Scientists around the world are currently hard at work genetically engineering new strains of mosquitoes that are poor hosts for diseases such as malaria, dengue and yellow fever, in the hopes of cutting down the spread of these germs. New research suggests, however, that although these insects might succeed in reducing the number of infections, they might also inadvertently boost the severity of remaining ones.

Researchers at Yale University and their colleagues investigated dengue, a mosquito-borne virus for which there is no vaccine or cure. Roughly 50 million cases of dengue occur per year, leading to some 500,000 hospitalizations and several thousand deaths annually.

Insecticides are currently used to control dengue, a practice which runs the risk of breeding insecticide-resistant bloodsuckers. One potential alternative strategy, first proposed more than four decades ago, is to introduce genes into mosquitoes to prevent infections—for instance, by interfering with the spread of germs from the insect's gut to its saliva.

After using mathematical models to investigate the potential evolution of dengue virus in response to genetic modifications of mosquitoes, the researchers found that strategies to block transmission of the virus between mosquitoes and humans may create an evolutionary pressure for the dengue to become more virulent—that is, drive up the extent to which the virus exploits its victims. This in turn could make remaining infections more deadly for humans.

The researchers investigated other transgenic approaches toward mosquito control, as well. Ones that reduce mosquito biting—perhaps by decreasing host-seeking behavior—can also run the risk of increasing dengue virulence and thus severity. Strategies that increased mosquito death rates, however, such as use of genes that kill off females or decrease mosquito immunity to infections, do not seem to impact virulence in humans. Similar findings apply to other lethal mosquito-control techniques, such as by infecting them with diseases that bring an early death, the researchers add.

Although the researchers focused on dengue, "nothing about the mathematical model is specific to dengue, so its results can apply to malaria and other mosquito-borne diseases, as well," says researcher Jan Medlock, an applied mathematician now at Clemson University in South Carolina.

"If we could stop malaria, dengue and yellow fever, there would be incredible benefits for humanity, so these strategies are definitely worth exploring very slowly and carefully," Medlock adds. "On the other hand, there is a long history of introductions of new animals to ecosystems that have been disastrous and often irreversible, such as the introduction of mongooses onto the Hawaiian Islands, and there's always the concern that modified mosquitoes might lead to a similar story by making dengue irreparably worse. This is something scientists are aware needs careful thought."

Medlock notes there are still points of uncertainty in the work, such as whether the dengue virus could evolve fast enough to avoid eradication. "There's tons of research yet that could be done," he says.

Geneticist Luke Alphey of the University of Oxford and its spin-off biotechnology company Oxitec in England who did not participate in this study notes, "It is extremely helpful to have theoretical analyses to give a view of the evolutionary responses that viruses might have, so you can monitor for incipient changes and hopefully do something about the design of your strategy." He and his colleagues are pursuing a transgenic mosquito strategy that employs sterile males as competition for wild males over mates, hopefully leading to smaller populations of the insects.

Alphey adds: "There are insecticides and other interventions used against dengue that are not transgenic but apply similar kinds of evolutionary pressures on the virus, and it would be very interesting to see if there's any data from the field with those interventions for the kinds of effects this study here suggests might appear."

Medlock and his colleagues will detail their findings in the October issue of The American Naturalist.