Field validation of a transcriptional assay for the prediction of age of uncaged Aedes aegypti mosquitoes in Northern Australia

Abstract

Background: New strategies to eliminate dengue have been proposed that specifically target older Aedes aegypti mosquitoes, the proportion of the vector population that is potentially capable of transmitting dengue viruses. Evaluation of these strategies will require accurate and high-throughput methods of predicting mosquito age. We previously developed an age prediction assay for individual Ae. aegypti females based on the transcriptional profiles of a selection of age responsive genes. Here we conducted field testing of the method on Ae. aegypti that were entirely uncaged and free to engage in natural behavior.

Methodology/principal findings: We produced "free-range" test specimens by releasing 8007 adult Ae. aegypti inside and around an isolated homestead in north Queensland, Australia, and recapturing females at two day intervals. We applied a TaqMan probe-based assay design that enabled high-throughput quantitative RT-PCR of four transcripts from three age-responsive genes and a reference gene. An age prediction model was calibrated on mosquitoes maintained in small sentinel cages, in which 68.8% of the variance in gene transcription measures was explained by age. The model was then used to predict the ages of the free-range females. The relationship between the predicted and actual ages achieved an R(2) value of 0.62 for predictions of females up to 29 days old. Transcriptional profiles and age predictions were not affected by physiological variation associated with the blood feeding/egg development cycle and we show that the age grading method could be applied to differentiate between two populations of mosquitoes having a two-fold difference in mean life expectancy.

Conclusions/significance: The transcriptional profiles of age responsive genes facilitated age estimates of near-wild Ae. aegypti females. Our age prediction assay for Ae. aegypti provides a useful tool for the evaluation of mosquito control interventions against dengue where mosquito survivorship or lifespan reduction are crucial to their success. The approximate cost of the method was US$7.50 per mosquito and 60 mosquitoes could be processed in 3 days. The assay is based on conserved genes and modified versions are likely to support similar investigations of several important mosquito and other disease vectors.

Figure 1. Variation with age in total RNA yield from sentinel cage and free-range Aedes aegypti females.

Lines indicate mean (±95% CI) total RNA quantity obtained from head and thorax tissue of individual mosquitoes.

Figure 2. Age predictions of free-range Aedes aegypti from the sentinel-cage derived calibration model.

Values are age predictions for individual free-range females (n = 145) derived from nonparametric bootstrap methods. Solid line, least squares linear regression line. Dashed line, predicted equals actual age line.

Figure 3. Assessments of the capacity of transcriptional age grading to differentiate two mosquito populations with differing mean life expectancies.

A. Experimental age predictions showing interpolated mean (solid line) and standard deviations (broken lines) that were used to generate cumulative normal distributions for each age. Ages were then randomly selected from these distributions and assigned to each individual sampled from a population defined by a given e_x value. Random samples of 100, 200, 300, 400 and 500 female mosquitoes were modeled. Symbols indicate mean e_x values for 999 replicates and bars show 95% confidence limits. B. Estimated population e_x values generated from random samples of different sizes when ages were predicted without error. C. Estimated population e_x values from different sample sizes when mosquito ages were predicted from the experimental error distributions for age predictions achieved in this study. Symbols indicate mean e_x values and bars show 95% confidence limits.