluctuating environments across diverse organisms due to different time-dependent temperature effects introduced by organismal responses to chronic vs. acute exposure to temperature, and whether chronic exposure induces thermal acclimation vs. thermal stress responses. Our objective is to use existing data from a variety of ectothermic systems to develop new theory that modifies rate summation for these time-dependent lags and that can be easily applied across a diversity of vector-borne disease systems to increase the predictive power of current modeling approaches.
We will fit a variety of thermal performance curves (symmetrical vs. asymmetrical) to the constant temperature data and then use rate summation along these relationships to infer how performance might vary in a diurnally fluctuating environment. We then will compare whether or not rate summation can predict thermal performance in a fluctuating environment. If rate summation fails to predict performance, we will identify where rate summation fails (e.g. CTmin, CTmax, non-linear portions of the thermal performance curve), determine if this failure appears to be due to thermal acclimation or thermal stress responses (e.g. rate summation over predicts performance in a constant temp environment), and identify if there are any condition /trait covariates that are predictive of when rate summation is successful and when it is not. We then plan to develop new theory to account for these time-dependent. The end product will be a research article highlighting the utility of rate summation, and the implications of rate summation failure for current modeling approaches and predictions for vector-borne disease transmission. We also aim to suggest experimental approaches that can be adopted in vector-borne disease systems to assess and validate these new theoretical approaches.
5. Density-dependence in size-structured mosquito populations
Proposed activities/objectives:
Understanding variation over time and space in the transmission potential of vector populations relies on a thorough understanding of the relation between environmental traits (such as the temporal availability of breeding sites, temperature, housing conditions, etc.) and the individual vector traits which determine both population dynamics and vectorial capacity. However, our understanding of mosquito population dynamics is hampered by our relative lack of knowledge regarding the strength and shape of density-dependence acting on the larval stages, although progress has been made in this area in recent years. A complication is that competition for resources not only affects larval survival and development rates, but also adult body mass, and the latter is linked to lower lifetime fecundity and possibly adult survival. Thus, there are potentially complex density-dependent feedbacks on population-level processes that are influenced by mosquito body size. Static descriptions of mosquito fitness (e.g., r’) which take body size into account exist and have proven very useful. Likewise, complex individual-based stochastic simulation models of Aedes aegypti population dynamics have been developed that do include larval growth and implications of adult size for certain vital rates in great detail. However, due to their complexity, using and certainly building upon such models is likely beyond the capability of most vector ecologists. Here, we propose developing a mathematical model that includes larval and adult size and a relation between size and vital rates. We believe this would allow for more comprehensive investigations into environmental effects on mosquito population regulation, while helping to make such demographic models accessible to a wider range of researchers.
While integral projection model (IPM) have become much more common and there are now a number of introductory ‘how to’ papers, often including code to get researchers started, these still are based on relatively simple approaches. Constructing a robust IPM for mosquitoes adds density-dependence and additional life stages, which is a step beyond these introductory papers. This probably means the approach is outside of the comfort zone of many vector ecologists. With a base IPM which captures the salient demographic elements of mosquito populations developed, the number of possible follow-up investigations is, we suspect, only limited by time and resources. A few examples that come to mind are:
- Structuring models by sex to explore SIT/wolbachia/genetic control strategies in more detail
- Considering additional state variables, such as age or infection status
- Embedding within a SEI-SEIR style epidemiological model to explore impact on transmission
- Exploring environmental covariates (e.g., temperature) to develop habitat suitability maps8
- Exploring population dynamics in a seasonal context
- Exploring the impact of vector control tools in relation to density dependence
- Eco-evolutionary dynamics of vital rates or vector traits using an adaptive dynamics approach.
Many of these might require further experimental data and this could therefore lead to a number of worthwhile projects based on a mix of empiricism and theory. Developing a robust IPM for mosquito populations, leading to development of a paper and making the code available therefore seems like it could open up a promising new direction in vector ecology.
6. Developing individual-based vector population models
Proposed activities/objectives:
Individual-based models represent a fundamentally different approach to modeling than the
classical approach of systems of differential equations commonly used for ecological and
epidemiological studies, in that they track attributes and interactions at an individual level. They are
increasingly used to study complex ecological systems and would allow more intuitive modeling
of several classes of important questions about the ecology of vector-borne diseases. For instance,
different individual vectors following the same set of rules about resource investment and risk
appetite are likely to have different life history experiences, such as feeding success at the larval
stage. These are likely to arise, at least initially, through stochastic differences in opportunities to
food, but may propagate as unsuccessful foragers use a higher proportion of their intake for survival
and less for investment into growth and development. This will lead to differences between adults in
terms of energy reserves, which in turn may affect starvation tolerance and therefore risk appetite
when feeding (possibly in turn driving host preference), and simultaneously affect factors such as
antiviral response. Simple stochastic variation in larval life history may, therefore, lead to an adult
vector population that is highly heterogeneous in a way that is not readily captured by models
consisting of systems of differential equations. In addition, agent-based approaches release
modelers from the tendency to simplify the nonlinearities present in transmission processes to
facilitate ODE-system approaches. Despite this, models of this type are still rare for vector
insects and have so far mainly been used to model spatial dispersion.
We intend to organize a working group to on agent-based modeling for vector-borne diseases. This will
consist of three classes of activities:
1) identifying classes of problems which are best addressed by agent-based approaches (for
instance complex heritability, highly-correlated effects of larval environment on components
of vectorial capacity);
2) developing simple coding frameworks for such models to share within the group and
(eventually) publically, to allow theoretical predictions to be made;
3) Demonstrating that the behavior of such systems is reproducible enough that known
parameter values can be recovered by analyzing simulated data sets;
4) Designing simple laboratory studies to test the predictions of the models.