Very little is known about the effects of viral infection on Aedes
mosquitoes. Several authors have shown that Dengue virus exhibits a remarkable tropism for the mosquito nervous tissues. Linthicum et al 
studied the tropism of DENV-3 in parenterally infected female Aedes aegypti
mosquitoes using immunocytochemical methods and observed that the nervous tissues were among the first tissues to be infected. In fact, these authors suggested that the nervous system is the primary site of virus amplification in mosquitoes infected using this method 
. Several years later, Salazar et al 
corroborated these findings by showing that in mosquitoes orally infected with DENV-2, the nervous tissues are among the first to be infected, presenting detectable levels of viral antigens 5 days after an infective blood meal. Interestingly, these authors also showed that heads and salivary glands were the only tissues where viral antigens continued to accumulate throughout the 21 days observed in their study. All other mosquito infected tissues presented a decrease in the infection rate.
This remarkable tropism of Dengue virus for the insect nervous tissues led us to hypothesize that the infection might have some role in modulating the vector locomotor activity behavior, since it is known that activity rhythms in Drosophila
and other Diptera are regulated by circadian clock neurons in the brain (reviewed in 
). In fact, our results show that although the daily activity patterns of DENV-2 infected and uninfected mosquitoes are similar, the total level of activity is clearly increased upon infection. This increase is most evident in the light-on/light-off transition (), an observation that is particularly interesting considering that the visual system is also highly infected 
. However, it is important to mention that this effect is also clearly detected in the “natural” activity peak occurring during the last hours of the photophase (), which is under circadian control 
, indicating that a similar effect is likely to occur in nature.
Other authors have already observed alterations in Aedes
behavior induced by virus infection. Grimstad et al 
studied the feeding behavior of Ae. triseriatus
females infected with La Crosse virus and reported that infected mosquitoes tend to probe more and engorge less than uninfected females. These results are in accordance with those obtained by Platt et al 
, who showed that the time required for feeding by DENV-3 infected mosquitoes was significantly longer than that required by uninfected mosquitoes. In contrast, Putnam and Scott 
observed that DENV-2 infection did not alter Ae. aegypti
female blood-feeding duration and efficiency in an uninfected host. An explanation for this difference might be that these authors infected mosquitoes with different Dengue virus (3 and 2, respectively) and that Putnam and Scott 
fed mosquitoes 14 days after an intrathoracic infection while Platt et al 
only observed significant differences in mosquitoes fed 5, 8 and 11 days after infection. In our study, we observed locomotor activity differences in DENV-2 infected mosquitoes 2 to 6 days after intrathoracic infection.
A considerable amount of information is currently available on the Aedes aegypti
immune response to Dengue virus infection 
. Several authors have shown an association between circadian rhythms and infection/immunity in insects (e.g. 
). For example, Shirasu-Hiza et al 
showed that Drosophila
infected with bacterium exhibit disrupted circadian activity rhythms and that clock gene mutants are more susceptible to infection than wild-type flies. Also, Lee and Edery 
showed that Drosophila
's ability to fight infections is under circadian control and that flies are significantly more resistant to bacterium when infected in the middle of the night than during the day.
It has been shown that several genes from Aedes aegypti
are up or down-regulated upon Dengue virus infection, and in DENV-2 infected mosquitoes at least one orthologue (AAEL012562) of a Drosophila
gene involved in the control of circadian rhythms, Clock
, has its expression nearly doubled after infection 
. We believe this variation in a gene probably central to the control of mosquito circadian rhythms could also contribute to the observed changes in activity behavior and we are currently investigating whether Dengue virus infection alters the circadian expression patterns of other clock genes 
We are aware that our study suffers from possible caveats. For example, we see a large variation in behavioral effects of Dengue infection between experiments that we cannot explain at the moment. Nevertheless, our study shows that Dengue infection increases mosquito locomotor activity. Changes in vector behavior caused by infection can have potential epidemiological implications. Our results encourage further studies to assess whether increased locomotor activity could have an impact on virus transmission dynamics and Dengue epidemiology.