In this post Penelope Hancock discusses her paper ‘Density-dependent population dynamics in Aedes aegypti slow the spread of wMel Wolbachia‘ published in Issue 53:3 today.

Aedes aegypti mosquitoes, the primary vector of dengue and zika, are the current target of a novel biocontrol strategy involving Wolbachia bacteria. Mosquitoes infected with Wolbachia are less able to transmit viruses to humans. Releases of Wolbachia bacteria into field populations of Aedes aegypti mosquitoes are currently underway in several countries, including Australia, Indonesia, Vietnam, Brazil and Colombia. The releases aim to cause the Wolbachia to spread through the target vector population rendering it refractory to infection with (and transmission of) human pathogens.

To support the real world application of this new biocontrol method, we need to develop our understanding of the ecological factors affecting Wolbachia invasion in wild mosquito populations. We know surprisingly little about the ecology of Wolbachia–mosquito interactions, partly because essential demographic traits in wild mosquito populations, such as survival, development rate and fecundity are not well characterized. This is a major barrier to understanding Wolbachia invasion, because the Wolbachia only spreads between mosquito hosts by maternal transmission. Therefore the reproductive fitness of its hosts is critical to allowing the Wolbachia to transmit through successive generations of offspring and invade the population.

Our study investigates the process of Wolbachia invasion in mosquito populations housed in a “semi-field” cage environment that aims to represent the natural mosquito habitat:

Clockwise from left: the semi-field cage, Aedes aegypti larva, Aedes aegypti blood-fed adult.

In addition to creating realistic extrinsic conditions, such as climate and spatial dimension, we aimed to simulate intrinsic levels of competition amongst individuals for limited food resources that are representative of conditions experienced in nature. Natural mosquito populations are known to face potentially strong density-dependent competition for food during their juvenile development phase, which can greatly affect their demographic traits. Body size, development rate, fecundity and survival all depend strongly on how much food is available to the mosquitoes in their larval form. We aimed to determine how well Wolbachia spreads through a mosquito population that experiences food-resource competition, determined by natural variation in larval density, and the consequent demographic impacts.

A challenge in quantifying mosquito demographic variation is that some demographic traits cannot be directly observed by monitoring a population over time. Larval development time, from when an egg first hatches to maturation of the larva into the pupal form, is an important determinant of population dynamics. However, natural mosquito populations have overlapping generations, so that adults create several new cohorts of developing larvae during their lifetime, creating a complex mixed age-class population structure. Because we cannot observe how long individual larvae take to develop, we required a Bayesian modelling approach to estimate larval development times for each cohort from observable quantities.

Estimated larval development time distributions for different larval cohorts

Our semi-field caged mosquito population showed wide variation in multiple demographic traits over time in response to changing larval densities, and the consequent variation in food competition. In order to investigate how this demographic variation impacted the spread of Wolbachia, we introduced mosquitoes carrying Wolbachia into the population, and monitored the Wolbachia frequency over time.

Wolbachia frequency over time

We found that Wolbachia spread much more slowly than would be expected based on the predictions of models that do not incorporate density-dependent variation in mosquito demography. We calculated the expected Wolbachia frequency in the population over time using our estimated larval development times and our observed Wolbachia frequencies in each cohort of newly hatched larvae. This showed that Wolbachia spread slowly because larval development times were very extended in our population due to the high levels of food competition that the larvae experienced.

Our results inform the prediction of Wolbachia invasion in field populations and the development of Wolbachia release strategies for biocontrol of mosquito-borne diseases. Firstly, we showed that the strong demographic variation in mosquito populations arising from food resource competition can have substantial impacts on the rate of Wolbachia spread. Density-dependent protraction of larval development times slows spread, therefore release strategies may need to ensure that Wolbachia frequencies are sustained at high levels during field release periods in order to establish Wolbachia in field populations.  Secondly, estimation of the fitness of Wolbaachia-infected mosquitoes in field environments will need to account for the strong variation in fitness that can be caused by density-dependent effects.