Does membrane feeding compromise the quality of Aedes aegypti mosquitoes?

Abstract

Modified Aedes aegypti mosquitoes are being mass-reared for release in disease control programs around the world. Releases involving female mosquitoes rely on them being able to seek and feed on human hosts. To facilitate the mass-production of mosquitoes for releases, females are often provided blood through artificial membrane feeders. When reared across generations there is a risk that mosquitoes will adapt to feeding on membranes and lose their ability to feed on human hosts. To test adaptation to membrane feeding, we selected replicate populations of Ae. aegypti for feeding on either human arms or membrane feeders for at least 8 generations. Membrane-selected populations suffered fitness costs, likely due to inbreeding depression arising from bottlenecks. Membrane-selected females had higher feeding rates on membranes than human-selected ones, suggesting adaptation to membrane feeding, but they maintained their attraction to host cues and feeding ability on humans despite a lack of selection for these traits. Host-seeking ability in small laboratory cages did not differ between populations selected on the two blood sources, but membrane-selected females were compromised in a semi-field enclosure where host-seeking was tested over a longer distance. Our findings suggest that Ae. aegypti may adapt to feeding on blood provided artificially, but this will not substantially compromise field performance or affect experimental assessments of mosquito fitness. However, large population sizes (thousands of individuals) during mass rearing with membrane feeders should be maintained to avoid bottlenecks which lead to inbreeding depression.

Fig 1. Establishment and selection of the laboratory and field Aedes aegypti populations for feeding on human arms or membrane feeders.

(A) The eight populations underwent selection for at least 11 generations on each blood source before conducting experiments (except for the semi-field experiment). (B) Direct human arm blood feeding. (C) Completed membrane feeding apparatus, showing a blood-filled reservoir covered with a collagen membrane and sealed by a rubber ring (image credit: Veronique Paris).

Fig 2. Fecundity (A, B) and egg hatch proportions (C, D) of Aedes aegypti populations derived from the laboratory and field populations and selected for feeding on human arms or membrane feeders.

All populations were then fed on either human arms (A, C) or human blood through a membrane feeder (B, D) and isolated for oviposition. Bars are medians with 95% confidence intervals.

Fig 3. Feeding duration (A) and blood meal weight (B) of membrane- and human-selected Aedes aegypti populations when females were fed on a human volunteer.

Bars are medians with 95% confidence intervals.

Fig 4. Proportion of Aedes aegypti females selected on human arms or membrane feeders that were visibly engorged when provided access to a human arm (A) or membrane feeder (B,C) for 10 min.

For experiments with membrane feeders, human-selected and membrane-selected populations were tested in separate cages (B) or in mixed cohorts in the same cage (C) where populations were marked with different colors of fluorescent powder. Data for all four populations selected for feeding on each blood source were pooled. Bars are medians and 95% confidence intervals.

Fig 5. Host-seeking ability of human-selected and membrane-selected Aedes aegypti females in BugDorm cages in the laboratory (A-B) and in a semi-field cage (right).

(A) and (C) show the cumulative proportion of human- and membrane-selected females landing over time, where shaded areas are 95% confidence intervals. Landing times for individual mosquitoes from each population tested in laboratory cages are shown in (B) with medians shown, where error bars are 95% confidence intervals.

Fig 6. Attraction of female Aedes aegypti selected on human arms or membrane feeders to host cues in a two-port olfactometer.

(A) Schematic of olfactometer showing the location of the stimulus and control ports (which were alternated between experiments). Arrows indicate the direction of airflow through the olfactometer. (B-C) The proportion of mosquitoes trapped in the control port (B) and stimulus port (C) for human-selected (gray) and membrane-selected (red) populations. Bars show median proportions and 95% confidence intervals while dots show proportions for individual trials.